Does the size of the male penis, in terms of length or width, make a difference in female sexual satisfaction?
Method
To study the effect of penis width vs. length on female sexual satisfaction, 50 sexually active female undergraduate students were asked which felt better, i. e., was penis width or length more important for their sexual satisfaction.
Results
None reported they did not know, or that width and length were equally satisfying. A large majority, 45 of 50, reported width was more important (p < .001).
Conclusion
Implications are discussed, including the fact that the data seem to contradict Masters and Johnson about penis size having no physiological effect on female sexual satisfaction.
Background
When people speak of penis size, they typically refer to length. Thus, a man with a short but wide penis would probably think of himself as having a small penis, and would be so thought of by others, too. However, width is part of size, although usually not acknowledged. Does width contribute to female sexual satisfaction? Is length more important? Or, perhaps size is unrelated to female sexual enjoyment.
The famous sex researchers Masters and Johnson [1,2] have concluded that size of the male penis can have no true physiological effect on female sexual satisfaction. They base this conclusion on their physiological studies that show that the vagina adapts to fit the size of the penis. Because of this vaginal adaptation, they refer to the vagina as a potential space rather than an actual space. Thus, despite the worries of many males about the size of their penis, Masters and Johnson concluded that any size penis will fit and provide adequate sexual stimulation to the female. The present study was conducted to see if female college students would report their sexual satisfaction related to penis length, width, or neither.
MethodProcedure
To test the notion of the possible importance of length vs. width and female sexual satisfaction, two male undergraduate college students - both popular athletes on campus - surveyed 50 female undergraduate college students, considered by the two males to be sexually active, based on the males' prior social experience and knowledge of the females.
Subjects
The female students ranged in age from 18 to 25 years old. In person or via telephone, the females were asked "In having sex, which feels better, length of penis or width of penis?" In half the cases, the word "width" was used before the word "length," but there were no order effects. There were also no effects for telephone vs. personal interview. All female participants answered the question, perhaps because they knew the student asking the question.
Results and Discussion
Of the 50 females surveyed, 45 reported that width felt better, with only 5 reporting length felt better (chi square = 32.00, df = 1, p < .001). No females reported that they could not tell any difference. Some did report that sex in a relationship was better than sex without commitment.
Masters and Johnson [1,2] have said that penis size should have no physiological effect on female sexual enjoyment, since the vagina adapts to fit the size of the penis. The current results call this conclusion into question, and point to the importance of penis width. However, Masters and Johnson could be correct if the present subjects are only reporting their psychological preference, and not showing a true physiological preference. In other words, the present study solely assessed females' perceived level of sexual satisfaction, which might differ from actual physiological arousal and satisfaction.
It is not obvious why a wide penis would be preferred to a long penis, but speculation would suggest the following. Penis width may be important due to a penis thick at the base providing greater clitoral stimulation as the male thrusts into the female during sexual intercourse. That is, a wide penis would seem to offer a greater degree of contact with the outer part of the vagina, including the clitoral area. If this is correct, then Masters and Johnson are wrong about penis size being unrelated, physiologically, to female sexual satisfaction. Masters, Johnson, and Kolodny [3] do not totally rule out penis size being relevant, but they suggest that it is likely of minor importance for female sexual satisfaction (see especially pages 509-510 in Masters, Johnson, and Kolodny [3]). Another possibility is that a wider penis provides the woman with a greater feeling of fullness, which is psychologically, and perhaps physiologically, satisfying.
Further research on sex is necessary to understand the various influences on sexual attitudes and behavior, including how attitudes influence behavior, if, in fact, they do [4,5,6,7,8,9,10,11,12,13]. Different samples could be studied, as well as using different methods of investigation. One might have women rank order different aspects of sexual satisfaction, including such things as physical attractiveness of the partner, romantic feelings, love, and other things, as well as penis size. This would give an understanding of where the different attributes rank in women's stated preferences. But, width vs. length deserves study.
Conclusion
Women reported that penis width was more important for their sexual satisfaction than penis length. The results were statistically significant. Penis width needs to be given more consideration, and taken into account when one discusses penis size. Also, it may be that Masters and Johnson [1,2,3] were wrong about penis size having little or no physiological effect on women's sexual satisfaction. However, the current data cannot provide a final answer, since they are based on self reports of women surveyed about penis length vs. width, and their sexual satisfaction. The results reflect either a psychological preference or a true physiological reality, but we cannot say which, with the present method that was employed.
Pre-publication history
The pre-publication history for this paper can be accessed here:
Acknowledgement
I am grateful to the two reviewers, Charles Negy and Robert M. Gordon, for their excellent suggestions. I have incorporated all of their suggestions into my article.
After 5 years, most reports show that less than 10% of people maintain a 5% loss from initial body weight. Weight maintenance after 10 years is rarely assessed, especially in commercial programs. The current article reports weight maintenance in individuals who had participated 2 to 11 years earlier in a popular commercial weight loss program based on Canada's Food Guide called Mincavi.
Methods
Randomly picked subjects answered a telephone questionnaire. Participants, 291 adult women from various regions of the province of Quebec, had followed the program 2 to 11 years earlier for at least a month. Body weight at the beginning and at the end of treatment was recorded as well as actual weight, age and height. Existing records allowed partial verification of the sample.
Results
Based on corrected weights, percentage of women who maintained at least 5% of their initial weight loss are as following; 2 years = 43.6% (n = 55), 3 years = 33.3% (n = 42), 4 years = 23.8% (n = 42), 5–6 years = 38.2% (n = 55), 7–8 years = 29.4% (n = 51), and 9–11 years; 19.6% (n = 46). Five to eleven years after they had participated in the program 29.1% of all women maintained a weight loss of at least 5%, while 14.3% maintained a loss of at least 10%.
Conclusions
Even though success rate is not as high as could be wished for, results show that participation in the Mincavi program can lead to effective weight maintenance long after individuals have left it. These findings suggest more thorough studies should be conducted on this weight loss program.
Background
Most studies assessing long-term weight loss maintenance have yielded disappointing long-term results, showing that almost all individuals regain lost weight after 3 to 5 years [1-6]. Follow-up on longer periods are rare, and tend to confirm that maintaining substantial weight loss is something that very few people achieve [7-10]. For example, Sarlio-Lahteenkorva and colleagues have recently reported that after 6 and 15-year follow-ups, only 5,1% of all women maintained a loss of at least 5% of their baseline body weight [7].
The majority of these studies have been conducted on hospital and university weight loss programs. Because overweight individuals who seek treatment in such settings display more psychopathology and binge-eating [11], it has been proposed that such studies may have produced overly pessimistic conclusions [12].
All around the world, and especially in North America, commercial weight loss programs have been established in great numbers to help obese and overweight individuals. However, little is known about their long-term effectiveness. In a recent report assessing weight maintenance 1 to 5 years after a commercial program, Lowe et al. have shown that such programs can yield encouraging results [12].
The present paper reports weight maintenance in women 2 to 11 years after their participation in a popular commercial program in the province of Quebec, Canada. This program, called Mincavi (meaning "thin for life" in French), has been enforcing Canada's Food Guide recommendations since 1983.
Upon entry in the program, participants, mostly women, receive a recipe book and are told about the importance of eating at least 3 meals a day and choosing from a variety of foods in the four major food groups (grain products, fruits/vegetables, milk/milk products, meat/meat alternate). Recipes are based on inexpensive, readily available whole grains products, vegetables and fruits, lean meats, low-fat dairy products and legumes. Participants decide themselves how much weight they want to lose.
Using a variety of nutritious, well balanced, family-friendly meals, women and teenagers, in the weight loss phase, are taught to eat approximately 1400 kcal a day and men, around 1800 kcal. During that phase, on average, 50% of the energy comes from carbohydrates, 25% from protein and 25% from lipids. In the maintenance phase, participants are encouraged to increase their caloric intake by 50 kcal per week, in a minimum of 8 weeks, to eventually reach a daily intake of 1800 kcal for women and 2200 kcal for men. During this second phase, diet composition changes slightly, with a decrease in protein and an increase in lipid content (carbohydrates:50%, protein: 22%, lipids: 27%).
In groups of 50 to 100, participants are taught how to record everything they eat in diaries designed for that purpose and are invited to show them to their group leader every week for feed-back. Group leaders are women who have lost weight and kept if off for at least two years by following the Mincavi program. Weigh-in sessions followed by 30 to 45 minute-conferences on various topics (ex. weight loss, nutrition, motivation) and recipe sampling take place on a weekly basis. Additional support from a dietician and a psychologist is available through a toll-free phone line and internet.
Participating in the program involves a one-time fee of 25$ (Canadian dollars), and a 7$ fee per week during the weight loss phase. Once a participant has reached her goal weight, she is given free access to weekly sessions for as long as she maintains her goal weight. A fee of 7$ will be charged on weighing sessions if she is found to have gained weight.
Modest losses such as 5% of initial body weight have been shown in the past decade, to induce significant health benefits such as improvements in lipid profile, glycemia, blood pressure, self-esteem and other health related indicators [13-16]. For that reason, maintenance of a 5% decrease from pre-treatment weight has been recognized in 1995 as the standard for success by the Institute of Medicine [17]. It was used here, as in other studies [7,18] as the cut-off point to determine successful vs unsuccessful weight-maintainers.
MethodsSubjects
Two hundred and ninety one (291) women participated in the present study. Subjects were randomly picked using the company's list of clients. In order to assess long-term weight maintenance, only individuals who had entered the program at least 2 years earlier were contacted. Pregnant women at the time of interview and individuals who had followed the program for less than a month were excluded from the analysis.
Data collection
Subjects were contacted by telephone. Those agreeing to take part in the study, representing 90% of the individuals contacted, were asked a series of questions. Age, height, date of entry in the program, body weight at the beginning of the program, amount of weight loss, and actual body weight, were noted. Body Mass Index (BMI) was calculated for each subject using height and before- and after-treatment weight. Existing records allowed us to verify body weights, weight loss and height on 11% of the sample (n = 31). Date of entry was available from records for all women (n = 291).
Since it has been demonstrated that people tend to underreport their actual body weight, especially if given by telephone, results were adjusted for the magnitude of the discrepancy. Tell and colleagues 19 have shown that on average, people reported a body weight 2.9% lower than the measured weight (mean= 2 kg). A similar discrepancy was observed in our sample. For that reason, a 2.9% increase in body weight was added to all subjects for whom present weight records were not available.
Statistics
Standard methods were used to calculate descriptive statistics and values are presented as means ± SD. Analysis of variance (ANOVA) was used to analyze quantitative variables. Using the ANOVA table, a Bonferroni post hoc test was performed to examine comparisons between groups. Paired t-test was used to evaluate differences between BMIs before the program and at follow-up. For all tests, p < 0.05 was accepted as the significant level.
Subjects characteristics. Average (± SD) age, weight loss, BMI before, after the program and at follow-up and weight loss maintained in terms of percentage of initial body weight, are given here for the entire group, as well as for each follow-up category. When BMIs before the program and at follow-up are compared, a significant difference could be found only in the 2-year follow-up group (t = 2.919, P = 0,0051).
Follow-up
Age
Weight
BMI
BMI
BMI at
Weight
loss
before
after
follow-up
loss
(kg)
program
program
maintained
(% initial
body
weight)
2 years
43,3 ± 12,7
10,4 ± 6,3
30,6 ± 4,5*
26,6 ± 3,9
29,2 ± 5,0*
7,0 ± 8,3
(n = 55)
3 years
45,8 ± 12,5
12,2 ± 7,6
31,5 ± 5,0
27,0 ± 4,1
30,8 ± 5,4
4,5 ± 6,1
(n = 42)
4 years
46,1 ± 14,6
10,2 ± 5,8
30,2 ± 6,1
26,3 ± 5,9
30,0 ± 7,0
3,4 ± 5,5
(n = 42)
5–6 years
43,4 ± 13,4
10,6 ± 7,0
29,6 ± 5,0
25,5 ± 3,5
29,2 ± 5,8
4,7 ± 6,3
(n = 55)
7–8 years
40,4 ± 13
11,2 ± 6,6
28,0 ± 3,4
23,3 ± 4,1
28,5 ± 4,8
3,5 ± 5,7
(n = 51)
9–11 years
39,8 ± 9,7
12,8 ± 9,1
29,3 ± 3,8
24,8 ± 4,4
29,7 ± 4,5
3,4 ± 6,7
(n = 46)
Entire
43,0 ± 12,8
11,2 ± 7,0
29,8 ± 4,7
25 ± 4,5
29,5 ± 5,4
4,5 ± 6,6
group
(2–11 yrs)
(n = 291)
ResultsSubjects characteristics
Mean age for the entire group was 43 yrs ± 13 upon entry in the program. Mean BMIs before, after the program and at follow-up were respectively; 29.8 ± 4.7, 25.5 ± 4.5 and 29.5 ± 5.4. When BMIs before the program and at follow-up are compared, a significant difference could be found only in the 2-year follow-up group (t = 2.919, P = 0.0051). On average, subjects lost 11.1 kg ± 7.1 and at follow-up maintained a mean loss of 4.5% ± 6.6 of initial body weight. At time of follow-up, most subjects were no longer enrolled in the program with only 18 individuals still participating in it.
Weight maintenance
Two to eleven years (2–11 y) after participating in the program, 49.5% (n = 144) of the women had either returned to their initial body weight or gained back additional weight, and 50.5% (n = 147) weighed 1 to 32% less than at the beginning of the program. As can be seen on Table 2, after 2 years, 43.6% of the subjects were found to maintain a weight loss of at least 5% of their initial body mass, whereas 29.1% maintained a loss of 10% or more. After 5–6 years, these numbers were respectively 38.2% and 16.4%. Almost twenty percent (19.6%) of subjects in the 9–11-year follow-up category maintained a weight loss of at least 5% of their initial mass while 10.9% were found to maintain a loss of 10% or more. Of the 18 subjects who were still enrolled in the program at follow-up, 12 (67%) maintained a 5% loss from initial body weight. Average loss maintained in that subset was a 16% decrease in body weight.
Percentage of subjects according to weight category at follow-up. Depending on their body weight at follow-up (2 to 11 years after beginning the program), subjects are placed in the present table in categories ranging from "Heavier than before program" to "Weight loss of more than 25% of initial body weight". For each follow-up period, percentage of subjects who maintain a 5% or a 10% loss from initial body weight is indicated.
Follow-
Same or
Weight
Weight
Weight
Weight
At
At
up
heavier
loss
loss
loss
loss
least a
least a
than
0,1-
5–9,9%
10-
15% or
5%
10%
before
4,9%
14,9%
more
weight
weight
program
loss
loss
2 years
34,5%
21,8%
14,5%
9,1%
20,0%
43,6%
29,1%
(n = 55)
3 years
42,9%
23,8%
14,3%
14,3%
4,8%
33,4%
19,1%
(n = 42)
4 years
52,4%
23,8%
11,9%
4,8%
7,1%
23,9%
11,9%
(n = 42)
5–6 years
50,8%
10,9%
21,8%
9,1%
7,3%
38,2%
16,4%
(n = 55)
7–8 years
58,9%
11,8%
13,7%
11,8%
3,9%
29,4%
15,7%
(n = 51)
9–11
58,7%
21,7%
8,7%
2,2%
8,7%
19,6%
10,9%
years
(n = 46)
2 to 11
49,5%
18,6%
14,4%
8,6%
8,9%
31,9%
17,5%
yrs (all
(n = 144)
(n = 54)
(n = 42)
(n = 25)
(n = 26)
(n = 90)
(n = 48)
subjects)
(n = 291)
5 to 11
56,1%
14,8%
14,7%
7,7%
6,6%
29,1%
14,3%
years
(n = 85)
(n = 22)
(n = 23)
(n = 12)
(n = 10)
(n = 45)
(n = 22)
(n = 152)
Age
Some studies have shown a positive correlation between age and weight maintenance 20 . In the present work, no correlation was found between the subjects' age at the beginning of the program and weight loss maintenance (P = 0.0651, N.S). However, the relatively narrow age spectrum represented among the Mincavi participants (43 ± 12,8) may limit the interpretation of the current results.
Discussion
The vast majority of weight loss programs reported in the literature show poor long-term efficiency. In the recent years however, a few studies have reported a relatively high level of weight maintenance. In the following section, methodological aspects of these studies are discussed in the light of our current results.
Sample characteristicsDuration of treatment
An intensive weight loss program in Slovenia including behavioral, psychological, cognitive and physical elements has shown promising long-term results on 48 subjects [21]. Median weight loss of completers when they left the program was 11.5 kg. At least 5 years later, 13 of them still maintained the reduced weight.
It is important to note that only participants who had successfully completed at least 4 months of treatment were included in the analysis. This criteria probably allowed selection of individuals already more successful or motivated than the ones who had quit the program after less than 4 months. In comparison, subjects were included in the present study after being enrolled for a minimum of one month in the Mincavi program representing the majority of individuals entering this program. In addition, it has been found that treatment duration is significantly correlated with weight loss after treatment and at follow-up – the longer the treatment, the better the results – [22].
Nevertheless, results provided by this Slovanian general practitioner are valuable as his study implied regular follow-ups and weight measurements of participants for 5 years. His study also confirm the importance of a comprehensive approach in the treatment of obesity.
Complementary treatment
One of the rare studies on weight loss maintenance after 5 years has been conducted by Björvell and Rössner, a Swedish team. A 10-year follow-up has indicated a maintenance of weight losses averaging 10.5 kg after a 4 year continuous treatment [23]. However, in an earlier report, the authors have indicated that 36% of their subjects had their jaws fixed from the start, a factor than could have possibly enhanced the results [24].
Selection of subjects
A recent follow-up of individuals who had successfully completed a popular commercial program has shown that 42.6% of the subjects still maintained a 5% weight loss after five-year, while 18.8% maintained a loss of 10% or more [12]. These promising results suggest that some commercial programs can generate effective long-term results. It has to be kept in mind, however, that this study was conducted on successful participants who had reached their goal weight and achieved Lifetime Member status. As the authors state in their article, these individuals only represent a fraction of those who enter this particular commercial program. For that matter, it cannot be assumed that the rest of the participants who had only progressed part way to their goal would have demonstrated similar weight maintenance.
Another such example is seen in a report on The Trevose Behavior Modification Program, a self-help weight loss program offering continuous care. Latner et al. have shown that members who had completed 5 years of the Trevose program were still 17.3% below their pre-treatment weight, showing considerable weight maintenance [25].
Again, it is important to note that the Trevose Program participants were selected upon entry and throughout the weight loss process, starting with 329 applicants and ending with 37 participants at the end of the 5-year treatment period. Therefore, only about a tenth of the participants, all highly motivated and successful at weight maintenance, were available for this particular analysis. This may explain, in part, such outstanding results.
However, even though efficient such continuous and strict treatments may not correspond to the needs and preferences of a majority of people. For example, failure to meet attendance or weight loss requirements results in immediate dismissal from the Trevose program with no possibility of re-entering it. Such programs may suit people who need a strict and highly structured environment to succeed but discourage those who need more flexibility.
Body Mass Index at the onset of treatment
It is known that larger weight losses are associated with greater weight maintenance because more weight is lost to begin with [25]. For that matter, average BMI of participants should be taken into account when comparing weight loss programs' efficiency. For example, average BMI at the onset of treatment was 41.5 for the Swedish program [24] and 34 for the Trevose program [25]. Participants of the Mincavi weight loss program were relatively light, with a mean BMI of 29.8 at the start of the program.
Amount of weight loss
Anderson and colleagues [20] have recently studied participants who had lost at least 10 kg through an intensive very-low-calorie diet. Forty percent (40%) of their subjects maintained a weight loss of at least 5% of their initial body weight after a 5-year follow-up (n = 112). In the present study, participants remained for analyses regardless of how little weight they had lost through the Mincavi program. Nevertheless, when only those women who had lost at least 10 kg were considered for a 5 to 6-year follow-up (n = 43), 55.8% maintained such a weight loss.
In addition, mean weight loss of their subjects was 29.7 kg, while among Mincavi's subjects who had lost at least 10 kg, mean weight loss was only 12,0 kg. As mentioned earlier, the greater the weight loss, the more frequent it has been shown to maintain a substantial portion of it over time.
Other factorsAdjustment for self-reported information
Among the few studies showing relatively high success rates after 3, 5 or 10 years, three relied mostly on self-reported body weight [20,23,26]. In these three cases, discrepancy between self-reported and measured weight was not adjusted for, suggesting that weight maintenance may have been over-estimated for these programs.
Effect of aging on weight maintenance
An additional factor that needs to be taken into account when assessing long-term maintenance of lost weight is the effect of aging on body weight. It has been demonstrated that body weight set-point increases steadily with age in animals [27]. The same phenomenon is believed to exist in humans [28] making it increasingly difficult to maintain one's young adult body weight. It has been demonstrated that an average weight gain of 11 kg occurs between the age of 25 and 65 in women [28]. A decrease in energy expenditure is believed to contribute to this weight gain [29]. Therefore, some of the weight gain observed in longer follow-ups (>10 years) could be attributed in part to the effect of aging rather than poor weight control. Consequently, it is possible that subjects who have returned to their initial body weight after a decade may in fact be leaner than what they would have been if they had not followed the weight loss program.
Conclusion
The present work is one of the rare existing studies on weight maintenance 10 years following a weight loss program, whether commercial or not. Another unique feature of the current report is that subjects were included in the study whether they had reached their goal weight or not while in the program. By doing so, results provide a picture of long-term weight maintenance in most individuals who enter this particular commercial program rather than focusing on successful individuals only.
Limitations of the present study include a small number of subjects in each follow-up category. While the initial number of participants is decent at 291, the subsets that are subsequently used in the analysis become small, thus eroding confidence in the results. Another limitation of this study is the use of self-reported data. Because the present work relied mostly on such data, interpretation calls for caution. For that matter, it was necessary to adjust for the discrepancy often seen between self-reported body weight and measured weight.
Once corrected, results show that 5 to 11 years after the program, 29% of women still maintained a weight loss of at least 5% of their initial weight (n = 45). While lower than what could be wished for, these results are more encouraging than those generally found in the literature.
These preliminary results suggest that Mincavi, a weight loss program that encourages participants to eat a variety of nutritious, well-balanced, family-friendly meals, can be a useful tool for the long-term treatment of overweight and obese individuals. Prospective studies involving a greater number of subjects and repeated measures of body weight should be conducted in order to better assess long-term effectiveness of the Mincavi program and understand factors contributing to weight maintenance.
Declaration of competing interests
Have you received reimbursements, fees, funding, or salary from an organization that may in anyway gain or lose financially from the publication of this paper in the past 5 years? If so, please specify.
Yes. Dr Gosselin is now scientific manager of the company. However, at the time the present study was conducted she was at the Faculty of Medicine of Sherbrooke University and not employed nor funded by Mincavi. No such competing interests for Ms. Cote.
Have you held any stocks or shares in an organization that may in any way gain or lose financially from the publication of this paper? If so, please specify.
No, for both authors.
Do you have any other competing interests? If so, please specify.
No, for both authors.
Are there any non-financial competing interests you would like to declare in relation to this paper?. If so, please specify.
No, both authors.
Pre-publication history
The pre-publication history for this paper can be accessed here:
Acknowledgements
The authors would like to thank Lyne Martineau and Caroline M. Gauthier, president and vice-president of Mincavi inc., as well as Véronique Gilbert, BSc, Danielle Dubois, Dtp and the Mincavi group leaders for their valuable assistance.
This investigation was undertaken to describe patient perception and awareness of the polycystic ovary syndrome (PCOS), the most common cause of anovulation/oligoovulation among women of reproductive age.
Methods
Fifteen parameters were evaluated by a computer-based research instrument accessed by a large, unscreened population. Incomplete questionnaires were not entered, and responses were electronically tabulated to block duplicate submissions.
Results
From 657 participants, the majority (63%) were between 26–34 years old; mean BMI was 30.4 kg/m2. 343 of 657 had at least one pregnancy and 61% of the study group had taken fertility medicine (any type) at least once. Physicians were the most common provider of PCOS information for all study participants, irrespective of age. Patient emotions associated with the diagnosis of PCOS included "frustration" (67%), "anxiety" (16%), "sadness" (10%), and "indifference" (2%). Self-reported patient aptitude regarding PCOS was scored as high or "very aware" in >60% of women. Respondents were also asked: "If your PCOS could be safely and effectively helped by something else besides fertility drugs or birth control pills, would that interest you?" Interest in alternative PCOS treatments was expressed by 99% of the sample (n = 648).
Conclusions
In our study population, most women associated negative emotions with PCOS although the self-reported knowledge level for the disorder was high. While these women regarded their obstetrician-gynecologist as integral to their PCOS education, traditional PCOS therapies based on oral contraceptives or ovulation induction agents were regarded as unsatisfactory by most women.
Background
Little is known of how patients with the polycystic ovary syndrome (PCOS) acquire information about their condition, whether they feel their level of understanding about PCOS is adequate, or what emotions are associated with this diagnosis. Some of the components of PCOS are known to be profoundly stressful, yet effective and systematic efforts to characterize the subjective awareness of this disorder have only recently been developed [1]. Mainstream treatments for PCOS have remained fairly fixed over the past 30 years, with the two principal non-surgical approaches being oral contraceptives and clomiphene citrate. PCOS therapies based on orally-active insulin sensitizers account for limited clinical use at present, although preliminary data suggests an acceptable safety and efficacy profile for such treatment [2-4].
In this study, we wished to describe the self-reported knowledge base for PCOS in a large, unscreened patient group, as well as to identify where patient information about PCOS originates. We were especially interested to learn which emotions PCOS patients ascribe to their problem once the diagnosis is made, and how they feel about the treatments proposed for their condition. Additionally, this study sought to measure women's receptiveness to alternative treatment strategies not predicated on oral contraceptives or ovulation induction when queried in a patient-centered, confidential and anonymous manner.
Methods
A multidisciplinary team developed a 15-item research questionnaire specifically targeted to women with PCOS. Queries were structured as single-answer from multiple choice responses and data tabulated accordingly for analysis; the questionnaire was then configured electronically for internet access [see link: ]. Although participants were required to register via computer in order to access the questionnaire site, there was no cost to do so. When a group of volunteers (n = 12) pre-tested the study questionnaire, the mean time required for completion was 8 min (including registration). The PCOS questionnaire was posted in English and continuously maintained on a free public-access medical internet site (OBGYN.net; Austin, TX) throughout the month of January, 2000. The questionnaire site was identifiable and its recognition confirmed via three major internet search engines directed to the terms "PCOS" or "PCOS studies".
Participation in this open study was not restricted to current or former patients of the investigators, but was made available to all women who had been diagnosed with PCOS from any center. Participants received no compensation or acknowledgement for their reply to this study questionnaire. Incomplete questionnaires were not tabulated, and responses were electronically tracked to block duplicate submissions from the same individual. Height and weight information was received from participants either in Imperial (U.S.) or SI (metric) units, with automated metric conversion for analysis. Only investigators were permitted access to running totals of the questionnaire during the study.
Our research design was submitted to the Institutional Review Board of Atlanta Medical Center prior to questionnaire implementation. As our study required no direct patient contact or medical record review, the research protocol was judged as "no risk to human subjects"; IRB oversight was therefore offered but not required.
Chi-square/contingency table test was used to determine dependence between measured variables in the data array. P values <0.05 were considered significant.
Results
During the study interval, 657 unique responses to the questionnaire were registered. More than 97% (n = 638) of the respondents indicated that they were familiar with PCOS, while 1.9% had not been told about PCOS, and <1% were uncertain. Approximately 61% of women regarded their level of understanding for PCOS as "very aware" (n = 403), but 9.1% and 2.9% of women considered themselves to be "minimally aware" or "not aware at all" regarding PCOS, respectively. The proportion of the sample registering as ever-pregnant was 52.2% (n = 343), and the average number of pregnancies reported (including miscarriages) was 2.08/patient. Median [IQR 25;75] height and weight in this study population was 165.1 [160;170] cm and 86.0 [70;104] kg, respectively (median BMI = 31.9 kg/m2).
A summary of responses stratified by age and other parameters is given in Figures 1,2,3,4,5,6,7. When these demographic data were related to the self-reported "awareness" of PCOS among these women, those subjects between age 26–34 were significantly more "aware" of PCOS than any other age group (p = 0.038). This age group was also more likely to have their most recent physician appointment occur within the past six months (p = 0.036). Analysis of matched-data from each age category supported a dependence between patient age and preferred information resource, as the physician was identified as responsible for a significant amount of counseling in the 26–34 year-old group (p = 0.022).
Age distribution of polycystic ovary syndrome study participants.
Educational resource most likely to be accessed first by respondents when additional information about PCOS is sought.
Initial source of information about PCOS among women with self-reported awareness of the disorder.
Time interval since most recent physician office visit.
Self-reported reason for most recent appointment to see physician.
Patient emotions/feelings attributed to their diagnosis of PCOS.
In contrast, distribution of self-reported emotional reaction to the diagnosis of PCOS could not be predicted by any clinical parameter, and the feelings experienced about PCOS did not follow any age-specific pattern (p = 0.268). There were, however, expected statistical dependencies between increasing age and ever-pregnant status (p < 0.001), as well as between age and history of receiving fertility therapy (p < 0.001). When asked: "If your PCOS could be safely and effectively helped by something else besides fertility drugs or birth control pills, would that interest you?", interest in alternative PCOS treatments was expressed by 99% of the sample (n = 648). Assuming normal distribution of PCOS in the general population, power analysis defined the probability of detecting a significant (p < 0.05) relationship between dependent and independent variables at .80 with a sample of n = 657.
The total number of page-views specifically attributed to the PCOS questionnaire site during the study period was 1,585 (41.5% response rate). Although completely voluntary, any prospective questionnaire viewer was given an opportunity to critique the instrument and post any concerns to the host. During the study there were no complaints or adverse outcomes regarding questionnaire content/format.
Discussion
Polycystic ovary syndrome (PCOS) describes a convergence of chronic multisystem endocrine derangements, including irregular menses, hirsutism, obesity, hyperlipidemia, androgenization, large and polycystic-appearing ovaries, insulin resistance and infertility [5]. Although PCOS is the most common hormonal disorder among reproductive age women [1], is the leading cause for anovulatory/oligoovulatory infertility [6], and has the potential for serious long-term health effects [7-10], paradoxically the condition is often underdiagnosed. The absence of uniform diagnostic criteria for PCOS [11] has probably contributed to its low clinical diagnosis rate, although organized media campaigns [12] aimed to increase PCOS awareness (both among physicians and patients) should result in more women seeking medical evaluation specifically because PCOS is suspected. As these laudable education efforts gain momentum, however, the current state of patient comfort with their "level of awareness" regarding PCOS, where their facts come from, and how they feel about their diagnosis are imprecisely characterized. If the objective of heightened public interest in PCOS is indeed achieved through wide public-service announcements or other structured media exposure, then a contemporary (baseline) measurement of patient perception and awareness of PCOS as reported here would be helpful.
Although we could not verify the self-reported cognitive claims made by study participants, more than half of respondents considered themselves to be "very aware" of PCOS. These data confirm an intuitive association between physicians and PCOS patient teaching through office-based patient education and counseling. That most women would turn first to their doctor for further guidance regarding PCOS is not a new finding, although these data indicate that the messages conveyed by doctors appears to be poorly received by patients.
Specifically, the strongly negative emotions study participant identified with PCOS management by physicians demonstrated a profound gap between patient expectations and actual (prescribed) therapy. Indeed, our study population appeared unsatisfied with oral contraceptives or clomiphene citrate when prescribed as treatment for PCOS. In this study we were careful not to call specific attention to any substitute therapy, but rather to describe patient acceptance of these two existing treatments. This investigation detected a strong receptiveness among PCOS women regarding any safe and effective alternative to PCOS management. "Frustration" and "anxiety" associated with PCOS was observed across all age groups in this sample, perhaps suggesting either that a PCOS diagnosis was delayed in some older patients, or that negative emotions persist in some PCOS patients for many years after they are informed of their condition.
As with any questionnaire-based research, there were some important limitations with our study that should be acknowledged. Our research method relied on an unscreened but motivated audience and a functional computer interface, with the result that PCOS patients lacking the means and/or ability to access the study questionnaire were excluded from our sample. The fact that study participation depended on internet use likely explains the high self-reported familiarity and knowledge level for PCOS in this group. While PCOS preferentially affects women of reproductive age, and patients in this age group may tend to have a greater fluency with computer-based research tools, a full exploration of this demographic association was beyond the scope of our investigation. Conversely, the possibility also existed that some of our respondents had not been properly diagnosed with PCOS and therefore inappropriately inflated the sample. We regarded these two balanced sources of selection bias to be of essentially equivalent magnitude and therefore mutually negating. Whether or not the observations reported by computer-assisted questionnaire are representative of all PCOS patients is difficult to establish, although this represents the focus of ongoing research at our institutions. It may be that results obtained exclusively from an anonymous, confidential computer-accessed questionnaire are more likely to depict extreme views not typical of those encountered clinically, yet the sample size registered here was considered sufficiently large to attenuate this effect.
Conclusion
Efforts to heighten the profile of PCOS among the general public remain critical, as "minimal awareness" or "no awareness" of PCOS was present among >10% of this motivated sample of unscreened patients. Although many women are affected by the polycystic ovary syndrome, difficulties associated with its uniform diagnosis persist. Given the diagnostic challenges of PCOS, there is perhaps little wonder that confusion also accompanies its treatment. Our work suggests that the present state of mainstream clinical management for this common endocrine disorder (birth control pills or clomiphene) appears to fall short of patient expectations, and therefore may represent a ready clinical niche for newer therapeutic modalities.
Competing interests
None declared.
Pre-publication history
The pre-publication history for this paper can be accessed here:
Acknowledgements
The authors are grateful to the computational informatics division at OBGYN.net for their work in facilitating data collection for analysis. We are especially appreciative of the many women who participated in this questionnaire and made our research possible.
Fallopian tube prolapse as a complication of abdominal hysterectomy is a rare occurrence. A case with fallopian tube prolapse was managed by a combined vaginal and laparoscopic approach and description of the operative technique is presented.
Case presentation
A 39-year-old woman with vaginal prolapse of the fallopian tube after total abdominal hysterectomy presented with an incorrect diagnosis of adenocarcinoma of the vaginal apex. The prolapsed tube and cystic ovary were removed by vaginal and laparoscopic approach. The postoperative course went well.
Conclusions
Early or late fallopian tube prolapse can occur after total abdominal hysterectomy and vaginal hysterectomy. Symptoms consist of persistent blood loss or leukorrhea, dyspareunia and chronic pelvic pain. Vaginal removal of prolapsed tube with laparoscopic surgery may be a suitable treatment. The abdominal or vaginal approach used in surgical correction of prolapsed tubes must be decided in each case according to the patient's individual characteristics.
Background
Fallopian tube prolapse (FTP) is an unusual complication of both abdominal and vaginal hysterectomies. Sporadic cases have been added to the literature about predisposing factors and techniques of management [1]. The diagnosis of FTP needs postoperative observation and careful examination of the vaginal vault. A biopsy is necessary to differentiate between this lesion and the presence of vaginal granulation tissue. The presence of cellular atypias may result in an unnecessary additional operation. These changes are probably secondary to the severe underlying inflammatory reaction. Vaginal infection, postoperative bleeding, a defective surgical technique, and poor physical condition of the patient are the most important predisposing factors for intravaginal prolapse of the fallopian tube [2]. An excisional biopsy constitutes the definitive diagnostic procedure and adequate treatment for this posthysterectomic vaginal complication.
We have managed a woman with FTP on combining a simple vaginal approach and laparoscopic total salpingectomy and oophorectomy. The operative technique described in this case with this condition provides the advantages of total salpingectomy without laparotomy.
Case presentation
A 39-year-old woman, gravida 9, para 8, has been performed a total abdominal hysterectomy and left salpingo-oophorectomy in two years ago for menometrorrhagica. She admitted complaining of vaginal spotting and pain with sexual intercourse of 3 months' duration. Pelvic examination revealed a mass at the apex of the cuff, with an erroneous diagnosis of adenocarcinoma of the vaginal apex. On ultrasonographic examination, the right ovary was cystic appearance in diameter 38 mm × 40 mm. Microscopic examination of a biopsy specimen from the mass revealed the prolapsed fimbrial end of a fallopian tube. Acute and chronic inflammation were present. After her admission medical and surgical treatment options were discussed with the patient. The right fallopian tube and ovary were removed together with granulation tissue, as described below.
Methods
Simultaneous vaginal and standard laparoscopic technique with additional puncture in the suprapubic region and at McBurney's point were used because of cystic appearance of the right ovary. These sites were used for the insertion of the bipolar forceps, and laparoscopic scissors. Thick and thin pelvic adhesions were incised, freeing the tube from the pelvic sidewall and minor bowel attachments and mobilizing the mesosalpinx. We made an elliptical incision around the vaginal mucosa and by sharp dissection, freed the tube from vagina after laparoscopic mobilization of the tube and ovary. The tube and ovary were withdrawn through transvaginally when freed from attachments. The pelvic peritoneum and vaginal mucosa were closed. Prophylactic antibiotic was administered.
Discussion
FTP may cause symptoms serious enough to warrant treatment. The fallopian tube is unlike other abdominal viscera in that it is sensitive to touch, cutting and crushing [3]. Because dense adhesions may involve the tube, ovary, bladder, and bowel, most authors have performed partial salpingectomy [1]. However, total salpingectomy is preferred, for the reason that continued pain by traction on the remaining portion of the prolapsed fallopian tube.
Factors that may contribute to FTP include pelvic infection, poor hemostasis, intraperitoneal vaginal drains or packs, and failure to close the vaginal cuff [1,3]. These may lead to the development of a defect between the pelvic peritoneum and the vagina [1,4]. As has been note by others, prolapse of the fallopian tube following vaginal or abdominal hysterectomy is probably much more common than the numbers of reports in the literature indicate [3,5]. FTP after vaginal hysterectomy is more common than abdominal hysterectomy [6,7]. However, it was also reported that most cases occurred after abdominal hysterectomies [5].
In present case, because the right ovary was bound by dense adhesions to the right fallopian tube, a right salpingo-oophorectomy was performed. Postoperatively, she did very well and within a few days after surgery, she was pain free. The exact incidence of FTP is difficult to estimate, because cases may go undiagnosed and may resolve before detection. In the present woman, factors associated with development of FTP include grand multiparity and low socioeconomic status. The diagnosis of FTP should be suspected when a red, granular, polypoid mass or lesion is seen at the vaginal cuff protruding into the vagina after hysterectomy. The differential diagnosis that must be considered includes proliferative granulation tissue related to surgery, malignant lesion. The definitive identification is made by histopathologic examination. In the present case, the pathological result of right fallopian tube and ovary was revealed as tube and cystic ovary with inflammation. Wetchler and Hurt describe their technique for total excision of the prolapsed fallopian tube via the vaginal approach [8]. Letterie and associate operative technique combines a vaginal approach with laparoscopic total salpingectomy [9].
Conclusions
The described method should be applicable in most clinical circumstances when salpingectomy is indicated, and may be performed on an outpatient basis. This combined technique allows adhesiolysis and full mobilization of the adnexa. However, dense bowel, bladder, and adnexal adhesions that prevent full mobilization of the involved tube often prevent total excision by a laparoscopic approach. Under these circumstances as partial salpingectomy or an abdominal approach must be considered. Following surgery, complete symptom resolution is usually observed and no recurrence has been reported [10].
Competing interests
None declared.
Pre-publication history
The pre-publication history for this paper can be accessed here:
Acknowledgments
The author wish to thank Professor Engin Aydin, MD, Department of Pathology, Inonu University Hospital, Malatya, and Associate Professor Cemal Gundogdu, MD, Department of Pathology, Ataturk University Hospital, for histopathologic evaluation. Written consent was obtained from the patient or their relative for publication of the patient's details.
Our investigation sought to compare changes in sexual function following supracervical hysterectomy (SCH) and total abdominal hysterectomy (TAH).
Methods
A retrospective chart review was performed to identify all patients who underwent supracervical hysterectomy or total abdominal hysterectomy at a tertiary care center. Patients who met criteria for participation were sent a one page confidential, anonymous questionnaire to assess sexual function experienced both pre- and postoperatively. A total of 69 patients in each group were eligible for participation. A multiple logistic regression model was used to analyze measured variables.
Results
Forty-eight percent (n = 33) of women undergoing a SCH returned the questionnaire, while 39% (n = 27) of those undergoing a TAH chose to participate. There were no significant demographic differences between the two groups. Patients who underwent TAH reported worse postoperative sexual outcome than SCH patients with respect to intercourse frequency, orgasm frequency and overall sexual satisfaction (P = 0.01, 0.03, and 0.03, respectively). Irrespective of type of hysterectomy, 35% of patients who underwent bilateral salpingoophorectomy (BSO) with hysterectomy experienced worse overall sexual satisfaction compared to 3% of patients who underwent hysterectomy alone (P = 0.02).
Conclusions
Our data suggest that TAH patients experienced worse postoperative sexual function than SCH patients with respect to intercourse frequency and overall sexual satisfaction. Irrespective of type of hysterectomy, patients who underwent bilateral salpingoophorectomy experienced worse overall sexual satisfaction.
Introduction
Hysterectomy is the most common gynecologic operation performed in the United States. It is estimated that by the age of 65, 1 of 3 women will have undergone a hysterectomy [1,2]. Current practice in the United States strongly favors total abdominal hysterectomy (TAH) over supracervical hysterectomy (SCH) for benign gynecologic disease. However, there has been a small but steady rise in the number of supracervical hysterectomies performed both regionally and nationally [3,4]. The reasons for this shift in hysterectomy type are multifactorial and poorly understood, but may reflect changing attitudes on the part of both patients and surgeons favoring less invasive operations. Potential advantages of SCH over TAH include decreased operative morbidity and reduced risk of urinary and sexual dysfunction [5-10]. Patient concerns about sexual functioning after hysterectomy may also play a role in the decision to undergo a supracervical hysterectomy versus total abdominal hysterectomy.
Various mechanisms have been proposed to explain why cervical conservation (SCH) may have a less detrimental effect on sexual function than TAH. Early pioneering work from the 1960's [11] described elevation of both cervix and uterus during the excitement and plateau phase followed by fundal uterine contractions progressively involving the lower uterine segment as orgasm developed. Cervical os dilatation occurred immediately afterwards, implicating a role for the cervix in the female sexual response.
Another theory postulates that the ability to achieve internal orgasm depends on the stimulation of nerve endings of the uterovaginal (cervical) plexus of Frankenhauser [6,12]. This plexus is a matrix of afferent nerve fibers intimately surrounding the cervix. Stimulation of the cervix may contribute to a pleasurable sensation, ultimately experienced as orgasm. Excision of the cervix that occurs during TAH necessarily results in the loss of a major portion of this plexus.
Studies examining sexual function after hysterectomy are contradictory and the controversy surrounding this topic is not new. Several investigators have reported a measurable decrease in sexual response after total abdominal hysterectomy [13-16] while others have found sexual functioning improved after hysterectomy [17,18]. Data regarding sexual function after an aggregation of all types of hysterectomy (TAH, SCH, vaginal and laparoscopy assisted vaginal hysterectomy) are especially difficult to interpret [18]. One intuitive mechanism to account for improved sexual functioning post hysterectomy is the relief from dyspareunia caused by excision of uterine pathology (i.e., fibroids, menorrhagia).
Although there have been many studies exploring the general impact of total hysterectomy on sexual function, few have compared particular types of hysterectomy and their associated effect on subsequent sexual function. Indeed, only one investigator to date has explicitly examined the effect of type of hysterectomy (SCH versus TAH) on sexual function [9,13]. A statistically significant reduction in orgasm frequency among TAH patients when compared to SCH was reported in this population [13]. Ours is the first investigation in over a decade to offer a comparative analysis of sexual response as a function of hysterectomy type (SCH versus TAH).
Methods
A retrospective review was conducted of all patients who had undergone a TAH or SCH at The New York-Presbyterian Hospital Weill Medical College of Cornell University from January 1, 1994 to December 31, 1994. Patients with a diagnosis of cancer (gynecologic or non-gynecologic), psychiatric illness, neurologic disease, diabetes, chronic pelvic pain, endometriosis or a debilitating medical condition were excluded. Patients greater than age sixty and non-English speaking patients were excluded. Institutional review board approval was obtained before data collection.
A total of 333 hysterectomy patients were identified for the study period, and 160 of these did not meet criteria for participation as defined above. An additional 17 charts could not be located. This resulted in 156 patients eligible for inclusion in the study. Demographic data were obtained from the chart review of these patients and were entered into a computer database.
All hysterectomies were performed intrafascially using the clamp-cut-ligate method as described previously [19]. For SCH patients, the corpus was amputated below the level of the internal cervical os. The cervical stump was closed using interrupted No. 0 delayed absorbable suture (Polyglactin 910).
Patients were sent a confidential questionnaire developed by a multidisciplinary team comprised of psychiatrists, statisticians, gynecologists, pathologists and a sexual therapist. It was pretested on a group of hysterectomy patients identified from a 1993 chart review at the same institution. Patient response and feedback from that survey resulted in the questionnaire being modified and condensed to one page. In its final form, the survey included 16 questions evaluating demographic data and measuring the following sexual variables: libido (desire), frequency of intercourse, dyspareunia, orgasm frequency, multiple orgasm frequency and overall sexual satisfaction. All questions were in a multiple-choice format although patients were allowed to add additional comments. In the questionnaire, patients were asked to describe their sexual function before and after their operation in absolute rather than relative terms in order to minimize recall bias (Table 1). Preoperative to postoperative change was then determined for each sexual variable and recorded as "better", "unchanged", or "worse". For example, a patient who preoperatively reported her frequency of intercourse (per month) as 1–3 and postoperatively as 4–6 was identified as "better". Non-respondents were sent a second mailing approximately one month later. After this second questionnaire, no further attempt was made to contact patients. Patients who did not wish to participate in the study could so indicate, and they were not contacted further.
Outcome variables measured by questionnaire
Sexual variable
Response choices
libido (desire)
absent, moderate, strong
frequency of intercourse (per month)
0, 1–3, 4–6, 7–9, >10
dyspareunia (thrusting pain)porgasm frequency (ability to achieve orgasm) multiple orgasm
never, infrequently, frequently, always
overall sexual satisfaction
worse, unchanged, better
Data were analyzed using the JMP statistical package (version 3.2.2 SAS Institute; Carey, North Carolina USA). Chi-square or Student's-t test was used to evaluate categorical or continuous demographic data and preoperative sexual function variables, as appropriate. A multiple logistic regression model was used to analyze independent predictor variables. Percent distributions were calculated for all variables. Statistical significance was defined as P < 0.05.
Results
The patient-partitioning schema for the study population is shown in Figure 1. The proportion of responders was not significantly different between the two surgical subgroups (Table 2). Forty-eight percent (n = 33) of women undergoing a SCH returned the questionnaire, while 39% (n = 27) of those undergoing a TAR chose to participate. While all respondents answered the majority of questions, not all study patients answered every question. Nonetheless, there was no question in the survey answered by fewer than 74% of the respondents. The mean time (+SD) from surgery to questionnaire was 19.2 months (± 3.4 months) in the SCH group and 19.8 months (± 3.9 months) in the TAH group (range = 14–26 months; P > 0.05).
Group characteristics
SCH (n = 33)
TAH (n = 27)
Significance
Age (mean, +/- SD)
45+/-5.0
46+/-4.8
NS*
Race
white
21(64)
16(59)
black
6(18)
6(22)
NS
hispanic
3(9)
5(19)
asian
3(9)
0
Marital status
married
19(58)
19 (70)
single
8(24)
6(23)
NS
divorced
6(18)
2(7)
Length of stay (days)(mean, +/- SD)
4.3+/-1.5
4.7+/-1.8
NS*
Uterine mass (grams)(mean, +/- SD)
665+/-766
400+/-273
NS*
Indications
leiomyoma
26(79)
16(60)
menometrorhagia
3(9)
6(22)
NS
cyst
3(9)
3(11)
other
1 (3)
2 (7)
Incision type
pfannensteil
25 (76)
23(85)
NS
vertical
8(24)
4(15)
Previous abdominal surgery
yes
17 (52)
10(37)
NS
no
16(48)
17(63)
Parity
parous
25 (76)
20(74)
NS
nonparous
8(14)
7(16)
Same partner
yes
27(84)
19 (79)
NS
no
5(16)
5(21)
Oophorectomy
yes
10(30)
17(63)
p < 0.05
no
23(70)
10(37)
Hormone replacement
yes
12(36)
16(59)
NS
no
21 (64)
11 (41)
Data are presented as n (%); NS = not significant; chi square test; NS* students t test
The demographic data obtained from the chart review and from the questionnaire are summarized in Table 2. The data illustrate that the two groups do not differ significantly except for oophorectomy status. Significantly more women underwent bilateral salpingoophorectomy (BSO) in the TAH group than in the SCH group (63 percent versus 30 percent, respectively; P < 0.05). Since removal of the ovaries may have a deleterious effect on sexual function [20], the possible effects of oophorectomy on outcome measures were considered in the analysis. In addition, we considered the possible effects of hormone replacement therapy (HRT) of current users on sexual function, even though the two surgical groups showed no significant difference in reported use of HRT.
Outcome variables in which a pre-surgical subjective rating was elicited were evaluated and compared between the two groups (Table 3). In the TAH group, three women reported the absence of intercourse before surgery (versus no women in the SCH group), while in the SCH group proportionately more women reported intercourse >6 times/month. Although the possible effect of this finding on our ultimate outcomes was unclear, we included it as a possible confounder in our subsequent modeling process to insure that the results were not biased. We also added consideration of prior expectations regarding the likely effect of the surgery on sexual function, in an attempt to control for as many potential confounding variables as possible.
Baseline (preoperative) sexual function & expectations
SCH
TAH
Significance
Libido (desire)
absent
1 (3)
2(8)
moderate
21 (64)
18 (72)
NS
strong
11 (33)
5(20)
Dyspareunia
never
19 (58)
8(35)
infrequently
7(21)
8(35)
NS
frequently
6(18)
5(22)
always
1 (3)
2 (9)
Multiple Orgasm
never
8(26)
4(17)
infrequently
9(29)
9(37)
NS
frequently
10 (32)
10 (42)
always
4(13)
1 (4)
Intercourse frequency (per month)
0
0
3(12)
1–3
11 (33)
4(16)
4–6
8 (24)
13 (52)
p < 0.05
7–9
8(24)
2(8)
>10
6(18)
3(12)
Orgasm frequency
never
1 (3)
0
infrequently
5(16)
4(17)
NS
frequently
18 (58)
12 (52)
always
7(23)
7(30)
Expectations
worse
13 (39)
5(22)
unchanged
18 (55)
15 (65)
NS
better
2(6)
3(13)
Data are presented as n (%); Not all patients answered every question; NS = not significant, chi square test
The results of our initial analysis led to the construction of a multiple stepwise logistic regression model in which type of procedure, oophorectomy, use of HRT, frequency of intercourse prior to surgery, and preoperative expectations were used to predict the various sexual function outcome measures. Since our hypothesis centered on loss of sexual function, each of the outcome measures was categorized according to a binomial classification of function relative to pre-surgical status.
The relevant findings of this modeling procedure are summarized in Table 4. Changes in libido, dyspareunia, and multiple orgasm frequency were not predictable based upon any of our independent variables. However, intercourse frequency, orgasm frequency, and overall sexual satisfaction were all significantly related to type of procedure (P = 0.01, 0.03, and 0.03, respectively). Regarding intercourse frequency, 42% (n = 10) of TAH patients experienced worse outcome compared with 15% (n = 5) of SCH patients. Forty-three percent (n = 9) of TAH patients experienced a decrease in the ability to achieve orgasm compared with 6% (n = 2) of SCH patients. Overall sexual satisfaction was worse for 33% (n = 8) of TAH patients compared with 6% (n = 2) of SCH patients. Only for overall sexual satisfaction did oophorectomy add significant predictive value (P = 0.02). Thirty-five percent (n = 9) of patients who had a BSO with their hysterectomy had worse overall sexual satisfaction when compared to 3% (n = 1) of patients who had hysterectomy only without BSO. With respect to HRT, there was no significant difference between sexual function of patients who were receiving hormone replacement and those who were not. Frequency of intercourse prior to surgery was not associated with postoperative change in any of the sexual variables analyzed (P < 0.05).
Sexual function results
Hysterectomy
BSO
HRT
SCH (N = 33)
TAH(N = 27)
OR95% CI
+ (N = 27)
- (N = 33)
+ (N = 28)
- (N = 32)
Libido
same/better
27(87)
17(68)
3.2
16(64)
28(90)
19(70)
26(87)
worse
4(13)
8(32)
(0.8–12.2)
9(36)
3(10)
8(30)
4(13)
Dyspareunia
same/better
30(91)
18(90)
1.1
22(92)
27(90)
24(92)
25(89)
worse
3(9)
2(10)
(0.2–7.3)
2(8)
3(10)
2(8)
3(11)
Multiple orgasm
same/better
29(91)
16(73)
7.8
16(67)
29(97)
19(73)
26(93)
worse
3(9)
6(27)
(1.5–40.9)
8(33)
1(3)
7(27)
2(7)
Intercourse frequency
same/better
28(85)
14(58)*
3.6
19(73)
23(74)
20(74)
22(73)
worse
5(15)
10(42)
(0.8–16.5)
7(27)
8(26)
7(26)
8(27)
Orgasm frequency
same/better
31(94)
12(57)**
11.6
14(61)
29(94)
17(68)
26(90)
worse
2(6)
9(43)
(2.2–61.8)
9(39)
2(6)
8(32)
3(10)
Overall sexual satisfication
same/better
31(94)
16(67)***
7.8
17(65)
30(97)†
20(74)
27(90)
worse
2(6)
8(33)
(1.5–40.9)
9(35)
1(3)
7(26)
3(10)
Data are presented as n(%) except OR (95%CI); Not all patients answered every question; *p = 0.01, **p = 0.03, ***p = 0.03, †p = 0.02, all others are not significant
Discussion
The results presented here support a role for the cervix in the female sexual response. In particular, our finding of a statistically significant decrease in the ability to achieve orgasm among TAH patients when compared to SCH patients (43% vs. 6%, P = 0.03) is noteworthy. Our results support and expand upon the work by Kilkku [13] who reported a greater reduction in orgasm following TAH when compared to SCH (P < 0.05).
If patients derive less satisfaction from sexual intercourse by experiencing a reduction in their ability to achieve orgasm, then the frequency of intercourse might decrease. Indeed, in this study population both frequency of intercourse and overall sexual satisfaction were worse among TAH patients when compared to SCH patients. Previous investigators [14] found that 37% of patients after TAH/BSO had a reduction in frequency of intercourse, similar to the 42% in our study group. The decline in overall sexual response after hysterectomy in our population was also consistent with results published by others [15,16].
Dyspareunia was the only sexual variable where the results after surgery revealed no trend favoring SCH. Nine percent (n = 3) of SCH patients and 10% (n = 2) of TAH patients experienced worsening of dyspareunia after the operation. Sixty six percent of TAH patients and 42% of SCH patients experienced some degree of dyspareunia prior to surgery. Perhaps the dyspareunia was caused by intrauterine pathology, such as leiomyoma. Since uterine leiomyoma was the main indication for hysterectomy in 79% of SCH patients and 60% of TAH patients, hysterectomy could have resulted in a similar reduction of pain in the two groups. That hysterectomy (SCH or TAH) is associated with improvement in dyspareunia supports the findings of others [9,18,21,22].
Although the two surgical groups in this study are very similar, they do differ with respect to intercourse frequency and BSO status (Table 2 and 3). Because the independent variable in the statistical model was change in frequency of intercourse (postoperative frequency minus preoperative frequency) and not the magnitude of intercourse frequency, this disparity in the preoperative characteristics does not adversely impact our findings. Indeed, realizing that a sexual frequency of zero cannot become worse, this may have in fact biased the results in favor of the TAH cohort.
As demonstrated in the literature, the analysis of the relationship of BSO and postoperative sexual function is complex. Although one investigator reported no difference in libido between patients who underwent TAH and TAH/BSO [15], our data are consistent with the findings of a more recent investigation [20] where BSO status did influence sexual outcome in the study population.
In accord with previous investigations [14,15], we determined that HRT did not affect sexual outcome in any of the variables tested although the majority (86%) of patients who were on HRT had undergone BSO. Perhaps this is because standard HRT may be insufficient to replace the full hormonal milieu produced by the ovary for some women, and other endogenous hormones may play a part in the female sexual response. For example, testosterone production continues even in the postmenopausal ovary, albeit at a markedly reduced rate [23]. A prospective, randomized study of TAH/BSO patients showed that those receiving combined testosterone and estrogen therapy had greater sexual desire than women receiving estrogen alone [24].
Menopausal status could have been directly determined from our questionnaire although it was found that the majority (71%) of patients greater than age 45 in the study population underwent surgical menopause (BSO) and were taking HRT at the time of questionnaire completion.
Conclusion
In summary, we found that patients who underwent a TAH experienced worse sexual outcome than SCH patients with respect to frequency of intercourse, frequency of orgasm and overall sexual satisfaction. Our results suggest that type of hysterectomy does predict overall postoperative sexual satisfaction. Adding in the consideration of whether a BSO was performed significantly enhances this prediction. Prospective studies are needed to confirm these findings and determine the clinical impact that cervical and ovarian conservation may have for the patient.
Competing interests
None declared.
Pre-publication history
The pre-publication history for this paper can be accessed here:
This study was designed to evaluate the effects of a 24-month period of moderate exercise on serum lipids in menopausal women.
Methods
The subjects (40–60 y) were randomly divided into an exercise group (n = 14) and a control group (n = 13). The women in the exercise group were asked to participate in a 90-minute physical education class once a week and to record their daily steps as measured by a pedometer for 24 months.
Results
Mean of daily steps was significantly higher in the exercise group from about 6,800 to over 8,500 steps (P < 0.01). In the control group, the number of daily steps ranged from 5,700 to 6,800 steps throughout the follow-up period. A significant interaction between the exercise group and the control group in the changes og total cholesterol (TC), high-density lipoprotein cholesterol (HDLC) and TC : HDLC ratio could be observed (P < 0.05). By multiple regression analysis, the number of daily steps was related to HDLC and TC : HDLC levels after 24 months, and the changes in TC and HDLC concentrations.
Conclusions
These results suggest that daily exercise as well as increasing the number of daily steps can improve the profile of serum lipids.
Background
It is known that lipid metabolism rapidly deteriorates in women when they reach menopause. In addition, it has been reported that the morbidity due to hyperlipidemia and coronary heart disease rises in women of menopausal age. This is mostly based on the decrease in estrogen, which has the action of controlling LDL production, advancing of HDL production and antioxidation [1,2]. In Japan, Western dietary habits, especially increased fat intake and decreased carbohydrate intake, are becoming one of the causes of the deterioration of the serum lipids [3]. This fact suggests that preventing the deterioration of serum lipids during menopause is very important.
Exercise is one of the methods to prevent the deterioration of serum lipids. It has been clarified that exercise can bring the serum lipids to an acceptable range [4-8]. However, with respect to the sex difference, the effect of exercise is more difficult to be detected in women compared with men [6,9,10]. Also, Motoyama et al.[6] reported that in elderly persons long-term low intensity aerobic training improved the profile of serum lipid and lipoprotein concentrations, while detraining returned the profile to that of the pre-training levels. In elderly women, it has also been demonstrated that the serum lipid levels of a person who continues exercise arc clearly different from those of a person who does not [11]. Hence, it is essential for a woman in her menopausal age to continue a long-term exercise schedule to improve her serum lipid profile.
Dose-response improvement in performance capacity after exercise training has been the subject of a number of studies [12,13]. On the basis of such scientific reports, conclusions were taken as the positive effects of endurance of high-intensity exercise training on increasing maximum oxygen consumption. For instance, according to the guidelines of the American College of Sports Medicine [14], the conditions for effective exercise are: 20–60 minuets of moderate- to high-intensity endurance exercise (50–85% of maximal oxygen uptake (VO2max) performed three or more times per week. However, it has been reported recently that intermittent bouts of physical activity, as short as 8 to 10 min, totaling 30 min or more on most days provide beneficial health and fitness effects [15]. In addition, it is recommended that the intermittent activity is not necessarily exercising training [16,17]. For example, walking up the stairs instead of taking the elevator, waking instead of driving short distance, gardening, housework, and so on. Therefore, it can be considered that the amount of activity is more important than the specific mode of activity performed. It is reported that the current low-participation rate of general population in sports activities may be due, in part, to the misperception of many people that to reap health benefits they must engage in vigorous, continuous exercise [15]. This view is of practical importance and is consistent with health recommendations for menopausal women, for whom the issues of moderate and short bouts of physical activity are of greater interest.
In the present study, in order to ensure that the subjects would continue their exercise program, only the increase in the number of their daily steps was requested for submission, and the subjects were asked but not obliged, to participate in the physical education class once a week. This 24-month follow-up study examined whether low intensity physical activity that seems to produce endurance gains is similarly effective in improving the blood lipid in menopausal women.
MethodsSubjects
A written explanation of the experimental procedures (physical education class) and potential risks was presented to each subject and informed consent was obtained.
The subjects were forty-eight menopausal women (age: 40–60 years) who participated in this study. None of them had engaged in any regular exercise program prior to this study. Then they were randomly divided into two groups, one consisting of 32 persons who consented to participate in the physical education class (exercise group), and the other consisting of 16 persons who did not (control group). As mentioned before, the subjects were not obliged to participate in the physical education class, hence we anticipated that some of them might discontinue attending the program over the 24-month follow-up investigation. On the basis of this assumption, the number of subjects in the exercise group was two times of that in the control group. The 32 subjects in the exercise group fell to 20 persons over the 24-month period, and 14 of them were selected to comprise the final exercise group based on the following criteria: (a) ages of < 55 years (48.6 ± 4.2 years), (b) no special diseases and medication, and (c) complete information available for each individual. Among the 16 subjects in the control group, 13 of them were selected to serve as the final control group based on the same criteria (48.0 ± 3.6 years). The subjects in both groups were asked to provide samples for blood examinations for 24-month. The profiles of the subject are listed in Table 1.
Characteristics of the subjects in the exercise group and the control group
Exercise (N = 14)
Control (N = 13)
Start
24 M
Start
24 M
Age (year)a
48.6 ± 4.2
48.0 ± 3.6
Weight (kg)a
54.1 ± 4.3
53.2 ± 3.9
55.2 ± 5.3
55.6 ± 4.2
BM1a
22.3 ± 1.6
22.0 ± 1.8
22.6 ± 1.9
22.7 ± 1.4
Menopausal statusb
Postmenopause
5
9
4
6
Menopause
6
4
4
5
Premenopause
3
1
5
2
aValues are means ± SD. bValues are number of the subjects. BMI: body mass index, Start: the start of this study, 24 M: after 24 months
The conditions of menstruation were defined as follows. Postmenopausal women: those who have no menstruation over half a year, and whose balance of follicle-stimulating hormone and estradiol 2 shows the pattern characteristic of postmenopause. Menopausal women: those whose menstrual cycle is irregular, yet have menstruation within the half year, and whose pattern of hormone balance is not characteristic of postmenopause. Premenopausal women: those whose menstrual cycle is regular.
An outline of the physical education class for menopausal women
A 90-minute physical education class was carried out once a week based on the program shown in Figure 1. Walking, jogging or rhythm dance was conducted at 40–60% of VO2max for each subject, as calculated by Karvonen's formula [18]. Participation in the physical education class was not forced. On the average, each subject participated in about 80% of the exercises throughout the study over 24-month period.
Basic exercise training program in the physical education class for menopausal women.
In the beginning, all subjects were each given a pedometer and asked to record the number of their daily steps for a week, under the conditions of not having any other kind of exercise and continuing their ordinary lifestyles. After the week was up, they brought their records of their calculated average daily steps. Each participant of the exercise group was asked to increase the number of her steps per day at least 2,000 to 3,000 more than her calculated average daily steps.
Method of evaluation
Evaluations were carried out in the participation at the beginning and after 6 months, 15 months, and 24 months. Body mass index (BMI) was calculated from height and weight. The blood examination was carried out during 7–8 p.m., before eating. As for premenopausal women, the schedule of blood collection was adjusted in order to complete blood collection in the luteal phase. The serum was recovered following centrifugation at 3,000 rpm for 15 minutes and then stored at 4°C. All analyses were completed within 48 hours. Samples were analyzed for estradiol (E2), follicle stimulating hormone (FSH), total cholesterol (TC), high-density lipoprotein cholesterol (HDLC),triglyceride (TG) and lipid peroxidation (LPO). E2 and FSH were assayed by radioimmunoassay. The TC in the serum was assayed by the cholesterol oxidase method (Cholesterol E-test Wako, Wako Co., Osaka, Japan). The HDLC in the serum was assayed by the phosphotungstic acid-magnesium chloride precipitation method (HDL cholesterol E-test Wako, Wako Co., Osaka, Japan). The TG in the serum was assayed by glycerol-3-phosphate oxidase method (Triglyceride E-test Wako, Wako Co., Osaka, Japan). The LPO in the serum was measured using the assay system described by Ohkawa et al.[19]. The TC: HDLC ratio was adopted according to the atherogenetic index.
Analysis method
Changes in the number of daily steps and the blood analysis data over the four evaluations were detected by 2 × 4 repeated measures of ANOVA. Difference in mean values at the same time between the exercise group and the control group were compared. Multiple regression analysis was conducted for all exercise subjects between: (a) change of TC, HDLC, TC: HDLC and LPO for 24 months and age, BMI and average number of daily steps for 24 months; and (b) value of TC, HDLC, TC: HDLC and LPO at each evaluation time and age, BMI, average number of daily steps for each of the intervals between the evaluation times and the initial value. This was done by increasing and decreasing variables by AIC. Comparisons between the two groups at certain time points were made using t-test with a Bonferroni correction for multiple measures. The level of significance in each analysis was P < 0.05.
Results
The average number of daily steps at entry (for a week) was 6,740 ± 1,326 steps in the exercise group and 7,149 ± 1,202 steps in the control group (P > 0.05). Figure 2 shows the average number of daily steps taken at 1,6,15 and 24 months in the both groups. The average daily steps of the exercise group at mentioned points were between 8,500 and 11,000 steps. Compared to the first week, the exercise group's average daily steps at 1,6,15 and 24 months after the moderate exercise started was significantly higher (P < 0.01) than the first week. In the control group, the number of daily steps ranged from about 5,700 to 6,500 steps. Compared with the first-week value, no significant difference at the taken intervals could be detected (F = 1.9, P > 0.05). The average daily steps of the exercise group (at 1,6,15 and 24 months after the moderate exercise started) were significantly (P < 0.01) higher than those of the control group.
Changes in average of number of daily steps in the exercise and the control groups at 1,6,15 and 24 months after the moderate exercise started. †: A significant difference at P < 0.01 was detected between the average daily steps during a week preceding the physical education class and that of the selected months during the follow-up period. **: Significantly different from the control group. **P < 0.01.
Table 2 shows changes of TC, HDLC and LPO concentrations and TC : HDLC ratio in both groups for 24 months. In the control group, the levels of TC, HDLC, TG, LDLC, TC : HDLC ratio and LPO did not change significantly for 24 months. By applying 2 × 4 repeated measures of ANOVA, a significant interaction between the exercise group and the control group in the changes of TC (F = 3.92, P < 0.05), HDLC (F = 4.08, P < 0.05) and TC : HDLC ratio (F = 4.27, P < 0.05) could be observed. HDLC at 24 months in the exercise group was significantly higher (P < 0.05) than that in the control group. In the exercise group, TC : HDLC ratio at 24 months in the exercise group was significantly lower (P < 0.05) than those in the control group.
Change in the value over 24 months in the exercise group and the control group
Start
6 M
15 M
24 M
mean
(SD)
mean
(SD)
mean
(SD)
mean
(SD)
Exercise group (N = 14)
Total cholesterol (mg/dl)
218.2
(52.2)
200.6
(31.8)
197.3
(38.2)
193.3
(24.8)*
HDL cholesterol (mg/dl)
59.0
(13.5)
62.7
(13.0)
64.9
(17.9)
68.4
(12.9)*,†
Triglyceride (mg/dl)
133.8
(51.4)
100.9
(35.6)
89.0
(40.5)
87.3
(41.8)
LDL cholesterol (mg/dl)
109.7
(29.7)
104.1
(29.2)
94.1
(34.1)
90.8
(27.3)
TC:HDLC ratio
3.91
(1.46)
3.33
(0.81)
3.35
(1.35)
3.00
(1.03)*,†
Lipid peroxidation (nmol/ml)
3.36
(1.10)
3.21
(0.96)
2.89
(0.79)
2.70
(0.78)
Control group (N = 13)
Total cholesterol (mg/dl)
210.0
(24.2)
194.9
(24.5)
220.5
(32.2)
224.4
(32.3)
HDL cholesterol (mg/dl)
56.7
(16.6)
56.6
(13.5)
57.3
(13.5)
56.7
(14.6)
Triglyceride (mg/dl)
140.9
(40.0)
121.8
(39.8)
140.7
(36.4)
138.2
(36.2)
LDL cholesterol (mg/dl)
119.2
(28.4)
108.9
(26.1)
118.6
(31.4)
124.2
(36.1)
TC:HDLC ratio
4.11
(1.31)
3.52
(1.15)
3.95
(0.98)
4.16
(1.13)
Lipid peroxidation (nmol/ml)
4.15
(1.02)
3.24
(0.88)
3.30
(1.07)
3.58
(0.79)
The significant differences were detected using 2 × 4 repeated measures of ANOVA:*P < 0.05, †P < 0.05 compared to the control group by t-test with a Bonferroni correction.
The results of multiple regression analysis of the values of TC, HDLC and LPO and TC: HDLC ratio at the start and 24 M evaluations are shown in Table 3. There was no effect of the considered factors on each value at the start of the study. The quantity of number of daily steps affected negatively the value of TC (F = 13.4, P < 0.01) and the TC: HDLC ratio (F = 12.8, P < 0.01) and positively the value of HDLC (F = 8.7, P < 0.05) at 24 M evaluations. The postmenopausal group was negatively related to the HDLC value (F = 6.9, P < 0.05). Age and postmenopause were positively related to the TC: HDLC ratio (F = 22.9, P < 0.01).
Multiple regression analysis of the value of TC, HDLC, TC:HDLC ratio and LPO at the start (Start) and after 24 months (24 M) in the exercise group
TC
HDLC
TC:HDLC
LPO
Start
24 M
Start
24 M
Start
24 M
Start
24 M
Explanatory variable
Age
0.004(0.000)
0.883(3.182)
-0.619 (1.450)
-1.138 4.971)
0.759 (2.141)
1.502** (15.78)
0.876 (0.317)
-0.299 (0.148)
BMI
0.086 (0.088)
0.327 (0.278)
-0.607 (8.543)
-0.234 (1.076)
0.574 (5.322)
0.334 (3.965)
0.166 (0.794)
0.296 (0.735)
Number of daily steps
-0.496 (2.615)
-0.919** (34.74)
0.182 (0.531)
0.848* (8.720)
-0.601 (5.178)
-1.019** (12.86)
0.470 (0.271)
0.595 (1.884)
Postmenopausal Yes(1)/No(0)
-0.347 (0.516)
0.333 (1.041)
-0.542 (1.730)
-0.984* (6.913)
0.402 (0.939)
1.002**(22.92)
0.123 (1.139)
-0.627 (1.207)
Premenopausal Yes(1)/No(0)
0.035 (0.003)
0.859 (3.160)
-0.340 (0.351)
-0.014 (3.364)
-0.674 (1.248)
-1.470**(13.57)
0.574 (0.320)
-0.330 (0.153)
Proportion
0.457
0.781*
0.597
0.714*
0.597
0.843**
0.416
0.334
F value
1.349
5.715
2.376
3.992
2.376
8.576
1.142
0.804
Statistically significant: *P < 0.05, **P < 0.01, Values are standard partial regression coefficient (F value). BMI: Body mass index, Start: at the start of this study, 24 M: after 24 months
Table 4 presents the result of the multiple regression analysis of change in TC, HDLC, TC : HDLC ratio and LPO over 24 months. The values were mostly influenced by the amount of TC (F = 89.8, P < 0.01), the TC: HDLC ratio (F = 23.2, P < 0.01) and the LPO (F = 12.0, P < 0.05) by each initial value. The number of daily steps affected the change in the value of TC (F = 6.8, P < 0.05) and HDLC (F = 7.2, P < 0.05).
Multiple regression analysis of change volume of TC, HDLC, TC:HDLC ratio and LPO over 24 months in the exercise group
TC
HDLC
TC : HDLC
LPO
Age
-0.095 (0.901)
-0.156 (0.402)
0.055(1.270)
-0.207 (0.056)
BMI
-0.125(1.231)
-0.433 (1.954)
0.190 (0.126)
-0.062 (0.559)
Number of daily steps
0.346 (6.807)*
-0.702 (7.169)*
0.283 (0.469)
-0.139(1.006)
Initial value
1.506(89.83)**
0.239 (0.610)
0.999 (23.27)**
0.898 (12.04)**
Proportion
0.919**
0.572
0.618*
0.896**
F value
25.710
3.005
3.643
9.136
statistically significant: *P < 0.05, **P < 0.01, Values are standard partial regression coefficient (F value), BMI: Body Mass Index, : the value at the start subtracted from the vale at 24 months.
Discussion
It has been reported that TC, LDLC and TG concentrations increase from 30–35-year-olds to 50–60-year-olds, and the serum lipid profile becomes almost constant after then in women [20,21]. The rises have been reported to be 20–30 mg/dl in TC, 30–40 mg/dl in LDLC and 35–40 mg/dl in TG [20]. Especially, TC and LDLC concentrations remarkably increase between 4 years before and 1 year after menopause [22], and there is a large difference between premenopausal and postmenopausal women [23-25]. HDLC concentration decreases a little with aging from age 20 [20,26,27], and the characteristic change around menopause is not observed. In this study, all subjects were around menopausal age; in fact, six subjects changed to postmenopausal from menopausal status and five subjects changed to menopausal from premenopausal status during the course of the study: Taking the age and the menstruation status into consideration, the subjects were put in the stage in which serum lipids were most easy to deteriorate. It is a primary goal to stop the deterioration of the lipids in women in this particular stage. Our results showed that HDLC and TC : HDLC ratio were significantly improved in the exercise group. HDLC after 24 months in the exercise group was higher compared with the control group. Regarding TC: HDLC ratio after 24 months, the obtained values in the exercise group were significantly lower than that in the control group. It is evident from these that exercise stops the deterioration of the lipid profile in women around menopause. There are some reports with results similar to ours [5,24,28,29]. It is possible to assume that a woman close to menopausal age would not obstruct the improvement of serum lipids through exercise. Moreover, the results obtained showed that the most influential factor was each initial value marking the improvement of TC: HDLC ratio and LPO; there are similar reports by others [6,10,30-32]. Consequently, we arrive at the conclusion that exercise is very useful for women around menopause not only to prevent the value from deteriorating but also to return the levels that have already deteriorated.
Consider now the factors related to the improvement of the serum lipids. The postmenopausal state affected the values of HDLC and TC: HDLC ratio. This fact is known as the effect of estrogen decrease. It was shown that the values of TC, HDLC and TC: HDLC ratio were related to the number of daily steps, which in turn affected the changes of TC and HDLC for 24 months. These results demonstrate that increasing of the number of daily steps work improves or at least maintains a good levels of the serum lipids. Irie et al.[33] reported that an increased number of steps was significantly correlated with increased HDLC, and this fact supports our view. In our study, the average of the number of steps per day for the first week was about 6,500. This is almost the same as the average of the number of daily steps in Japanese homemakers [34], and it reflects the lifestyle of those who do not move very much. By encouraging the increase in the number of daily steps taken, our subjects increased them to about 9,000. Duncan et al. [35] found that HDLC concentration and TC : HDLC ratio improved not in their aerobic group or brisk group, but in their strolling group. Reaven et al.[36] also noted that HDLC concentrations were the highest in the group in which the physical activity level was light. It is reasonable to suppose that the improvement can be expected even in exercise of low intensity. In the view of Blair et al.[37], the response to exercise depends on neither the intensity nor the length of the training period, duration of exercise session, or the frequency, etc. When considering the exercise conditions applied until now by research on women, the frequency was mostly 3–5 times/week and the duration of an exercise session was 20–60 minutes. There is a difference between longitudinal research and cross-sectional research in the training period; it is between 10 weeks and 12 months and between 3 years and 5 years, respectively. The effect of exercise on the lipids is different among various studies conducted in longitudinal research. On the other hand, the value of lipids in trained women is obviously better than in untrained women in cross-sectional research. This fact indicates that the lipids are surely improved owing to long-term exercise. In our study, it seems that suitable results were obtained because the period of exercise was long, though the intensity of exercise was low.
Although the physiological mechanisms that involved in the decrease of TC and the increase of HDLC due to low intensity physical activity were not directly examined in the present study, several possible mechanisms have been proposed to explain them. Some investigators have demonstrated that aerobic exercise increases both plasma lipoprotein lipase and lecithin cholesterol acyltransferase activities [38,39]. It would be expected that the same mechanisms might also been presented in the menopausal women. To clarify the related mechanisms, more investigations need to be carried out.
Turn now to the role of the antioxidation that estrogen and LPO had. LPO oxidized the low-density lipoproteins (LDL), and the oxidation of LDL is incorporated by macrophages, becomes a foam cell, and then it becomes deposited in the arterial vessel wall. It is said that it worsens arteriosclerosis, and it is considered in the meantime that estrogen suppresses the LPO. Since estrogen is failing around menopause, it appears that LPO increases. It is necessary to control the increase of LPO and LDL, and toward this end, exercise is expected to play an important role. There is a report on the effect of exercise on the oxidation of LDL, whose production could be suppressed by 10-month exercise [40]. There is no other survey examining on the effect of long-term exercise. In many surveys, it was found that acute exercise accelerated the generation of LPO [41,42]. In this study, LPO concentration decreased, though not significantly, and the factor that affected the alteration of LPO could not be clarified by multiple regression analysis. In the future, it will be necessary to sufficiently examine the effect of long-term exercise on the generation of LPO and oxygen action including the effects of menstruation.
In follow-up studies involving physical education class, the number of persons who participate throughout the entire term of investigation is an important factor. Participation rate of about 60% has been reported [43]. In the current study, moderate exercise regimen had a practice rate 62.5% for 24 months. Thus, we can speculate that moderate exercise, as applied in our study, may improve the practice rate.
We have undertaken another study a large number of controls and exercise group. Our goal is to overcome a number of weakness points we have been facing in the present investigation including: small subjects' numbers and lack of dietary control. In addition, it would have been desirable to perform analyses by dividing the subjects into the same status of menopause. However, in the present study the number of subjects was not enough to do so. The results of upcoming research would further our understanding of the effects of moderate exercise and serum lipids status in women.
Conclusion
On the basis of the materials discussed here, it was concluded that daily exercise as well as increasing the number of daily steps as a moderate exercise can improve the profile of serum lipids in menopausal women. However, it is difficult to clarify the relationship between blood lipids and the low intensity physical activity in humans. More work needs to be done to understand blood lipid profiles associated with exercise.
Competing interests
None declared
Pre-publication history
The pre-publication history for this paper can be accessed here:
Acknowledgments
The authors would like to thank the Takatomi town (Gifu, Japan) office for cooperating with our study and Dr. Toshiya Itoh from Gihoku hospital for the technical support in this study. This study was supported in part by a Grant-in-Aid for Scientific Research from the Ministry of Education, Science, Sports and Culture of Japan (Grant no. 08780091 and 07772270).
Women's health, traditionally defined, emphasises reproductive and maternal conditions without consideration of social contexts. Advocates urge a broader conceptualisation. The medical literature influences the definitions and delivery of women's health care. We compared how women's health was represented in leading general medical (GM) versus women's health specialty (WS) journals.
Methods
Original investigations published between January 1 – June 30, 1999 in leading GM (n = 514) and WS (n = 82) journals were compared. Data were collected from 99 GM and 82 WS articles on women's health. Independent reviewers conducted content analyses of sample characteristics, study design, and health topic. Each article was classified as "Traditional" (e.g. menstruation, breast cancer), "Non-traditional" (e.g. abuse, osteoporosis), or "Both."
Results
Of the GM articles, 53 (53.5%) focused solely on a traditional women's health topic; half were reproductive and half female cancers. In contrast, 22 (26.8%) WS articles were traditionally focused. A non-traditional topic was the sole focus of 27 (27.3%) GM articles versus 34 (41.5%) WS articles. One-fifth of GM and one-third of WS articles addressed both. RCTs dominated the GM articles, while 40% of WS articles used qualitative or mixed study designs. Leading sources of women's death and disability were not well covered in either type of journal.
Conclusions
Most GM articles drew on a narrow definition of women's health. WS journals provided more balanced coverage, addressing social concerns in addition to "navel-to-knees" women's health. Since GM journals have wide impact, editorial decisions and peer review processes should promote a broader conceptualisation of women's health.
Background
Women represent over half the population and use more than 50% of health care resources [1,2]. Peer-reviewed journals are health care practitioners' major sources of information about women's health. The leading general medical journals are particularly important because they are prestigious, widely read across clinical, research, and policy disciplines, and elicit news coverage [3,4]. Thus, their content is instrumental in defining "health."
Leading causes of women's death and disability include heart disease, lung and breast cancers, depression, and abuse [1,5]. Women's health advocates argue that reproductive and maternal conditions have been over-emphasised in clinical research at the expense of the broad range of health issues contributing to women's disease burden [5]. Indeed, non-traditional women's health concerns have significant impacts on population health and health care costs. Optimal health care for all is said to draw upon a broad definition of women's health that takes into account social, economic, and political contexts [6-8]. This is accomplished by both quantitative and qualitative research methods [9].
Information published in the general medical literature defines the parameters of women's health and fosters its relative importance as a topic of clinical and scholarly concern. Women's health journals, similarly to clinical specialty journals, devote focused attention to a particular area, but have less impact on conventional criteria (Table 1). Women's health journals provide a reflection of the state of the field and their portrayal of women's health may be instructive to general medical journals.
Quality indicators of leading general medical and women's health specialty journals
Journal (Location)
Impact Factor *
Citation rate Immediacy Index *
Current Circulation †
Inclusion in MEDLINE
General medical
Annals (U.S)
10.097
1.959
91,097
Yes
BMJ (U.K.)
5.143
1.992
117,000
Yes
JAMA (U.S.)
11.435
3.728
3,705,000
Yes
Lancet (U.K.)
10.197
2.634
45,000
Yes
NEJM (U.S.)
28.857
6.445
183,000
Yes
Women's health speciality
HCWI (U.K.)
NR
NR
612
Yes
JWH (U.S.)
1.038
0.128
5,000
Yes
W&H (U.S.)
0.974
0.186
1,225
Yes
WHI (U.S.)
0.404
0.031
3,000
Yes
* Institute for Scientific Information or Social Science Citation Index Journal Citation Reports, 1999. Impact factor is a measure of the frequency with which the "average article" in a journal has been cited in a particular year, calculated by dividing the number of current citations to articles published in the two pervious years by the total number of articles published in the two previous years. Immediacy Index is a measure of how quickly the "average article" in a journal is cited, calculated by dividing the number of citations to articles published in a given year by the number of articles published in that year. Higher values indicate higher impact. † Current circulation obtained from Ulrich's International Periodicals Directory, 39th Edition, 2001 and/or confirmed by communication with publisher. NR = not ranked Annals = Annals of Internal Medicine, BMJ = British Medical Journal, JAMA = Journal of the American Medical Association, NEJM = New England Journal of Medicine HCWI = Health Care for Women International, JWH = Journal of Women's Health & Gender-based Medicine, W&H = Women & Health, WHI = Women's Health Issues
This study compared the representation of women's health in general medical journals with its portrayal in women's health specialty journals.
MethodsSample selection
Original investigations published in the leading general medical (GM) journals Annals of Internal Medicine, BMJ, JAMA, Lancet, and New England Journal of Medicine (NEJM) published between January 1 and June 30, 1999 were compared to original investigations published in leading women's health specialty (WS) journals Health Care for Women International, Journal of Women's Health & Gender-based Medicine (JWH), Women & Health, and Women's Health Issues for the same time period (Table 2).
Sections containing original investigations in leading general medical and women's health specialty journals
Journal
Section
General medical
Annals
Articles
Academia and Clinic
BMJ
Papers
General Practice
JAMA
Original Contributions
Caring for the Critically Ill
Preliminary Communication
Clinical Investigations
Lancet
Articles
Early Reports
NEJM
Articles
Special Articles
Women's health specialty
HCWI
Articles
JWH
Original Articles
W&H
Articles
WHI
Articles
Only original investigations studying adult human populations were included, providing an initial sample of 514 GM and 82 WS articles.
Identification of women's health articles
All WS articles were included and three strategies were used to identify GM articles with a women's health focus. First, we (JPC, PAR) identified all 88 articles that studied women-only samples. Second, three independent reviewers (JPC, PAR, PdN) read all 514 titles and conducted searches for keywords: "woman," "women," "female," "sex," or "gender," those related to female-specific conditions (e.g., breast cancer, fertility, estrogen, etc.), and keywords related to social determinants of health generally (e.g., education, social, equity, poverty, etc.). Results of the keyword searches were compared and consensus reached. Eighty-nine articles were identified, 11 of which were supplementary to the first method of identification. Third, a MEDLINE search was conducted for the following medical subject headings: "women's health," "women's health services," and "women." Limiting the MEDLINE search to the journals in our sample and the time period of study, 16 articles resulted, only 4 of which were original investigations; none of these were supplementary. Using these three search strategies, 99 GM articles with a women's health focus were identified. A more detailed review (JPC) confirmed all GM and WS articles contained women's health content and met the original inclusion criteria.
Content analysis of articles
We developed a semi-structured content analysis instrument to assess sample characteristics, study design, and health topic evaluated in each article. Study design was identified and classified according to a standard taxonomy [10,11] modified to include qualitative research (i.e., randomised controlled trial (RCT), cohort, case-control, cross-sectional, case report, or qualitative) and health topic was identified and examined using an approach consistent with other investigators [12-14]. All health topics were then classified according to definition of women's health used [9]: Traditional, non-traditional, or both. Traditional topics involved reproductive conditions. i.e., fertility, pregnancy, childbirth, breastfeeding, menstruation, menopause, hormone replacement therapy, and female-specific cancers such as breast, cervical, endometrial, and ovarian. Non-traditional topics were all other health conditions, including those which afflict women to a greater extent than men such as abuse, osteoporosis, and eating disorders. Articles classified as drawing on "both" definitions evaluated, as an example, depression in pregnancy.
The instrument was pilot tested on 10 articles. Two independent coders (JPC, GDF) reviewed articles. Interrater reliability ranged from 91% to 100% agreement per item. Coders had 91% agreement on health topic classification. Discrepancies were resolved by consensus.
Results
Among the 514 GM articles in our sample, 99 (19.2%) were related to women's health. Most of these 99 appeared in Lancet (26.3%), BMJ (23.2%), and NEJM (22.2%). Eighty-eight (88.9%) of the GM women's health articles reported on women-only study samples; the remaining 11 (11.1%) were of mixed-gender. Most of the 82 WS articles appeared in JWH (37.8%) and Women & Health (29.3%). Similarly to the GM articles, 87.8% of the WS articles reported on women-only study samples.
Representation of women's health issues (Table 3)
Comparison between women's health original investigations in leading general medical versus women's health specialty journals
GENERAL MEDICAL JOURNALS
WOMEN'S HEALTH SPECIALTY JOURNALS
Characteristic
N
Characteristic
N
Journal
Journal
Annals
10 (10.1%)
HCWI
20 (24.4%)
BMJ
23 (23.2%)
JWH
31 (37.8%)
JAMA
18 (18.2%)
W&H
24 (29.3%)
Lancet
26 (26.3%)
WHI
7 (8.5%)
NEJM
22 (22.2%)
Total
99
Total
82
Representation of Women's Health
Representation of Women's Health
Traditional
53 (53.5%)
Traditional
22 (26.8%)
Reproduction
26
Reproduction
15
Female cancer
27
Female cancer
7
Non-traditional
27 (27.3%)
Non-traditional
34 (41.5%)
Heart disease
6
Obesity/physical activity
7
Health care
3
General health
5
HIV/AIDS/STDs
3
Musculoskeletal
4
Musculoskeletal
3
Depression/mental health
3
Obesity/physical activity
3
Heart disease
3
Other
9
HIV/AIDS/STDs
3
Other
9
Incorporating Both Traditional and Non-traditional
19 (19.2%)
Incorporating Both Traditional and Non-traditional
26 (31.7%)
Reproduction & other
16
Reproduction & other
24
Female cancer & other
3
Female cancer & other
2
Total
99
Total
82
To compare how women's health issues were represented, we categorised articles into three mutually exclusive groups: Traditional, non-traditional, or both. The distribution of topics was significantly different between GM and WS journals (X2 = 13; p = 0.0013). Fifty-three (53.5%) of the 99 GM articles addressed solely a traditional women's health topic. Of these, 26 (49.1%) related to reproductive health issues and 27 (50.9%) studied female-specific cancers. In contrast, 22 (26.8%) WS articles focused solely on traditional women's health. Of these, 15 (68.2%) related to reproduction and 7 (31.8%) to female cancers.
Twenty-seven (27.3%) of the GM articles concerned a women's health topic more broadly defined and were categorised as non-traditional women's health: heart disease (n = 6), health care delivery (n = 3), HIV/AIDS/STDs (n = 3), musculoskeletal conditions such as osteoporosis (n = 3), obesity/physical activity (n = 3), and a range of other women's health topics (n = 9). With respect to WS articles, 34 (41.5%) addressed solely a non-traditional women's health topic: obesity/physical activity (n = 7), general health (n = 5), musculoskeletal (n = 4), depression (n = 3), heart disease (n = 3), HIV/AIDS/STDs (n = 3), and a range of topics (n = 9).
Nineteen (19.2%) of the GM women's health articles addressed both a traditional and non-traditional women's health topic, compared to 26 (31.7%) WS articles which incorporated both types of women's health. While overall more than 40% of both GM and WS articles related in some way to women's reproduction, WS articles much more frequently combined this narrow women's health topic with a non-traditional issue (e.g. HIV in pregnancy). Specifically, 42 GM articles focused on reproduction, of which 16 (42.4%) incorporated a non-traditional issue, while 24 of 39 (61.5%) WS articles combined both perspectives. This combined approach reflects the broad, contextual definition of women's health advocated by scholars, practitioners, and patients [8,9].
Only one (1.0%) article published in GM journals used a non-quantitative research method. The others reported on RCT (n = 30), cohort (n = 38), case-control (n = 11), cross-sectional (n = 14), and case report (n = 5) designs. In contrast, a full 40% of WS articles involved qualitative or mixed methodologies, as advocated by women's health scholars [5,15]. The remaining WS studies were dominated by the cross-sectional (n = 30) design (Figure 1).
This figure compares the proportions of women's health articles published in general medical (GM) versus women's health specialty (WS) journals according to the study design used. It demonstrates that GM journals are more likely to publish articles using study designs higher up the conventional hierarchy of evidence (i.e., RCT), and WS articles incorporate qualitative and mixed study designs.
Discussion
Our study offers valuable insight into the presence and nature of women's health content in general medical versus women's health journals. Among the GM journals, each devoted approximately 20% of their original investigations content to women's health. This is encouraging and may reflect burgeoning scholarship of women's health researchers, government initiatives, and leadership of medical editors to include comprehensive approaches to health care. However, most GM articles drew on a narrow definition of women's health. WS journals provided more balanced coverage of women's health, publishing articles that addressed key women's health topics in context, more broadly defined, and based on mixed quantitative-qualitative methodologies. These integrated approaches are in keeping with principles designed to promote women's health across all clinical and research disciplines.
Leading sources of women's death and disability were not well represented across either set of journals. Lung cancer is the leading cancer killer of women but was not the focus of any articles published during the first half of 1999; in contrast, breast and cervical cancer articles together accounted for one-fifth of the entire sample. No GM articles and only five (6.1%) WS articles addressed violence against women. Likewise, heart disease produces significant illness burdens for women, but accounted for only 6% of articles in our entire sample.
The representation of women's health in WS journals is an important yardstick for the general medical literature which has greater influence on the health message communicated to practitioners and the public. Recently, JAMA published a second special theme issue on women's health (the only leading GM journal thus far) that included a range of topics such as heart disease, ovarian cancer, and HIV [16]. This is a significant leadership move because theme issues draw attention to health topics and imply topic importance; indeed, readers appear to prioritise women's health [17,18]. In addition, Lancet recently published a six-part series on violence against women that will surely raise the profile of violence's impact on women's health. Where possible, a better strategy might be to profile the range of women's health topics regularly, rather than under the auspices of special issues and series.
On conventional criteria, GM journals clearly have more impact (i.e., circulation, citation rate) than WS journals, appear to publish articles higher up the hierarchy of evidence (i.e., RCTs and intervention studies), and thus likely publish articles of superior quality. This may mean that health topics less amenable to RCT design (e.g., abuse) stand a better chance of publication in non-GM journals. Regardless, the focus of GM journals on "navel-to-knees" women's health may contribute to an uneven evidence base regarding women's health care. GM journals are challenged to maintain scientific excellence while incorporating the range of topics reflective of women's lives and health care needs. Precisely because GM journals have wide impact, it would be of benefit for editorial decisions and peer review mechanisms to promote a broader conceptualisation of women's health.
Limitations
We chose a particular time period, but our study provides an important preliminary audit of a developing field. Second, our results may not be generalisable to all medical and women's health journals, but it is unlikely that less prestigious journals provide better women's health coverage. We only evaluated original investigations and it is possible that issues relevant to broadening the definition of women's health are discussed in reviews and editorials. However, original investigations are important for leading new scientific understanding. Finally, we relied upon a primarily quantitative method. Additional in-depth and qualitative analysis of the medical literature, including its quality across types of journals, would enhance explorations of the representation of women's health.
Conclusions
Most women's health articles in the general medical literature represented traditional conditions, while specialty journal articles were more likely to address non-traditional topics. Neither type of journal well represented leading sources of women's death and disability. Since women's reproductive and maternal capacities represent only a fraction of women's health issues, leading journals should include articles about women's health that incorporate social, economic and political contexts. Because the leading general medical journals have wide impact, it would be of benefit for editorial decisions and peer review to promote a broader conceptualisation of women's health.
Competing interests
None declared.
Authors' contributions
JPC conceived of the study, collected data, and drafted the manuscript. GDF and PAR contributed to study design and acquisition of data. All authors obtained funding, contributed to analysis and interpretation of data, and read and approved the final manuscript.
Pre-publication history
The pre-publication history for this paper can be accessed here:
Acknowledgements
JPC is a doctoral fellow of the Canadian Institutes for Health Research (CIHR) and the Inner City Health Research Unit, University of Toronto, and is the 2001 Mitchell/Venn Fellow in Women's Health. PAR is supported by an Investigator Award from CIHR. The work was also facilitated by the National Network on Environments and Women's Health with funding from the Centres of Excellence Programme, Health Canada. Financial support from these sources is gratefully acknowledged. We also wish to thank M. Binns, P. de Nobrega, J. Dergal, D. Laxer, and A. Misra for their assistance.
A version of this paper was presented at the Fourth International Congress on Peer Review in Biomedical Publication, Barcelona Spain, September 16, 2001.
Over 300 therapies have been proposed for premenstrual syndrome. To date there has been only one survey conducted in the UK of PMS treatments prescribed by GPs, a questionnaire-based study by the National Association of Premenstrual Syndrome in 1989. Since then, selective serotonin re-uptake inhibitors have been licensed for severe PMS/PMDD, and governmental recommendations to reduce the dosage of vitamin B6 (the first choice over-the-counter treatment for many women with PMS) have been made. This study investigates the annual rates of diagnoses and prescribing patterns for premenstrual syndrome (1993–1998) within a computerised general practitioner database.
Methods
Retrospective survey of prescribing data for premenstrual syndrome between 1993–1998 using the General Practice Research Database for the West Midlands Region which contains information on 282,600 female patients
Results
Overall the proportion of women with a prescription-linked diagnosis of premenstrual syndrome has halved over the five years. Progestogens including progesterone were the most commonly recorded treatment for premenstrual syndrome during the whole study period accounting for over 40% of all prescriptions. Selective serotonin-reuptake inhibitors accounted for only 2% of the prescriptions in 1993 but rose to over 16% by 1998, becoming the second most commonly recorded treatment. Vitamin B6 accounted for 22% of the prescriptions in 1993 but dropped markedly between 1997 and 1998 to 11%.
Conclusions
This study shows a yearly decrease in the number of prescriptions linked to diagnoses for premenstrual syndrome. Progestogens including progesterone, is the most widely prescribed treatment for premenstrual syndrome despite the lack of evidence demonstrating their efficacy.
Background
It is estimated that between 85% to 97% of women of reproductive age experience some symptoms in the premenstrual phase of the cycle and about 30–40% of these women will seek help from their GP [1]. Premenstrual syndrome is defined as the recurrence of adverse physical and behavioural symptoms which recur in the luteal phase and which remit during, or after, menstruation; the symptoms are severe enough to significantly disrupt every day life for 3–5% of women [2]. The underlying cause of premenstrual syndrome remains unclear and speculative; many hypotheses have been suggested and numerous treatments advanced. The uncertainty as to the aetiology of premenstrual syndrome coupled with the very high placebo response has resulted in a large number of claims for potential therapies. There have been as many as 327 different treatments for premenstrual syndrome [3]. The variety of treatment approaches include hormonal preparations (oestrogen, oral contraceptives, GnRH analogues, danazol, progestogens and progesterone), diuretics, antidepressants, nutritional therapies (vitamin B6, mineral supplements and evening primrose oil), bromocriptine, surgery, complementary therapy and more recently light therapy and psychotherapy (cognitive and behavioural therapy) [4].
Despite this wide range of possible therapies for premenstrual syndrome no recent information concerning prescribing habits for premenstrual syndrome has been conducted in the UK. A survey of 658 women with self-reported PMS by Corney and Stanton [5] in 1990 found that vitamin B6 was the most widely used treatment, followed by evening primrose oil and then hormonal preparations. A similar survey of 220 women with self-reported PMS in the US found that dietary supplements and exercise were the most frequently suggested treatment option followed by progesterone and pain relievers [1]. Another recent survey found that of the 31% of women who reported severe premenstrual symptoms less than half sought any help for their condition. One of the reasons given was the belief that no effective treatments were available. Of those who used prescription medicines, paracetamol, vitamin B6, progesterone and oral contraceptives were the most frequently used treatments in the UK [6].
However, whilst these studies provide valuable information on what women are using to gain relief for their symptoms this does not necessarily reflect what General Practitioners are prescribing for this syndrome. A survey of diagnostic and treatment practices of US and Canadian primary care physicians reported in 1984 found that progesterone suppositories were the most widely prescribed treatment; 70% of all the surveyed physicians prescribed progesterone [7]. The National Association for Premenstrual Syndrome in 1989 assessed the treatment practices of 273 general practitioners in the UK by questionnaire. They found that 68% of general practitioners reported prescribing vitamin B6 and over half reported that they prescribed progestogens or progesterone for premenstrual symptoms [8].
In the last three years premenstrual syndrome has gained considerable media attention. In 1997, the UK Department of Health proposed to limit the sales of vitamin B6, a very popular self-help remedy for premenstrual syndrome, because of possible neurotoxic side effects at higher doses [9]. In 1999, the Medicines Control Agency granted the selective serotonin re-uptake inhibitor, fluoxetine, a license for use in severe premenstrual syndrome/premenstrual dysphoric disorder. This was the first time a licence has been granted specifically for premenstrual dysphoric disorder. There has also been heavy media and Internet promotion of 'natural' topical progesterone creams for PMS [10].
One of the issues highlighted in the 1989 UK postal survey of general practitioners [8] was the wide variety of treatments used by GPs to manage premenstrual syndrome. In order to determine current prescribing patterns, we used the General Practitioner Research Database (which records prescribing patterns and referrals) for the years 1993–1998 to identify prescriptions associated with the diagnosis of premenstrual syndrome.
Methods
The General Practitioner Research Database (GPRD) is an UK database, which records prescribing data and referrals and provides a valuable resource for analysing prescribing patterns. Anonymised records of individual patients are allocated a unique patient number. Data on medical events, patient problems and other doctor-patient interventions are captured in the database through the OXMIS (Oxford Medical Information System) dictionary. The OXMIS dictionary is an amalgamation of the eighth revision of the International Classification of Diseases (ICD-8) and the surgical operation codes used by the Office of National Statistics. During 1998 some practices started using Read codes. The Read codes were developed by James Read in the early 1980s to enable GPs to code and record relevant information from a patient encounter. Read codes are cross-referenced to the main national and international classifications. General Practitioners who provide data for the database have agreed to record the information in a standard manner to permit its use in research. The General Practitioner Research Database for the West Midlands former region contains 33 million records for prescribing or diagnosis for a population of 612 700 patients. The age sex profile of the patients for the West Midlands region matches that of England and Wales and represents 10% of the UK population [11].
For individual patient records, a gynaecologist (PMSOB) and a researcher (KW) reviewed the OXMIS and Read codes to select the codes which define premenstrual syndrome. Table 1 details the General Practice Research Database diagnoses used to identify premenstrual syndrome. All diagnoses of premenstrual syndrome with same-day prescriptions were extracted from the General Practice Research Database.
GPRD diagnoses for PMS
OXMIS
6269a
Tension premenstrual
6269f
Syndrome premenstrual
6269mt
Tension menstrual
READ
k584.00
Premenstrual tension syndrome
k584.11
Migraine menstrual
Prescriptions were linked to diagnoses by patient identifier, date of prescription and diagnosis. Prescription rates for premenstrual syndrome were calculated for 1993–1998 (the prescription rate is defined as the number of prescriptions for a specific drug group linked to a diagnosis of premenstrual syndrome, divided by the total number of prescriptions linked to diagnosis of premenstrual syndrome for that year). Prescriptions were identified as either repeat or acute and were determined for each drug group.
The Medicines Control Agency was contacted for a list of drugs, which have been licensed for premenstrual syndrome, premenstrual tension, premenstrual mastalgia, premenstrual oedema or premenstrual dysphoric disorder. The British National Formulary (BNF March 2001) and Monthly Index of Medical Specialities (MIMS; October 2000) and the Compendium of Datasheets (1999–2000) were searched for indicated treatments for the same disorders. Any identified treatment indicated for premenstrual symptoms was subsequently cross-checked in the other publications.
Table 2: (see Additional File) shows the list of drugs which have either been licensed for premenstrual symptoms or which appear in the data sheet compendium, MIMS or the BNF as indicated for premenstrual symptoms.
Statistical analysis
The yearly trends in prescribing for the individual drugs were analysed using Armitage trend test, P < 0.05 was considered to be significant [12]. Armitages trend test is used when one wishes to compare proportions between groups where the groups are ordered, to establish whether there is an upward or downward trend in the proportions. Here the ordered groups are years and the data was analysed to determine whether the proportion of prescriptions linked to diagnoses of PMS is increasing or decreasing over time.
Results
A total of 5,891 women were diagnosed with premenstrual syndrome between 1993–8 and issued with 6,172 same-day prescriptions. Table 3 shows the rate of diagnosis of premenstrual syndrome fell from 0.92% (of the total female population in the West Midlands GPRD) in 1993 to 0.42% in 1998. The rate of prescribing fell from 1.01% to 0.67%. Prescribing fell for all the British National Formulary sub-groups except selective serotonin reuptake inhibitors.
Rates of diagnosis of premenstrual syndrome
Year
Number of diagnoses
Number of same day prescriptions
Female population (GPRD)
% diagnoses
% same day prescriptions
93
1533
1688
166845
0.92
1.01
94
1231
1325
172003
0.72
0.77
95
1178
1230
182204
0.65
0.68
96
863
861
131226
0.66
0.66
97
645
667
121754
0.53
0.55
98
441
401
105571
0.42
0.38
93–98
5891
6172
879602
0.67
0.70
Figure 1 shows the rates of prescribing for the individual drug groups for 1993–1998. There was a significant downward trend (P < 0.05) in the prescription rates for vitamin B6, progesterone and progestogens and gonadotrophins and a significant upward trend over the study period for selective serotonin reuptake inhibitors. During the study period progestogens (including progesterone) were the most widely prescribed treatment comprising 44% in 1993 and still accounted for 42% of all prescriptions in 1998. There were significantly more prescriptions for progesterone and progestogens over the five-year period than any other drug type (P < 0.05). The most marked increase in prescriptions over the study period was for selective serotonin reuptake inhibitors; they represented only 2% of all prescriptions for premenstrual syndrome in 1993 but by 1998 accounted for 16% of all prescriptions. Similarly, the most marked fall in prescriptions was for prescriptions of vitamin B6 which fell from 22% in 1997 (the second most frequent prescription) to 11% in 1998.
Figure 2 shows a detailed yearly breakdown of the different treatments, which make up the progestogen/progesterone category. There was a significant drop in the level of prescriptions for progesterone (32% in 1993; 14% in 1998 of prescriptions in this category) and a corresponding increase in the number of prescriptions for dydrogesterone (19% in 1993; 40% in 1998). Prescriptions for norethisterone remained nearly constant accounting for approximately 45% of all progestogen and progesterone prescriptions. There were very few prescriptions of medroxyprogesterone acetate for premenstrual syndrome during this study period (1–4%). Of the progesterone prescriptions all were for Cyclogest® which is progesterone administered as a suppository or pessary. There were no recorded prescriptions for Crinone®, the vaginal progesterone gel or for Gestone® (progesterone administered by injection). Natural progesterone is not recorded, as is it not a prescription medicine.
The proportions of repeat and acute prescriptions were determined for each drug group over the study period. The proportions could provide an indication of whether the women perceive the treatment to be efficacious; the greater the number of repeat prescriptions, then presumably the more successful the treatment is perceived as being. The variation in the ratio of repeat and acute prescriptions remained consistent across the study period for each group of drugs apart from diuretics. (The only significant difference in the number of acute and repeat prescriptions over time was for the subgroup of diuretics which showed a significant increase in the number of repeat prescriptions over the study period). However, there was some variation in the ratio of acute and repeat prescriptions when the yearly proportions of acute and repeat prescriptions for the individual drug groups were compared. Figure 3 shows the proportions of new and repeat prescriptions for the individual drug groups for 1996. Prescriptions linked to the diagnosis of premenstrual syndrome for tricyclic antidepressants were most likely to be repeated whereas prescriptions for oral contraceptives were the least likely to be repeated. Figure 3 also shows the general proportions of acute and repeat prescriptions, taken from reference [22].
Discussion
There is a vast array of suggested treatments for premenstrual syndrome. The purpose of this study was to look at the diagnoses and associated prescription rates for the various medical treatment options for women presenting at a general practice with premenstrual syndrome
This study has revealed a surprising drop in the number of prescription-linked diagnoses of premenstrual syndrome. It is not immediately apparent why fewer women are being diagnosed with premenstrual syndrome, particularly as an analysis of the OXMIS code for another menstrual cycle disorder, heavy bleeding, showed that this diagnosis has remained relatively constant over the same period. A possible explanation is the recent increase in popularity of alternative, non-prescription treatments for premenstrual syndrome. A survey of medical herbalists in 1998 showed that the second most commonly treated condition was premenstrual syndrome [13]. It is therefore feasible that women are experiencing sufficient relief from these non-prescription treatments such that they no longer seek medical intervention.
A recent meta analysis of progesterone and progestogens has shown that neither treatment is efficacious in the management of PMS [14], however it is not surprising that progesterone and progestogens are the most commonly prescribed treatment for premenstrual syndrome. Dalton, who has been largely responsible for premenstrual syndrome becoming a recognised disorder, has enthusiastically advocated progesterone as a successful treatment for many years [15]. Progesterone is available in the UK, on prescription, as a pessary or suppository, vaginal gel or injection. Oral micronised progesterone is available in mainland Europe and the United States but not in the UK. Progestogens, synthetic analogues of progesterone, have been developed as oral preparations of progesterone. Of the progesterone preparations only Cyclogest®, (progesterone pessary/suppository) has a license for PMS in the UK. Dydrogesterone, norethisterone and an ethinyloestradiol/levonorgestrel combination (Ovranette®), all have a license for PMS (Table 2: see Additional File). Medroxyprogesterone acetate does not have a licence and is not recommended for PMS in MIMS, BNF or the drug datasheets and this could account for its low prescription rate, (less than 3% of the total hormonal prescriptions (figure 2)), although it has been trialed as a treatment for premenstrual syndrome [16].
The steady increase in the proportion of prescriptions of selective serotonin re-uptake inhibitors is in agreement with an overall increase in the number of prescriptions for this type of drug for other indications [17]. We anticipate this increase to continue as fluoxetine received a license for the treatment of severe premenstrual syndrome/premenstrual dysphoric disorder in 1999, and a recent meta analysis has shown it to be a highly effective treatment for this disorder [18].
It is possible that the sudden drop in vitamin B6 prescriptions from 19% in 1997 to 11% in 1998 results from proposals from the UK Department of Health to restrict the dosage available because of potential, although reversible, neurotoxic effects at very high doses [19].
Research into potential treatments for premenstrual syndrome has been restricted by the lack of accurate diagnostic techniques of the disorder in the study population. It is now widely accepted that the diagnosis of premenstrual syndrome can only be obtained through the structured evaluation of 1–3 months of prospective symptom recording and with the exclusion of any underlying psychiatric or medical disorder. Recent trials for premenstrual syndrome/premenstrual dysphoric disorder are conducted using these strict diagnostic criteria and therefore any reported efficacy can be considered to be proven for a woman with premenstrual syndrome. In practice this is not the same population that presents with the complaint of premenstrual syndrome. It has been estimated that between 25–75% of women who present with premenstrual syndrome actually have another medical condition [20]. The lack of suitable, easy to administer, diagnostic tools mean that virtually all diagnoses of premenstrual syndrome are made by self-assessment and up to half of these self-assessments have been shown to be incorrect [21]. This high level of incorrect assessment coupled with the known high placebo response [22] has probably resulted in the widespread use of inappropriate treatments and the inappropriate licensing of certain treatments. It is interesting that the number of repeat prescriptions for any drug group is so low in this study. A recent study estimated that repeat prescriptions generally account for 75% of all prescriptions [23]. The highest proportion of repeat prescriptions in this study was 50%, which was the proportion of repeat prescriptions for tricyclic antidepressants in 1997. It is possible that the lack of correct diagnosis and the failure to exclude other medical disorders are reflected in a low number of successful treatments and thus, repeat prescriptions.
An obvious limitation of this study is that premenstrual syndrome may be entered under a different diagnostic code to these searched on. There is frequent confusion in the clinical situation between premenstrual syndrome and other gynaecological disorders such as, dysmenorrhoea and the perimenopause, and in psychological disorders such as depression or anxiety. Similarly, this study can only reflect on what women are being prescribed and there is a large amount of evidence to suggest that most women with premenstrual syndrome will, at least initially, self-medicate [1,6,8].
Despite the recent concerns surrounding the interpretation of data from the General Practitioner Research Database [24], it remains a unique resource for providing information on the number of diagnoses and prescribing patterns for many disorders including premenstrual syndrome. This study has highlighted a dramatic drop in the number of women being given a diagnosis of premenstrual syndrome. It has also demonstrated the wide range of therapeutic techniques, which are currently being used in the management of this disorder. The development of effective treatments for premenstrual syndrome has been hampered by imprecise diagnostic standards, poorly controlled trials and the promotion of therapies which lack scientific support. We believe that the evidence for all possible treatment options for premenstrual syndrome should be assessed by critical appraisal and meta analysis to permit definitive statements to be made on the most efficacious management of premenstrual symptoms. Whilst this statement is true for all disorders, premenstrual syndrome treatments in particular highlight the result of prescribing practice based on inappropriate evidence.
Conclusions
• Prescription linked diagnoses of premenstrual syndrome fell four fold over the study period (1993–1998)
• Progestagens including progesterone are the most widely prescribed treatment for premenstrual syndrome
• Vitamin B6 was the second most frequently recorded prescription in 1993 but prescriptions halved between 1997 and 1998
• Prescriptions for selective serotonin re-uptake inhibitors increased dramatically between 1993 and 1998 and were the second most frequently prescribed treatment in 1998
• Significantly more prescriptions for all drug groups were acute rather than repeat, in contrast to other disorders.
Pre-publication history
The pre-publication history for this paper can be accessed here:
The objective of this study was to study the association of early age at menopause with pulse pressure (PP), a marker of arterial stiffness, and PP change.
Methods
The effect of natural menopause was studied in 2484 women from the Atherosclerosis Risk in Communities (ARIC) Study who had not used hormone replacement therapy and who had not had a hysterectomy. The cross-sectional association of age with PP was evaluated in the entire cohort. The cross-sectional association of recalled age at menopause was evaluated in the 1688 women who were postmenopausal at baseline. PP change over 6 years was assessed in relation to menopausal age separately in women who were postmenopausal at baseline and in those whose menopause occurred during the 6-year interval.
Results
Chronological age was strongly and positively associated with PP in cross-sectional analyses, but not independently associated with PP change. While menopausal age was not associated cross-sectionally with PP, early age at menopause (age<45) was significantly and independently associated with a slightly larger increase in PP (8.4, 95% CI 7.0–9.8) than later menopause (6.5, 95% CI 5.8;7.2). However, among normotensive women the difference was not statistically significant (p = 0.07, 6.1 vs 4.7).
Conclusions
Early age at menopause may be related to a greater increase in arterial stiffness, but the effect appears to be small and further evidence is needed.
Background
Pulse pressure, i.e. the difference between systolic (SBP) and diastolic blood pressure (DBP), is a crude but readily acquired measure of arterial stiffness. Recent findings from the Framingham study have suggested pulse pressure being superior to SBP and DBP in predicting CHD risk [1]. Increased arterial stiffness has been proposed as a marker or mechanism for initiation or progression of atherosclerosis and/or structural arterial changes due to hypertension [2]. Arterial stiffness estimated with ultrasound measured arterial diameter change, adjusted for blood pressure, was found to be higher in boys and men than in girls or women up to the age of 50–54 years, whereas these differences disappeared thereafter [2,3]. Menopausal factors, or loss of ovarian function, have been proposed to close the gap of the gender difference in arterial stiffness after age 50. Early menopause may hasten the increase in arterial stiffening with age, and lead to an increased risk of cardiovascular mortality. A Dutch study showed a 2 % decrease in cardiovascular mortality associated with each year's delay in menopause [4]. A Norwegian study estimated a substantial protective effect of late menopause, with 60% fewer cardiovascular deaths among women reaching menopause at age 47 years or later [5].
In the current study we hypothesized that early age at menopause is associated with increased arterial stiffness as indicated by pulse pressure and with increasing pulse pressure over time.
Material and methods
The ARIC Study [8] is a prospective study of 15,792 45–64 year old men and women in four US communities: Forsyth County, North Carolina; Jackson, Mississippi; selected suburbs of Minneapolis, Minnesota; and Washington County, Maryland. Initial participant response rates were 46% for Jackson and 66% for other three communities. The baseline examination of the ARIC study was conducted in 1987–89, and follow-up examinations occurred in 1990–92 and 1993–95. At the baseline, sitting blood pressure, anthropometry, venipuncture (12-hour fasting) and interviews concerning medical and reproductive history and medication use were performed. The accuracy of information pertaining to medication use was enhanced by asking the participants to bring containers of all medications used in the previous two weeks to the ARIC clinic to be transcribed by interviewers.
Three sitting blood pressure readings were recorded after 5 minutes of rest using a random zero sphygmomanometer. The systolic and diastolic blood pressure values used in the analysis as outcomes were averages of the second and third readings.
The total population at baseline included 8710 women. Women were classified as pre-, peri- or postmenopausal at baseline. Premenopausal women were those who reported that they had menstrual periods in the past 2 years, and had not reached menopause, or who reported no missing periods during last 2 years. Perimenopausal women were defined as those who reported that they had reached menopause but also reported any menstrual periods in the past 2 years. Women were defined as naturally postmenopausal if they had not been menstruating for the past 2 years and reported their menopause was natural. Similar menopausal definitions have been applied and described in earlier ARIC studies [6,7].
We excluded 686 women with unknown menopausal status at any visit, 21 women of ethnic origin other than African-American or white, 548 women with missing information on 3-year and/or 6-year follow-up, 2765 women who had ever used hormonal replacement therapy (HRT), 302 women coded as perimenopausal at both baseline and 6-year follow-up or those who underwent hysterectomy during the follow-up and 588 women who had missing values for other analysis variables. All women who had undergone hysterectomy prior to baseline (N = 1316) or during follow-up were excluded leaving 2484 women in the study population. The rationale for excluding women who had ever used HRT, those who had a hysterectomy, and those who remained perimenopausal at the end of the study, was to study the effect of untreated natural menopause. A greater proportion of women in the excluded sample were black, fewer had a high school education, and fewer had used antihypertensive medications than women in the included sample.
In the final study sample, age at natural menopause was reported at baseline by 1,688 postmenopausal, non-hysterectomized women. Since there might be differences by the centers in the time between follow-up visits and other factors, adjustments by ARIC field center were performed by using three dummy variables in most of the models.
Because pulse pressure is affected by hypertension per se and hypertensive medication, we divided women into consistent users of hypertensive medication and all others. Consistent users included women who reported hypertensive medication use at baseline and each follow-up visit.
Chronological age was evaluated in the entire study population (2484 women) in relation to PP level at baseline and at the 6-year follow up examination (cross-sectional analysis) and also in relation to the 6-year change in PP (longitudinal analysis). Both PP level and PP change were studied similarly in relation to recalled age at menopause in the 1688 women who were menopausal at baseline. Finally, PP change was analyzed in relation to recorded age at menopause in the 259 women who underwent menopause between the baseline examination and the 6-year follow up examination.
In order to adjust the 6-year change in pulse pressure for confounding factors, analysis of covariance was performed. Pulse pressure change was the dependent variable; age at menopause was the independent variable and baseline age, baseline pulse pressure, race, hypertensive medication use, ARIC field center and smoking were the covariates considered.
When fitting by ordinary techniques models of change in an outcome variable as a function of some exposure and the baseline level of the outcome variable, it is well known that the presence of measurement error in the baseline outcome variable can cause substantial bias in the estimates of all model coefficients [9]. One method to correct the measurement error is to replace the baseline values of the outcome variable by Stein estimates of the true values of that variable, and then apply ordinary least squares [9]. In correcting for measurement error we have used estimates for repeatability of SBP, DBP and PP from a sample of 363 ARIC female participants with repeat measurements 1–2 weeks apart during ARIC visit 3. The correlation coefficients for repeat measurements of SBP, DBP and PP were 0.75, 0.62 and 0.66, respectively.
Due to the possibility of varying prevalence of hypertensive subjects in different age-groups to confound pulse pressure change estimates, we also studied PP change as % from baseline. Secondly, we considered the possibility of age at menopause being significant variable when using it as continuous in the multivariate models. Thirdly, formal analyses to consider potential interaction between hypertension status at baseline or use of hypertensive medication and menopausal age have been performed also.
Results
Among the women who were postmenopausal at baseline, there was no cross-sectional association of pulse pressure with recalled age at menopause (Table 1). Longitudinally, the 6-year increase in PP was greatest among those who had undergone menopause at the earliest age, 25 to 35 years, though the overall linear trend was not statistically significant. Women recalling their menopausal age as younger than 45 years had a significantly higher PP increase (8.0) than women recalling their menopausal age as 45 years or older (6.6), after adjusting for age at baseline, baseline PP, race and hypertensive medication use. PP change for earlier (< 45 years) menopausal women was significantly (p = 0.014) higher (8.4) than PP change for women menopausal at age 45 or older (6.5) after taking into account measurement error. However, among normotensive women the difference was not statistically significant (p = 0.07, 6.1 vs 4.7). Age at menopause was also used as continuous variable in the models explaining pulse pressure change, but it was not significant.
Systolic (SBP) and diastolic (DBP) blood pressure, pulse pressure (PP) and PP change by recalled age at menopause among non-hysterectomized women postmenopausal at baseline. N = 1688.
Baseline
6-year follow-up
Recalled age atmenopause
N
Mean age atbaseline
SBP
DBP
PP
SBP
DBP
PP
PP change
Adj1 PP change
95% CI
25–35
40
57.6
122.9
72.3
50.6
129.5
68.4
61.1
10.5
10.1
6.1–14.2
36–40
155
56.8
122.2
71.1
51.1
128.9
69.5
59.4
8.3
8.7
6.6–10.7
41–44
167
56.5
118.9
70.3
48.6
124.5
68.5
56.0
7.4
7.3
5.3–9.3
45–48
501
57.0
123.1
71.9
51.2
127.3
69.7
57.5
6.3
6.6
5.5–7.8
49–51
469
58.1
123.4
71.7
51.7
128.1
70.0
58.1
6.4
6.2
5.0–7.3
52+
356
59.8
125.4
71.3
54.0
129.9
68.7
61.2
7.2
6.8*
5.4–8.2
1 Adjusted for baseline age, baseline PP, race, hypertensive medication use and measurement error. * Adjusted test for linear overall trend p = 0.20, for women menopausal at age <45 (PP change 8.4, 95% CI 7.0–9.8) vs. age = 45 (PP change 6.5, 95% CI 5.8–7.2), p = 0.014.
When pre- or perimenopausal women at least 45 years of age at baseline were followed for 6 years, those with lower estimated age at menopause did not have greater PP increase over time compared to women with higher estimated age at menopause (Table 2).
Systolic (SBP) and diastolic (DBP) blood pressure, pulse pressure (PP) and PP changes by estimated age of menopause. N = 259. Only women pre- or perimenopausal at baseline included.
Baseline
6-year follow-up
Estimated age atmenopause
N
Average age atbaseline
SBP
DBP
PP
SBP
DBP
PP
PP change
Adj.1 PP change
95% CI
45–50.4
151
47.8
115.8
71.7
44.1
121.4
71.6
49.7
5.7
5.7
3.4–8.0
50.5–52.4
43
49.1
117.1
71.9
45.2
125.3
73.2
52.1
6.8
6.5
2.7–10.4
52.5–54.4
37
50.3
117.8
72.5
45.3
122.9
73.4
49.5
4.2
4.1
-0.2–8.4
54.5–65.5
28
52.0
123.2
74.4
48.8
128.3
73.6
54.7
5.92
6.73
1.3–12.1
1 Adjusted for baseline age, baseline PP, race, hypertensive medication use and ARIC field center. 2 Test for overall linear trend p = 0.87. 3 Test for overall linear trend p = 0.74.
In the cross-sectional analyses, at baseline and at the 6-year follow-up, systolic blood pressure was strongly positively associated with chronological age, but diastolic BP varied little with age (Figure 1). Thus, the difference between SBP and DBP, i.e. pulse pressure, increased with age. In the baseline data, average pulse pressure was 42.3 mmHg in women aged 45 to 46 years and increased monotonically to 56.5 mmHg in the oldest women aged 62–64 years (Table 3). Linear trends in both cross-sectional associations between baseline age and baseline PP as well as between baseline age and PP at 6-year follow-up, were highly significant (p=.0001) even after adjusting for race and hypertensive medication use. Longitudinally there was a significant relationship (p=.0001) between 6-year change in pulse pressure and chronological age at baseline after adjustment for baseline PP, race, hypertensive medication use, ARIC field center and smoking (Table 1). This relationship did not remain significant after additional adjustment for measurement error (p = 0.23).
Systolic (SBP) and diastolic (DBP) blood pressure at baseline and 6-year follow-up visits by chronological age at baseline among non-hysterectomized women without HRT.
Pulse pressure (PP) and PP changes by chronological age at baseline among non-hysterectomized women. N = 2484.
Age atbaseline
N
Adjusted PP atbaseline1
Adjusted PP at 6-year follow-up1
AdjustedPP change2
Adjusted PP change corrected for measurement error3
95% CI
45–46
206
42.3
47.4
5.1
6.2
4.5–8.0
47–49
266
44.0
48.7
4.7
5.2
3.6–6.7
50–52
364
45.9
51.1
5.3
5.4
4.1–6.7
53–55
398
48.2
53.5
5.3
5.1
3.9–6.4
56–58
442
50.6
57.6
7.0
6.6
5.5–7.8
59–61
418
53.7
61.0
7.2
7.0
5.8–8.3
62–64
390
56.51
64.12
7.63
7.3
6.0–8.6
1 Test for overall linear trend, adjusted for race and hypertensive medication use, p = 0.7149 2 Test for overall linear trend, adjusted for baseline PP, race, hypertensive medication use, ARIC field center and smoking, p = 0.64 3 Additionally corrected for measurement error, p = 0.23.
The proportional changes in pulse pressure were nearly similar in both hypertensive and non-hypertensive subjects, but the absolute values differed. The different prevalence of hypertensive subjects in age-groups did not explain the results. Additionally, the interactions between hypertensive status at baseline or use of hypertensive medication and menopausal age were not significant in any subgroup, with or without adjustments for other variables (results not shown in tables).
Discussion
Pulse pressure, a marker of arterial stiffness, is clearly related to older chronological age. Our study suggests that PP may also be related to older "biological age", if one considers early age at natural menopause as a sign of biological aging. Women whose menopause occurred before age 45 experienced a greater increase in PP than women with menopause at later ages. This association remained significant when taking into account chronological age, baseline PP, race, and use of antihypertensive medications. However, the association was limited to PP change and was not seen for PP level at baseline or 6-years later. The PP effect appears to be slight in magnitude, approximately 2 mm Hg, so it may be detected better in longitudinal than cross-sectional analyses. Longitudinal analysis reduces the effects of extraneous variables, since each woman serves as her own control when her baseline PP is subtracted from her 6-year PP. Furthermore there was no association of PP with later menopause, suggesting that age at menopause after age 45 is not an effective marker of biological aging. Arterial stiffening, as assessed by pulse pressure, may offer one explanation for the findings of others relating cardiovascular risk to menopause before age 45[4,5].
A change in PP in relation to menopause could reflect the gradual reduction in circulating sex hormones which occurs at the time of menopause. However, our earlier findings provide some evidence against that interpretation [7]. There we showed that women currently undergoing the menopausal transition had no more perimenopausal change in BP than women of the same age not undergoing the transition. Our current findings are different. They show a greater PP increase after a baseline examination in women who recalled their menopausal age as younger than 45 years compared with similar women with later menopausal ages. We interpret this postmenopausal PP increase as suggesting that early menopause may be an indicator of a more rapid biological aging process, i.e. that women with an early menopause are generally aging more quickly than expected for their chronological age, as evidenced by the relatively rapid changes in arterial stiffness observed postmenopausally. Earlier studies have shown that estrogen deficiency may induce functional changes in large arteries, but structural changes are more likely to be seen in long-term effects [10-15].
Ours is not the first study to evaluate the effect of menopausal age on pulse pressure, a surrogate marker for arterial stiffness, which may serve as an indicator of preclinical atherosclerosis. Menopause per se has been shown to be associated with increasing stiffness of the aorta, later leading to dilatation of the common carotid artery [16]. Most earlier cross-sectional studies during the 1990's, which have addressed the relations between timing of menopause, mortality, cardiovascular mortality and myocardial infarction, have shown significant association between age at menopause and cardiovascular outcomes [4,5,17,18]. Among Seventh-Day Adventists, age-adjusted OR of death in women with natural menopause before age 40 was 1.95 (95% CI 1.24–3.07), and the OR decreased with increasing age at natural menopause until age 55[17]. In two studies, cardiovascular mortality has been found to be greater in women with age at menopause lower than 45 years as compared to women with later age at menopause [4,5]. One recent study reported slightly increased risk of total mortality (adjusted mortality rate ratio 1.50, 95% CI 0.97–2.34) among women with a natural menopause at age 40 or before as compared to women who were menstruating to age 50 or later [18]. Spontaneous cessation of ovulatory function before age 45 was associated with increased risk of myocardial infarction (RR 2.1, 95% CI 1.3–3.2) as compared with women who had a natural menopause at age 50 or older [19]. In Nurses' Health Study, age at menopause lower than 35 years was associated with increased risk of myocardial infarction, but natural menopause at later ages was not [20].
Earlier longitudinal studies [21-25] have addressed the question of menopause and hypertension as well. According to the Framingham study menopause was not accompanied with changes in blood pressure [21]. In a Swedish and Pittsburgh cohort studies [22,23] the findings correlated Framingham study results. A Dutch cohort including a selected sample of women also concluded that menopause is not a possible cause of hypertension [24,25], although ovarian failure seemed to reverse temporarily the increase in blood pressure due to aging. In a recent longitudinal study by Scuteri et al [26], postmenopausal women using HRT had a smaller increase in systolic blood pressure than women not using HRT.
Opposite to the findings of longitudinal studies, a large number of studies measuring either diameter, compliance, distensibility of large arteries or aortic pulse wave velocity have concluded that menopause per se is associated with loss of aortic elasticity, over and above the effects of aging [27-33].
On the basis of studies using other measurements described above, there are limitations in the use of pulse pressure as an index of stiffness. Pulse pressure is determined by a number of factors apart from arterial stiffness including heart rate, stroke volume, and ejection time. Results might have been different with a more sensitive measurement of arterial stiffness. Studies, which have assessed menopause and arterial stiffness using carotid ultrasound measurements, provide some support for our findings, although the relationship between timing of menopause and change in stiffness was not assessed in any of the described studies [34-37]. Stiffness was found to be significantly greater in postmenopausal than age-matched premenopausal women in the Dutch study [34]. Menopause had an independent significant association to the stiffness index in the Swedish study [35]. In the Healthy Women Study from Pittsburgh [36] premenopausal values of pulse pressure were independently predictive of plaque.
Since the ARIC cohort was limited to persons aged 45 or older, our findings relating to menopause before age 45 had to depend on recall, which may be imprecise. Recall errors may increase with time from the menopause [37], but appear not to be related to educational status or age at interview [38]. Imprecision in recalled age at menopause may have attenuated our findings somewhat.
Factors known to be associated with early age at menopause are smoking [39,40], low sociodemographic status [41], nulliparity [42,43] and leanness [44], although the effect of the latter may be partlyexplained by larger proportion of smokers among lean women. In our study, smoking did not explain the association between early age at menopause and pulse pressure change. Multiple triggers are known to lead to reproductive decline, but there is no consensus about an ultimate pacemaker of the menopausal transition [45]. One view maintains that the hypothalamic-pituitary changes that accompany menopause are a consequence of compromised ovarian function, whereas another view emphasizes age-related changes in the nervous system as initiators of menopausal transition [42]. Our study suggests that, whatever the reasons for early age at menopause, it appears to be independently and significantly related to a small subsequent increase in PP, but clearly, confirmation of this interesting finding is required.
Competing interests
None declared.
Pre-publication history
The pre-publication history for this paper can be accessed here:
Acknowledgements
ARIC-study: University of North Carolina, Chapel Hill: Phyllis Johnson, Catherine Paton, James Pankow, Sharon Pope. University of North Carolina, Forsyth County: Melinda Cochran, Shirley Cothern, Amy Haire, Delilah Posey. University of Mississippi Medical Center, Jackson: Bobbie Alliston, Agnes Hayes, Penny Lowery, Stephanie Parker. University of Minnesota, Minneapolis: Todd Avant, Joseph Bjorklund, Dorothy Buckingham, Carolyne Campbell. Johns Hopkins University, Baltimore:Pam Bowers, Joyce Chabot, Carol Christman, Dorrie Costa. University of Texas Medical School, Hemostasis Lab, Houston. Chul Ahn, Nena Aleksic, Ashley Ewing, Harinder Juneja. The Methodist Hospital, Atherosclerosis Clinical Laboratory, Houston: Wanda Alexander, Christine Ballantyne, Charles E. Rhodes, Andre Surguchov. Bowman-Gray School of Medicine, Ultrasound Reading Center, Winston-Salem: Carolyn Bell, Delilah Cook, Bob Ellison, Kathy Joyce. University of North Carolina, Chapel Hill, Coordinating Center: Myra Carpenter, Barbara Dennis, Tom Goodwin, Steve Hutton, Doris Jones.
Grants: Academy of Finland, Finnish Cultural Foundation.
This study assessed women and providers' satisfaction with a new evidence-based antenatal care (ANC) model within the WHO randomized trial conducted in four developing countries. The WHO study was a randomized controlled trial that compared a new ANC model with the standard type offered in each country. The new model of ANC emphasized actions known to be effective in improving maternal or neonatal health, excluded other interventions that have not proved to be beneficial, and improved the information component, especially alerting pregnant women to potential health problems and instructing them on appropriate responses. These activities were distributed within four antenatal care visits for women that did not need any further assessment.
Methods
Satisfaction was measured through a standardized questionnaire administered to a random sample of 1,600 pregnant women and another to all antenatal care providers.
Results
Most women in both arms expressed satisfaction with ANC. More women in the intervention arm were satisfied with information on labor, delivery, family planning, pregnancy complications and emergency procedures. More providers in the experimental clinics were worried about visit spacing, but more satisfied with the time spent and information provided.
Conclusions
Women and providers accepted the new ANC model generally. The safety of fewer visits for women without complications with longer spacing would have to be reinforced, if such a model is to be introduced into routine practice.
Background
In spite of the increasing involvement of technology in routine antenatal care in both developed and developing countries, the clinical encounter between patient and caregiver still represents the core of the current health care paradigm. At least in theory, any care offered should be acceptable for the recipients. However, the importance of allowing for patients' views, alongside medical and economic considerations regarding care assessment during pregnancy and childbirth, wasn't stressed until the late 80's and almost only in developed countries [1].
The importance of caregivers' views has been acknowledged even less frequently, even though it is a crucial component of any attempt to change institutional protocols. [2] In fact, such views about their own clinical work can strongly influence daily performance and acceptance of institutional protocols and norms [3]. For example, physicians' attitudes appear to be the most important factor influencing the rate of Caesarean sections [4].
Undoubtedly, patients and caregivers' perspectives mirror the quality of the care received and provided. However, quality of care has been traditionally a difficult concept to operationalize. As a reflection of the emphasis on the application of advanced technology and specialized training, quality of care has been largely defined in terms of clinical aspects, neglecting social interaction and the subjective dimension of the patient [5]. Only in the last decade and based on Donabedian's work [6] did Bruce's framework highlight the importance of stressing not only the technical but also the interpersonal domain in the field of family planning [7].
Measuring quality of care conceptualized in such a broad manner represents a true challenge. While the technical quality of a health service can be assessed by evaluating the outcomes of the care provided, the subjective dimension of quality of care (interpersonal relationship with the provider and the system's responsiveness to the expectations of the population) can only be assessed through interviews that are strongly influenced by the cultural milieu and the circumstances under which they are conducted. In the field of antenatal care, recent efforts have been made to sort out these various influences [8-13]; however, the knowledge about users' views is still very limited, especially in developing countries [14,15]. This paper describes women and providers' perceptions of the quality of antenatal care (ANC) and their degree of satisfaction with it, alongside a large randomized controlled trial [16-18].
Methods
The project reported here was a special component of a large multicenter randomized controlled trial, to evaluate a new ANC program [18]. The primary hypotheses tested were that a new ANC model, consisting only of actions scientifically proven to improve maternal and newborn outcomes, was as effective as the traditional model with regard to specified maternal and perinatal end-points among singleton pregnancies, was not more expensive, and was accepted by women and providers [16]. Conducted by WHO and collaborating organizations, fifty-three randomly allocated (cluster randomization) ANC clinics in Argentina, Cuba, Saudi Arabia and Thailand participated in the study, providing either the new program or the traditional program in use. There were 12,568 women randomized to the new model and 11,958 to the standard ANC model [18].
The model in the control clinics was the antenatal care currently offered, following guidelines formally recommended by the local health authorities, based on the traditional western model. In general, women made visits once a month during the first six months, one every 2–3 weeks for the next 2 months, and one every week until delivery. Clinical activities, urinary tests, syphilis screening, hemoglobin measurement, and blood group typing were done routinely.
In the intervention clinics, women judged not to need further assessment or special care at the time of the first visit according to predefined risk criteria were assigned to the new ANC model, which required fewer visits (usually four) with longer spacing between them than the standard ANC model recommended. Activities in the new program included: 1) screening for health conditions likely to increase the risk of adverse pregnancy outcomes; 2) providing interventions known to positively impact pregnancy outcomes, and excluding other common interventions that have not proved beneficial to pregnant women (e.g., maternal weight was measured only during the first visit; subsequent measurement was limited to patients with low weight); and 3) alerting pregnant women to potential health problems, especially emergencies, and instructing them on appropriate responses (e.g., recommendations for emergencies were provided in each visit; at the third visit instructions for delivery and suggestions for breastfeeding and contraception were included) [16]. Results of the general effectiveness of the new ANC model (measured by examining low birth weight for the fetal condition and rates of preeclampsia/eclampsia, severe postpartum anemia, and treated urinary tract infection or pyelonephritis for the maternal conditions) have been published elsewhere [18]. The assessment of women and provider satisfaction with the standard model and the new model is described here.
Study Design
The assessment of women and providers' perception of the quality of both models of antenatal care was organized in two stages. First, we used an ethnographic approach, including focus group discussions and in-depth interviews with women and health personnel to assess the culture-related values in each country. During the qualitative stage we addressed general topics on health care provision and prenatal programs to gain initial understanding of the way health care was perceived in each specific cultural context [17]. The findings of this in-depth study [19] were incorporated in the second stage (quantitative), which used standardized, closed-ended questionnaires that were prepared based on the most relevant categories obtained at the qualitative stage and the aspects of antenatal care that were expected to change as a result of the intervention (i.e., number of visits, spacing, time with provider, and information provision regarding maternal and perinatal health and complications). Both instruments (one for women, one for providers) were developed in English, translated into Spanish and Arabic, and piloted in each country. Changes suggested in each site were incorporated into the final version of the instruments by the investigators responsible of this component of the trial and then reviewed and approved by a WHO special technical review group. These final English versions (see Additional File 1 [Women's questionnaire] and Additional File 2 [Providers' questionnaire]) were again translated into local languages.
The questionnaire for women consisted of 24 questions addressing patients' preferences about the number of antenatal care visits; time spent in the waiting room and with the caregiver; and the amount and appropriateness of the information received during the visits. Women were also asked about their worries concerning their health status and their babies', and the reassurance they received from the provider. Because of the known limited validity of questions that include the word "satisfaction", we decided to include only one direct question ("Are you satisfied with the antenatal care you have received in this clinic so far?"), adapted from a previous antenatal care trial [20] to facilitate their meta-analysis, and two indirect ones ("Would you come back to this clinic?" and "Would you recommend this clinic to a relative or friend?"), which were developed based on information gathered from the focus groups (see Additional File 1 [Women's questionnaire]). We expected the "satisfaction" variable to synthesize women's overall perceptions of the quality of antenatal care [17].
The questionnaire for providers included 15 questions, probing the same issues as the patients': number and spacing of antenatal visits, time spent with the woman, information provided, perception of the quality of antenatal care, and recognition of women's satisfaction. (see Additional File 2 [Providers' questionnaire])
While some questions were worded per the terminology used in each country, their meaning was retained in the four settings. Both questionnaires were piloted in the four participating areas and adjusted accordingly [17].
Sample size and sampling strategy
The sample size was estimated to detect a minimum difference between a dissatisfaction rate of 5% in one arm and 10% in the other, with a two-sided test at a significance level of 5%, and with 80% power. The sample size obtained with standard formulas to compare two proportions for individually randomized design was multiplied by a design effect of 1.7 to account for the decrease in efficiency of the cluster randomized design. A sample of 1600 women (800 per arm of the study, 400 per country) was deemed necessary. A design effect of 1.5 had been previously calculated for the outcome of low birth weight [21]. Since this design effect does not necessarily apply to a different outcome, and there was no information regarding design effects or intraclass correlation coefficients from other studies, we arbitrarily increased it to 1.7. The survey used clinics as strata, and women were sampled proportionally to each clinic's number of women per year.
The interviewers started surveying all eligible women on a randomly selected day and continued during working days until enrolling the estimated sample size. Because women needed to be sufficiently exposed to ANC in order to form their own opinion on the quality of care they had received, we administered the questionnaire only to patients that were at 32 weeks of gestational age and had attended the health care facility for their second or subsequent antenatal visit. The women were surveyed in a private environment, in approximately 15 minutes. We did not request an individual separate informed consent for this component of the trial but we did have special institutional consent. Therefore, the questionnaire was administered to all women that met the inclusion criteria (i.e., 32 weeks of gestational age and two or more antenatal care visits) in both clinics' groups until completion of the sample size.
We asked all 174 ANC providers from both intervention and control clinics to complete a self-administered questionnaire that took approximately 10 minutes. We recruited 92 caregivers from the experimental institutions (57 physicians, 33 nurses, and 2 midwives) and 82 from the routine care arm (54 physicians, 25 nurses, and 3 midwives). No provider refused to fill the instrument.
Outcomes
Regarding the women's survey, affirmative answers to questions about satisfaction measured overall satisfaction (primary outcome). Other satisfaction dichotomous outcomes were satisfaction with number of visits, spacing between visits, waiting time, and time spent with provider.
Additionally, the following four summary indexes were constructed as outcomes for the women's survey: 1) Information received, defined as the number of 'as much as you wanted' answers provided by a woman to the six questions on information received about looking after own health, tests during pregnancy, any treatment that might be needed during pregnancy, labor, breastfeeding and family planning; 2) Information about how to recognize problems, defined as the number of 'yes' answers provided by a woman to the six questions on: whether she was told how to recognize the following pregnancy-related problems: rupture of membranes, hemorrhage, premature contractions, dizziness and fainting, fever, and other; 3) Information about what to do in the presence of the above-described problems; and 4) Information about how to recognize and handle these problems. For every woman, each index summarized six questions of the survey, thus the numerical value could vary from 0 to 6.
For the providers' questionnaire, information given was measured through an index defined as the number of 'yes' answers to the six questions about health, tests and treatments during pregnancy, labor and delivery, breastfeeding and family planning. This index also ranged from 0 to 6.
Data analysis
Percentages or mean and standard deviations, as appropriate, were computed by group for baseline variables for the women interviewed, by arm, and checked for imbalance between groups. Baseline statistics for the sub-sample of women interviewed were compared with those for all participants to confirm that they were representative of the main trial population.
For the women's survey, the average values of the event rates of satisfaction outcomes were compared between arms, using a rate difference and a t-test at the cluster level, obtaining the standard errors for the difference from a variance analysis adjusted for strata. The indexes were analyzed as numeric outcomes using a random model approach, with clinic and subject as random factors, and arm and strata as fixed factors. Outcomes were adjusted for baseline variables showing a prognostically important imbalance to detect a possible confounding effect.
Since all the clinics' antenatal caregivers were interviewed, and as they constituted a fixed population, the providers' questionnaire was analyzed descriptively by computing percentages or mean and standard deviations, as appropriate.
Ethical review
This component of the study was reviewed and approved as part of the overall ethical review of the WHO trial, which was approved by the Scientific and Ethical Review Group of the UNDP/UNFPA/WHO/World Bank Special Programme on Research, Development and Research Training in Human Reproduction, the WHO Secretariat Committee for Research into Human Subjects, the Institutional Review Boards of the individual participating centres, and corresponding health authorities of the regions where the trial was implemented.
Results
Since this was a randomized controlled trial, the analysis focused on the differences between women and providers in the control clinics as compared to those that offered the new model of ANC.
The comparison between the sub-sample of women recruited for this study and the total trial population showed no significant differences in age, height, weight at first visit, marital status, schooling, proportion of nulliparae and primigravidae, and smoking. However, women in the sub-sample study were slightly better off than women in the main trial regarding prior low birthweight (LBW) (total population: 5.8% vs. sub-sample: 4.0%), gestational age at first visit (total population (mean): 16.2 weeks vs. sub-sample (mean): 12.7 weeks) and previous hospital admissions (total population: 1.4% vs. sub-sample: 0.8%).
Women in both arms of the study had similar characteristics at trial entry (Table 1). There were no differences in prior LBW, stillbirths, neonatal losses and conditions of current pregnancy, but women in the standard ANC model (i.e., control clinics) had a slightly higher proportion of previous abortions (34.2% vs. 30.4%). Regarding providers in both trial arms, age and years since graduation were similar.
Baseline characteristics of women enrolled in the satisfaction study, according to ANC model
Women's Characteristics
New ANC (n = 790)
Standard ANC (n = 748)
%
Mean (STD)
%
Mean(STD)
Married/stable union
95
93
Education (less than primary)
16
17
Smoking during pregnancy
9.5
11
Substance abuse
0.4
0
Ratio of persons/room
2.4 (1.3)
2.3 (1.1)
Maternal age (years)
25.5 (5.8)
25.8 (5.6)
Surgery
1.8
0.8
Any previous LBW (<2500 g)
3.9
4
Any previous abortions
30
34
Any previous stillbirths or neonatal losses
3.8
4
PRESENT PREGNANCY:
Iso-Immunization Rh (-)
1.3
0.3
Vaginal bleeding first trimester
2.4
1.7
LMP unknown
4.3
3.3
Nulliparae
25
27
Primigravidae
29
29
Maternal height (cm)
157 (6.6)
156 (6.4)
Maternal weight at first visit (kg)
59.2 (12.6)
58.9 (11.8)
Gestational age at first ANC visit (weeks)
13 (5.7)
12.4 (5)
Late booking for ANC (>28 weeks at first visit)
2.7
1.7
For women in the new ANC model, the median was five ANC visits (1st quartile: 4; 3rd quartile: 6) while for those in the standard model it was nine (1st quartile: 6; 3rd quartile: 12). The median waiting time to see a doctor or midwife was shorter (Median: 30'; 1st quartile: 20, 3rd quartile: 60) for patients in the new model than for those under the standard model (Median: 45'; 1st quartile: 30, 3rd quartile: 75). There were no differences, however, in time spent with a doctor or nurse for the two arms (median: 15'; 1st quartile: 10, 3rd quartile: 20).
Women under the new ANC model were slightly less satisfied with the number of visits (77.4% vs. 85.2%; 95% CI of the difference: -16.0% to 0.2%) and visit spacing (72.7% vs. 81.0%; 95% CI of the difference: -16.8% to 0.3%) than their counterparts in the intervention clinics, although these differences were not statistically significant (Table 2). Women in both arms of the trial were equally satisfied with waiting time, but those in the new model were more satisfied with the time spent with their provider, although the difference was not significant (85.7% vs. 79.1%; 95% CI of the difference:-0. 5% to13.7%) (Table 2). Adjusting all these outcomes for the baseline abortion rate (which was slightly different among women in the intervention and control groups) did not change the results.
Women's satisfaction with antenatal care, according to ANC model*
Satisfaction with:
New ANC (%)
Standard ANC (%)
Adjusted mean difference (%)
95%CI
Number of visits
77.4 (789)
85.2 (744)
-7.9
-16.0
0.2
Spacing between visits
72.7 (782)
81 (744)
-8.3
-16.8
0.3
Waiting time
78.3 (780)
77.6 (743)
0.7
-7.4
8.8
Time spent with provider
85.7 (789)
79.1 (747)
6.6
-0.5
13.7
(Number of women) *Mean differences and 95% CIs adjusted for strata.
Women in both trial arms were equally satisfied with the information provided by the caregiver about their health, tests during pregnancy and treatment they might need (Table 3) [22]. However, women in the new ANC model were substantially more satisfied with the information received about normal labor and delivery processes, breastfeeding, family planning, and danger signs (Table 3).
Women's satisfaction with information on ANC, labor, delivery, and postpartum care, according to ANC model
Women satisfied with information received about:
New ANC
Standard ANC
N
%
N
%
Their own health
789
79.7
744
79.5
Tests during pregnancy
789
86.8
745
83.2
Treatment they might need
788
62.0
744
68.0
Labor and delivery*
785
70.0
745
59.5
Breastfeeding*
789
76.1
743
67.9
Family Planning*
788
65.9
744
51.1
Rupture of membranes*
787
64.3
740
50.5
Hemorrhage*
785
73.0
741
57.4
Premature contractions*
783
73.8
742
59.4
Dizziness and fainting
782
53.3
742
55.9
Fever
779
49.2
741
51.2
* Significant at 5% (adjusted for simultaneous inferences using Bonferroni method) [22]
During ANC visits, health professionals are usually expected to focus on issues that may worry their patients, a component specially stressed in the new ANC model. Therefore, we asked those women that said they were worried about a specific condition what reassurance they had obtained from their ANC providers (Table 4) [22]. In general, a similar proportion of women in both trial arms worried about fetal position, size and possible abnormalities, risk of prematurity, and other complications. Providers reassured a higher number of women in the new ANC model clinics, although these differences were not significant (Table 4).
Women who were worried about maternal/perinatal conditions and were reassured by provider
Perinatal/maternal conditions
Women who were worried*
Worried women who were reassured*
New ANC
Std ANC
New ANC
Std ANC
N
%
N
%
N
%
N
%
Fetal position
788
56.0
788
52.2
436
87.2
395
79.8
Fetal size
788
52.2
740
48.8
406
81.0
355
78.6
Prematurity
787
49.7
740
48.5
384
81.5
354
73.7
Fetal abnormalities
785
60.6
738
59.1
468
76.7
431
68.5
Other complications
762
31.5
720
29.3
328
85.7
402
86.6
Mother's own health
788
42.6
738
55.4
287
88.9
330
87.9
Mother's own weight
787
37.1
737
45.2
188
89.9
166
87.4
* All comparisons non-significant adjusting for simultaneous inferences by Bonferroni method [22]
To assess women's overall satisfaction with the information received during ANC, we compared the composite indexes. Overall, women in the experimental clinics were statistically significantly more satisfied with the general information they received (4.4 vs.4.0; 95% CI of the difference: 0.1 to 0.7), with information on how to recognize problems (3.0 vs. 2.4; 95% CI of the difference: -0.0 to 1.1) and what to do in an emergency (3.0 vs. 2.2: 95% CI of the difference 0.1 to 1.3), and on how to recognize problems and what to do (3.4 vs. 2.9: 95% CI of the difference 0.1 to1.3.)
Overall satisfaction was measured in the women's survey by three affirmative answers to the questions "If you were pregnant again, would you come back to this clinic?", "Would you recommend this clinic to a relative or friend for their antenatal checkups?" and "In general, are you satisfied/very satisfied with the ANC you have received in this clinic so far?." Women in both arms of the study showed very high levels of satisfaction, and there were no statistically significant differences between groups. The expressed levels of satisfaction were similarly high when women were asked direct and indirect questions. The overall satisfaction index showed that more than 90% of women in both ANC models said that they were "very satisfied".
In the case of the providers we did not have a sample; rather it was a census. We distributed the self-administered questionnaire to all health professionals of the clinics where the study was conducted. Providers were slightly more satisfied with the number of visits under the new ANC model (68.5 vs. 64.6%); less satisfied with the spacing between visits (60.9 vs. 69.5%); and substantially more satisfied with the time spent with each woman (85.9 vs. 69.5%). Concerning the information component, providers in general gave themselves higher scores in both ANC models (New ANC: 5.6 STD: 0.9; Standard ANC: 5.2 STD: 1.3) than women did. Finally, most of the health professionals surveyed in the new ANC model qualified the care they provided as "good" or "very good" (82.7%), while a higher proportion in the standard ANC clinics gave themselves that same score (95.1%).
Discussion
Women in the new ANC model clinics were, in general, as satisfied as their counterparts in the standard model. Furthermore, women in both arms were equally satisfied with waiting time and information provided about their health, tests during pregnancy, and treatment they might need. There were also no significant differences regarding what women worried about and whether the caregiver reassured them. Yet, women in the new ANC model were more satisfied with the time spent with the provider and with the information they received. Providers were more satisfied with the new ANC model with regards to number of visits, time spent with the patient, and information provided, but they were less satisfied with the spacing between visits. More providers rated the overall care provided under the standard model as good or very good than under the new ANC model.
Overall, these results show that both ANC models were equally well accepted by women and providers, suggesting that the adoption of the new antenatal care model would not face major obstacles derived from women or providers' perception of ANC and their satisfaction with it.
Within this framework, specific issues deserve special attention. In terms of the number of visits and spacing, the qualitative stage findings of our study [19] and those of several previous trials conducted to evaluate ANC models that reduced the number of visits [15,20,23-26] showed that more women in the intervention groups reported dissatisfaction with a reduced number of visits and longer spacing between them [20,25]. However, our study only demonstrated a trend towards patients' dissatisfaction with the changes introduced by the new ANC model, as no statistically significant differences between the trial arms were found. In another study conducted in a developing country results were similar to ours: there was no change in patients' satisfaction with a smaller number of ANC visits and longer spacing between them [27].
Still, our study findings suggest that number of visits and spacing are potential areas of concern for women. Providers could address these concerns by giving women information on the safety of these protocol changes, as was demonstrated by the results of the large WHO trial [18] and the systematic review of all randomized controlled trials [28]. Other needs that work as incentives for women to attend ANC clinics, such as socialization and social support, should be addressed through other activities that do not necessarily involve formal encounters with medical providers.
Regarding time spent with the provider, women in the new model had a higher level of satisfaction with the time spent with the provider than those in the standard model clinics, although the actual duration of the clinical encounter was similar. This positive impression may have resulted from an improvement in the quality of the patient-provider interaction. It is interesting to highlight that although waiting time was effectively reduced, women's satisfaction did not reflect the difference (Table 2).
One of the main goals of the new model was to strengthen the information component [16]. The fact that a larger proportion of women in these clinics perceived that their information needs were satisfactorily met even if there were only five visits to the clinics reveals that the new model was effective in reinforcing this aspect of care. In the Sikorski et al trial conducted in London, which achieved only a small reduction in number of visits, provision of information was also stressed; however, they did not find any difference in satisfaction among women in both arms of the trial [20].
The summary questions used to explore overall women's satisfaction with ANC showed surprisingly high levels among patients in both models, especially considering that women from the same clinics had expressed concerns about the quality of care during the focus groups and personal interviews conducted during the first stage of this study [19]. A hypothesis to explain this difference is that qualitative techniques capture the feelings of few more outspoken women and may provide a biased perception of the group. This could also be due to a "courtesy bias", which usually affects the answers to inquiries about satisfaction with care received, especially when women are asked in clinical settings [29]. In our study, qualitative techniques allowed to discriminate better among women with different levels of satisfaction than close-ended questions, especially the summary ones. This may be due to the wording of questions meant at exploring overall satisfaction; in fact, those that addressed specific issues (such as number of visits, spacing between them, information provided, etc) received answers with more variability.
There is another interesting hypothesis to consider as well. One study in Scotland found that pregnant women are fairly uncritical of health care, accepting whatever care they receive as appropriate. [30] The authors suggest that it would not be surprising to see high levels of overall satisfaction in a controlled study comparing two ANC models, and that it would be important to examine the differences between the two groups studied in their expressed preferences rather than the absolute magnitude of the expressed satisfaction. This was the case with our study, where we were able to differentiate women's satisfaction between the two models. However, women in the clinics of both models of ANC seemed to be equally uncritical.
Another difficulty in interpreting our findings derives from the variability in views and expectations originated by various cultural and socio-economic settings. In a study conducted in Chile, for instance, low-income urban women defined high quality as "being treated as a human being"; technical quality was not even mentioned [31]. Village women in Thailand identified inequalities of power fundamental to gender, class and ethnic relations as dimensions that crucially affect the client-provider interaction [32]. This was an important challenge in our study. The satisfaction questionnaire we used in each country was standardized with adaptations of terminology only and therefore did not provide any detailed clues about what aspects of ANC women of different cultural backgrounds appreciated more.
Women's satisfaction is a sensitive indicator that responds to changes in quality of care, even before changes in health status are detected, [33] but its measurement remains an important challenge. Qualitative methods allow women to reveal their feelings in greater depth than survey research methods [34]. In fact, most studies aimed at exploring women's views about quality of reproductive health care resort to interviews and focus groups [12]. However, results obtained with these techniques cannot be extrapolated and have low external validity. Yet, although data collected through questionnaires usually offer more superficial insights and do not reflect cultural nuances, when administered to a representative and large sample they can be safely extrapolated to the population from which the sample was obtained. In an attempt to overcome these limitations, we combined both methodologies [34].
Although our study makes important contributions to the area of users' perception on changes introduced into ANC models, it does not address methodological issues involved in the measurement of clients' satisfaction, which other authors have extensively addressed in observational studies [35,36]. In fact, we analyzed the differences between the perspectives of women in the intervention and control clinics, focusing only on those specific aspects that changed as a result of the introduction of the new ANC model (number and spacing of visits, information provided, etc.) Our study did not explore women's satisfaction with any other aspects of ANC such as technical quality, physical environment, access and continuity of the provider [37,38] that were not modified with the intervention, or the differences in users' satisfaction associated with ANC received in different types of facilities (i.e. private or public.) [39]
While users' perspective of quality of care has been assessed relatively often, the perspective of health professionals has been assessed occasionally at best [27,40]. In our study, while some degree of resistance to the new ANC model was expected, doctors and midwives did not have strong views against it. For instance, providers' satisfaction with the number of antenatal visits was similar in clinics of both arms of the trial. The reason for this may be that all providers worked at public health institutions, where the number of visits does not have a serious impact on their workload or income. Similar results were obtained in the study conducted in public hospitals in Harare, where the assessment showed that staff wished women made fewer visits to ANC clinics [27]. In the Sikorski trial [40] doctors were in favor of a reduced number of visits, but the average number under routine circumstances was much higher than in the four countries that participated in the WHO trial.
Regarding the information, our study confirmed an imbalance between women's expectations and providers' responses: providers scored themselves higher than their patients did in relation to the information they provide during antenatal check-ups. There appears to be a mismatch between doctors and nurses' perception of the quality and quantity of the information they provide and the users' needs. Furthermore, providers should be aware of the importance of meeting women's information needs during ANC visits, and thus be prepared to satisfy them.
In matters of overall satisfaction with ANC, although the proportion of providers that said care offered in their clinics was good or very good was high in both arms of the trial, those working in the standard ANC model clinics were more satisfied. This difference could be interpreted as a sign of discontent with the new ANC model.
Conclusions
Increasing attention is given to patients' views in health care evaluation. Policymakers and program managers should know that women's views are determinant in greater acceptance and sustained use of services. Additionally, health professionals' perspective needs careful evaluation before and during translating new care models into institutional protocols; being a conscious player in the process of change would certainly contribute to improving providers' commitment to their clinical work.
Competing interests
None declared.
Authors' contributions
Author 1 and 4 participated in the design of the study and coordinated it. Author 1 elaborated the different versions of the manuscript. Author 2 was the PI of the WHO randomized trial, coordinated the study's design and implementation, and made essential contributions to the different versions of the manuscript. Author 3, 6, 7 and 8 coordinated project implementation in each country. Author 5 participated in the study design and performed the statistical analysis. The rest of the authors participated in the WHO randomized trial and provided input to this specific component of the study. All authors read and approved the manuscript.
Pre-publication history
The pre-publication history for this paper can be accessed here:
Supplementary Material
Additional file 1
Women's Questionnaire - This file contains the English version of the questionnaire for users that was administered in the four participating countries after translating it into local languages.
Click here for file
Additional file 2
Providers' Questionnaire - This file contains the English version of the questionnaire for providers that was administered in the four participating countries after translating it into local languages.
Click here for file
Acknowledgements
This trial was supported by the UNDP/UNFPA/WHO/World Bank Special Programme of Research, Development and Research Training in Human Reproduction of WHO. Additional support was provided for the implementation of the study by: Municipal Government, City of Rosario, Argentina; Ministry of Health, Cuba; National Institute of Public Health, Mexico; The Population Council – Regional Office for Latin America and the Caribbean; Ministry of Health, Saudi Arabia; Swedish Agency for Research Cooperation with Developing Countries (SIDA/SAREC); Ministry of Public Health and Faculty of Medicine, Khon Kaen University, Thailand; Department for International Development (DFID) of the United Kingdom; Mother Care – John Snow Inc.; National Institute for Child Health & Human Development (NICHD), National Institutes of Health (NIH), USA; and The World Bank. For the preparatory phase: University of Western Ontario, Department of Epidemiology & Biostatistics, Canada; National Institute of Public Health, Norway; United Nations Development Programme; and the University of Uppsala, Department of Obstetrics & Gynaecology, Sweden.
We would like to thank specially the women and their babies who participated in this trial and the many doctors, nurses, and other staff of the clinics and hospitals that made the implementation of this project possible.
Special thanks are given to Drs G. Lindmark and V. Wong for their active participation as members of the Steering Committee and their continuous support during the trial, to Dr M. Koblinsky for her personal interest and support for this project, to Dr O. Meirik for his continued encouragement and support, to Dr D. Khan for editing the trial's Newsletter and to Ms Carol Peters for her help in the preparation of the manuscript.
Spontaneous premature ovarian failure presents most commonly with secondary amenorrhea. Young women with the disorder are infertile and experience the symptoms and sequelae of estrogen deficiency. The mechanisms that give rise to spontaneous premature ovarian failure are largely unknown, but many reports suggest a genetic mechanism in some cases. The small family size associated with infertility makes genetic linkage analysis studies extremely difficult. Another approach that has proven successful has been to examine candidate genes based on known genetic phenotypes in other species. Studies in mice have demonstrated that c-kit, a transmembrane tyrosine kinase receptor, plays a critical role in gametogenesis. Here we test the hypothesis that human KIT mutations might be a cause of spontaneous premature ovarian failure.
Methods and Results
We examined 42 women with spontaneous premature ovarian failure and found partial X monosomy in two of them. In the remaining 40 women with known 46,XX spontaneous premature ovarian failure we evaluated the entire coding region of the KIT gene. We did this using polymerase chain reaction based single-stranded conformational polymorphism analysis and DNA sequencing. We did not identify a single mutation that would alter the amino acid sequence of the c-KIT protein in any of 40 patients (upper 95% confidence limit is 7.2%). We found one silent mutation at codon 798 and two intronic polymorphisms.
Conclusion
Mutations in the coding regions of the KIT gene appear not to be a common cause of 46,XX spontaneous premature ovarian failure in North American women.
Background
Premature ovarian failure is a condition characterized by cessation of normal ovarian function before the age of 40. In most cases a precise mechanism for the ovarian failure is not identified; however, there is evidence suggesting a genetic mechanism in at least some cases. [1-4]
Normal ovarian function in reproductive-aged women is dependent upon the presence of a store of functional primordial follicles. This store develops from primordial germ cells, a small population of cells that differentiate from other cell lineages in very early embryonic life. [5] Premature depletion of primordial follicles is a known mechanism of premature ovarian failure.[6,7]
Studies in mice have demonstrated that c-kit, a transmembrane tyrosine kinase receptor, plays a critical role in hematopoeisis, melanogenesis, and gametogenesis. The c-kit receptor kinase and its ligand KL (kit ligand, stem cell factor) are encoded at the white spotting (W) and steel (Sl) loci of the mouse, respectively. Various mutations at the W and the Sl loci have been shown to cause anemia, pigmentation defects, and sterility.[8] Point mutations in the Kit receptor tyrosine kinase in mice can selectively impair fertility without inducing detectable abnormalities in hematopoesis or pigmentation.[9] Studies employing cultured mouse primordial germ cells have demonstrated that the Steel/Kit interaction is required for germ cell survival. [10-12]
These observations in mice suggest that mutations in the human KIT gene (MIM 164920) might be a cause of premature ovarian failure in women. The human KIT gene is located on chromosome 4 at map locus 4q12. Mutations in the human KIT gene have been identified as a cause of Piebaldism, a rare autosomal dominant disorder of melanogenesis characterized by patchy absence of pigmentation of the skin and overlying hair.[13,14] Here we test the hypothesis that specific human KIT mutations might be a cause of spontaneous premature ovarian failure even in the absence of detectable disorders of pigmentation or hematopoesis.
MethodsSubjects
Our institutional review board approved the study and all participants signed an informed consent and the procedures of the study were in compliance with the Helsinki Declaration. The study population comprised 42 women with premature ovarian failure and 10 normal healthy control women who had regular menses and proven fertility. Referring physicians made the diagnosis of premature ovarian failure based on the following criteria: development of at least 4 months of amenorrhea before age 40 associated with two serum FSH levels above 40 IU/L (drawn at least 1 month apart). Women with premature ovarian failure as a result of surgery, radiation, chemotherapy, or known karyotype abnormalities were not included in the study. Patients who had not previously had a karyotype analysis were included pending our analaysis. There were 34 Caucasian, 6 African American and two Hispanic women. The median age at the onset of menstrual irregularity was 25 years (range 14 to 39, 25th percentile 18.5, 75th percentile 34). Five women had a family history of premature ovarian failure (two patients had sisters with premature ovarian failure, one patient's mother had premature ovarian failure, one had a paternal aunt with premature ovarian failure, and one patient had a paternal aunt, a grandmother, and a great-grandmother who were thought to have the condition). All patients underwent a history and physical examination and laboratory screening to confirm the diagnosis of premature ovarian failure.
Cytogenetics
Karyotypes were obtained on all patients. Peripheral blood specimens were cultured for 72–96 hours using standard methods and methotrexate-thymidine synchronization.[15] Harvesting and GTW-banding were also performed according to standard cytogenetic protocols.[16] Fifty metaphases were examined for each patient and scored for numerical or structural chromosome abnormalities. Band resolution was at the 600 band stage or higher.
Fluorescence in situ hybridization (FISH) was performed on metaphases from cultured lymphocytes of two patients with abnormal G-band findings. Chromosome and probe denaturation, hybridization, detection, and counterstaining were all done according to standard protocols as recommended by the manufacturer.
DNA extraction and polymerase chain reaction (PCR)
Genomic DNA was extracted from peripheral blood using standard procedures. Specific primers for PCR amplification of the exons were designed based on the human KIT genomic sequence (Genebank accession number AC006553, Human chromosome 4) (Table 1). In exon 21 the reverse primer is located in the 3' noncoding region, in all other cases the primers are located at introns flanking each exon. The PCR amplification was carried out in a total volume of 25 μl reaction mixture containing 1.5 mmol/l of MgCL2, 0.2 mmol/l of each dNTP, 50 ng of genomic DNA, 0.5 μm of each primer, and 1.25 IU of Taq DNA polymerase per manufacturer's instructions (Invitrogen, Carlsbad, CA). The PCR began with an initial denaturing at 95°C for 5 minutes. Thereafter the cycling profile consisted of 35 cycles of denaturing at 95°C for 30 seconds, annealing of the primer pair at the appropriate temperature for 30 seconds (Table 1), and then extension at 72°C for 60 seconds. This was followed by a final stabilization step of 10 minutes at 72°C. After PCR, 5 μl of the amplified product was examined by 1.5% agarose gel electrophoresis.
Primer sets and conditions for amplification of human KIT exons. Forward and reverse primers flanking each of the 21 exons are shown. The size of each exon-coding region, the size of each PCR product generated by each set of primers, and the respective annealing temperatures are indicated. Sequences marked by a * are derived from Spritz et al..[28]
Analysis of DNA polymorphism was conducted using a method of "Cold SSCP" as described previously.[17] Briefly, the PCR products (4 μl) were mixed with SSCP sample buffer (6 μl, formamide with 0.05% bromophenol blue and xylene cyanol). After denaturing at 95°C for 5 minutes, the DNA samples were placed on ice immediately and loaded onto precast 20% TBE acrylamide gels (Tris Base, Boric Acid, EDTA, 4% glycerol, and 20% acrylamide, Invitrogen, Carlsbad, CA). Gel elecrophoresis was run in 1 X TBE buffer at 200 V with the circulator set at 4°C (using a Penquin Water-cooled Dual-Gel Electrophoresis system attached to a thermostatically controlled refrigerated circulator, Amersham & Pharmacia, Uppsala, Sweden). Gels were stained with ethidium bromide (1 μg/ml) for 20 minutes at room temperature. DNA bands were visualized and photographed under UV light (340 nm).
Direct DNA sequencing
PCR products were either purified using ExoSAP-IT (USB Co, Cleveland, OH) or subcloned into a TA cloning vector (Invitrogen, Carlsbad, CA) per the manufacturer's instructions. DNA sequencing was conducted by PCR (26 cycles of 96°C, 10 seconds; 50°C 5 seconds; 60°C 4 minutes) using a dRhodamine terminator cycle sequencing kit (PE Applied Biosystems, Foster, CA) and 10 μl of purified PCR product or 1 μg of purified plasmid DNA. The sequencing samples were separated by 5% Long Ranger gel electrophoresis on the ABI Automatic 310 Sequencer (Applied Biosystems, Foster, CA)
ResultsClinical
None of the 42 women were found to have clinical findings to suggest Turner syndrome. None of the patients had clinical or laboratory findings to suggest a stem cell deficiency in melanogenesis or hematopoiesis. Two women were found to have partial monosomy of the X chromosome that was felt to be related to the presentation of spontaneous premature ovarian failure. For this reason they were excluded from further analysis. These findings highlight the usefulness of careful cytogenetic screening of women with spontaneous premature ovarian failure. The remaining 40 women had a normal 46,XX karyotype analysis.
One woman of normal stature had a karyotype analysis showing: 46,X, ?rec(X) dup(Xp) inv(X)(p11.4 q24). ish rec(X) dup(Xp) inv(X)(p11.4 q24) (KAL++). In all cells examined there was one normal X chromosome and one X chromosome that had most of Xp (Xp11.4→Xpter) replacing a portion of Xq (Xq24→Xqter). This rearrangement was confirmed with a FISH study. A probe to the Kallman syndrome region (Oncor Inc., Gaithersburg, MD) at Xp22.3 was hybridized to patient's metaphases and showed signal on both ends of the abnormal X chromosome (as well as the normal X chromosome) confirming the G-band impression that both ends of this chromosome was comprised of Xp. In summary, the patient had partial trisomy Xp and partial monosomy Xq as demostrated by G-banding and FISH.
A second woman of normal stature had a karyotype analysis showing: 46,X, der(X) t (X;13) (q22.3;q14.1). All cells examined in this patient had extra material on Xq which was 13q in origin. FISH was performed using whole chromosome probes to the X and 13 (Oncor, Gaithersburg, MD). The X probe showed signal covering the entire normal X as well as Xp and the proximal half of Xq on the derivative chromosome. Whole chromosome probe 13 showed signal on the normal 13 as well as on the derivative X chromosome. These results are consistant with partial monosomy Xq from q22→q terminus and partial trisomy 13 from q 14.1→q terminus. Given the lack of symptoms/characteristics associated with trisomy 13 (multiple congenital anomolies with severe cognitive delays) we hypothesize that the derivative X chromosome is the inactivated X, thus sparing the patient the expected phenotype.
Screening of the KIT gene
All the exons of the gene (Figure 1A) were successfully amplified by PCR in all 10 controls and all 40 patients with 46,XX spontaneous premature ovarian failure. Except for product 21, which carried only the coding region, each PCR product contained one complete exon. All PCR products were of the expected size. As shown in Figure 1B, the size of the 21 PCR products ranged from 198 to 438 bp and all appeared as a single clear band.
Human KIT organization and PCR amplification of its exons. A. Schematic representation of exon-intron map of human KIT. It is composed of 21 exons and 20 introns. Vertical bars and horizontal solid lines represent the exons and introns, respectively. The // indicates that these introns are not in scale. Arabic numbers above each of the vertical bars indicate the exon number. Modified from Giebel et al.[27]. B. Analysis of PCR products of the human KIT gene. Human genomic DNA spanning each of the human KIT exons were amplified using specific primer pairs. PCR products for each of the exons were separated by 1.5% agarose-electrophoresis and stained with ethidium-bromide. The Arabic numbers above each of the PCR products represent the exon number. DNA size markers (φ X174 RF DNA/HaeIII) are shown at the left of the panel (M).
The PCR products were screened for mutations by SSCP analysis. We found mobility shifts in the PCR products from two of the patients and none of the controls. As shown in Figure 2, patient 4 had a mobility shift in the products from two exons, 10 and 16; patient 5 had a mobility shift in the product from exon 17. In patient 4 both the polymorphisms were determined by sequencing to be within intron regions, and would thus not change the sequence of the c-KIT protein. In the polymorphism 71 nucleotides upstream of exon 10, substitution at nucleotide 71 replaced the normal T with an A [IVS9-71T>A]. In the polymorphism downstream of exon 16 there was a deletion of 5 T nucleotides from within a normal series of 22 consecutive T nucleotides [IVS16 +66_+70delTTTTT] (Figure 3). In patient 5 sequencing uncovered a mutation at codon 798; however, the mutation C798T was silent with no change in amino acid sequence (Isoleucine).
SSCP analysis of KIT PCR products The PCR products were separated on 20% TBE-acrylamide gel electrophoresis and stained with ethidium bromide. SSCP analysis of samples from three normal control women and three patients with spontaneous premature ovarian failure are shown for exon 10,16, and 17. DNA mobility of PCR products from exons 10 and 16 was altered in one patient (Lane 4). One patient showed an alternation of DNA mobility in the PCR product from exon 17 (Lane 5). Also shown is the DNA mobility of samples from normal women (Lanes 1–3) and from a representative patient with no changes in DNA mobility (Lane 6).
DNA sequence analysis of a KIT intronic region variant downstream of exon 16 Electropherogram displaying the sequence of the KIT variant of the intronic sequence downstream of exon 16 compared to the wild-type sequence. Arrows indicate a deletion of 5T nucleotides in the variant as compared to a normal series of 22 consecutive T nucleotides in the wild-type.
Restriction Fragment Length Polymorphism Analysis
The C798T substitution in exon 17 of patient 5 generates a new restriction site for SspI (AAT ATT). To determine if this mutation was homozygous or heterozygous we analysed restriction fragment length polymorphism (RFLP) on PCR products. As shown in Figure 4, the mutant PCR product (390 bp) for exon 17 was digested with SspI to yield two shorter DNA fragments only (207 and 183 bp). These data suggest that this mutation is homozygous. We did not perform RFLP analysis on the mutant PCR products from exons 10 and 16 because no new restriction sites were generated by the mutations. Other methodologies such as forced RFLP and sequencing of multiple cloned products could be applied to determine if these mutations are homozygous or heterozygous.
Restriction fragment length polymorphism analysis of a KIT variant in exon 17 using Ssp I Undigested DNA and wild-type DNA digested with Ssp I gives a single band of 390 bp. Homozygosity for the variant yields two fragments of 207 and 183 bp. A heterozygote carrier would have demonstrated all three of the fragment lengths. Therefore, this patient is homozygous for this variant.
Discussion
Primordial germs cells originate from a small number of progenitor cells that arise in very early development. Thereafter, complex factors that are not well understood direct their proliferation, survival, migration to the genital ridge, differentiation into oocytes, and the subsequent expenditure of oocytes via ovulation or apoptotic death.[5,7][18] Mutations or deletions in any of the genes involved in this complex process could theoretically be a cause of premature ovarian failure on the basis of germ cell deficiency, either due to a deficient initial endowment of primordial follicles or an accelerated expenditure of this endowment.
Evidence that genetic factors play a role in some cases of premature ovarian failure has been around for a long time [19,20]. However, the small pedigrees necessarily associated with this condition make it extremely difficult to conduct genetic linkage analysis that can be used to identify candidate genes.[4] One approach to this dilemma has been to evaluate candidate genes based on existing knowledge of ovarian physiology, as has been reported in the case of inhibin.[21] Other candidate genes are those that when mutated are known to cause germ cell deficiency in other species, for example the diaphanous gene of drosophila.[22] Numerous growth factors have been associated with primordial germ cell behavior during migration in mice using in vitro studies, but in most cases it is unknown whether these factors function in vivo.[5] One exception is the Steel/Kit interaction. Steel and W are genetic loci identified in strains of mice that exhibit sterility on the basis of germ cell deficiency. The products of these two loci are now known to be the tyrosine kinase receptor Kit (W) and its ligand Steel factor (Steel). It was on this basis that we set out to examine patients with premature ovarian failure for mutations in KIT.
The SSCP technique has been used successfully for mutation searching in other genes. [23-26] Using the PCR-SSCP technique we were able to detect only one polymorphism in the entire coding regions of the KIT gene, and this polymorphism did not change the amino acid sequence at codon 798. Therefore, these results suggest that mutations in the coding regions of the KIT gene are unlikely to be a common cause of premature ovarian failure. The upper 95% confidence limit of the proportion 0/40 is 7.2% (no mutations found in 40 women with 46,XX spontaneous premature ovarian failure). The DNA polymorphisms that we identified within intron regions upstream of exon 10 and downstream of exon 16 would not change the amino acid sequence of the c-KIT protein. However, we cannot exclude the possibility that these changes in nucleotides that are close to exons might alter c-KIT transcription, or that mutations in the regulatory regions of the KIT gene might be a cause of spontaneous premature ovarian failure. Furthermore, it should be noted that SSCP methodology does not detect all mutations of PCR products. Although in this report using this technique we were able to detect a single nucleotide mutation in a PCR product of 390 bp, most laboratories would suggest that SSCP detects approximately 70 to 80% of mutations. Screening a larger group of patients with a more sensitive technique such as denaturing high-performance liquid chromatography might detect more mutations.
Conclusions
We were unable to detect any significant mutations in the entire coding region of the KIT gene in 40 patients with 46,XX spontaneous premature ovarian failure. Therefore, mutations in the coding regions of the KIT gene do not appear to be a common cause of 46,XX spontaneous premature ovarian failure in North American women.
Competing interests
None declared.
Authors' contributions
K.S., Z-B.T, and K.V. participated in the design of the study and carried out the molecular genetic laboratory studies. K.S. and Z-B.T. drafted the manuscript. V.H.V. recruited the patients and helped characterized them clinically. L.M.N. conceived the study, participated in its design, coordination, analysis and manuscript preparation. All authors read and approved the final manuscript.
Pre-publication history
The pre-publication history for this paper can be accessed here:
Acknowledgements
We thank Jeanne Meck, Ph.D. and the staff of the Georgetown University Hospital Cytogenetics Laboratory for classical and molecular cytogenetics expertise. We thank Constantine Stratakis, M.D., D.Sc. for reviewing our manuscript prior to submission.
The aim of this study was first, to investigate whether women starting oral contraceptive (OC) use at a young age and before first birth have an increased risk for breast cancer and second, to report difficulties encountered in studying long-term health impacts of medical technologies.
Methods
Breast cancers occurring up until 1997 among 37153 Helsinki students born between 1946 and 1960 were identified by record linkage from the Finnish Cancer Registry; for each cancer case, five age-matched random controls were picked from the same student population. Those who had used the Helsinki Student Health Service (HSHS) at least three times (150 cases and 316 controls) form the final study subjects. Data on OC use and background characteristics were collected from patient records, and data on live births were derived from the population register. Odds ratios (OR) were adjusted for number of births, smoking and sports activity.
Results
Compared to the few non-users, OC users had a higher risk of breast cancer: the adjusted OR was 2.1 (95% confidence interval 1.1–4.2). Among OC users, no statistically significant differences in risk of breast cancer were found in regard to starting age or first birth, but small numbers made confidence intervals wide. Even though we had chosen students to be our study group, the population turned out to be unsuitable to answer our research question: most women had started their OC use old (at the age of 20 or later) and there were very few unexposed (almost all had used OC and before their first birth).
Conclusions
Because adoption of the modern pattern of OC use was not common among students, it is unlikely that the impact of early and extended OC use can be studied before 2010, when women born in the 1960s are 40 to 50 years old.
Epidemiological studies give varying results in regard to whether oral contraceptive (OC) use is a risk factor for breast cancer. A thorough meta-analysis based on epidemiological studies up to the mid-1990s concluded that women who have used combined OCs in the past 10 years are at a slightly increased risk of having breast cancer, although the excess cancers diagnosed tended to be localised [1,2]. Some of the later studies concluded that OC use is a risk factor for breast cancer [3-10], or is for some subgroups [11,12], while other studies have not [13-16]. Young women having used the pill for a long time before their first pregnancy have been identified as a potential risk group [6,14,17,18], but currently there are not enough studies to prove or disprove the increased risk of breast cancer [19].
OCs were introduced into clinical use in the early 1960s. Their use spread rapidly, but in the beginning they were used mainly by women older than 20 years, after the first birth, and intermittently; that is, for short periods of time [20]. Over time the OC use patterns have radically changed in many countries: start in the teen-years, use for a long time before the first pregnancy, and continuous use have increased and the type of pill has changed [6,20-23]. In Finland, 7% of 16-year-old girls and 22% of 18-year-old girls used OCs in 1981; by the year 1989 figures had increased to 17% and 38% [24] and have remained high [25]. In 1997 one third of girls had experienced their first sexual intercourse by the age of 16 years [26] and in 1998 the mean age at having the first child was 28 years [27]
The original purpose of this study was to see whether women who started OC use young and before their first birth have an increased risk for breast cancer before menopause. The study was done among Finnish students who used the Helsinki Student Health Service (HSHS), and cancers were identified by record linkage to the Finnish Cancer Registry. However, as shown below, we ended up with a low study power, and a second purpose of this article is to report the reasons to it to help others designing studies on OC use and breast cancer.
Subjects and methods
This is a case control study among defined student populations. The first study base consisted of all Finnish female Helsinki University students from 1965 onwards who were born in the 1946–1960 period (n = 28109) and were customers of the HSHS. The second study base was female students of other higher educational institutions in the Helsinki area who used the HSHS in 1980 or later (n = 9044). Most Helsinki students used the HSHS because belonging to the programme was compulsory and everyone had to pay the health service fee, and municipal public health care, with the exception of emergencies, was restricted to municipality residents. Students could use private services. The HSHS had very low fees for visits, and once students entered the system they usually used its services. In 1971, the first gynaecologists were employed by the HSHS.
To identify those who later got breast cancer, the data on the women in the study bases were linked to data in the Finnish Cancer Registry for 1967 to 1997 using unique personal identification (ID) numbers; since 1967 these ID-numbers have been given to everyone who has lived in Finland. The Finnish Cancer Registry has operated since 1953 and is based on compulsory notifications by physicians who diagnose a malignant disease; the data in this register were completed with information from hospitals, laboratory records and death certificates. The notification rate and accuracy of the registry are high [28]. Through the linkage, 396 cases were found. For each cancer case, five random controls (total 1980), matched for year of birth, were chosen from the same study bases.
A requirement of at least three visits to HSHS, other than the initial health check-up or a visit to a dentist or an ophthalmologist, was set to exclude women who had sought their health care and OCs from other sources. Number of visits was determined by manually searching the HSHS archives; 153 cases and 314 controls had at least three visits. For 14 cases no controls remained. For them, one new control, born in the same year, was randomly picked from the 388 leftover controls (i.e. controls whose matched case had not used services at least three times). The dates of live births and women's vital status were obtained by record linkage to the population register. Cases whose first HSHS visit was after the breast cancer diagnosis (n = 3) and controls who had died before or whose first visit was after the time of the diagnosis of breast cancer in the case (n = 12) were excluded. Altogether 150 cases and 316 controls remained for analysis. Two cases had a carcinoma in situ, others a malignant tumour; 80% were ductal carcinomas. By the end of 1997, 17 women in the case group (11%) had died of breast cancer.
Data on exposure to OCs, reproductive history and background characteristics were collected from the HSHS patient records. A trained research nurse blind to the case-control status abstracted the data. To identify OC use, a list of all OCs on the Finnish market in the 1962–1985 period was made with the help of old drug catalogues. The HSHS kept continuous patient records in which all prescriptions were marked. Information on OC use while an HSHS customer was searched for from physicians' notes and referral notes to the laboratory for PAP smears (at that time, PAP smears were recommended to be regularly done for OC users). In 1971 the HSHS introduced health status questionnaires to be filled in at the first visit, including a standard question on current and past OC use. Overall, 29% of women (25% of cases and 30% of controls) had a note indicating OC use before entering HSHS. To describe the women, the duration of OC use was calculated by summing up the months from OC prescriptions, and the estimates given in the records on prior use and on OCs prescribed outside the HSHS.
Using all available data in HSHS documents, the year of start of OC use was calculated. Exposure data were collected for all years up to the until one year prior to the breast-cancer diagnosis of the case. For each woman the following exposure data were constructed: OC use, age starting OC use, and OC use before first live birth. Use of OCs was classified into three classes: no (no, likely not), yes (yes, likely yes), and not recorded. Women with any documented use were included in the 'yes' class; the shortest duration was one month. The group 'not recorded' was a mixed bag of women for whom there was some indication of OC use, but no details of use ware available. It included women who had a private gynaecologist or had received an OC prescription, but who in later visits indicated no OC use.
For one analysis, OC users were classified into the following group: 1) started before the age of 23 years and before first pregnancy, 2) other users (started older or after first pregnancy or information on these characteristics was lacking), 3) others. The cut-off point of 23 years was chosen to get a reasonable number of women in each group.
Odds ratios (OR) and 95 % confidence intervals (CI) were estimated using conditional logistic regression [29] with GLIM 4 software [30]. The following data on potential confounders were available in the patient records: study discipline, weight and height, menarche age, menstruation cycle (average length and duration of bleeding), smoking, alcohol use, participation in sports (either competitive or other) at the first visit, and pregnancies at any time while using the HSHS. Data on live births were obtained from the population register. The odds ratios were adjusted for smoking (never, ever, unknown), sports (no sports, sports, unknown), and for parity and age at delivery.
ResultsBreast cancer
The cases and controls were similar in terms of most studied background characteristics, but there were more controls who were active in sports (Table 1). The proportions of women having had births were relatively similar among cases (73 percent) and controls (67 percent) and the cases and controls had their first child about at the same age (Table 1). In addition to the variables shown in Table 1, we compared the cases and controls in terms of how soon they entered the HSHS after starting their studies (for 60 % of cases and 64 % of controls the difference between the first visit and starting year was less than 2 years), last visit to the HSHS (56% and 53% before 1980), study institute (University of Helsinki for 90% and 93%), distribution of weight at first visit (similar), distribution of height (similar), distribution of menarchial age (similar), distribution (similar) and mean (both 28.5 days) of the length of menstrual cycle, and distribution (similar) and mean (5.1 and 5.0 days) of the duration of menstrual period. Record linkage to the population register showed that cases somewhat more often had children (not statistically significant).
Percentage distributions or means (standard deviation, SD) of cases with breast cancer and their controls by study characteristics in the HSHS cohort
Study characteristic
Percentage or mean
Cases n = 150
Controls n = 316
Year of birth, %
1946–1950
48
47
1951–1955
29
30
1956–1960
23
24
First visit at the HSHS, %
1965–1969
19
14
1970–1974
39
41
1975–1995
43
45
Smoking, %
Current
23
29
Quitted
3
6
Never
58
55
No information
16
10
Alcohol use, %
Moderate or heavy
17
18
Slight
37
38
No
9
15
No information
37
29
Sports, %
Active
21
35
No
53
44
No information
26
21
Weight (kg), mean (SD)
57.0 (7.1)
56.2 (7.6)
Height (cm), mean (SD)
166 (5.2)
165 (5.6)
Age at menarche (in years), mean (SD)
12.8 (1.2)
12.7 (1.2)
Contraception *, %
94
88
Any births ‡, %
73
67
Age at first delivery § (in years), %
18–24
17
20
25–29
48
49
30–34
29
26
35–41
5
5
Mean (SD)
28.1 (3.7)
27.7 (4.1)
*Any method recorded by the HSHS at any time. ‡ Births in Central Population Register before year of case's diagnosis. §Delivery before year of case's diagnosis.
Most study women had used OCs while customers of the HSHS or before that, and started the use at the age of 20 or later. The mean age was 22.6 years (SD = 3.2) among cases and 22.3 (3.2) among controls. Starting young increased by time: 3% of the women born between 1947 and 1950 (n= 220), 10% of those born between 1951 and 1955 (n = 138), and 24% of those born between 1956 and 1960 (n = 110) started before 20 years of age. Most women had started their OC use before first birth. 80% of OC users had made last visit to HSHS when they were 25 years or younger. The length of OC use as noted in the records was 3 years or less among most women, with means 2.6 years (SD = 2.5) for cases and 2.9 (2.3) for controls.
Compared to those few women who had not used OC, OC users had a higher risk of breast cancer (Table 2). Adjustment for confounders did not abolish the difference. Among OC users, those who had started young did not have a higher risk of breast cancer than those who had started older (25 or more) (Table 3). There was also no difference in the risk of breast cancer in regard to whether the use was started before or after the first birth, or in regard to having not given birth, Table 4. There was no difference in the risk of breast cancer among women who had been pregnant while in the HSHS in regard to whether they had used OCs before or only after the first pregnancy.
Odds ratios (OR) with 95% confidence intervals (CI) for breast cancer related to the use of oral contraceptives (OCs) while customer of the HSHS or before that.
OC user
Cases
Controls
Crude
Adjusted*
n
%
n
%
OR
OR
95% CI
No †
15
10
57
18
1
1
Yes
114
76
222
70
2.1
2.1
(1.1, 4.2)
Not recorded
21
14
37
12
1.9
1.6
(0.6, 4.0)
Total
150
100
316
100
*Adjusted for parity and age at delivery (no births before year of case's diagnosis, birth before 25 years, other), sports (no sports, sports, unknown), and smoking (never, ever, unknown). † Reference group.
Odds ratios (OR) with 95% confidence intervals (CI) for breast cancer related to age at start of oral contraceptive (OC) use‡.
Age at start of OC use (in years)
Cases
Controls
Crude
Adjusted*
n
n
OR
OR
95% CI
16–19
11
36
0.6
0.5
(0.2, 1.6)
20–24
77
140
0.9
0.9
(0.5, 1.9)
≥ 25†
21
40
1
1
Age missing
5
6
..
..
‡ Women missing information about OC use or (21 cases and 37 controls) or not having used OCs while in the HSHS (15 cases and 57 controls) were excluded. *Adjusted for parity and age at delivery (no births before year of case's diagnosis, birth before 25 years, other), sports (no sports, sports, unknown), and smoking (never, ever, unknown). † Reference group.
Odds ratios (OR) with 95% confidence intervals (CI) for breast cancer associated with use of oral contraceptives (OCs) in relation to the first birth‡.
Use of OC
Cases
Controls
Crude
Adjusted*
n
n
OR
OR
95% CI
User after 1st birth †
7
12
0.8
0.8
(0.2, 2.5)
User before 1st birth
74
143
1.1
1.0
(0.5, 1.7)
User, no births
32
62
1
1
Order of use and births** not known
1
5
‡ Women missing information about OC use or (21 cases and 37 controls) or having not used OCs while in the HSHS (15 cases and 57 controls) were excluded. *Adjusted for sports (no sports, sports, unknown) and smoking (never, ever, unknown). † Reference group. **Chronological order of the first delivery and start of OC use.
When OC users were compared in terms of the time interval from the first use to the first birth, no statistically significant differences in the risk of breast cancer were found. When users having the time interval 1 to 5 years were used as the reference category, the adjusted odds ratio for the 6- to 10- year interval was 0.7 (95 % CI 0.4, 1.3) and for the 11- to 20- year interval it was 0.8 (0.3, 1.9).
When we compared OC users by the age at start of OC use and relation to first pregnancy, no statistically significant differences were found. Women who started young (less than 23 years) and before their first pregnancy did not have a higher risk (OR 1.1, 95% CI 0.7, 1.8) than other OC users.
Feasibility
University students were chosen as the study population because we assumed them to be the first to adopt the modern pattern of OC use: starting young and having long use before first birth. Furthermore, many of them had used only one service that has feasible archiving system, as well as computerised registers with unique personal ID-numbers, making record linkage easy. We matched the cases and controls by age and they all were university students. Our assumption about there being many young starters in this data set turned out to be wrong: only among the youngest cohort, born in the latter half of the 1950s, was OC use before the age of 20 years common. We also had very few non-users. OC use seems to have spread among students very rapidly and almost everyone at lest experimented with them.
Discussion
The hypothesis that OC use is related to breast cancer in young age received limited support. But no clear relation was found with starting age or starting before first birth. We could not study the impact of the length of OC use because we had no data on the women's OC use after their student years. A side finding was that cases were less likely to have been active in sports, either competitive or recreational, at their first visit. This is in accordance with earlier literature [30,31].
Several methodological problems, however, weaken these findings. The main problems were, as described above, small numbers of young starters, few non-users, and few starters after first birth. Furthermore, because of our criterion of at least 3 visits to HSHS, we very likely proportionally selected out more non-users than users of OC: one of the important reasons for HSHS visits was to get OC. Because more control (84%) than case women (63%) were excluded because of this limit, we were likely to proportionally exclude more non-users among the controls. Accordingly our analysis may underestimate the relation between OC use and breast cancer. In hindsight, the drawbacks of the potential bias caused by this criterion surpass the benefits, and we should not have used such a criterion.
A strength of our study was that the data on exposure were based on patient records. Because of the eligibility requirement of at least three visits, it is likely that we have relatively complete information on OC use while the women were students in Helsinki. The information on their status before they entered the HSHS was based on women's interviews as noted in the patient records, and may be incomplete. These inaccuracies in exposure data weaken the comparisons between different user groups. The comparison between cases and controls was made with the assumption that the omissions were similar in these two groups. We excluded women who had breast cancer before their first HSHS visit, but women who had it between the first and the third visit were included. Inclusion of these women makes our finding of an increased risk for breast cancer due to OC use conservative. On the other hand, if the women classified as 'OC use not recorded' were in fact non-users, the association between OC use and breast cancer becomes much smaller.
Most earlier case-control studies on OC use have relied on women's (or their relatives') memory of OC use. Compared to such an approach, our information on exposure is more accurate in terms of starting age and relation to the first birth (and type of drug, not analysed here). But we could not adjust for current OC use, which has been found to be a risk factor for breast cancer [1]. Nor do we have data of the length of OC use.
Because adoption of the modern pattern of OC use was not common among students, it is unlikely that the impact of early and extended OC use can be studied before 2010, when women born in the 1960s are 40 to 50 years old. Studies before that have to focus on either the impact on very early breast cancer, which is rare and may be different (e.g. due to genetic predisposition) from breast cancer emerging around menopause, or focus on women whose OC-use patterns are very different from those of other women. Such women are likely to differ from others on many dimensions and confounding by unknown factors may be large.
The case of OC use well illustrates the difficulty of evaluating the long-term health impact of medical technologies. The true impacts may be revealed only after decades of use, and often only after the technology is no longer in use. In the case of OCs, the technology itself is still in wide use, but the use patterns, user groups and chemical composition of the drugs have changed. The results of our study concern a population of women who started OC use before their first birth and pregnancy, who were much older than the current OC-user population, and who used OCs differing in composition and chemical strength from those used today.
Conclusions
The hypothesis that OC use is related to breast cancer in young age received only limited support. Our study women commonly used OC before first birth, but started their use at a later age. The impact of young starting age in this population can be studied only later, e.g. after a decade or so.
Competing Interests
None declared
Authors' Contributions
EH: Planned the study and was responsible for data collection and reporting.
TL: Did the statistical analyses
EP: Participated in the planning of the study and statistical analyses
DA: Participated in the planning of the study and data collection
TH: Participated in the statistical analyses
All authors participated in writing the manuscript and have read and approved the final version.
Pre-publication history
The pre-publication history for this paper can be accessed here:
Acknowledgements
We thank Susanna Kaipainen and Nina Apter for their help in collecting data, and Kati Tanninen for her secretarial help.
Abnormal uterine bleeding (AUB) is a common problem that affects one in five women during the pre-menopausal years. It is frequently managed by family physicians, especially in northern, rural and isolated areas where severe shortages of gynecologists exist.
Methods
We surveyed 194 family physicians in northern, rural and isolated areas of Ontario, Canada to determine their educational and resource needs for the management of AUB, with a specific focus on the relevance and feasibility of using clinical practice guidelines (CPGs).
Results
Most physicians surveyed did not use CPGs for the management of AUB because they did not know that such guidelines existed. The majority were interested in further education on the management of AUB through mailed CPGs and locally held training courses. A major theme among respondents was the need for more timely and effective gynecological referrals.
Conclusion
A one-page diagnostic and treatment algorithm for AUB would be easy to use and would place minimal restrictions on physician autonomy. As the majority of physicians had Internet access, we recommend emailing and web posting in addition to mailing this algorithm. Local, hands-on courses including options for endometrial biopsy training would also be helpful for northern, rural and isolated physicians, many of whom cannot readily take time away from their practices.
Background
Abnormal uterine bleeding (AUB) is a common problem that affects one in five women during the pre-menopausal years [1]. The bleeding may be unmanageable, disruptive to daily activities and damaging to self-esteem and personal relationships. Primary care physicians (family doctors) are often responsible for the management of women with heavy bleeding, especially in northern, rural and isolated areas where severe shortages of gynecologists exist. It is vital to ensure that these physicians have the tools to provide optimal management. This study addresses the educational and resource needs of family physicians in northern, rural and isolated areas of Ontario, Canada with respect to the management of AUB with a specific focus on the relevance and feasibility of using clinical practice guidelines (CPGs).
Methods
We developed a five-page questionnaire entitled "Management of Heavy Vaginal Bleeding: Northern and Rural Primary Care Physicians" composed of four parts, of which parts: I. Physician Information, III. Clinical Practice Guideline Use and IV. Education and Resource Needs Assessment, are relevant to this paper (see accompanying article "Management of Abnormal Uterine Bleeding by Northern, Rural and Isolated Primary Care Physicians: Part I: How Are We Doing?" for survey instrument – Appendix 1). We collected information on physicians' sex, age, history of clinical practice, Internet access (for availability of information), previous obstetrical experience (possible indication of comfort level with managing gynecological problems) and Canadian College of Family Physicians (CCFP) membership status (access to continuing medical education resources). Questions on CPG use were developed on the basis of a literature review in June, 2000 on clinical practice guideline dissemination and adherence. Medline keywords were clinical practice guideline, adherence and family physician [2-8]. The education and resource needs assessment consisted of both "mark your choices" and open-ended questions.
A package including a cover letter, stamped return envelope and the questionnaire was mailed to 676 physicians in Northern Ontario. Reminder letters and replacement questionnaires were sent 3 and 7 weeks respectively after the original, using a modification of the Dillman (TDM) approach [9]. Due to incorrect address or specialty information (i.e. respondents were not family physicians), fifty-four surveys were returned. From the remaining 622 potential participants, there were 194 completed responses (Response Rate = 31%).
The Statistical Package for the Social Sciences (SPSS, version 10) was used for statistical analysis. Descriptive analyses of all continuous and categorical variables were conducted. Chi square tests of independence, independent sample t-tests and analyses of variance (ANOVA) investigated whether responses were dependent on physician variables.
ResultsPhysician information
Detailed physician information has been reported in the accompanying article. Significantly, the average age of the responding physicians was 43 years, with an average of 15 years in family practice. In accordance with the gender distribution of physicians in Ontario, 37.5% of the respondents were female, 62.5% were male, with males being significantly older and with more years in family practice than females. All physicians practiced in Northern Ontario, with 45.1% also designating their practice population as "rural" and 17.5% as "isolated". These categories were not mutually exclusive (i.e. a physician could practice in an area that is rural and isolated). Half of the respondents were practicing in areas with populations of less than 15 000 people, with the remainder in larger towns (25%) or small cities (25%). Three quarters (75%) were members of the Canadian College of Family Physicians (CCFP) with a higher proportion of women being members than men. Almost all (92.7%) of respondents had Internet access either in their homes or offices.
Clinical practice guidelines
Most physicians (91.2%) did not use a CPG for the management of AUB. Reasons included not knowing that such a guideline existed (63.3%), not being familiar enough with it to use it (19.8%) and feeling confident managing AUB without a CPG (14.7%) (Table 1). There were physicians who were "somewhat" (13.0%), "fairly" (10.4%) or "quite/very" (2.0%) familiar with a guideline for the management of AUB but chose not to use it. They thought that the guideline would "require a change in practice habits", "restrict their autonomy", was "not applicable to their practice" or was "not consistent with their clinical experience". Guideline sources were continuing medical education (CME) programs (47.5%), journals (36.2%) and medical school/residency training (35.0%). Only 8.6% cited the Internet as their source. The most popular guideline used was from the Society of Obstetricians and Gynecologists of Canada (SOGC) (N = 7), with others using American College of Obstetricians and Gynecologists, textbooks or unspecified resources.
Physicians reasons for not using a CPG for AUB or hysterectomy (N = 177)
Deterrent
N (%)
I did not know that such a guideline existed
112 (63.3)
I am not familiar enough with the guideline
35 (19.8)
I am confident in managing abnormal uterine bleeding without a guideline
26 (14.7)
I don't have time to look through the guideline
9 (5.1)
I have searched, but was unable to find such a guideline
6 (3.4)
I never see cases of abnormal uterine bleeding
2 (1.1)
I do not agree with the guideline(s)
1 (0.6)
I do not believe in CPGs in general
1 (0.6)
Seventeen responding physicians (8.8%) used a CPG for AUB or hysterectomy in their practices, citing "relevance to practice" (92.9%), "easy to use" (85.7%), "evidence-based" (78.6%) and "consistent with clinical experience" (78.6%) as the greatest incentives for use.
Education and resource needs
Almost all physicians (92.7%) were interested in receiving continuing medical education on the management of AUB, with popular formats being mailed copies of a CPG, tips for medical/hormonal management of AUB, journal articles and courses held locally (Table 2). Several physicians suggested training in in-office endometrial biopsies.
Continuing medical education: desired format (N = 179)
Format
N (%)
Copies of a clinical practice guideline
141 (78.8)
Tips for medical/hormonal management of AUB
90 (50.3)
Journal articles
69 (38.5)
Courses held locally
63 (35.2)
Pamphlets
38 (21.2)
Web-based
36 (20.1)
Hands-on instruction (e.g. endometrial sampling)
33 (18.4)
Videotapes
27 (15.1)
Audiotapes
17 (9.5)
Telemedicine
17 (9.5)
Calendar charts for quantitative measurement of bleeding
17 (9.5)
Courses held at a medical school
15 (8.4)
Increasing access to gynecologists was a common suggestion for improving management of AUB. Family physicians reported that the average length of time that a woman in their practices such as the one in our clinical scenario would wait for a gynecological consult was 7.1 +/- 4.7 weeks (range 1–28). Less than half (45.4%) of the physicians could refer such patients to gynecologists in their hometowns and 42.3% of physicians' reported that patients would have to commute between 1 and 3 hours to see a gynecologist while 8.8% reported commutes of over three hours. Physicians were also concerned with the "quality of the gynecological referral", wanting "more and better gynecologists to work locally", more "visiting specialist programs" and "easier referral with more informative consultant reports".
Discussion
The moderate response rate to this survey limits the generalizability of the results as we discuss in the accompanying article. Although the creation of strategies for improving the management of AUB will be based on the responses to this questionnaire, it is hoped that the development of convenient, easy-to-use educational tools and resources will also appeal to those family physicians who did not respond to the survey.
It appears that most family physicians did not use CPGs for the management of AUB because they did not know of such guidelines. These guidelines are published by specialist (Obstetrics and Gynecology) organizations [10-13]. It is unreasonable to expect primary care physicians to peruse all specialty journals that could have some relevance to their practice. However, due to the prevalence of AUB, there is a need for a mass awareness campaign for physicians on where to find information on the management of AUB. Clearly, the internet is an untapped resource as most of the existing CPGs are available online. However, many physicians suggested mailed clinical practice guidelines and some physicians urgently requested that the researchers mail them a copy of a CPG for AUB. Therefore, mailings (by physician organizations with accurate lists of northern family physicians and their web addresses) may be a useful option for practitioners who do not access guidelines electronically or who prefer hard copies of guidelines.
Another vital question is whether physicians would use guidelines if they were more readily available. Respondents who commented on the content of current AUB guidelines cited incentives and barriers to guideline use that are consistent with the findings in the general literature on guideline use-relevance, ease of use, consistency with their clinical experience and a lack of restriction to their autonomy [2-8]. With this in mind, we recommend a one page mailing, emailing and web posting of the salient management points for AUB including a diagnostic approach with specific treatment options. This would be easy to use, would place minimal restrictions on physician autonomy and could be easily derived from existing CPGs on AUB. Once physicians become more familiar with a CPG for AUB management, specific incentives and barriers to guideline use for AUB by family physicians can be explored more fully.
The recurrent theme of more timely and effective gynecological referrals is a difficult issue. There is no immediate way to increase the number of gynecologists in these regions. Augmentation of visiting specialist programs would be difficult in an already understaffed specialty. The development of a triage system for gynecological referral is not an imminent possibility. For now, it appears that primary care physicians, who are also in short supply in many northern areas, must become more comfortable managing abnormal bleeding while waiting for, or instead of, a gynecological consultation. As several physicians suggested, clinical practice guidelines and locally held training courses would be an appropriate start.
Conclusion
The needs outlined in this study should be considered for their applicability when developing educational programs and materials on abnormal uterine bleeding for family physicians and patients in northern, rural and isolated communities. We recommend that a one page summary of the necessary investigations, treatment options and their costs be mailed and also be made available online (with the CPG web address on an adhesive label in the mailed package). Educational initiatives for physicians and patients are also necessary and should include local, hands-on courses (including options for endometrial biopsy training) where requested by physicians, most of whom cannot readily take time away from their practices.
Competing interests
Competing interests: None declared.
Authors' contributions
Dr. Stewart conceived of the project, supervised the work, and fine-tuned the write-up.
Ms. Vigod conducted the literature review, developed the questionnaire, conducted the survey, analyzed the data and drafted the report.
Pre-publication history
The pre-publication history for this paper can be accessed here:
Acknowledgments
We are grateful to Dr NA Leyland for his suggestions on this project and manuscript.
Canadian hysterectomy rates have declined in recent years. However, hysterectomy rates for discretionary indications, principally abnormal uterine bleeding (AUB), remain high in some regions. In northern Ontario, hysterectomy rates for women aged 34–45 are almost triple the rates in southern, urban areas. Primary care physicians (family doctors) usually manage AUB initially in these northern areas where a severe shortage of gynecologists exists.
Methods
We surveyed 194 family physicians in northern Ontario with a case scenario of a pre-menopausal woman with heavy vaginal bleeding to characterize management and to gain physicians' perspectives on the factors that affect it.
Results
To investigate her heavy vaginal bleeding, only 17% of physicians recommended a pelvic examination for the woman in our case scenario. Most physicians advocated a course of medical therapy before referral to a gynecologist, for whom the average waiting time was seven weeks. However, most physicians recommended referral after only one failed trial of medical treatment. Physicians felt that major deterrents to medical treatments were patient desires for immediate relief and/or permanent solutions, poor patient compliance and the high cost of medication. Only 25% of respondents indicated that they would perform an endometrial biopsy prior to referral.
Conclusions
Family physicians would benefit from further education on appropriate investigations for AUB, primarily training in pelvic examination and endometrial biopsy techniques, as well as appropriate treatment algorithms. Further research into patient perspectives on treatment options is needed.
Background
Although hysterectomy rates have declined in developed countries over the past ten years, several countries including the United States, Canada and Australia have much higher rates than most European countries and report large variations in rates by region [1,2]. In Canada, for example, the age-adjusted hysterectomy rate for women aged 34–45 in some northern and rural areas of Ontario is nearly triple that in urban areas such as Toronto. The differences between regions are largely attributable to a higher rate of hysterectomies being performed for discretionary indications, primarily for abnormal uterine bleeding (AUB) that is unrelated to a systemic cause, anatomical lesion or malignancy [2]. Clinical practice guidelines conclude that hysterectomy is indicated primarily for women with AUB who have not had an adequate response to medical treatment or other surgical interventions [3-5]. The reasons for the increased regional hysterectomy rates are likely to involve physician, patient and economic factors. AUB is a common problem and accordingly primary care physicians (family doctors) are often responsible for its management, especially in northern and isolated areas where a severe shortage of gynecologists exists. It is essential that family doctors are confident in providing initial investigation and treatment for women who present with abnormal uterine bleeding. The present study characterizes the management patterns of primary care physicians in northern Ontario who responded to a survey and gains their perspectives on the factors that affect their management of AUB.
Methods
A five-page questionnaire entitled "Management of Heavy Vaginal Bleeding: Northern and Rural Primary Care Physicians" was developed in June, 2000 on the basis of the current accepted standards for the management of AUB and recommendations for hysterectomy (Keywords: Menorrhagia, Hysterectomy andAbnormal Uterine Bleeding) [6]. Clinical practice guidelines on AUB and hysterectomy were sought from the Society of Obstetricians and Gynecologists of Canada (SOGC) and the American College of Obstetricians and Gynecologists (ACOG). The guidelines available at the time of survey development had been published in 1996 (SOGC) and 1994 (ACOG) [3,7,8]. Input from family physicians, gynecologists, sociologists, psychologists and other women's health experts was incorporated during the survey development and piloting. The survey (see appendix – additional file 1) was composed of four parts, of which I. Demographic Information and II. Clinical Scenario, Investigations and Treatment are reported in this paper. Parts III. Clinical Practice Guideline Use and IV. Education and Resource Needs Assessment are described in the accompanying article ("Part II: What Do We Need?").
We collected information on physicians' sex, age, history of clinical practice, Internet access (for availability of information), previous obstetrical experience (possible indication of comfort level with managing gynecological problems) and Canadian College of Family Physicians (CCFP) membership status (requirements for continuing medical education (CME) and access to CME resources).
In the clinical scenario, a 39 year old woman presents with heavy cyclic vaginal bleeding of 3 months duration which impacts her daily function. Family physicians were asked what investigations they would perform and assuming that the results of all investigations were normal (i.e. that the cause of the bleeding was not identifiable), to outline their first, second and third lines of treatment.
The questionnaire was mailed to 686 physicians in the northern Ontario regions in the three district health councils that have the highest rates of hysterectomy[2]. Physicians were identified with the help of the Northern Shores District Health Council, the College of Physicians and Surgeons of Ontario (CPSO) web page and the local Yellow Pages. Each package included a cover letter explaining the objectives of the study, a questionnaire and a stamped return envelope. A modification of the Dillman (TDM) approach was used with two reminders, sent 3 and 8 weeks respectively after the original, each including a replacement questionnaire and a reminder letter [9]. Fifty-four surveys were returned due to incorrect address or specialty information (i.e. respondents were not family physicians). There were 194 completed responses from the remaining 622 potential participants (Response Rate = 31%). Response rates were evenly distributed among the three health regions and all respondents were family physicians currently practicing in northern, rural and/or isolated areas.
The Statistical Package for the Social Sciences (SPSS, version 10) was used for statistical analysis. Investigation and treatment recommendations generated by physicians were coded into individual variables. Descriptive analyses of all continuous and categorical variables were conducted. Chi square tests of independence investigated whether responses were dependent on sex, Internet use, previous obstetrical experience or Canadian College of Family Physicians (CCFP) membership status. Independent sample t-tests and analyses of variance (ANOVA) were used to determine any differences in responses by age or the population of the physician's practice location.
ResultsPhysician information
Responding physicians, mean age 43.4 +/- 10.5 years, had been in family practice an average of 14.9 +/-10.8 years. Corresponding to the gender distribution of physicians in Ontario, 37.5% of the respondents were female, 62.5% were male. Males were significantly older with more years in family practice than females (t = -4.612, p = 0.00; t = -4.324, p = 0.00). All physicians practiced in northern Ontario, with 45.1% also designating their practice populations as "rural" and 17.5% as "isolated". These categories were not mutually exclusive (i.e. a physician could practice in an area that is rural and isolated). Fifty percent of physicians were practicing in areas with populations of less than 15 000 people, with the remainder in larger towns (25%) or small cities (25%). Three-quarters of the respondents were members of the Canadian College of Family Physicians (CCFP) indicating completion of a family practice residency, successful passing of an examination in family practice and participation in continuing medical education (CME). A higher proportion of women were members of the CCFP than men in keeping with their younger age (χ2 = 13.8, p = 0.00). With respect to obstetric experience, 41.4% had previously delivered babies in their practices, 32.5% were currently doing deliveries and 26.2% had never delivered babies in their practices. Almost all (92.7%) respondents had Internet access either in their homes or offices.
Investigations and treatment
Investigations suggested by physicians are reported in Table 1. The tests recommended most frequently were pelvic ultrasounds (87.1%), complete blood counts (75.3%) and thyroid function tests (41.2%). Endometrial biopsy (EB) was recommended by 25.3%, and dilatation and curretage (D&C) by 24.2%, of physicians either during the initial investigation or after treatment failure. Older physicians were more likely to recommend D&C than younger physicians (t = 2.28, p = 0.02) and women physicians were more likely to recommend EB than men (χ2 = 5.13, p = 0.02). Seventeen percent of physicians indicated that they would perform a pelvic exam on the patient in the clinical scenario, with a greater proportion of non-CCFP members recommending pelvic exams than CCFP members (χ2 = 6.83, p 0.009). Some respondents (15.5%) recommended coagulation testing. These physicians were significantly younger than physicians who did not order the aforementioned tests (t = -2.51, p = 0.015).
International normalized ratio (INR)/ prothrombin time (PT)
46 (23.7)
Platelets
46 (23.7)
Pap smear
45 (23.2)
Pelvic exam
33 (17.0)
Partial thromboplastin time
30 (15.5)
Follicle stimulating hormone (FSH)
22 (11.3)
Luteinizing hormone (LH)
18 (9.3)
Glucose
10 (5.2)
Abdominal exam
8 (4.1)
Estrogen
8 (4.1)
Pregnancy test (beta HCG)
8 (4.1)
Prolactin
7 (3.6)
Progesterone
4 (2.1)
Referral (no investigations)
4 (2.1)
* requested either during initial investigation or during course of treatment
Table 2 presents physician treatment patterns for first, second and third lines of treatment. Medical options were the most highly recommended first and second lines of treatment. Physicians who chose medical options as a first line of treatment were significantly younger than those who did not (mean age 42 versus 50, t=-4.1, p = 0.00) and had been in family practice for less time (13.6 versus 22.5 years, t = -4.1, p = 0.00). Oral contraceptives (54%) and non-steroidal anti-inflammatory drugs (26%) were the most popular recommendations. A second trial of medical therapy was recommended most frequently by female physicians and CCFP members. Family doctors who practice obstetrics were more likely than those who did not, to recommend a third trial of medical therapy (χ2 = 6.47, p = 0.01).
Treatment Pattern for AUB.
Treatment
First LineFrequency (%)
Second LineFrequency (%)
Third LineFrequency (%)
Watchful Waiting
7 (3.6)
1 (0.5)
0 (0)
Medical
Oral Contraceptive
105 (54)
51 (28)
3 (2.4)
NSAID
50 (26)
8 (4.4)
1 (0.8)
Progesterone alone
22 (11)
33 (18)
6 (4.8)
Estrogen alone
2 (1.0)
0 (0)
0 (0)
Anti-fibrinolytic
2 (1.0)
2 (1.1)
4 (3.2)
Gonadotropin Agonist
3 (1.5)
6 (3.3)
6 (4.8)
Referral
18 (9.3)
88 (48)
106 (84)
Surgical
Dilatation and Curettage
6 (3.1)
31 (17)
10 (7.9)
Endometrial Ablation
1 (0.5)
10 (5.5)
13 (10)
Hysterectomy
2 (1.0)
10 (5.5)
25 (20)
Total Respondents
194
183
126
Physicians who recommended immediate referral to a gynecologist (9.3%) were significantly older, and had been in practice longer, than those who did not. Male physicians and non-CCFP members were more likely to refer after the failure of only one trial of medical treatment. Upon failure of a second medical treatment, 84% of physicians who had not yet referred the patient to a gynecologist did so. Of these referring physicians, 20% specifically recommended hysterectomy (note that physicians were not asked to speculate on the outcome of a referral). Average waiting time for a gynecological referral for a woman with AUB was approximately 7 weeks (range = 28 weeks) and more than 50% of patients would have to commute for at least one hour to attend a gynecologist.
Physicians identified several patient factors as deterrents to medical and/or hormonal management of AUB (Table 3). Major deterrents to medical management were a desire for immediate relief (70.8%), poor compliance (51.9%) and that it might not be a permanent solution (49.7%). Another deterrent was financial impact on patients (31.9%) and this was cited more often by physicians who practiced in areas with smaller populations (t = -3.30, p = 0.001).
Deterrents to Medical or Hormonal Management of AUB: Patient Factors (N = 185)
Deterrent
N (%)
Desire for immediate relief
131 (70.8)
Poor compliance
96 (51.9)
May not be a permanent solution
92 (49.7)
Financial impact
59 (31.9)
Limited patient comprehension
52 (28.1)
Patient's partner's preferences
23 (12.4)
Distance patient lives from medical care
15 (8.1)
Transportation problems
13 (7.0)
Embarrassment or offence of patient
12 (6.5)
Other: contraindications, side effects, cultural/religious beliefs, family planning, health beliefs (i.e. that medication won't work)
N/A
Discussion
The study's response rate (31%) illustrates one of the greatest challenges in improving the management of AUB, that is, to communicate with all family physicians who manage this problem, both to determine their needs and to disseminate information. Current lists of active family doctors practicing in the North are difficult to obtain as there is a rapid turnover of physicians and this presents a challenge for surveys and education. The physicians who responded to this questionnaire may be more highly motivated and a more geographically stable group than the average. As three-quarters of them were members of the College of Family Physicians they accordingly may represent a more highly trained, younger and possibly more knowledgeable subset of the practicing regional general practitioners. With this in mind, the authors take caution in interpreting the results of this questionnaire. However, as physicians who are practicing medicine in the northern, rural and isolated regions are all affected by problems such as time pressure and lack of resources, the issues identified by our respondents may be the tip of the iceberg.
To identify areas where responding physicians can improve their management of AUB, we compared their recommended investigation and treatment responses to best evidence guidelines or algorithms available at the time that the survey was conducted. Since this time, clinical practice guidelines on the investigation and management of AUB have been updated [5,6]. However, as the majority of respondents were unaware of any existing guidelines at the time of our survey, it is unlikely that they have significantly changed their practices in light of the new ones (see accompanying article).
Current clinical practice guidelines (CPGs) recommend a pelvic exam to visualize the cervix and palpate pelvic masses, however, only 17% of respondents indicated that they would do this. These physicians may not be confident in their pelvic exam skills or perhaps believe that a pelvic exam can be replaced by pelvic ultrasound (suggested by 87% of physicians). Pelvic exam training and identification of cervical lesions and uterine fibroids would allow further specific investigation and treatment to be instituted earlier for affected women thereby decreasing the need for unnecessary investigations and costs to the patients and the health care system.
Appropriately, the majority of family physicians suggested a complete blood count to look for an anemia. Thyroid function tests, recommended by almost half of the responding physicians, and coagulation tests (recommended by 15.5%) should be considered when there is clinical suspicion of thyroid dysfunction or bleeding disorders respectively. Identification and appropriate treatment of thyroid dysfunction and bleeding disorders may reduce the necessity for surgical intervention in AUB. Tests of follicle stimulating hormone (FSH) and luteinizing hormone (LH), recommended by 11.3% and 9.3% of respondents respectively, are more useful in the clinical context of examining ovarian function.
Whether endometrial sampling to look for endometrial hyperplasia or neoplasia was necessary in this scenario is a controversial point. Guidelines in existence at the time of the study stated that endometrial sampling should be done in any pre-menopausal woman with heavy vaginal bleeding who is refractory to treatment [3]. More specifically, recent guidelines (2002) support endometrial sampling in premenopausal women with AUB who are over forty years of age, who have risk factors for endometrial cancer including new onset of heavy, irregular bleeding or whose bleeding is refractory to three months of medical therapy [4]. As the woman in our case scenario had cyclic bleeding, endometrial sampling may not have been necessary in the initial management plan, yet less than half the respondents recommended endometrial sampling at any time during her management, even after she did not improve with treatment.
Endometrial biopsy (EB) is the initial diagnostic endometrial sampling procedure of choice for detecting endometrial hyperplasia or neoplasia (sensitivity 67–95%). Diagnostic D&C has a lower sensitivity and higher procedural risk (e.g. uterine perforation, hemorrhage, anesthetic complications) and it is only recommended when no other sampling methods are feasible. Pelvic ultrasound is neither sensitive nor specific for endometrial carcinoma in premenopausal women and can therefore not be used as a form of endometrial sampling [10]. EB can be performed in-office by family physicians and having the ability to rule out endometrial hyperlasia and neoplasia will allow medical treatments to be attempted, possibly reducing gynecological referral rates in the long-term.
Most respondent physicians advocated for at least one course of medical therapy before referring the patient to a gynecologist and/or recommending surgical intervention. It is not surprising that younger family doctors were most likely to recommend initial medical treatment as they have the greatest proximity to their medical training and AUB management has changed in recent years. Further, CCFP members (who have continuing medical education requirements) and those practicing obstetrics (who likely have a high proportion of women in their practices and may be more confident and up-to-date on obstetrical and gynecological issues) were most likely to recommend subsequent medical treatments. This highlights the impact of knowledge and confidence on physician management patterns.
Many physicians reported that patient desire for immediate relief and permanent solutions were deterrents to medical or hormonal treatment. However, the waiting time for a gynecological referral could encompass several cycles of heavy bleeding and missed work. By this point, an untreated woman experiencing significant frustration and distress may have an increased wish for a definitive treatment, such as hysterectomy [11]. Poor compliance was anticipated as a deterrent to medical management by over half of responding physicians. Therefore, it is important to educate women about the medications and their effects. Physician perceptions of patient deterrents to medical management must be further confirmed with patients in order to design appropriate interventions.
Conclusions
Evidence and opinion on the management of AUB changes rapidly and the efforts in the management of AUB by already overburdened primary care physicians in northern, rural and isolated areas should be commended. However, family physicians would benefit from further education on appropriate investigations and treatment for AUB, training and practice in pelvic exam and endometrial biopsy techniques. As well, women may not have all of the facts about medical and surgical treatment options and this is an opportunity for primary care physicians to help women obtain the best information to make effective treatment decisions. Specific suggestions for the form and dissemination of these recommendations are found in the accompanying article.
Competing interests
None declared.
Authors' contributions
Dr. Stewart conceived of the project, supervised the work, and fine-tuned the write-up. Ms. Vigod conducted the literature review, developed the questionnaire, conducted the survey, analyzed the data and drafted the report.
AppendixClinical scenario
A 39-year old woman presents with a three-month history of cyclic, heavy vaginal bleeding every 24 days. She is without pain, bowel or bladder complaints. However, her heavy bleeding prevents her from going to work 4–6 days per cycle. She has no history of medical problems and is not taking any medications. She has only been hospitalized once, for the uncomplicated birth of her only child, 3 years ago. For the past 5 years, on pelvic examination, she has had a fibroid the size of an 8-week pregnancy.
Pre-publication history
The pre-publication history for this paper can be accessed here:
Acknowledgements
We are grateful to Dr. Leyland for his suggestions on this project and manuscript.
Cervical cancer can often be prevented by screening and may be curable if identified and treated in its early stages. However, 80% of new cases occur in less-developed countries where cervical cancer screening programmes are small-scale or non-existent. This is a human tragedy of great proportion, with many of those affected being young mothers. There is some evidence that cancerous or precancerous lesions may be detected by visual inspection with acetic acid (VIA) and field studies indicate that this technique is effective, safe and acceptable to women. However, the provision of a light source for inspection of the cervix presents a major problem in less-developed countries, where candles and torches often provide the only means of illumination. Our objective was to develop a light source based on clockwork technology, that required no batteries or external power source.
Methods
We adapted the design of a commercially available clockwork torch to provide a light source for cervical inspection. The light source was then tested under laboratory conditions in a comparison with other illumination methods typically used in this application.
Results
The light source gave illuminance levels greater than those produced by any other method tested, and also had considerable advantages in terms of ease of use and safety.
Conclusion
This design is small, compact, effective and safe to use and promises a better and more affordable means of visualising the cervix. Further field trials of VIA are now required which incorporate this light source.
Background
Cervical cancer can often be prevented by screening, and may be curable if identified and treated in its early stages. However, in most less-developed countries, cervical cancer screening programmes are small-scale or non-existent. Thus there are few opportunities for the diagnosis and treatment of precancerous cervical disease. It is estimated that in excess of 470,000 new cases of cervical cancer are diagnosed each year around the world, with more than 80% occurring in less developed countries [1]. More than 233,000 women are dying from this disease each year and the morbidity is highest in those countries that are least equipped to deal with the problem. This is a human tragedy of great proportion, and many of those affected are young mothers.
There is some evidence that cancerous or precancerous lesions may be detected by simple inspection of the cervix [2]. The cervix is first washed with acetic acid (vinegar) solution that causes precancerous changes in the cervix to turn white. Community (village) health workers are trained to detect the abnormality with the naked eye [3]. This technique of visual inspection with acetic acid (VIA) may have potential advantages over traditional screening techniques, providing an immediate feedback of results and enabling prompt treatment [4]. Field studies indicate that this technique is both sensitive and specific, and it may provide a cheap and acceptable means of screening for cervical cancer in poor-resource locations [5]. A test and treat approach using VIA and Cryotherapy appears to be both safe and acceptable to women and providers, and has the potential to be an efficient method of cervical cancer prevention in these settings [6]. However, the provision of a light source for inspection of the cervix presents a major problem in many rural communities. Many of these rural communities are without electricity, and candles and torches offer the only means of illumination (Personal Communication). In addition, portable power supplies and batteries are a desirable commodity and are rapidly diverted in other settings and applications.
Our objective was to develop a clockwork power source and light that required no batteries or external means of power. The light source needed to be small and compact enough to allow the health worker to illuminate and view the cervix, yet provide sufficient illumination to enable the identification of precancerous lesions. It was essential that the light source did not pose any hazard to either the patient or the health worker and that it was robust, and relatively inexpensive to deploy in less-developed countries. In addition, as it clips into the speculum it is likely to be contaminated by vaginal discharge and so needed to be waterproof to enable sterilisation (see figure 4). We believe that with simple adaptations to existing technology, we have achieved our objective.
Materials and Methods
Recent developments in dynamo-powered devices have brought us the clockwork radio, and more recently the clockwork torch. We adapted the design of commercially available clockwork torch technology to provide a light source which could be used independently of any external power supply and required no batteries.
The components used to adapt the clockwork torch are widely available and the prototype added less than €2 (Sterling) to the cost of the Torch. (Figure 1).
The clockwork power source takes approximately 6 hrs to fully charge the batteries, giving approximately 2 1/2 hours of continuous use. It can also be used directly from the clockwork spring in an emergency setting, whereby 60 turns on the handle will provide sufficient charge to operate the light source for around 3 minutes.
In order to assess the effectiveness of the light source, a comparison was made with a range of other illumination methods that may be typically used for this application: namely a 100 W-desk lamp; a 40 W-desk lamp; a torch (0.5 W); a small compact torch (0.75 W); and a candle with and without a metal reflector.
For the experimental set-up, consideration had to be given to both the anatomy of the vagina and the relative positions of the light source and health worker (Figure 2). A cylinder 4 cm in diameter and 11 cm long was used to mimic the anatomy of the vagina (reference man ICRP 23). A calibrated photodiode light detector with photopic filter (SED038 with Y filter) was sealed at the far end, where in effect the cervix would lie. A plastic speculum was introduced into the other end of the cylinder and illuminance measurements at the 'cervix' were made. In order to accurately reproduce clinical conditions, measurements were made using the light source whilst simultaneously viewing the detector or 'cervix'.
Experimental set-up for laboratory testing
Results
The human eye responds to a narrow band of the electromagnetic spectrum, typically 400–700 nm with a peak response at 555 nm. In order to compare the different light sources it is essential that account be taken of the response of the human eye.
Illuminance (measured in lux) is the electromagnetic flux density falling on unit area, corrected for the response of the human eye. Typical flux densities encountered in daily life are shown in Table 1 and cover a large dynamic range. The human eye is able to accommodate a wide range in illuminance (Table 1).
Typical illuminance values
Illuminance (typical lux)
Sunny Day
~100,000
Office Lighting
~1000
Full Moon
~0.1
The measured illuminance levels are shown in Figure 3. The light source gave illuminance levels greater than that of the desk lamps or torch, both of which are commonly used for cervical inspection as part of a routine gynaecological examination in the UK. Figure 3 also clearly demonstrates the poor performance of a candle in this situation, an illumination method which is commonly used in less-developed countries and which carries potential hazards both for the patient and clinician.
Measured illuminance levels of light sources
By comparing Table 1 with Figure 3, we can see that the light levels achieved at the cervix when using the light source are broadly comparable to normal office lighting.
Clockwork light source speculum in use (simulation model) a. speculum b. light source
Discussion
Cervical cancer is the third most common cancer in the world. Screening programmes aim to detect cervical cancer in its early stages and make successful treatment more likely, with the possibility of reducing the incidence and mortality. In countries where regular cervical cancer screening is available, death rates have been reduced. However, cervical cancer remains the leading cause of death from cancer among women in less developed countries, where cervical cancer screening programmes are non-existent or fragmented at best [1].
Cervical cancer screening is a complex process. Less developed countries often face economic, social and cultural constraints which prevent the implementation of organised cervical cytology screening. A recent analysis of cervical cancer screening strategies for low-resource settings found VIA to be cost-effective ($39 per year of life saved) and when coupled with immediate treatment it reduced the incidence of cervical cancer by 26% [7]. However, compared to cytology, VIA is less specific with low positive predictive value for high-grade disease [2]. In their report from Zimbabwe, The University of Zimbabwe/JHPIEGO Cervical Cancer Project suggested that higher test quality for VIA could be achieved with better lighting, standardised training and better service delivery conditions [3].
The light source was the best amongst the illumination methods tested. Its illuminance was greater than that of the torch, which is still commonly used for gynaecological examination even in UK hospitals. It is not simply the power output of the light source but also how compact it is that determines the measured illuminance at the cervix. The 100 W and 40 W lamps are certainly the brightest sources used but they are large, unwieldy and must be placed behind the observer. The light source has the bulb detached from the power source, so that its size is only that of the clip and bulb. Hence it does not impede the view of the health worker. In addition the speculum clip means that the light source does not need to be held, leaving the health worker's hands free (Figure 4).
Conclusion
This design promises a better and more affordable means of visualising the cervix, and further field trials of VIA are required which incorporate this light source.
List of Abbreviations
VIA visual inspection (of the cervix) with acetic acid
Competing interests
None declared.
Authors' Contributions
This project was instigated by RJ, who directed it through to completion and wrote the first draft of this paper but died before it could be submitted for publication. Co-authors are responsible for the final form of this manuscript. All authors contributed to the development of a working prototype. Laboratory testing of the light source was undertaken by BB and CK.
Pre-publication history
The pre-publication history for this paper can be accessed here:
Acknowledgements
This paper is dedicated to Richard Johanson who died prior to publication. We are grateful to the Rotary Club for their support and encouragement. Linda Lucking prepared the manuscript.
There is little existing research to guide researchers in estimating the minimum number of measurement occasions required to obtain reliable estimates of serum estrogens, progesterone, gonadotropins, sex hormone-binding globulin (SHBG), and urinary estrogen and progesterone metabolites in premenopausal women.
Methods
Using data from a longitudinal study of 34 women with a mean age of 42.3 years (SD = 2.6), we calculated the minimum number of measurement occasions required to obtain reliable estimates of 12 analytes (8 in blood, 4 in urine). Five samples were obtained over 1 year: at baseline, and after 1, 3, 6, and 12 months. We also calculated the percent of true variance accounted for by a single measurement and intraclass correlation coefficients (ICC) between measurement occasions.
Results
Only 2 of the 12 analytes we examined, SHBG and estrone sulfate (E1S), could be adequately estimated by a single measurement using a minimum reliability standard of having the potential to account for 64% of true variance. Other analytes required from 2 to 12 occasions to account for 81% of the true variance, and 2 to 5 occasions to account for 64% of true variance. ICCs ranged from 0.33 for estradiol (E2) to 0.88 for SHBG. Percent of true variance accounted for by single measurements ranged from 29% for luteinizing hormone (LH) to 92% for SHBG.
Conclusions
Experimental designs that take the natural variability of these analytes into account by obtaining measurements on a sufficient number of occasions will be rewarded with increased power and accuracy.
Background
Several active research programs are investigating the risk associated with serum estrogens, gonadotropins and urinary sex hormone metabolites for a variety of diseases including breast cancer [1], endometrial cancer [2], and osteoporosis [3]. The results of the few published studies suggest that the natural temporal variability (true variation over time, not variation due to storage or other factors) of some serum estrogens, gonadotropins and urinary sex hormone metabolites is sufficiently great that a single measurement occasion may be inadequate to ensure a reliable estimate [4-6]. Published intraclass correlation coefficients (ICC) vary between 0.06 and 0.62 for estradiol (E2) and between 0.52 and 0.69 for estrone (E1) [4]. Only the percent of free E2 and of SHBG-bound E2 have been found to be sufficiently reliable to account for as much as 50% of the variance in the true mean (ICC > 0.7).
The term reliability can refer either to the consistency of a measuring procedure or to the temporal stability of the target of measurement [7]. The definition of temporal reliability used in this study includes both those dimensions, but emphasizes the latter. While researchers can control error due to insufficient repeated measures by increasing the number of measurement occasions, obtaining measurements is expensive. It is therefore useful to have evidence-based guidelines for estimating the minimum number of occasions required to obtain a given degree of reliability for a particular analyte.
All types of measurement error distort, confound, or attenuate the tests of association that constitute one of the primary products of research [8,9]. Figure 1, though not exhaustive, shows the sources of variance in a measurement and the interrelationships between error and tests of model fit or significance.
Total observed variance
The relation of a measurement to the object being measured can be represented as: σO = σT + σE, where σO = variance in the observed measurement of the target, σT = variance in the true value of the target, and σE = random variance, or error. If the true value of the target is invariant across measurements, i.e., if σO = σE, the observed variance will be purely a function of the unreliability of the measuring instrument. Conversely, if perfectly error-free measurement of the target could be assumed, i.e., if σE = 0, then σO = σT and the observed variance would be purely a function of the temporal stability of the target. If σE ≠ 0 and σT ≠ 0, the observed variance will be a function of both the temporal stability of the target and of the unreliability of the measuring instrument.
Measurement error can result from a variety of factors, including true variance not captured by a particular measurement strategy, which may complicate the interpretation of temporal reliability estimates. These other factors include variance due to: fluctuations across cycle phases within each woman's menstrual cycle [10]; duration of sample storage prior to analysis [11]; limitations of the assay; multiple analysis batches [10]; multiple types of assays [12]; and multiple laboratories [10]. Ideally, estimates of as many sources of error as possible should be included when considering the impact of temporal reliability on measurement strategy. The objective of this study was to determine the following for various serum estrogens, gonadotropins, and urinary sex hormone metabolites: the minimum number of repeated measurements required for reliable estimates; the ICCs; and the amount of true variance accounted for by single measurements.
MethodsExperimental design
The data for this study come from a randomized double-blind study investigating the effects of a 100 mg/day soy isoflavone regimen on estrogen levels in 34 premenopausal women. A detailed description of the study design and the results of the intervention were reported in Maskarinec et al., 2002).)[13]. The Committee on Human Studies at the University of Hawaii approved the study protocol. Written informed consent was obtained from each subject, prior to participation. The study group consisted of 17 premenopausal women per group. Four women left the study before the end of the year and another was able to give only four blood draws for health reasons. Eligibility criteria included: an age range of 35–46 years; an average intake of less than 7 servings of soy foods per week; no prior cancer diagnosis (except basal cell skin carcinoma); no use of oral contraceptives or hormone preparations within the past three months; no intention of becoming pregnant within the next year; an intact uterus and ovaries; self-defined regular menstrual periods; no serious medical condition. Subjects had a mean age of 42.3 years (SD = 2.6), and a mean weight of 65.6 kg (SD = 12.8). Subjects were ethnically diverse: 18 were Caucasian; 6 were Chinese; 5 were Japanese; 5 were Hawaiian.
Sample collection
Subjects were asked to donate 5 urine and blood samples, one at baseline and one after 1, 3, 6, and 12 months of participation. All samples were collected approximately 5 days after the ovulation (approximately day 19 in a 28 day cycle). Subjects used ovulation kits (Ovuquick test kits from Quidel, La Jolla, CA) to determine the time of ovulation. This kit detects the mid-cycle rise of LH using morning urine with a sensitivity of 35 mIU/mL of LH and its predictive validity with respect to ovulation has been estimated as 93% [14]. Although the use of a minimum progesterone value to exclude data from anovulatory cycles from the analyses helped ensure acquisition of the mid-luteal phase samples, only 52% of samples were obtained on exactly the 5th day from ovulation. Ninety-one percent were obtained between the 4th and the 6th day from ovulation. Blood samples were drawn at a commercial laboratory, in the morning between 7 and 9 o'clock to control for circadian rhythm in hormone levels. Serum and urine samples were stored at -80°C after separation and aliquoting.
Serum analysis
Hormone assays were conducted at the Department of Obstetrics and Gynecology, University of Southern California (Los Angeles, CA) in the Reproductive Endocrine Research Laboratory. The analyses for E2, free E2, E1, E1S, progesterone, SHBG, follicle stimulating hormone (FSH), and LH were conducted in 2 batches. Samples of these analytes collected at baseline, month 1 and month 3 were analyzed in batch 1, and 6-month and 12-month samples were analyzed in batch 2 one year later. E2, E1, progesterone, FSH, LH, and SHBG were quantified in serum by specific and sensitive radioimmunoassays (RIAs). Prior to RIA, E1 and E2 were first extracted with ethyl acetate: hexane (2:3) and then purified by Celite column partition chromatography, using ethylene glycol as stationary phase [15]. E1 and E2 were eluted off the column with 15% and 40% toluene in isooctane, respectively. 3H-E1 and 3H-E2 were used as internal standards to follow procedural losses. FSH and LH levels were determined using an immunoradiometric assay (IRMA). E1S, progesterone and SHBG were measured by direct RIAs using kits obtained from Diagnostic Systems Laboratories, Webster, Texas. Free E2 (non-SHBG or albumin-bound-E2) was determined by calculation using a computerized algorithm described previously).)[16]. The majority of intra-assay CVs for all analytes were below 10% (Table 1) indicating good quality control in the laboratory. They ranged from <0.5% for SHBG to 13.0% in the low concentration range of batch 1 for E1.
Urine samples were analyzed for estrone-3-glucuronide (E1-G), pregnanediol-3-glucuronide (PDG), 16α-hydroxyestrone (16α-OHE1) and 2-hydroxyestrone (2-OHE1). E1-G and PDG were measured directly in urine by enzyme immunoassay [17]. Commercially available enzyme-linked immunosorbent assay kits (Estramet: Immuna Care Corporation, Bethlehem, PA) were used to determine levels of 16α-OHE1 and 2-OHE1 in urine [18]. All results are relative to creatinine excretion.
Statistical analysis
The SAS statistical software package version 8.2 (SAS Institute Inc., Cary, NC, 1999–2001) was used to perform the statistical analyses. All statistics were computed using logged values when raw values were not normally distributed. To ensure that all measurements in the analysis were from the same time in the menstrual cycle, observations were only included if the concurrent progesterone values were at least 5 ng/mL, a minimum value after an ovulation has occurred. Because analyses for 8 of 12 analytes were conducted in two batches, we included consideration of error due to between batch variance in our analysis of the temporal stability of these analytes. Therefore, estimates of temporal stability for the 8 analytes were calculated for the total number of samples and for the first and second batches separately.
Two types of estimates of the number of measurement occasions (O) necessary to obtain an adequately reliable estimate were computed. The first, the relative type (OR) includes the between-subject variance. OR was computed using the formula proposed by Nelson et al. [19]: where r is the correlation between the observed and the true mean analyte values for an individual over a year, sW2 is the within-subject variance, and sB2 is the between-subject variance. Setting r to 0.9 results in a calculation of the number of measurement occasions required to obtain an estimate that would account for 0.92 or 81% of the true variance in the target. Ninety-five percent confidence intervals (95% CI) for OR were computed using a published method).)[20].
The second estimate of the number of measurement occasions necessary to obtain an adequately reliable estimate, the absolute type (OA), includes only within-subject variance. OA was calculated as , where σw is the within-subject variance [21]. By adjusting the denominator, this method allows for the desired approximation to the true mean to be specified as a percentage. Setting the denominator to 0.2 results in a calculation of the number of occasions required to obtain an estimate that is within 20% of the true mean. A SAS macro using Proc Varcomp and Proc Means to produce estimates of OR, OA, and related statistics is available from the authors.
ICCs measure the proportion of variance attributable to targets of measurement as a ratio of within-subject variance to total variance [22] and are suitable to compare variables of the same measurement class [23]. We computed two types of ICCs using the notation developed by Shrout and Fleiss [22]: ICC(2,1) was computed for each analyte using all 5 measurement occasions to estimate the temporal reliability of the analyte; ICC(2,k) was computed between batches to estimate the contribution of between-batch variance to the temporal reliability estimate. ICC(2,1) was computed as ICC(2,1) = , where BMS is the between-subjects mean square, EMS is the error mean square, k is the number of observations, OMS is the observations mean square, and n is the number of subjects [22]. ICC(2,k) was computed as ICC(2,k) = . We applied the formulas by Shrout and Fleiss [22] to obtain 95% CIs.
To estimate the percentage of true variance accounted for by a single measurement, we assumed that the best available estimate of the true variance was the total variance for all occasions.
After calculating the Pearson correlation of each occasion with all other occasions, we considered the squared average of these correlations as the estimate of the most likely percent of true variance for which a single occasion could account. We used the formula , where % σT is the percent of true variance, rT is the Pearson correlation of each occasion with the total of all other occasions, and o is the number of occasions.
Results
Overall means, number of samples, and means by measurement occasion for all analytes (Table 2) indicate the overall stability for the analytes over one year. Although estrogen and progesterone levels were on the average 7% higher and gonadotropins and urinary sex hormone metabolites 10% lower in the intervention than in the control group (data not shown), none of the differences was even close to statistical significance (p values ranged from p = 0.16 to p = 0.90 for Estrone-sulfate and Estrone respectively). Because of this homogeneity, results in this study were collapsed across experimental groups. The decrease in E2 and E1 are the result of laboratory drift and were independent of intervention status).)[13].
Basic descriptive data for all measurement occasions of all analytes
All Measurements
Means for Each Measurement Occasion
Analyte
N
M (SD)
1 (BL)
2 (1 mo)
3 (3 mo)
4 (6 mo)
5 (1 yr)
Estradiol (pg/mL)
162
133.32 (51.41)
141.44
140.12
151.78
114.70
116.66
Free Estradiol (pg/mL)
159
3.22 (1.21)
3.44
3.34
3.58
2.74
2.95
Estrone (pg/mL)
162
103.74 (34.77)
115.82
115.29
122.59
81.24
80.83
Estrone-sulfate (ng/mL)
160
4.20 (2.44)
4.74
4.81
4.16
3.80
3.41
Progesterone (ng/mL)
162
9.81 (4.89)
10.09
8.87
11.09
10.16
8.76
SHBG (nmol/mL)
159
48.56 (21.18)
49.81
50.01
50.49
46.15
45.73
FSH (mIU/mL)
159
4.78 (4.21)
3.68
4.89
4.77
5.83
4.73
LH (miu/ml)
161
5.12 (4.34)
4.46
6.46
5.56
4.82
4.18
E1-G (ng/ml)
164
27.74 (20.79)
28.34
31.04
26.06
27.50
25.43
PDG (ug/ml)
164
3.47 (2.15)
3.93
3.23
3.67
3.36
3.11
2-OHE1 (ng/ml)
164
13.76 (8.01)
13.32
14.55
13.07
15.10
12.65
16α-OHE1 (ng/ml)
164
6.94 (5.18)
6.78
6.97
6.42
8.77
5.62
Measurements 1–3 were analyzed in batch 1 and measurements 4 & 5 were analyzed in batch 2. SHBG = Sex hormone-binding globulin. FSH = follicle stimulating hormone. LH = luteinizing hormone. E 1-G = estrone-3-glucuronide. PDG = pregnanediol-3-glucuronide. 2-OHE1 = 2-hydroxyestrone. 16α-OHE1 = 16α-hydroxyestrone.
The measurement occasions required to obtain a reliable estimate differed considerably by analyte (Table 3). Using the relative method to account for 81% of the true variance, the number of occasions required ranged from OR = 0.48 to OR = 11.43 (for SHBG and E1 respectively). To account for 64% of the true variance, the number of occasions ranged from OR = 0.20 to OR = 4.77 (for SHBG and E1 respectively). Using the absolute method, the number of occasions required to obtain an estimate to within 20% of the true mean, ranged from OA = 0.34 to OA = 10.27 (for E2 and PDG respectively). It appears that, except for SHBG and E1S, using a single measurement for any of the analytes in this analysis may be problematic for the typical purposes of epidemiological research because the results of typical epidemiological research center on analyses of the mean value obtained from one group vs. the mean value obtained from another, e.g. a group of cases or an intervention group vs. a control group.
Minimum occasions required to obtain a reliable estimate, intraclass correlation coefficients, and percent of true variance accounted for by single measurements
Analyte
Occasions (samples)
Measurement Occasions Required
ICC(2,1) (95% CI)
% of True Variance Accounted for by a Single Occasion (Range)
To Account for 81% of True Variance
To be Within 20% of True Mean
Relative Method (95% CI)
Absolute Method
Estradiol
5 (100)
8.26 (4.53–13.88)
0.34
0.33 (0.18-0.51)
37 (18-74)
Free Estradiol
5 (85)
5.32 (2.92-8.94)
2.00
0.41 (0.26-0.59)
48 (29-72)
Estrone*
3 (60)
11.43 (6.26-19.20)
0.37
0.51 (0.30-0.69)
50 (14-56)
Estrone-sulfate
5 (90)
1.42 (0.78-2.38)
2.03
0.71 (0.58-0.82)
74 (53-77)
Progesterone
5 (100)
5.15 (2.82-8.65)
8.90
0.40 (0.25-0.58)
40 (27-50)
SHBG
5 (85)
0.48 (0.26-0.80)
1.78
0.88 (0.81-0.93)
92 (84-96)
FSH
5 (85)
5.11 (2.80-8.59)
2.63
0.40 (0.25-0.58)
37 (22-66)
E1-G
5 (100)
3.88 (2.13-6.52)
0.92
0.47 (0.32-0.64)
45 (29-55)
LH
5 (95)
8.38 (4.59-14.07)
8.08
0.30 (0.15-0.48)
29 (17-62)
PDG
5 (100)
5.17 (2.84-8.69)
10.27
0.40 (0.25-0.58)
32 (15-56)
2-OHE1
5 (100)
4.21 (2.31-7.07)
1.57
0.46 (0.30-0.63)
16 (4-40)
16α-OHE1
5 (100)
2.71 (1.48-4.55)
2.44
0.56 (0.41-0.71)
46 (14-79)
Note: includes only complete observations where progesterone > 5 ng/mL; logarithmic transformations of values were used if raw values were not normally distributed. * Calculations based on batch 1 only due to high inter-batch variance. SHBG = Sex hormone-binding globulin. FSH = follicle stimulating hormone. LH = luteinizing hormone. E1-G = estrone-3-glucuronide. PDG = pregnanediol-3-glucuronide. 2-OHE 1 = 2-hydroxyestrone. 16α-OHE 1 = 16α-hydroxyestrone.
Figures 2 to 4 illustrate the different relationship of between- to within-subject variance and the corresponding difference between OR and OA.
Sex hormone-binding globulin values for all participants by measurement occasion
Pregnanediol-3-glucuronide values for all participants by measurement occasion
Logged estradiol values for all participants by measurement occasion
In the case of SHBG (Figure 2), within-subject variance is small relative to between-subject variance. There is little variation within subjects relative to the variation between subjects, resulting in small OR and OA estimates (0.48 and 1.78 respectively). The PDG values (Figure 3) illustrate the case in which within subject variation is high and overlap one another considerably, resulting in relatively large OR and OA estimates (5.17 and 10.27 respectively). Finally, Figure 4 depicts the case in which within-subject variance is small, but so is the variance between subjects. In this case, the small within-subject variance results in a small OA estimate (0.34), but because the within-subject variance is not small relative to the between-subject variance, the OR is relatively large (8.26).
Because ICCs include both within- and between-subject variance, ICCs closely followed OR rather than OA estimates.ICC(2,1) ranged from ICC(2,1) = 0.30 to ICC(2,1) = 0.88 (for LH and SHBG respectively, Table 3). The intraclass correlation coefficient ICCs for absolute agreement between the two analysis batches ranged from ICC (2,1) = 0.47 to ICC (2,1) = 0.96 (for E1 and SHBG respectively, Table 4). Estimates of ICCs were, generally, consistent across batches, with similar estimates based on analysis of all 5 occasions and for estimates based on each batch. The between batch ICC for E1, however, was less than 0.5, suggesting that the batch 1 ICC may be a better indicator than the ICC based on all samples. The percent of true variance accounted for by a single measurement ranged from 29% to 92% for LH and SHBG respectively.
Intraclass correlation coefficients between batches for analytes analyzed in 2 batches
We have provided estimates to the minimum number of measurement occasions required to ensure adequate reliability for two types of experimental aims. Analyses in epidemiologic studies involve calculations in which between-subject as well as within-subject variance is important. Therefore, OR will usually be the appropriate index of the minimum number of occasions needed to obtain a reliable estimate. Estimates of OR based on our sample suggest that only SHBG and E1S had sufficient temporal stability to be adequately reliable with a single measurement when the desired amount of variance to account for was set as low as 64%. A single measurement of any of the other analytes would be unlikely to account for even 50% of the true variance. For cases in which the within-subject variance is the only variance of interest, e.g., when the measured value of an analyte will be compared with a fixed standard, OA will be the appropriate index. The omission of between-subject variance from the formula for calculating this statistic produces very different results from OR. Several of the analytes that were adequately reliable with a single measurement or very few measurements, when between-subject variance was a factor, required higher numbers of measures when only within-subject variance was involved and vice versa.
This study confirms previous findings that SHBG may be reliably measured in premenopausal women using a single occasion. It also indicates that E1S may be reliably measured using one sample only. More importantly, our results suggest that none of the other analytes examined meet minimal reliability requirements that would permit confidence in single measures. These results are in agreement with the wide range if ICCs reported in previous studies [4-6]. Our conclusions are limited to the collection of samples at midluteal phase, however, and may not generalize to other phases of the menstrual cycle.
The use of ICCs to estimate the agreement between analysis batches differs from their use as an index of temporal reliability. The appropriate type of ICC for this purpose uses a mean of several values rather than single values and is typically higher than that calculated using single values. Though the ICCs between batches were higher than those estimating temporal reliability, they were relatively low, demonstrating the importance of measuring all samples in one batch when possible. As was previously noted [11], error due to time in storage will affect estimates of temporal reliability. Analyzing in multiple batches is one means of decreasing this source of error, but runs the risk of increasing error due to multiple batches. Until better estimates of the impact of storage time on each of these analytes are available, however, it will be difficult to draw conclusions about whether error due to multiple analysis batches or error due to storage time has the more detrimental effect on temporal reliability.
Several sources of error are effectively beyond researchers' capacity to control. For example, the validity and reliability of the best assay available for measuring a given analyte cannot be increased through improving study design. Other sources of error, however, can be dramatically reduced through the use of appropriate designs. These strategies may include, increasing the sample size to reduce the impact of random error, analyzing all samples in one batch, and using a sufficient number of repeated measures to obtain an adequately reliable estimate. It is also possible, though not uncontroversial, to control error statistically by correcting for attenuation using validation data [24].
Several improvements, in addition to a larger sample and more repeated measures, would have increased confidence in the results of our study. First, if the effects of storage time on the analytes were known, we could have taken into account the contributions of this source of variance to our temporal reliability estimates and distinguished its impact from that due to assay reliability. Second, obtaining blood and urine samples on day 5 following ovulation was most appropriate for the measurement of progesterone and near-optimal for SHBG, but may not have been the best day to obtain estimates of the other analytes [25]. Third, though our data were drawn from an intervention study in which no results approached significance, a more clearly homogeneous sample would have been preferable. Fourth, variation in menstrual cycle length and variance due to pulsatility of excretion were additional sources of error.
Finally, our estimates were based on targets that changed across measurements, and we could not assume error-free measurements. Consequently, we were not able to precisely distinguish between the contributions of assay reliability and the contributions of each analyte's natural variability to our estimates of temporal reliability. However, despite some limitations, this study provided significant new insights into the variability of sex hormones, gonadotropins, and urinary hormone metabolites in premenopausal women during a one-year period. Our estimates of temporal reliability represent the combined computation of the consistency of a measure across repeated measurements and the temporal fluctuations in the target of measurement.
Conclusions
Given the relatively large sample size for this analysis and the strictly controlled protocol to collect samples on the same day of the menstrual cycle, our results will be useful for designing future research projects exploring the role of sex hormones in the etiology of cancer and other diseases.
Competing interests
This project was supported by the Pharmavite Corporation in San Fernando, California and a Developmental Funds award from the Cancer Center Support grant to the Cancer Research Center of Hawaii (P30CA071789).
Authors' contributions
AW conceived of the study and performed the statistical analyses. GM was the primary investigator on the original study from which the data for this study was drawn and contributed to the design of this study. FS carried out the immunoassays and contributed to the writing-up of this study. AF participated in the study design and consulted with the authors.
Pre-publication history
The pre-publication history for this paper can be accessed here:
Acknowledgements
The authors gratefully acknowledge the valuable assistance, advice, and guidance provided by Lynne R. Wilkens, Dr PH, and Ian Pagano, MA, both of the Cancer Research Center of Hawaii. We are grateful to the women who donated their time and effort to participate in this study. The project was funded by a contract from the Pharmavite Corporation in San Fernando, California and by a Developmental Funds award from the Cancer Center Support grant to the Cancer Research Center of Hawaii (P30CA071789).
To investigate obstetric prognosis in sisters of preeclamptic women.
Methods
We identified consecutive 635 sib pairs from the Birth Registry data of Kuopio University Hospital who had their first delivery between January 1989 and December 1999 in our institution. Of these, in 530 pairs both sisters had non-preeclamptic pregnancies (the reference group), in 63 pairs one of the sisters had preeclampsia and the unaffected sisters were studied (study group I). In 42 pairs both sister's first delivery was affected (study group II). Pregnancy outcome measures in these groups were compared.
Results
Unaffected sisters of the index patients had uncompromised fetal growth in their pregnancies, and overall, as good obstetric outcomes as in the reference group. The data on affected sisters of the index patients showed an increased prematurity rate, and increased incidences of low birth weight and small-for-gestational age infants, as expected.
Conclusion
Unaffected sisters of the index patients had no signs of utero-placental insufficiency and they were at low risk with regard to adverse obstetric outcome, whereas affected sisters were high-risk. Clinically, affected versus unaffected status appears to be clear-cut in first-degree relatives regardless of their genetic susceptibility and unaffected sisters do not need special antepartum surveillance.
Background
Preeclampsia is a leading cause of maternal and perinatal morbidity and mortality[1] and the disease carries a tendency towards familial clustering [2-6]. Although the pattern of inheritance is not yet resolved, investigations into the genetic etiology of preeclampsia have yielded intriguing results implying that genes are responsible for the disease rather than shared environment [7]. Current concepts favor the hypothesis, that preeclampsia results from interplay of multiple genes and the environment, the disease being a polygenic trait with a strong maternal contribution [8]. First-degree relatives are known to have a fivefold increased risk of developing the disease compared with women with no family history of preeclampsia. Women above a certain threshold of this trait manifest the disease, whereas those below the threshold have a normal phenotype [9]. As far as we are aware of, studies looking at the obstetric outcome in unaffected sisters are still lacking. This study was undertaken to evaluate pregnancy outcome in unaffected sisters of preeclamptic women, to find out whether their genetic susceptibility is associated with adverse outcome. Such data is useful not only for scientific but also for counselling purposes.
Methods
The study was approved by the Research-Ethics Committee of Kuopio University Hospital.
The total number of deliveries during the study period was 23 772, and of those, 9576 women were primiparous and 14 196 women were multiparous. The study material comprised consecutive 635 sib pairs, both of whom had their first delivery in the Department of Obstetrics and Gynecology, Kuopio University Hospital, between January 1989 and December 1999. The data for this study were prospectively collected, and the case records were retrospectively analyzed. In 530 pairs both sisters had normotensive pregnancy and one sister of each sib pair (the one who gave birth later) was included in the analysis to constitute the reference group. In 63 sib pairs, one of the sisters had preeclampsia in her first pregnancy and study group I was derived from the unaffected sisters. In 42 pairs, both sisters had preeclampsia and one sister of each sib pair was incorporated (the one who gave birth later) to form study group II. All pregnant women were monitored in an identical manner as outpatients until the development of preeclampsia or another pregnancy complication requiring hospitalization. Basic clinical data were collected at prenatal visits and at delivery for all the target population by the team that took care of treatment.
Hypertensive complications of pregnancy were classified as advocated by the U.S. National Institute of Health Working Group on Hypertension in Pregnancy [10]. Preeclampsia wss defined as the development of hypertension and new-onset proteinuria (greater than 300 mg of urinary protein in 24 h) in women with no proteinuria at baseline. For those with a baseline diastolic pressure of 90 mmHg, hypertension was defined as a rise of at least 25 mmHg, measured on two consecutive occasions at least 24 h apart. For those with an initial diastolic pressure of 90 mmHg or above, an increament of at least 15 mmHg was required [11].
For data analysis, the following were used as evidence of fetal compromise: intrauterine death, admission to a neonatal intensive care unit, small-for-gestational -age delivery, low birth weight, prematurity, low Apgar scores at 1 and 5 min, and fetal acidosis at birth. Reference material values for the birth weight percentiles were obtained from own records [12].
Differences between study subjects and the reference group were tested for significance by using X2 statistics or Fisher's exact tests, as appropriate. Student's t-test was used to analyze continuous variables. Differences were considered to be significant when P < 0.05.
Results
The mean maternal age in study group I (± SD) was 24.3 years (± 4.7 years) and it was 25.5 years (± 4.5 years) in the reference group (P = 0.04). In study group II, the mean maternal age was 24.3 years (± 3.4 years) (P = 0.04). Characteristics of women in study groups I and II, who gave birth at our hospital, during the 10-year period are compared against those of the reference group in Table 1. Apart from maternal age, the demographic data investigated in this study were comparable in the three groups.
Maternal Risk Factors
reference group
unaffected sisters of preeclamptic index patients
affected sisters of preeclamptic index patients
risk factor
N = 530
N = 63
N = 42
Age < 18 years
14 (2.64%)
2 (3.17%)
1 (2.38%)
Age > 35 years
11 (2.08%)
0 P = 0.617a
0 P = 1.00a
Miscarriage
39 (7.34%)
4 (6.35%) P = 1.00a
5 (11.90%) P = 0.358a
Pregravid body mass index > 25
73 (13.73%)
4 (6.78%) P = 0.134
10 (22.5%) P = 0.129
Unemployed
74 (13.94 %)
7 (11.11 %) P = 0.537
2 (4.76 %) P = 0.092
Not married
244 (46.04%)
21 (33.33%) P = 0.055
14 (33.33%) P = 0.111
IUD before pregnancy
4 (0.75 %)
0 P = 1.00a
0 P = 1.00a
Infertility
27 (5.09 %)
2 (3.17 %) P = 0.758a
3 (7.14 %) P = 0.476a
Smoking (> 5 cigarettes/day)
49 (9.23 %)
2 (3.17 %) P = 0.105
1 (2.38 %) P = 0.162a
Alcohol consumption
17 (3.20 %)
0 P = 0.240a
1 (2.38 %) P = 1.00a
Maternal diabetes
7 (1.32 %)
1 (1.59 %) P = 0.594a
0 P = 1.00a
Chronic illness
25 (4.71 %)
3 (4.76 %) P = 1.00a
1 (2.38 %) P = 0.712a
a Fisher's exact test
Table 2. summarizes the frequencies of various pregnancy and delivery characteristics in the two study groups and in the control group. Preeclamptic women whose sisters also had preeclampsia in their first pregnancy underwent cesarean deliveries more often than the control women. However, the rate of vaginal operative deliveries did not differ between the groups. 1.59% of the women in study group I and 0.57% in the reference group had pregnancy-induced hypertension. Otherwise, the pregnancy characteristics were similar in these groups.
Pregnancy and Delivery Characteristics
reference group
unaffected sisters of preeclamptic index patients
affected sisters of preeclamptic index patients
characteristics
N = 530
N = 63
N = 42
Pregnancy-induced hypertension
3 (0.57 %)
1 (1.59 %) P = 0.365a
3 (7.14 %) P < 0.001a
Placental abruption
2 (0.38 %)
0 P = 1.00a
1 (2.38 %) P = 0.205a
Placenta previa
4 (0.75 %)
2 (3.17 %) P = 0.126a
0 P = 1.00a
Prolonged gravidarum (>42 w)
39 (7.36 %)
6 (9.52 %) P = 0.612a
1 (2.38 %) P = 0.346a
Female fetus
261 (49.25 %)
30 (47.62 %) P = 0.807
23 (54.76 %) P = 0.491
Isoimmunization (Rh)
1 (0.19 %)
0 P = 1.00a
0 P = 1.00a
Low hemoglobin concentration (< 100 g/L)
7 (1.32 %)
0 P = 1.00a
0 P = 1.00a
Cesarean delivery
97 (18.30 %)
9 (14.29 %) P = 0.432
14 (33.33 %) P = 0.018
Forceps/vacuum
47 (8.87 %)
7 (11.11 %) P = 0.558
2 (4.76 %) P = 0.566a
Bloody amniotic fluid
11 (2.07 %)
2 (3.17 %) P = 0.638a
2 (4.76 %) P = 0.246a
Meconium-stained amniotic fluid
73 (13.75 %)
9 (14.29 %) P = 0.907
4 (9.52 %) P = 0.440
Placental/fetal mass ratio (%)
16.6 %
16.2 % P = 0.51b
17.8 % P = 0.46b
a Fisher's exact test, b Student's t-test
The mean birth weight (± SD) among those delivering at term (after 37 gestational weeks) was 3530 g (± 444 g) in the reference group, 3471 g (± 378 g) in study group I (P = 0.33) and 3398 g (± 494 g) in study group II (P = 0.12). Table 3. shows the pregnancy outcome measures in the reference and study groups. As expected, the incidence of prematurity (P < 0.001), low birth weight (P < 0.001) and small-for-gestational age infants (P = 0.06) was increased in sisters affected by preeclampsia (group II), whereas there was no difference in the rate of fetal death between the groups. Obstetric outcomes in study group I were comparable with those in the reference group.
Obstetric Outcomes
reference group
unaffected sisters of preeclamptic index patients
affected sisters of preeclamptic index patients
outcome
N = 530
N = 63
N = 42
Admission to a neonatal intensive care unit
34 (6.42 %)
3 (4.76 %) P = 0.786a
9 (21.43 %) P = 0.002a
Intrauterine fetal death
3 (0.57 %)
0 P = 1.00a
0 P = 1.00a
Prematurity (delivery before 37 weeks)
25 (4.72 %)
1 (1.59 %) P = 0.344a
10 (23.80 %) P < 0.001
Low birth weight (<2500 g)
16 (3.02 %)
0 P = 0.398a
12 (28.57 %) P < 0.001
Small for gestational age (< 10th percentile)
61 (11.51 %)
8 (12.70 %) P = 0.781
9 (21.43 %) P = 0.059
Low Apgar score (<7 at 1 min)
31 (5.85 %)
4 (6.35 %) P = 0.780a
5 (11.90 %) P = 0.173a
Low Apgar score (<7 at 5 min)
7 (1.32 %)
0 P = 1.00a
0 P = 1.00a
Fetal venous pH <7.15 at birth
11 (2.1 %)
3 (4.76 %) P = 0.175a
2 (4.76 %) P = 0.250a
a Fisher's exact test
Discussion
The main finding of this study was, that unaffected sisters of preeclamptic index patients had normal outcomes in their first pregnancy, the course of pregnancy being comparable to that in the general obstetric population. Basically, no differences were noted in the reproductive risk factors of the groups studied. Among affected sisters, the rate of prematurity, low birth weight, and small-for-gestational age infants were increased, as expected [13]. Although a trial of this size cannot reliably detect differences in rare complications, such as neonatal death ascribable to familial risk, the number of cases in the present study is sufficient to make statistically valid comparisons with regard to commonly used outcome variables. However, it is not known whether similar changes are present in unaffected first-degree relatives during pregnancy because of their genetic susceptibility to preeclampsia which in turn would adversely affect their pregnancies. Any adverse effect was undetectable in the present study, since only one (1.59%) of the unaffected women with a first-degree relative with preeclampsia developed pregnancy-induced hypertension and there were no signs of chronic utero-placental insufficiency.
Caruso et al. demonstrated in their study of pregnant women that preeclampsia, but not gestational hypertension, was characterized by atherogenic metabolic features similar to those of patients with insulin resistance syndrome, such as lower insulin sensitivity, and higher levels of triglycerides and nonesterified fatty acids [14]. Thus, the clinical definition seems appropiate, and in clinical work, preeclampsia phenotypes defined by other criteria such as glomerular endotheliosis are too difficult to assess routinely [15].
Preeclampsia is a heterogeneous disorder, and women with various medical problems are at risk of developing the disease [16-18]. Many of these conditions are known to be governed by some components which may have genetic origin, e.g. genetic risks associated with essential hypertension and diabetes mellitus, and mitochondrial abnormalities, which in turn, are characterized by microvascular disease [19-21]. Our results may be useful in counselling patients and their first-degree relatives. Genetic susceptibility to preeclampsia has minor effects, if any, on pregnancy outcome in first-degree relatives of index patients, if they do not develop the disease. Basically, routine antenatal care in which blood pressure and urine dipstick are checked each visit is sufficient for women with an affected sister and there is no need to initiate special fetal monitoring in these pregnancies. Accordingly, the results may have implications in genetic linkage studies [22], since the phenotype in unaffected sisters of preeclamptic women can be considered normal in terms of clinical outcome measures. In other words, this observation provides justification to stratify pregnant women into the categories of affected and unaffected individuals which is the current practice in clinical work [23].
Conclusions
The sisters of preeclamptic women are at low risk with regard to adverse pregnancy outcome if they do not develop preeclampsia. They have pregnancy outcomes comparable to that of the general obstetric population, and routine antenatal follow-up and management is sufficient in these cases. Although the etiology and pathogenesis are as yet unresolved, the complexity of the disease phenotype supports the theory of a polygenic trait. Women belonging to the liability group could have considerable pathological changes in their placental tissue without developing the maternal syndrome, and this might explain, why such changes are occasionally observed in fetal growth retardation, which has also been called normotensive preeclampsia [24]. However, in the present study genetic liability in unaffected first-degree relatives of preeclamptic index patients was not associated with any clinical phenotype
Competing interests
None declared.
Authors' contributions
NHE participated in the design of the study and performed the statistical analysis and drafted the manuscript.
SHE participated in the design of the study, drafted the manuscript and coordination.
PKI conceived of the study and participated in its desingn and coordination.
All authors has read and approved the final manuscript.
Pre-publication history
The pre-publication history for this paper can be accessed here:
Ovarian antibodies as detected by indirect immunofluorescence have been used to detect ovarian autoimmunity, but to our knowledge the rate of false positive findings using this method has never been reported.
Methods
Here we examine whether a commercially available ovarian antibody test system, using cynomologous monkey ovary, might be useful in the diagnosis of autoimmune premature ovarian failure. The test was performed in a blinded manner in 26 young women with 46,XX spontaneous premature ovarian failure, in 26 control women with regular menstrual cycles (matched for age, race, and parity) and 26 control men (matched for age and race). We also compared the frequency of other autoantibodies associated with ovarian autoimmunity.
Results
As a group young women with premature ovarian failure had an increased incidence of thyroid and gastric parietal cell autoimmunity (p < 0.05). Unexpectedly, however, nearly one third (31%) of normal control women had ovarian antibodies using the commercially available test. One half of young women with premature ovarian failure were found to have ovarian antibodies (P = 0.26). In our own laboratory we found similar results and we were unable to improve the specificity of the test. None of 26 men were found to have ovarian antibodies (P < 0.001).
Conclusion
Since approximately one third of normal women were found to have ovarian antibodies using the system under study, we conclude that ovarian antibodies as detected by this indirect immunofluorescence method have poor specificity. The specificity of any ovarian antibody test should be established before it is used clinically.
Background
Autoimmunity is a well-established mechanism of premature ovarian failure. [1-4]. It has been suggested that the presence of ovarian antibodies may be helpful in the diagnosis of ovarian autoimmunity [5,6]. However, presently there is no validated serum marker that can establish a diagnosis of autoimmune premature ovarian failure with certainty [7]. Young women can experience ovarian failure by several mechanisms other than autoimmunity [8]. A false positive diagnostic test indicating autoimmunity as the mechanism of spontaneous premature ovarian failure could put young women at risk for inappropriate therapy. Such therapy could have serious consequences, such as the development of osteonecrosis[4]
Some women with premature ovarian failure have ovarian follicles that function intermittently [9,10]. and pregnancies have occurred after the diagnosis of premature ovarian failure [11-14]. Clinicians are in need of an accurate method for the diagnosis of autoimmune oophoritis, a test with proven sensitivity and specificity [15]. Here we conduct an investigation to determine whether a commercially available ovarian antibody test (Immunodiagnostic Laboratories, Inc., San Leandro, CA) might be useful in the diagnosis of autoimmune premature ovarian failure. We performed the test in women with premature ovarian failure, women with normal ovarian function, and in men. We found that this commercially available test has a high incidence of false positives. Therefore, this test would not be expected to be useful in the diagnosis of autoimmune premature ovarian failure.
MethodsPatients and Controls
By local advertisement we recruited 26 control women with regular menstrual cycles (matched for age and parity to our patients) and 26 control men (matched for age). The controls were also matched to patients for race. We recruited patients with spontaneous premature ovarian failure by letters to physicians and notices in medical journals. The National Institute of Child Health and Human Development Institutional Review Board approved the protocol. We diagnosed premature ovarian failure in women who before the age of 40 had experienced amenorrhea in association with serum FSH levels greater than 40 mIU/mL (confirmed on two separate occasions, at least one month apart). Twenty-six patients with premature ovarian failure participated in this study (median age of 33 years, range 18–39 years). The women had been diagnosed at a median age of 30 years (range 15–38 years). The median time since diagnosis was 2 years (range 0.33–12 years). All patients had a normal karyotype and had no history of chemotherapy or radiation. Six patients (26%) had hypothyroidism, one patient had Addison disease, and one patient had Raynaud syndrome.
Ovarian Antibodies
We sent blinded specimens to Immunodiagnostic Laboratories, Inc. (San Leandro, CA) to be tested for ovarian antibodies using an indirect immunofluorescence test system supplied by Scimedx Inc. (Denville, NJ). The kit includes frozen sections of cynomologous monkey ovary as tissue substrate, a positive control of human serum known to contain antibodies against the zona pellucida, a negative control serum, and fluorescein isothiocyanate (FITC) conjugated goat antibodies against human immunoglobulins including IgG, IgM and IgA. Binding of antibody to the zona pellucida at a 1:10 dilution was reported by the laboratory on a scale of 0 to 3+ according to the intensity of fluorescence: 0, negative; 1+, weak; 2+ moderate or 3+ strong fluorescence (Figure 1).
Indirect immunofluorescence using cynomologous monkey ovary. Shown are a representative sample of a positive control (A), a 3+ positive patient sample (B), and a negative control (C). The scale bar represents 50 μm. Each arrow points to a zona pellucida. Images were prepared in our laboratory using a commercially available test system (Scimedx, Inc., Denville, NJ).
We also tested serum in our own laboratory using this same test system (Scimedx, Inc., Denville, NJ). Aliquoted serum samples of patients and controls were stored at -80°C. After thawing at room temperature, serum samples were diluted 1:10 with PBS. We confirmed the adequacy of monkey ovary sections from individual lots in an unblinded manner before using them. For each experiment slides of monkey ovary with the same lot number were allowed to equilibrate to room temperature. Twenty microliters of blinded sample were applied to the tissue substrate slides. A slide containing positive and negative controls was run unblinded with each experiment. Slides were placed into a moist covered chamber and incubated for 30 minutes at room temperature. Next, the slides were washed in PBS for three 10-minute soaks. After blotting the slides, 20 uL of FITC-conjugated goat antibodies against human immunoglobulins was delivered to each sample and slides were again incubated in the moist chamber for 30 minutes. The washings and blotting were repeated as above. Mounting medium and a coverslip were applied. The slides were kept in a dark moist chamber until evaluated using a Zeiss Axiophot Fluorescent Microscope (Carl Zeiss Inc., Thornwood, NY) at 40× magnification with emission at 400 nM and excitation at 520–560 nM. We graded fluorescence of the zona pellucida on the same 0 to 3+ scale as noted above.
Reproducibility
To evaluate the reproducibility of the test we selected four patient samples previously determined by this technique to be 0, 1+, 2+, and 3+ for zona pellucida antibodies. In a blinded manner, we tested each sample 15 times (5 tests each on three separate days).
Other Autoantibodies
Antinuclear antibody (ANA) titer was determined by indirect immunofluorescence using Hep-2 substrate. Rheumatoid factor was determined by latex agglutination. Antiparietal cell antibodies were measured by indirect immunofluorescence using rat stomach (Smith Kline Beecham Clinical Laboratories, Van Nuys, CA). Antithyroid peroxidase (anti-TPO) binding activity was determined by a radiobinding assay using Iodine125 labelled human recombinant thyroid peroxidase (Nichols Institute, San Juan Capistano, CA).
Statistics
We used the Fischer exact test, chi-square, and the Spearman rank correlation test as appropriate. We used one-tailed tests to look for an increase in autoantibody as compared to controls and we set P < 0.05 as significant. Statistical analysis was performed using Sigma Stat Software (Jandel Scientific, San Rafael, CA).
ResultsOvarian antibodiesBy commercial laboratory
Eight of 26 control women (31%) and 13 of 26 women with spontaneous premature ovarian failure (50%) had ovarian antibodies as detected by the commercial laboratory (Table 1). The frequencies were not significantly different (P = 0.26, chi-square test).
Results of ovarian antibody tests in women with premature ovarian failure (POF) and control women (CTRL) matched for age and parity.
Commerciala
NIHb
NIH Revised Methodc
Sample
POF
CTRL
POF
CTRL
POF
CTRL
1
1+
-
1+
1+
+
-
2
-
-
-
1+
-
-
3
1+
1+
1+
1+
+
+
4
-
-
-
1+
-
+
5
-
1+
-
2+
-
+
6
-
-
-
-
-
+
7
-
2+
1+
3+
+
-
8
1+
-
1+
-
+
+
9
1+
-
2+
-
+
-
10
1+
-
-
-
+
-
11
1+
-
3+
-
+
+
12
-
1+
-
-
-
-
13
-
-
-
-
-
+
14
-
1+
-
1+
+
-
15
1+
1+
1+
2+
-
-
16
-
1+
-
-
-
+
17
-
-
-
-
-
-
18
1+
-
1+
2+
+
+
19
1+
-
1+
-
+
-
20
-
-
-
-
-
+
21
-
1+
-
-
-
-
22
1+
-
1+
1+
+
-
23
1+
-
-
-
-
-
24
-
-
-
-
+
-
25
2+
-
3+
-
+
+
26
1+
-
1+
-
+
-
Total (+)
13
8
12
10
14
11
aResults by commercial laboratory bResults by our laboratory using the commercial method cResults by our laboratory with outcome measure as simply positive (+) or negative (-)
By our own laboratory
When we evaluated these same samples for ovarian antibodies in our own laboratory we obtained similar results. Twelve of 26 patients (46%) with premature ovarian failure had ovarian antibodies versus 10 of 26 control women (38%) (P = 0.78, chi-square test; Table 1).
Reproducibility
We could not reliably reproduce the four categories of ovarian antibody test grading. Although our results correlated with the known graded samples (r = 0.72, P < 0.05), the 1+ readings were reproduced in only 8 of the 15 tests (53%) and the 2+ readings were reproduced in only 2 of the 15 tests (13%) (Table 2). In contrast, the 0 and 3+ readings were reproduced in 13 (87%) and 12 (80%) of the 15 respective tests (Table 2).
Reproducibility of the ovarian antibody grading system.
Known Graded Sample
0
1+
2+
3+
Test Result
0
13
6
2
2
1+
1
8
2
0
2+
1
1
2
1
3+
0
0
9
12
Percent Reproduced
87
53
13
80
Efforts to improve the test
Because the 1+ and 2+ readings had poor reproducibility, we were concerned that this inaccuracy might have impaired our ability to detect a difference between patients and control women. We therefore redefined the outcome measure to be read simply as positive (+) or negative (-). Positive was defined as the presence of homogeneous fluorescence on the zona pellucida above background and negative was defined as any fluorescence less than this. Tests were then repeated using serum from the 26 patients and 26 control women. We were unable to demonstrate a significant difference using the revised outcome measure. Patient serum was positive in 14 of 26 samples (54%) and control serum was positive in 11 of 26 samples (42%) (Table 1). When we tested the sera on three separate occasions with this method, we found 9 patient sera (35%) and 6 control sera (23%) to be consistently positive for zona pellucida antibody (P = 0.38).
We were unable to improve the test by using higher dilutions of serum (1:20, 1:40, and 1:80). All 26 patients and all 26 controls were negative for ovarian antibodies at 1:20. The positive control sample was positive at 1:40 and negative at 1:80.
Specificity
To evaluate the specificity of the test we compared sera from 26 normal women with sera from 26 normal men matched for age. The tests were performed at 1:10 dilution using the positive and negative outcome measure as discussed above. Women were significantly more likely to have ovarian antibodies than men. Eleven of 26 normal women (42%) had ovarian antibodies whereas none of the 26 normal men had these antibodies (P < 0.001). These findings were confirmed in a separate experiment (9/26 versus 0/26, P < 0.002).
Other Auto-antibodies
Results of other antibody tests are shown in Table 3. Women with spontaneous premature ovarian failure were significantly more likely than control women to have parietal cell antibodies (P < 0.05, Fischer's exact test) and thyroid peroxidase antibodies (P < 0.05, Fischer's exact test).
Frequency of other selected autoantibodies in women with premature ovarian failure (POF) and control women matched for age and parity.
Antibody Test
Women with POFa n = 26
Control Womena n = 26
Antithyroid peroxidase ≥ 0.9 u/mL
9 (35)b
3 (12)
Antiparietal Cell ≥ 1:10
4 (15)b
0 (0)
ANA ≥ 1:80
5 (19)
6 (23)
Rheumatoid Factor ≥ 1:320
2 (8)
0 (0)
aValues are the number of positive tests, in parentheses are percent positive bSignificantly different from controls (P < 0.05, Fischer's exact test).
Discussion
There is evidence to suggest that some patients with spontaneous premature ovarian may have clinically significant circulating ovarian antibodies. In a seminal study reported in 1979, Coulam and Ryan demonstrated that patients with premature ovarian failure, as a group, have ovarian antibodies present in their sera as determined by immunoprecipitation of radiolabeled human ovarian proteins [16]. However, the identity of these specific ovarian antigen(s) remains unknown. It should pointed out, however, that immunoblotting studies have failed to reveal a consistent pattern of binding using the sera of patients with premature ovarian failure[7].
Presently, there is no proven sensitive and specific serum test to confirm that a woman has ovarian failure on an autoimmune basis[7]. Histologic evaluation of the ovary is the only way to diagnose autoimmune oophoritis with certainty [17]. Nonetheless, some clinicians currently use commercial ovarian antibody tests for the diagnosis of autoimmune premature ovarian failure and make treatment recommendations based on these findings. This approach can have severe adverse consequences such as the development of osteonecrosis related to glucocorticoid therapy[4]. Here we demonstrate that detection of ovarian antibodies by indirect immunofluorescence using one specific commercial test system has poor specificity as a diagnostic test. The test is positive in nearly one-third of normal women. Therefore, the test is not a useful diagnostic marker of autoimmune premature ovarian failure.
We found that women with premature ovarian failure are significantly more likely to have positive thyroid peroxidase and parietal cell autoantibodies than women with normal ovarian function. In this regard our findings are in agreement with previous studies [18-20]. In contrast to one previous study, however [21], we did not find a significantly higher prevalence of antinuclear antibodies in women with premature ovarian failure.
Because little is known about the pathophysiology of human autoimmune oophoritis, experimental animal models have been used to gain insight into possible mechanisms [22,23]. The fact that we found an increased frequency of thyroid and gastric parietal cell antibodies in our patients suggests similarity to a model of autoimmune ovarian failure that can be induced in certain strains of mice by performing neonatal thymectomy. These mice also develop autoimmunity against thyroid and gastric parietal cells [22]. It is well established that these mice with ovarian autoimmunity also develop circulating autoantibodies against the oocyte cytoplasm, the zona pellucida, and against steroid producing cells [22]. While we found evidence that one half of our patients had specific antibodies directed against the zona pellucida, we did not find any consistent pattern of immunofluorescence to suggest that our patients had antibodies directed against the oocyte cytoplasm or the steroid producing cells of the ovary. It is noteworthy that immune serum from mice with ovarian autoimmunity reacts specifically with an oocyte-specific protein called MATER [24,25]. We have subsequently identified the homologous human MATER gene and protein[26] In ongoing work we are evaluating the value of MATER antibodies as a marker for autoimmune ovarian failure in women.
The fact that we did not detect ovarian antibodies binding to the zona pellucida in any of 26 men (p < 0.001) is intriguing. This finding supports proposals that pre-B cells undergo positive selection directed by the presence of surface heavy chains with low albeit sufficient affinity to the autoantigen. It appears then, that men, lacking the specific ovarian self-antigen, fail to provide positive selection for these pre-B cells clones. Also, this finding suggests that the detection system is specific for zona pellucida, and that further refinement of the assay system by using pure human recombinant zona pellucida protein might be useful. Sacco and Moghissi detected zona pellucida antibodies in both men and women by indirect immunofluorescence, but they used porcine rather than primate zonae pellicidae [27]. Normal animal and human serum are known to contain a wide range of low-titer autoantibodies that have been termed "natural autoantibodies" [28]. Even B cells from newborn mice and human cord blood produce these natural autoantibodies [29-31]. The low titer ovarian antibodies that we demonstrate here in the serum of normal women are likely natural autoantibodies, and not antibodies with any pathophysiologic significance.
Identifying women with autoimmune premature ovarian failure presents the opportunity to restore ovarian function by proper immune modulation therapy. However, at present we don't really know how many women with premature ovarian failure develop it due to autoimmunity. Currently, ovarian biopsy is the only way to diagnose autoimmune premature ovarian failure with certainty. However, because there is no treatment proven safe and effective to restore fertility, in our view, ovarian biopsy is not indicated outside a controlled trial [17].
Conclusions
We found that approximately one-third of women with normal ovarian function have ovarian autoantibodies detected by indirect immunofluorescence using monkey ovary as substrate. Hence, this ovarian antibody test as presently performed has poor specificity, and therefore appears to have no benefit in the evaluation and management of autoimmune premature ovarian failure. The specificity of any ovarian antibody test should be established before it is used clinically.
Competing interests
None declared.
Authors' contributions
J.A.N. and Z-B.T. participated in the design of the study and carried out the laboratory investigation. All authors participated in the analysis of the data and the preparation of the manuscript. L.M.N. conceived the study, participated in its design and coordination, recruited the patients, and was responsible for their evaluation and clinical care. All authors read and approved the final manuscript.
Pre-publication history
The pre-publication history for this paper can be accessed here:
Acknowledgements
We thank Rene Kimzey, RN for help in recruitment, patient care and protocol coordination. SNK was the recipient of a scholarship by the Alexandros S. Onassis Public Benefit Foundation
The Malaysian Ministry of Health promotes breast self-examination (BSE) for all women, and Pap smear screening every three years for all sexually active women ages 20 years and above. The objectives of this paper were to examine the practice of these two screening tests among women production workers in electronics factories, and to identify factors related to practice.
Methods
This was a cross-sectional survey of women production workers from ten electronics factories. Data was collected by a self-administered questionnaire from a total of 1,720 women. The chi-square test, odds ratio and binomial logistic regression were used in bivariate and multivariate analysis.
Results
Prevalence rates were 24.4% for BSE once a month, and 18.4% for Pap smear examination within the last three years. Women who were significantly more likely to perform BSE every month were 30 years and older, Malays, with upper secondary education and above, answered the BSE question correctly, and had a Pap smear within the last three years. The proportion of women who had a Pap smear within the last three years were significantly higher among those who were older, married, with young children, on the contraceptive pill or intra-uterine device, had a medical examination within the last five years, answered the Pap smear question correctly, and performed BSE monthly.
Conclusion
Screening practice rates in this study were low when compared to national rates. Socio-demographic and health care factors significantly associated with screening practice are indicative of barriers which should be further understood so that more effective educational and promotional strategies could be developed.
Background
The effectiveness of breast self examination (BSE) in lowering breast cancer mortality has been extensively studied. After reviewing studies published between 1966 to October 2000, the Canadian Task Force on Preventive Health Care found no evidence that routine BSE teaching or practice is beneficial in terms of lower risk of breast cancer death, or earlier stage of cancer at diagnosis [1]. Instead, they found that women who were taught BSE and practised it faced a higher rate of physician visits, and higher risk of anxiety and benign biopsies. Based on these findings, they recommended that routine teaching of BSE be discontinued, particularly for women aged 40 to 69 years. Following on this, the final results of a 10–11 year randomized controlled trial in Shanghai were published, confirming that intensive instruction in BSE did not reduce breast cancer deaths [2].
Nevertheless, there is disagreement with the recommendation of discontinuing BSE advocacy among women [3-5]. Those who disagree point out that many breast tumours are discovered by women themselves, and that the practice of BSE makes women more aware of their breasts and therefore more likely to detect tumours, even if they did not detect them while doing BSE. In developing countries, it is considered to be a simple, inexpensive, non-invasive, and non-hazardous intervention, which is not only acceptable, cost-effective and appropriate, but also encourages women to take an active responsibility in preventive health [6]. It should be pointed out, however, that spending health resources on BSE, if it is ineffective, detracts from focussing on trying to find other ways of prevention.
On the other hand, the utilization of the Pap smear for the early detection of cervical cancer is better established, although the reliability of the test varies and is dependent on the expertise of the health professionals who take the smear as well as those who examine it [7]. In developed countries such as the United Kingdom, the recognition that Pap smear screening could result in false alarms or false reassurance, overdiagnosis and overtreatment has led to a recent emphasis on 'informed uptake', that is, that women should be made fully aware of all positive and negative consequences before undergoing screening. In some developing countries, the cost-effectiveness of Pap smear screening programmes has been questioned because they have not resulted in declining cervical cancer rates due to inefficient screening programmes and poor quality tests [8].
In Malaysia, breast cancer is the leading cause of cancer deaths among women, accounting for about 11% of all medically certified deaths in the country [9,6]. The national incidence rate is not available, but it is estimated that 1,200 new cases occur every year, and an increasing trend of cases among younger women has been observed. The Ministry of Health promotes the practice of monthly BSE for women above the age of 20 years, and annual clinical breast examination by a medical or paramedical personnel [6]. Mammography is not widely available as a screening method, and it is only recommended for women at special risk of breast cancer, identified as those who have had breast cancer in one breast, and those with a mother or sister who have breast cancer. BSE is primarily taught by nurses to women who attend government health clinics, as well as family planning clinics run by the non-governmental family planning associations.
Unlike breast cancer, mortality from cervical cancer has declined in relative importance, although it is still the second most important cancer among women in Malaysia [10,11]. The national incidence rate for cervical cancer is not known, but over the last ten years, there have been 2,000–3,000 admissions per year to government hospitals. The policy of the Ministry of Health is to promote and provide Pap smear screening to all women between 20–65 years of age [11]. Official recommendation is for women to undergo the Pap smear test annually in the initial two years, and subsequently, once every three years, with priorities for sexually active women who are more than 35 years old, have more than five children, have practised contraception for more than five years or who are new acceptors of family planning services, and women diagnosed with sexually transmitted diseases. Women who attend postnatal and family planning services are primary targets.
Although the policy is to target all sexually active women, in practice, married women have more access due to a variety of reasons. One of the reasons is that this is a recent policy change, and prior to 1995, when a nation-wide campaign on prevention of breast and cervical cancer was carried out, the official policy was to provide cervical cancer screening for married women only.
There is currently no system in the country for collecting routine information on the practice of BSE among women, and available information on Pap smear screening is only from government facilities. As such, both were included in the scope of the country's Second National Health and Morbidity Survey (NHMS) 1996 [6,11].
Likewise, reflecting national concerns, a government-funded research project on health status and lifestyle of women workers in the electronics industry also included these two topics. This paper uses data from this recent study to examine the prevalence of BSE and Pap smear screening and the relationship of various socio-demographic and health care factors with the practice of the two screening tests among electronics women workers.
MethodsSelection of factories
This was a cross-sectional analytical survey of women workers in selected electronics factories carried out by four research centres – Universiti Sains Malaysia, Penang (USM Pg), Universiti Sains Malaysia, Kelantan (USM Kel), Universiti Kebangsaan Malaysia (UKM) and Universiti Putra Malaysia with the help of the National Institute for Occupational Safety and Health (UPM-NIOSH). The inclusion criteria used for selecting factories was that they must be electronics assembly factories, have at least 500 or more women production workers and have been in production for at least two years. Four geographical areas were identified for the research, and each research centre took responsibility for one area – USM Pg for Penang (a north-western state), USM Kel for Kota Baru (the capital of the north-eastern state of Kelantan), UKM for Hulu Kelang Free Trade Zone (FTZ) and UPM-NIOSH for Bangi FTZ (both FTZs are in the west-central state of Selangor). A target sample size of 2000 women workers, or 500 per centre, was determined as sufficient for analysis, based upon limited resources.
In Kota Baru, there were three factories which fulfilled the criteria. All three were invited to participate in the study, and one agreed. In the Hulu Kelang FTZ, there were two eligible factories, both of which were invited, but only one agreed. In the Bangi FTZ all seven eligible factories were invited but only five agreed to participate. The UPM-NIOSH centre was not able to meet the target number of 500 respondents from these five factories, and through personal contacts, managed to get the cooperation of another factory located in the Sungei Way FTZ (also in Selangor state). In Penang, 15 factories were purposively selected from 160 in the Penang Development Corporation (Corporate Investment Section) 1995 list based on the inclusion criteria. These 15 factories were invited to participate in the study, but only one agreed. In order to make up the target sample size, the USM Pg centre enlisted the cooperation of a factory in Sungei Petani (in the northwestern state of Kedah) through a personal contact.
In summary therefore, a total of 27 factories in four geographical areas were invited to take part in the study, but only eight factories agreed to participate. In order to achieve the target sample size, two more factories were included, both of which were not within the original four research areas. Finally, therefore, ten factories located in six areas participated in the study.
Selection of respondents and data collection
Varying levels of cooperation were obtained from the ten factories in the selection of respondents and in data collection. It was determined, however, that respondents had to be women between the ages of 17 and 55 years, Malaysian citizens, and had worked at least a year as production workers (below supervisory level) in the factory where they were presently working. The requisite age span was determined based upon the legal age for working and the retirement age. In Kota Baru (Factory A) (Table 1), full cooperation was obtained from the factory, and all the women workers who fitted the selection criteria were included in the study. In the Penang factory (Factory B), full cooperation was also obtained, and researchers were able to randomly select 250 workers from a list provided by the management. In these two factories, the selected workers were released during working time to take part in the survey, and there were no workers who declined to participate. The percentage of completed questionnaires out of the total number of workers in the sample were 97.6% for Factory A and 98.0% for Factory B.
The sample
Research Centre
Location
Factory
Estimated no. of women production workers in factory
No. of women workers in sample (No. of questionnaires obtained)
No. of questionnaires in dataset (after cleaning)
Selection of workers
Response rate (%)
Location for data collection
Time for data collection
USM Kel
Kota Baru
A
500
250
244
All eligible workers
97.61
Factory
Working time
USM Pg
Penang
B
800
255
250
Random Sampling
98.01
Factory
Working time
Sungei Petani
C
700
260
250
Volunteers
(35.7)2
Factory
Own time
UKM
Hulu Kelang FTZ
D
3500
505
490
Volunteers
(14.0)2
Factory
Own time
UPM
Bangi FTZ
E
500
526
486
Volunteers
(12.4)2
Factory and hostel
Own time
Bangi FTZ
F
1000
Bangi FTZ
G3
1100 (400)
Bangi FTZ
H3
600 (230)
Bangi FTZ
I3
1200 (500)
Sungei Way FTZ
J
1300
All
Total
112004
1796
1720
(18.2)4
1Response rate calculated from number of questionnaires in dataset (after cleaning) as percentage of women workers in sample. 2Response rate calculated from number of questionnaires in dataset (after cleaning) as percentage of estimated number of women workers in the population from which the volunteers were drawn, i.e. the catchment population. For Factory C, D, E, F and J, the catchment population is the estimated number of women workers in the factory, while for Factory G, H and I, it is the estimated number of women workers in the hostels. 3These factories only allowed data collection at their hostels, not on factory premises. Estimated number of occupants in their hostels are given in brackets. 4Although the total estimated number of women workers in all the factories is 11,200, however, if we consider the catchment population only, that is, for those factories that only allowed for data collection in hostels we only take into consideration the number of women workers in the hostels, then the total is 3,930 for UPM Centre, and 9,430 for all research centres. The overall response rate is therefore 1,720 from 9,430, which is 18.2%.
In all the other factories, the management gave their cooperation only to the extent of publicising the study (by posting notices and making announcements) and asking for volunteers. The volunteers had to participate in the study either during break or meal times, or in their own free time. Three factories did not permit data collection on factory premises, but allowed researchers access to their hostels; the catchment population in these factories were therefore limited to the hostel occupants. In all these cases, participation was much lower than in Factory A and Factory B, and the percentage of completed questionnaires from the catchment population ranged from 12.4% to 35.7% (Table 1). The overall response rate, strictly speaking, could not be calculated, but with the above caveats in mind, it may be estimated at 18.2%.
Data collection was carried out between March 1999 and September 2000, with much delay encountered while obtaining consent from factories. During the survey, batches of 2–10 respondents attended sessions where they were first briefed on the purpose of the survey, and assured confidentiality, before filling in a self-administered questionnaire in the presence of research assistants. The research assistants verbally interviewed those who faced language or literacy barriers in filling in the questionnaires themselves (but these cases were generally less than 5%).
Instrument and pre-test
The questionnaire was in Malay, the national language, and had been pre-tested among 60 women workers who were randomly selected from a factory in Penang (not included in the sample for the study). The pre-test was carried out in September 1998.
Definition of variables and data analysis
The practice of Pap smear screening was defined as having had a test done within the last three years, while the practice of BSE was defined as doing BSE at least once a month. Postnatal and family planning visits were factors of interest since Pap smear screening and BSE are advocated through these two channels in Malaysia. The variable of having at least one young child, that is, preschool age (6 years) or younger, was identified as a proxy measure for having had recent contact with postnatal health care service.
Currently using either the contraceptive pill or intra-uterine device (IUD) was used to reflect contact with family planning services because other methods of contraception used (condom, spermicidal cream, withdrawal, herbal medicine, rhythm) did not require contact with family planning services, while tubal ligation only required a one-off visit. Knowledge was tested by true or false answers to the questions "BSE should be done every month before the menstrual period" and "Pap smear is a test to identify cervical cancer". The correct answer for the BSE question was 'false', and for the Pap smear question, it was 'true'.
Data were coded prior to entry, merged and analysed using SPSS version 10.0. The chi-square test was used to test the association of sociodemographic and health care factors with the practice of BSE and Pap smear screening. Odds ratios (OR) were calculated and interpreted as positive odds ratios (POR), following the example of Ejlertsson et al. [12], who had defined the POR as an indicator of positive health or practices rather than the conventional odds ratio that is used as a measure of risk. For ease of interpretation, the OR was consistently calculated for the group with a higher proportion of positive health behaviour compared to the group with a lower proportion of positive health behaviour. Variables of interest were used in logistic regression to yield adjusted odds ratios.
ResultsSocio-demographic and health care factors
This was a fairly young group of women, with mean age of 30.1 ± 7.9 years, and more than half (53.7%) less than 30 years old (Table 2). The majority were Malays (78.9%), and most had reached at least a secondary level of education (31.3% lower secondary, 58.3% upper secondary, 5.8% higher). The sample was almost equally divided into single (48.7%) and married (47.0%) women, with a small proportion who were either divorced or widowed (4.4%). There were 5.9% who were pregnant at the time of the study, while 28.5% of the women had young children who were of preschool age (six years) or younger.
Background characteristics of study population (n = 1720)
Distribution of respondents
No.
%
Age group (yrs)
≤ 20
69
4.0
20–29
854
49.7
30–39
539
31.3
40–49
244
14.2
≥ 50
14
0.8
Mean age (yrs) (± SD)
30.1 ± 7.9
Ethnicity
Malay
1357
78.9
Non-Malay
363
21.1
Educational attainment
No formal schooling
7
0.3
Primary
74
4.3
Lower secondary
538
31.3
Upper secondary
1002
58.3
Higher than upper secondary
99
5.8
Marital status
Single
837
48.7
Married
808
47.0
Divorced/ widowed
75
4.3
Pregnancy Status
Ever pregnant
676
39.3
Currently pregnant
101
5.9
Never pregnant
943
54.8
Young children (preschool or younger)
At least one
491
28.5
None
1229
71.5
Contraception
Currently using
259
15.1
Currently not using
1461
84.9
Types of contraception1
Contraceptive pill
131
7.6
IUD
49
2.8
Tubal ligation
29
1.7
Condom
15
0.9
Withdrawal
14
0.8
Others (herbal, rhythm, spermicide)
38
2.2
Medical examination within last 5 years
Yes
775
45.1
No
945
54.9
1Respondents may be using more than one method
While 15.1% of the women were currently using contraception, the types used included condoms, spermicidal cream, herbal medicines, and others, all of which do not require attendance at health clinics. Among this 15.1%, 7.6% were on the contraceptive pill and 2.8% were using the IUD, both of which require women to be in contact with health care services. Women who had tubal ligation (1.7%) would not have had to make use of health care services after the operation had been completed. Among the women, 45.1% reported having had a medical examination within the last five years. The type of examination, however, was not specified, and it could encompass either a pre-employment examination, a specific exposure-related examination related to type of work, or a general medical examination sought by the woman herself.
Practice of BSE and pap smear screening
Although 79.1% of the women had heard about the BSE, and 53.0% knew how to conduct the examination, only 44.8% had ever done the examination, while 24.4% said that they do it once a month (Table 3). Likewise, although 25.3% had ever had the Pap smear, only 18.4% had their last examination within the last three years.
Practice of breast self-examination and pap smear screening among study population (n = 1720)
Distribution of respondents
No.
%
Breast self-examination
Have heard about it
1360
79.1
Know how to conduct the examination
911
53.0
Have ever done the examination
769
44.8
Frequency of examination
Once a month
419
24.4
Once every 2 months
111
6.5
Once every 3 months
81
4.7
Once every 6 months
79
4.6
Annually
79
4.6
Pap-smear
Have heard about it
1232
71.6
Have ever done the examination
434
25.3
Last examination
≤ 3 years
316
18.4
>3 years
118
6.9
Women who were significantly more likely to do BSE every month were older, more highly educated, ever married, ever pregnant, had young children, had a medical examination in the last five years, had a Pap smear within the last three years, and gave a correct response to the question on when to do the BSE (Tables 4 and 5). However, BSE practice was not significantly associated with ethnicity or the use of the contraceptive pill or IUD.
Practice of breast self examination (BSE) and Pap smear screening by sociodemographic variables (n = 1720)
Do BSE every month
Pap smear in last 3 years
Yes
No
χ2
p
Yes
No
χ2
p
Age
13.8
0.000***
204
0.000***
≤ 30 years (n = 1010)
213 (21.1)
797 (78.9)
72 (7.1)
938 (92.9)
> 30 years (n = 710)
206 (29.0)
504 (71.0)
244 (34.4)
466 (65.6)
Ethnicity
3.7
0.055
20.7
0.000***
Non-Malay (n = 363)
74 (20.4)
289 (79.6)
97 (26.7)
266 (73.3)
Malay (n = 1357)
345 (25.4)
1012 (74.6)
219 (16.1)
1138 (83.9)
Educational attainment
3.9
0.049*
10.3
0.001**
≤ Lower secondary (n = 619)
133 (21.5)
486 (78.5)
139 (22.5)
480 (77.5)
≥ Upper secondary (n = 1101)
286 (26.0)
815 (74.0)
177 (16.1)
924 (83.9)
Marital status
15.8
0.000***
350
0.000***
Never married (n = 837)
168 (20.1)
669 (79.9)
3 (0.4)
834 (99.6)
Ever married (n = 883)
251 (28.4)
632 (71.6)
313 (35.4)
570 (64.6)
Pregnancy Status
17.3
0.000***
340
0.000***
Never pregnant (n = 943)
193 (20.5)
750 (79.5)
26 (2.8)
917 (97.2)
Ever pregnant (n = 676)
198 (29.3)
478 (70.7)
256 (37.9)
420 (62.1)
Currently pregnant (n = 101)
28 (27.7)
73 (72.3)
34 (33.7)
67 (66.3)
Young children
7.4
0.006**
91.3
0.000***
At least one (n = 491)
142 (28.9)
349 (71.1)
160 (32.6)
331 (67.4)
None (n = 1229)
277 (22.5)
952 (77.5)
156 (12.7)
1073 (87.3)
*p < 0.05, **p < 0.01, ***p < 0.001
Practice of breast self examination (BSE) and Pap smear screening by health care and knowledge variables (n = 1720)
Do BSE every month
Pap smear within last 3 yrs
Yes
No
χ2
p
Yes
No
χ2
p
Contraceptive pill or IUD
1.0
0.320
92.7
0.000***
Currently using (n = 177)
49 (27.7)
128 (72.3)
80 (45.2)
97 (54.8)
Not using (n = 1543)
370 (24.0)
1173 (76.0)
236 (15.3)
1307 (84.7)
Medical Examination
5.5
0.019*
40.9
0.000***
Within last 5 years (n = 775)
210 (27.1)
565 (72.9)
194 (25.0)
581 (75.0)
Not within last 5 yrs (n = 945)
209 (22.1)
736 (77.9)
122 (12.9)
823 (87.1)
Pap smear within last 3 years
29.6
0.000***
Yes (n = 316)
115 (36.4)
201 (63.6)
No (n = 1404)
304 (21.7)
1100 (78.3)
Do BSE once a month
29.6
0.000***
Yes (n = 419)
115 (27.4)
304 (72.6)
No (n = 1301)
201 (15.4)
1100 (84.6)
Answer to BSE question
4.9
0.027*
Correct (n = 249)
75 (30.1)
174 (69.9)
Incorrect (n = 1471)
344 (23.4)
1127 (76.6)
Answer to Pap smear question
66.0
0.000***
Correct (n = 1334)
300 (22.5)
1034 (77.5)
Incorrect (n = 386)
16 (4.1)
370 (95.9)
*p < 0.05, **p < 0.01, ***p < 0.001
Similarly, women who had a Pap smear within the last three years were significantly more likely to be older, ever married, ever pregnant, had young children, had a medical examination in the last five years, answered correctly the question on the purpose of Pap smear, and did BSE every month. In contrast to women who practise BSE however, they were also significantly more likely to be non-Malays, less educated, and using the contraceptive pill or IUD.
Table 6 shows the crude and adjusted odds ratios for the practice of BSE and Pap smear screening by the factors of interest. Women who were more than 30 years old had 1.53 times higher odds of doing BSE monthly (95% CI = 1.23–1.91) and 6.82 times higher odds of having had a Pap smear within the last three years (95% CI = 5.13–9.07) when compared with women who were 30 years old or younger. The number of never married women who had practised Pap smear screening was so small as to render the analysis imprecise (OR 153, 95% CI = 49–478).
Crude and adjusted odds ratios for breast self examination (BSE) and Pap smear screening by selected variables (n = 1720)
Do BSE once a month
Have had Pap smear within last 3 years
OR1
95% CI
Adj OR2
95% CI
OR3
95% CI
Adj OR4
95% CI
Age (> 30 years, ≤ 30 years)
1.53
1.23–1.91
1.36
1.04–1.79
6.82
5.13–9.07
2.65
1.88–3.74
Ethnicity (Malay, Non-Malay)
1.33
1.00–1.77
1.55
1.13–2.12
1.90
1.44–2.49
1.31
0.93–1.86
Educational attainment (≥ upper sec, ≤ lower sec)
1.27
1.01–1.61
1.38
1.07–1.78
1.51
1.18–1.94
0.86
0.63–1.18
Marital status (ever married, never married)
1.58
1.27–1.98
1.25
0.92–1.71
153
48.7–478
66.6
20.8–213
Currently pregnant (yes, no)
1.22
0.78–1.89
1.10
0.68–1.78
2.41
1.57–3.69
1.28
0.79–2.07
Young children (at least one, none)
1.40
1.10–1.77
1.17
0.90–1.52
3.33
2.58–4.28
1.42
1.05–1.91
Contraceptive pill or IUD (currently using, not using)
1.21
0.86–1.72
0.79
0.54–1.17
4.57
3.29–6.33
1.62
1.12–2.33
Medical examination within last 5 years (yes, no)
1.31
1.05–1.63
1.17
0.93–1.47
2.25
1.75–2.89
1.97
1.47–2.64
Pap smear within last 3 years (yes, no)
2.07
1.59–2.69
1.70
1.24–2.31
–
–
–
–
Do BSE once a month (yes, no)
–
–
–
–
2.07
1.59–2.69
1.54
1.12–2.16
Answer to question on BSE (correct, incorrect)
1.41
1.05–1.90
1.42
1.05–1.92
–
–
–
–
Answer to question on Pap smear (correct, incorrect)
–
–
–
–
6.71
4.00–11.2
3.07
1.73–5.46
1For all variables, the last category was the reference category. 2From a logistic regression model that consists of all ten covariates with values given below. For all the covariates, the last category was the reference category. 3For all variables, the last category was the reference category, except for ethnicity and educational attainment, for which the first category was the reference category. 4From a logistic regression model that consists of all ten covariates with values given below. For all the covariates the last category was the reference category, except for ethnicity and educational attainment, for which the first category was the reference category.
The logistic regression model for BSE contained ten covariates, and after adjusting, the practice of BSE was found to be significantly higher among older women (Adj OR 1.36, 95% CI = 1.04–1.79), Malays (Adj OR 1.55, 95% CI = 1.13–2.12), women with upper secondary education and above (Adj OR 1.38, 95% CI = 1.07–1.78), women who had had a Pap smear within the last three years (Adj OR 1.70, 95% CI = 1.24–2.31), and those who gave correct responses to the BSE question (Adj OR 1.42, 95% CI = 1.05–1.92). Pap smear within the last three years was likewise tested in a logistic regression model containing ten covariates. After adjusting, having had Pap smear within the last three years remained significantly higher among women who were older (Adj OR 2.65, 95% CI = 1.88–3.74), with young children (Adj OR 1.42, 95% CI = 1.05–1.91), on the contraceptive pill or IUD (Adj OR 1.62, 95% CI = 1.12–2.33), had a medical examination within the last five years (Adj OR 1.97, 95% CI = 1.47–2.64), do BSE monthly (Adj OR 1.54, 95% CI = 1.12–2.16), and answered correctly on the Pap smear question (Adj OR 3.07, 95% CI = 1.73–5.46). Ever married women had significantly higher odds of having had Pap smear within the last three years, but the result was imprecise (Adj OR 66.6, 95% CI = 20.8–213). Educational attainment, ethnicity and pregnancy status were not significantly related.
Discussion
The selection of factories in this study was not representative of the electronics industry as a whole; moreover, the number of participating factories was small relative to the total number in the country. The women workers in the study were not all randomly selected; a large proportion were volunteers, and as such, this sample was not statistically representative of all the women workers in the participating factories. Voluntary participation could have led to a bias for greater inclusion of women who were more health conscious, or with a greater propensity to seek information on their own health; while women who felt that they had less free time to spare could have been less likely to volunteer. The participation rate in the various factories was uneven, and depended on whether the factory management allowed the workers to take time off from work for the study. In the factories where researchers had to rely on volunteers doing the study in their own time, the participation rate was generally low.
In spite of the weaknesses outlined above, this study is an important one because of the lack of research in Malaysia on health-seeking behaviour of women workers in general, and electronics factory women workers in particular. Although not statistically representative of the whole industry, the current study provides data from a relatively large number of women workers that could be compared to the nationally representative sample in the 1996 National Health and Morbidity Survey (NHMS), thereby marking its contribution to the existing literature.
Practice of BSE
The rate of ever doing BSE among the electronics women workers (44.8%) was higher than the rates reported from the NHMS for women above 20 years old (34.2%), urban women (36.3%), or women in production (33.8%) [6]. It is also higher, but closest to the NHMS rate for women aged 20–54 (39.8%). The NHMS sample included all women above the age of 20 years, while this sample only had women between the ages of 17 and 55 years. Doing BSE at least once a month, however, was found to be lower among the electronics women in this study (24.4%) compared to the NHMS rates for women above 20 years (26.5%), women in production (27.7%) and women aged 20–54 (30.9%).
Although a higher percentage of women workers in this study had ever done the BSE compared to the NHMS country-wide representative sample, more were also unable to sustain it on a regular basis. It would appear therefore that although this group of women had better than average awareness of and access to information on BSE, they also faced more barriers to practising it regularly.
The BSE rate found in this study, however, is sharply higher than another Malaysian study, where only 1.3% of 1,303 women were found to practise BSE regularly [13]. Drawn from women who had registered with the Well Person's Clinic at the Outpatient Department of Ipoh Hospital between April 1995 and March 1997, the women were primarily from the lower socioeconomic group, with a majority who were Chinese, between 30–59 years old, married, and housewives, and had never been taught BSE.
Nevertheless, the rate of BSE measured in this study (24.4%) as well as that from the NHMS (26.5%) were low when compared to the 1995 National Health Survey of Australia, where rates were in the region of 60% and higher (among a sample of 10,179 women aged 18 and above, those who do NOT regularly perform BSE was 38.0% among the 18–39 year age group, 27.4% for 40–49 years, and 25.0% for 50–59 years) [14]. The BSE once a month rate of 24.4% was also lower compared to 30.9% who had done BSE within the past one month from a study of 123 Korean women from a larger convenience sample of 1,202 Asian American women [15]; and compared to 30.6% who do BSE once a month among a nationwide representative sample of Norwegian female physicians [16].
Practice of pap smear screening
The rate of ever having had a Pap smear among the electronics women in this study (25.3%) was in between the NHMS rates for women aged 20–54 (29.8%) and production workers (22.0%). Likewise, the rate of having had Pap smear within the last three years for the electronics women (18.4%) was slightly lower than the NHMS rate for women aged 20–54 years (22.7%) and slightly higher than the rate for production workers (15.8%). The Pap smear rate found in this study (18.4%) was low compared to a study of 470 women staff members in a Malaysian university, where 27.7% had had a Pap smear within the last three years and 16.0% had one more than three years ago [17]. The difference could be due to differences in sociodemographic profile, in particular, the higher proportion of never married women in this study (48.7%) as compared to the university sample (26.2%), which was made up of academic, administrative, as well as support staff, who also had greater access to health care through the university health services.
A community-based survey in two residential districts in Singapore among 726 women aged 30–59 years had found a screening rate of 62.4%, with 42% reporting regular smears [18]. The Malaysian rates are lower compared to Singapore, and of course much lower when compared to those found in developed countries. In 1996, only 2% aged 25–59 had never had a Pap smear in Uppsala County, Sweden, while 41% aged 25–29 had not had a Pap smear within the last three years, and 21% aged 30–54 had not had one within the last five years [19]. Among the nationally representative sample of 284 Norwegian female physicians, the 54.6% found to have a Pap smear test once every third year was considered low by the researchers [16]. Likewise, the 55% Korean American women found to have had a Pap smear in the previous three years was considered a low rate [20].
The present study was limited in that it did not seek to verify self-reported rates of screening with clinical or laboratory records. Other studies have reported discordance between self-reports and medical records. In a random household survey in the Hunter region of New South Wales, Australia, for example, only 61.2% of the self-reported Pap smears in the last three years could be verified with pathology laboratory records [21]. In the study in Uppsala county, Sweden, those who ever had Pap smear screening were found to have a higher rate of recall accuracy (95%-99%) compared to those who never had, among whom 50% believed they already had one. Furthermore, among those who had a Pap smear within last three years (25–29 year age group) or last five years (30–59 year age group), 74% recalled accurately the time since their last Pap smear, while 57% of those who had not had a recent smear underestimated the time lapse since the last one.
Factors associated with BSE and Pap smear screening
In general, the associations found in this study were stronger for Pap smear screening than for BSE practice. Both BSE and Pap smear screening were more likely to be practised by women older than 30 years, but BSE was more commonly practised among Malays and more highly educated women, while Pap smear screening was not significantly associated with ethnicity and educational level after adjusting for other factors.
This BSE pattern is similar to the NHMS, where the practice of BSE was found to be higher among Malay women compared to other ethnic groups, and increased significantly by educational level [6]. It is also similar to the 1995 National Health Survey of Australia, where women who were less educated were found to be less likely to perform BSE regularly [14].
The Pap smear screening pattern in this study, however, is different from the NHMS, where it was found to significantly increase with higher educational levels, and to be higher for women younger than 35 years old compared to older women [11]. In the Singapore community-based study, women who had regular Pap smears were also more likely to be younger, and to have had more years of formal education [18]. Whereas women in these two other studies had varying employment status and were from a range of occupations, the current study only involved women working in factories. The common employment experience may be a factor in lessening the effect of educational attainment in relation to Pap smear screening.
Pap smear screening in this study was significantly related to marital status, but not to pregnancy status, after adjusting for other factors. In fact, the women workers who were never married almost never had the Pap smear. Out of 837 never married women, only three (0.4%) had done the Pap smear within the last three years, and nine (1.1%) had ever done the Pap smear at all (Table 3). Marital status was a confounding factor in the bivariate analysis of Pap smear screening with age and education, as the younger women tended also to be more highly educated and never married, and all three variables were inversely related to Pap smear screening.
Other Malaysian studies also show the dependence of Pap smear screening on marital status. In the NHMS, only 2.7% of single women had ever had a Pap smear, compared to 33.4% married women, 18.3% divorced women, 9.1 % widowed women, and 10.5% among women who were cohabiting [11]. In the study among staff members of a local university, Pap smear screening was also significantly associated with an older age, and being married [17].
Other than sociodemographic variables, Pap smear screening was found in this study to be tightly bound to health service delivery variables, being significantly associated with having young children, being on the contraceptive pill or IUD, and having had a medical examination within last five years. Studies in other countries have also found a close association between Pap smear screening and health care utilization indicators [16,18-20]. The Singapore community-based study, interestingly, did not find that women who had regular Pap smears were more likely to be married, but they were more likely to have had the last smear as a self-initiated screening test or part of a regular health check-up rather than as part of a postnatal or family planning visit, and also more likely to have visited a general practitioner or a gynaecologist rather than a government polyclinic [18].
In the population-based study in Sweden where attendees of Pap smear screening were compared to non-attendees, those more likely to be non-attendees were women who used condoms or no contraceptives compared to users of oral contraceptives [19]. Likewise, among the Norwegian female physicians, significant correlates of Pap smear screening were being married/cohabitant, having children under 12 years, and using oral contraceptives [16], while among the Korean-American women, the strongest correlates of Pap smear in the previous two years were being married and employed, and having had a routine check-up within the last two years, which indicated either access to health care or preventive health behaviour [20].
In contrast to the pattern of Pap smear screening, BSE among the electronics women workers of this study was not related to marital status, being currently pregnant, having young children, being on the contraceptive pill or IUD, or having had a medical examination in the last five years. It was found to be less dependent on health service utilisation, and therefore, also to have less bias against never married women.
Nevertheless, this constitutes a departure from the national trend in the NHMS, where the higher BSE rate among married women compared to single women was attributed to opportunity and access to information and being taught BSE, which in turn was attributed to programme policy [6]. In the 1995 National Health Survey of Australia, also, previously married and never married women were significantly less likely than currently married women to perform BSE regularly [14].
Knowledge questions
In this study, correct responses to the question on the correct time of the month to do BSE was significantly associated with practising BSE every month. However, on the whole, many more women gave incorrect (n = 1471) than correct responses (n = 249); and even among those who do BSE, more gave incorrect answers (82.1%) than correct answers (17.9%) (Table 4). In other words, most women, including women who do BSE, do not know when is the correct time for doing BSE, although compared to women who do not do BSE, they had better knowledge.
Knowledge that Pap smear is a test for cervical cancer was also significantly associated with having had one within the last three years, but in contrast to the BSE question, there were many more women who gave correct (n = 1334) than incorrect (n = 386) responses (Table 4). Even among those who had not had a Pap smear within the last three years, more gave correct answers (74%) than incorrect answers (26%).
This is congruent with the finding in the NHMS, where 86.6% of women who had Pap smear knew the reason for doing so [11]. Similarly, the study of university staff members also found that Pap smear screening was significantly associated with having a higher level of health knowledge, and showing positive self-interest towards health [17].
A significant association was found in this study between having had Pap smear within the last three years and practising BSE monthly. This association was also noted in the NHMS, where 74% of the women who had undergone Pap smear screening also practised BSE, as compared to 43% of those who had not done a Pap smear [11]. This may be explained as the tendency for individuals who are aware and concerned about their health to be motivated toward both BSE and Pap smear screening practice.
Conclusion
The group of electronics women workers in this study is drawn from an important segment of the Malaysian workforce. Although they constitute a group from the lower socio-economic strata, nevertheless, they are largely urban-based and draw a higher wage than women production workers in many other industries.
The sample was not representative of the electronics industry as a whole, and the large proportion of respondents who were volunteers could have led to a tendency to include women who were more active in seeking health information and in accessing health care. As such, the lower BSE and Pap smear screening rates found among these women as compared to the national average for women of a similar age group is a cause for greater concern.
In general, there were more women who know about BSE and the Pap smear than those who conduct them at the proper intervals as part of a regular preventive health strategy. There could be barriers to the regular practice of BSE and Pap smear that are specific to women production workers, and it would be important to identify these barriers, which may include psycho-social as well as economic ones, so that educational and promotional strategies could be more effective. Factory-based cancer screening and education programmes could contribute toward improving health knowledge and screening practices of women production workers.
The association between Pap smear screening and health service variables reflect the success of governmental strategy to disseminate it through postnatal and family planning service outlets. However, its close connection with marital status, to the practical exclusion of single women, is an indication that the health services have not been successful in widening its availability to all women.
The practice of BSE, on the other hand, was not found to be dependent on health service variables, but on educational level. This implies that there is wider interest in this than could be reached through postnatal and family planning services, and that this interest is linked to education. Adequate information and educational services should particularly reach out to women with lower educational levels, and young and single women.
Indeed, the large proportions who did not know the best time for doing BSE, even among those who were practising it once a month, reflect that the methods employed to disseminate information on this subject should be thoroughly reviewed. Many women are taught BSE during one-off postnatal visits, although those who are on oral contraceptives and the IUD receive reinforcement each time they visit the family planning clinic.
In any case, governmental policy on the promotion of BSE should be reviewed in light of recent research findings. Whether or not BSE should be taught and promoted, and how it should be done, should be ascertained through a thorough assessment of current practice and its effects. Information and knowledge about breast cancer could still be widely disseminated, whether or not BSE is advocated, and women should still be encouraged to be aware of their breasts, and any symptoms which could arise.
Competing Interests
None declared.
Author's Contributions
RS was the leading investigator and the coordinator of the USM-Kel Centre, CHL was the coordinator of the UPM Centre, KS was the coordinator of the UKM Centre, and IO was the coordinator of the USM-Pg Centre. All authors participated in the conception and design of the study. KS merged and cleaned the data, CHL analysed the data and wrote the draft manuscript, and all authors read, commented on and approved the manuscript.
Pre-publication history
The pre-publication history for this paper can be accessed here:
Acknowledgements
This study was funded by the Malaysian Ministry of Science, Technology and Environment Intensification of Research in Priority Areas (IRPA) (Medical Sector) Programme for the project entitled "A Study of Working Women's Health Status in Relation to Lifestyles". The participating institutions were Universiti Sains Malaysia (USM) Penang and Kelantan campuses, Universiti Kebangsaan Malaysia (UKM), Universiti Putra Malaysia (UPM), and the National Institute of Occupational Safety and Health (NIOSH).
A cross sectional study was conducted in Tehran Iran to examine the extent of patient delay and associated factors in the presentation of breast cancer.
Methods
A group of newly diagnosed breast cancer patients were interviewed and were asked about the period from first onset of symptoms to first medical consultation to indicate patient delay. This was studied in relation to patients' age, educational level, marital status, family history of breast cancer, history of benign breast disease, number of children and the nature of the first symptom seen.
Results
In all, 190 breast cancer patients were interviewed. Of these, 75% presented to physician within 3 months. Forty-two patients (25%) delayed more than 3 months. In multivariate regression analysis it was found that there was a risk for longer delay in widowed or divorced women (OR 3.7, 95% CI 1.5–9.7), women with a positive family history of breast cancer (OR 2.8, 95% CI 1.1–7.7), and less educated patients (illiterate: OR 5.2, 95% CI 1.5–17.7; primary schooling: OR 4.6, 95% CI 1.4–14.7). Significant associations also were found between delay presentation and the late stage disease (P = 0.01) and bigger tumor size (P = 0.004).
Conclusion
The findings suggest that one in four women with breast cancer present late and this has significant effect on their disease prognosis. To reduce patient delay health education programs regarding breast cancer should be implemented and target women who are at higher risk of delay.
Background
Delayed presentation of breast cancer is associated with lower survival [1,2]. Moreover the late stage of disease and high mortality are seen with delay in diagnosis and treatment of breast cancer [3,4]. There is evidence that smaller tumors are more likely to be treated successfully with limited breast surgery, and perhaps a better quality of life.
Delay in breast cancer is defined as patient delay (the interval between first detection of symptom and first medical consultation) and system delay (the interval between first presentation to a medical professional and initial treatment). Prolonged delays usually defined as intervals greater than 12 weeks [3]. Delay and late stage at diagnosis of breast cancer are related to socio-demographic factors such as age, education, marital status, economic status, history of breast disease, family history of breast cancer, the nature of the first symptom and many other factors [1,3,5-7].
In Iran, breast cancer is one of the most growing and important women's health problems, although its statistics is very similar to that of the regional countries [8]. It is estimated that the crude incidence rate of the disease is about 20 new cases per 100,000 women per year. Given that Iran has about 30 million female population, this corresponds to a total number of 6000 new cases of breast cancer annually [9]; many of whom because of advanced disease at presentation (70%) die within a short period of time [10]. Therefore to improve breast cancer care understanding the magnitude of delay in breast cancer diagnosis is an important issue.
The aim of this study was to examine the extent of patient delay in Iranian breast cancer patients. Since most literature on delayed presentation of breast cancer is from developed countries it was thought that a study from a developing country with a different culture might contribute to existing knowledge on the topic.
Methods
A group of newly diagnosed breast cancer patients who were admitted to a university hospital or attending a breast clinic in Tehran, Iran were interviewed following their surgery or first course of chemotherapy between September 2001 and March 2002. After obtaining oral consent from each patient, data were collected using a structured questionnaire including socio-demographic factors, date of first symptom recognition and first medical consultation. Delay was defined as time intervals of more than 12 weeks from first symptom recognition to first medical consultation and thus patients were divided into two groups: those who presented at three months or less and those who delayed more than three months. Patients were excluded if their data concerning delay were unreliable. If patients were not sure of the date of first symptom recognition or could not recall the date, data considered unreliable. Clinical data including stage of disease, tumor size and lymph node status also were extracted from patients' medical records. Statistical analysis was performed using univariate and multivariate logistic regression models to calculate odds ratios (OR). Age, marital status, educational level, family history of breast cancer, history of benign breast disease, number of children and the nature of first symptom have been selected as potential explanatory factors. The Statistical Package for Social Sciences (SPSS) was used to analyze data [11].
Results
In all, there were 235 eligible patients. Of these, 203 patients (86%) were recruited into the study and the remaining 32 patients were missed or refused to be interviewed. Thirteen cases were excluded because of unreliable data, leading to a final study population of 190 breast cancer patients. The mean age of patients was 47.0 (SD 11.3) years, and most were married (80%). Patient delay ranged from less than one week to 60 months (Mean 3.8, SD 8.6 months). Delay of more than three months was reported by 25% of patients. The characteristics of study population and patient delay are shown in Table 1.
The characteristics of breast cancer patients (n = 190)
No.
%
Age groups (years)
< 35
25
13
35–44
54
28
45–54
72
38
≥ 55
39
21
Mean (SD)
47.0 (11.3)
Range
24–82
Marital status
Single
8
4
Married
152
80
Widowed/divorced
30
16
Education levels
Illiterate
56
30
Primary
78
40
Secondary/Higher education
56
30
Number of children
None
20
11
1–3
88
46
≥ 4
82
43
Family history of breast cancer
No
146
77
Yes
44
23
History of benign breast disease
Yes
34
18
No
156
82
First symptom seen
Lump
167
88
Other*
23
12
Delay presentation (months)
≤ 3
142
75
> 3
48
25
Mean (SD)
3.8 (8.6)
Range
< 1–60
Stage of disease (n = 165)
I
9
5
II
94
57
III
51
31
IV
11
7
Tumor size (n = 170)
< 2 cm
21
12
2–5 cm
93
55
≥ 5 cm
56
33
Nodal involvement (n = 165)
No
46
28
Yes
119
72
* including discharge, pain, and skin problems.
Table 2 shows the result of univariate logistic regression analysis. There was a significant risk for patient delay by marital status and educational level. Widowed and divorced women had a significant delay compared to married women (OR 3.4, 95% CI 1.5–7.7). Also a significantly higher risk of more than three months delay was found among illiterate (OR 5.7, 95% CI 1.9–16.5) and primary educated women (OR 4.2, 95% CI 1.5–12.1). No significant differences were found among the other variables studied.
The result of univariate and multivariate logistic regression analysis on patient delay
≤ 3 months (n = 142)
> 3 months (n = 48)
Univariate analysis
Multivariate analysis
No. (%)
No. (%)
OR (95% CI)
P
OR (95% CI)
P
Age groups (years)
<35
20 (14.1)
5 (10.4)
1.0 (ref.)
1.0 (ref.)
35–44
43 (30.3)
11 (22.9)
1.0 (0.31–3.3)
0.97
1.5 (0.40–5.8)
0.54
45–54
55 (38.7)
17 (35.4)
1.2 (0.40–3.8)
0.71
1.2 (0.33–4.1)
0.82
≥55
24 (16.9)
15 (31.3)
2.5 (0.77–8.1)
0.13
1.3 (0.34–5.4)
0.67
Marital status
Married
121 (85.2)
31 (64.6)
1.0 (ref.)
1.0 (ref.)
Single
5 (3.5)
3 (6.3)
2.3 (0.53–10.3)
0.26
1.2 (0.15–9.5)
0.86
Widowed/divorced
16 (11.3)
14 (29.2)
3.4 (1.5–7.7)
0.003
3.7 (1.5–9.7)
0.007
Education levels
Secondary/Higher education
51 (35.9)
5 (10.4)
1.0 (ref.)
1.0 (ref.)
Primary
55 (38.7)
23 (47.9)
4.2 (1.5–12.1)
0.006
4.6 (1.4–14.7)
0.01
Illiterate
36 (25.4)
20 (41.7)
5.7 (1.9–16.5)
0.001
5.2 (1.5–17.7)
0.009
Number of children
None
12 (8.5)
8 (16.7)
1.0 (ref.)
1.0 (ref.)
1–3
71 (50.0)
17 (35.4)
0.36 (0.13–1.0)
0.05
0.43 (0.10–1.9)
0.26
≥ 4
59 (41.5)
23 (47.9)
0.58 (0.21–1.6)
0.30
0.48 (0.13–2.0)
0.32
Family history of breast cancer
No
105 (73.9)
41 (85.4)
1.0 (ref.)
1.0 (ref.)
Yes
37 (26.1)
7 (14.6)
2.1 (0.85–5.0)
0.11
2.8 (1.1–7.7)
0.04
History of benign breast disease
Yes
24 (16.9)
10 (20.8)
1.0 (ref.)
1.0 (ref.)
No
118 (83.1)
38 (79.2)
0.77 (0.34–1.7)
0.54
0.66 (0.25–1.7)
0.39
First symptom seen
Lump
122 (85.9)
45 (93.8)
1.0 (ref.)
1.0 (ref.)
Other
20 (14.1)
3 (6.3)
0.41 (0.11–1.4)
0.16
0.51 (0.13–2.0)
0.34
Performing multivariate logistic regression analysis entering all variables studied, marital status, education levels and family history of breast cancer showed significant results (Table 2). There was a risk for longer delay in widowed or divorced women (OR 3.7, 95% CI 1.5–9.7), women with a positive family history of breast cancer (OR 2.8, 95% CI 1.1–7.7) and illiterate (OR 5.2, 95%CI 1.5–17.7) or primary educated women (OR 4.6, 95% CI 1.4–14.7).
Finally a cross tabulation analysis showed that delay of more than three months was significantly associated with advanced disease (P = 0.01), and bigger tumor size (P = 0.004). However, nodal status did not show significant results. The results are shown in Table 3.
Association between delay presentation and clinical variables
≤ 3 months
> 3 months
P
No. (%)
No. (%)
Stage of disease (n = 165)
I and II
84 (67.8)
19 (46.3)
III and IV
40 (32.2)
22 (53.7))
χ2 = 6.02, df = 1
0.01
Tumor size (n = 170)
< 2 cm
19 (15.1)
2 (4.5)
2–5 cm
74 (58.7)
19 (43.2)
≥ 5 cm
33 (26.2)
23 (52.3)
χ2 = 11.1, df = 2
0.004
Nodal involvement (n = 165)
No
36 (29.0)
10 (24.4)
Yes
88 (71.0)
31 (75.6)
χ2 = 0.33, df = 1
0.56
Discussion
The study findings indicate that about 25% of patients with breast symptoms had a delay of more than three months before presenting to a health professional. This finding is comparable with other studies. Recent studies have shown a range of 19% to 32% for patient delay [1,3,4,12,13]. However the extent of patient delay can be different in different places. One explanation for such a difference might relate to the patients' health related behaviors and the social context they live in. It is argued that an intention to seek evaluation of breast symptoms is not merely a matter of education and economics but it is dependent on a complex picture of personal and social factors on the perceived amount of negative consequences of delaying diagnosis and on previous habit of health care utilization [5,14].
In this study, widowed and divorced women had a higher risk of delay. Perhaps one might argue that this could be explained by the fact that widowed and divorced women do not have enough motivation to seek help or care about themselves and lack support [15]. Socio-demographic factors and delayed presentation of breast cancer have long been studied. Earlier studies on patient delay showed that marital status was significant predictor of delay [16], and this was confirmed in a few later publications [4,7]. However, the association between marital status and delay remains controversial [3,17] and at present strong evidence exist that marital status is unrelated to delay by patients [1]. It seems that the findings from this study and other studies [18] clearly suggest that marital status and a positive family history of breast cancer are risk factors for both incidence of breast cancer and for delayed presentation in Iran.
Being less educated was a significant predictor of patient delay. The role of education and knowledge in decreasing delay has been confirmed in other studies [17,19-21]. The finding suggests that lack of knowledge about breast cancer is an important factor in Iran and there is a need for public educational programs especially for less educated women. However, in Iran social values and moral considerations limit the use of mass media for publicizing breast cancer awareness. Breast cancer is not taboo but because the breast is regarded as part of female sexual identity, people use the word chest instead. This is a cultural custom rather than a religious restriction. There is no evidence to suggest that religious beliefs interfere with early detection behaviors and contribute to subsequent delayed presentation of breast cancer in Iranian women [22].
The study did not demonstrate any association between age and patient delay, although the findings were in the expected direction. We suspect that the small number of cases in the reference category might attenuate a statistical significance. Studies have shown that older age is a predictor for patient delay [1,5,17]. A recent study on women's knowledge and beliefs regarding breast cancer concluded that since older age is a risk factor for both developing breast cancer and subsequent delayed presentation, any intervention program should target older women in particular [23]. In contrast it has been shown that women aged 50 years or younger had longer delays compared with older patients. The finding is explained by the fact that a higher index of suspicion of breast cancer exists for women older than 50 years than for younger women [24]. However, this could not be the case in the present study since the Iranian breast cancer patients are relatively ten years younger compared to their western counterparts [10].
In the present study the nature of the first symptom had no association with patient delay. It is argued that discovery of a breast lump reduces the patient delay and an association has been suggested in other studies [1,3,19]. Perhaps fear of cancer when a woman find a lump in her breasts, or lack of knowledge about common symptoms of breast cancer might explain why there was no association between the nature of the first symptom and delay in this study. However, the findings suggest that women need to be educated about the different types of breast cancer symptoms, especially the most frequent symptom, a non-tender breast mass. A qualitative study of delay among women reporting symptoms of breast cancer concluded that women need further information about the different types of breast cancer symptoms to assist symptom recognition as well as encouragement to seek medical advice if a symptom is ambiguous [12]. Similarly preliminary findings from the second phase of this study suggest that interventions to reduce delay behavior in help-seeking for breast symptoms should inform women of the diversity of breast cancer symptoms and do provide advice on how to obtain help for breast cancer symptoms [25]. As recommended health education programs should address both attention to cancer symptoms and appropriate help-seeking behavior otherwise if people do not intend to react when they detect such symptoms education becomes useless [26].
There were no associations between patient delay, history of benign breast disease and living in larger households. These factors may affect help-seeking behaviors in breast cancer patients. For example, it has been shown that women who live in larger households may have to care for children or other dependents and thus are at higher risk to present with late stage breast cancer [5,27]. Further studies are needed to investigate these associations, especially in Iran where strong family ties and cultural considerations exist.
Like other studies the findings indicated that those who presented late had significantly bigger tumor size and presented with an advanced stage of the disease. The influence of delay on tumor size and disease stage is well documented [1,2]. Although the present study do not provide information regarding the distribution of tumor differentiation, it is important to note that a substantial proportion of late stage diagnoses of poorly differentiated breast cancer cases could be avoided if patients with breast cancer presented to a doctor earlier [28].
Finally, it is worth noting that there were several limitations inherent in this study and the findings cannot be generalized beyond the study sample. The sample size was small and thus the predictive power of the study was limited. Additionally, the questions used to recall dates and times such as first presentation of the disease and first medical consultation might be biased, especially in cases with longer delay. Another limitation was that there were no data for patients' knowledge and attitudes towards breast cancer. Health beliefs and knowledge of breast cancer are two important factors that influence help-seeking behavior and delay [29]. More recent studies indicated that the likelihood of patient delay is more related to behavioral and knowledge variables, which are in turn linked with advanced breast cancer at diagnosis [19]. Furthermore, it seems that in addition to patient delay, system or provider delay is also an important issue that merits further investigation. Unfortunately this was not investigated in this study and as indicated in a meta-analysis of the literature on delayed presentation of breast cancer, provider delay appears to be both under researched and underestimated [15]. However, the study results do provide some understanding on the topic and indicate that Iranian women need more education on breast cancer care. The study findings suggest that patient delay is an important health problem, and can be reduced by educating women who are at higher risk of delay. The next step is to implement interventions to reduce delays and improve outcomes in breast cancer patients.
Authors' contribution
NM, MA, and AS contributed to the process of data collection. ME contributed to the study design, the data analysis, and wrote the first draft of the paper. AM contributed to the design, the data analysis and wrote the final draft of the paper.
Pre-publication history
The pre-publication history for this paper can be accessed here:
Urinary incontinence (UI) is a frequent public health problem with negative social consequences, particularly for women. Female susceptibility is the result of anatomical, social, economic and cultural factors. The main objectives of this study are to evaluate the prevalence of UI in the female population of Andorra over the age of 15 and, specifically, to determine the influence of socio-demographic factors. A secondary aim of the study is to measure the degree of concern associated with UI and whether the involved subjects have asked for medical assistance, or not.
Methods
Women aged 15 and over, answered a self-administered questionnaire while attending professional health units in Andorra during the period November 1998 to January 2000. A preliminary study was carried out to ensure that the questionnaire was both understandable and simple.
Results
863 completed questionnaires were obtained during a one year period. The breakdown of the places where the questionnaires were obtained and filled out is as follows: 32.4% – medical specialists' offices; 31.5% – outpatient centres served exclusively by nurses; 24% – primary care doctors' offices; 12% from other sources. Of the women who answered the questionnaire, 37% manifested urine losses. Of those,45.3% presented regular urinary incontinence (RUI) and 55.7% presented sporadic urinary incontinence (SporadicUI). In those women aged between 45 and 64, UI was present in 56% of the subjects. UI was more frequent among parous than non-parous women. UI was perceived as a far more bothersome and disabling condition by working, middle-class women than in other socio-economic groups. Women in this particular group are more limited by UI, less likely to seek medical advice but more likely to follow a course of treatment. From a general point of view, however, less than 50% of women suffering from UI sought medical advice.
Conclusion
The prevalence of UI in the female population of Andorra stands at about 37%, a statistic which should encourage both health professionals and women to a far greater awareness of this condition.
Urinary IncontinencePrevalenceWomenBackground
Urinary incontinence (UI) has been defined as the complaint of any involuntary leakage of urine[1]. In fact, this worldwide entity has negative influences on the quality of life as it may cause a socially unacceptable problem [2-5]. Since many women accept it as a "normal" condition, they are unaware that, in many instances, it can be successfully treated. Socio-cultural factors, especially in certain geographical areas, mean that affected women do not dare to ask for medical advice [2,6].
The prevalence of UI ranges from 10 to 60%, depending on the countries and populations studied [7-14]. Women are much more susceptible to UI than men. Anatomical and physiological differences [14-20], such as reproductive and hormonal changes associated with pregnancy and menopause, explain the differences prevailing between male and female. It is highly probable that socio-economic and cultural factors play a crucial role in UI. However the extent of the influence of these factors on women's health remains relatively unknown.
The Andorran Women's Research Group (WRG) has undertaken a task not only to evaluate the prevalence of UI in our country, but also to ascertain to what extent the socio-economic and cultural factors influence the outcome of UI in Andorran women.
MethodsPopulation and sample
The targeted population was the female population of Andorra aged over 15. The sample was obtained from women aged over 15 who voluntarily answered a self-administered questionnaire while attending health professionals' offices throughout Andorra between November 1998 and January 2000. The initial hypothesis was that through this population we could access a very large part of the female population of our country. We believe that the distinctive characteristics of the Andorran National Health Service allows us, to a very considerable extent, to extrapolate the results of our sample to the majority of the general female population. The questionnaire [Additional file: 1 and Additional file: 2] was widely distributed to every place were health advice might be sought: general practitioners and medical specialists' offices, and outpatient health centres exclusively served by nurses. People attending the latter facilities are not necessarily ill since most visits are to do with prevention programs, or even administrative issues. The Andorran National Health System is based on these health centres, on general practitioners who care for the majority of the population and on different surgical and medical specialists. All the health centres actively participated in this study. The majority (90%) of GPs also participated and the specialists we selected were those with the greatest number of patients (ophthalmologists, rheumatologists, urologist, gynaecologists, pneumologists, physiatrists, dermatologists). To avoid seasonal bias and repeat completion, all patients and their female companions were asked to fill out the questionnaire on a particular day of the month, throughout the year. Collection days were chosen at random.
At the beginning of the study, the female population of Andorra aged over 15 was 26,648 at the beginning of the study. 863 women completed the questionnaire. Of those, 320 confirmed involuntary urine leakage. The estimated margin of error is +/- 3,35% with a confidence level of 95,5% for the total sample, and +/- 5,6%, with the same confidence level, for estimates of women with urinary incontinence.
Questionnaire
To ensure that the questionnaire was easily understandable, a few months before the onset of the study, a preliminary questionnaire was distributed to a small group of women (n = 68). After reviewing these preliminary results, some questions were revised resulting in an improved, final version. The following socio-demographic and medical history risk factor variables were analysed in order to evaluate the possible predisposing factors for UI: age, education level, working activity, economic status, body mass index, childbirth, child's weight at birth, chronic diseases, and medication. The following aspects of UI were also monitored: symptom severity, disability degree, illness perception, and treatment strategies. Andorra is a multilingual country, with a high immigration rate, so the questionnaire was provided in several languages: Catalan (local national language), Spanish, French, Portuguese, and English.
Statistical analysis included descriptive and multiple logistic regression models to determine risk factors related to UI and to the effects (degree of concern, limitations, medical visits and treatment). As a multiple model, the risks obtained for every factor are adjusted by the other factors.
To construct models related to IU we started with a multiple model which include all possible factors and excluded variable by variable all those not representing a risk. Variables affecting UI included age, social class, UI type.
ResultsUrinary incontinence prevalence
Of the total of 863 women voluntarily answering the questionnaire, 320 (37%) manifested some kind of UI (Table 1). Unconcious urinary incontinence is defined as those women who answered NO to question number 5 (Have you ever experienced involuntary urine loss, at present or in the past, that was out of your control?), but who answered in the affirmative subsequent questions on UI's characteristics. Sporadic urinary incontinence (SporadicUI) and regular urinary incontinence (RUI) are categorized depending on the frequency of the UI episodes, RUI being at least once a week. Since unconcious urinary incontinence is nonetheless UI, we have included it. Moreover, if it is unconcious, we presume it must be uncommon, and therefore sporadic, and for the purposes of this study, it is included within Sporadic UI.
Prevalence and percentage distribution of the types of incontinence
Type of incontinence
Number
% total
% IU
Unwitting UI
79
9
Sporadic UI
96
11
SPORADIC UI (SUI)
175
20
55,7
REGULAR UI (RUI)
145
17
45,3
Total
320
37,1
100
Figure 1 shows the prevalence of UI in different age groups. As expected, the prevalence increases with age, especially RUI. In women with UI we observed that the percentage reporting regular urine leakage increased with age, rising from 15% in the youngest group, 32% in women between 25 to 44 years old, 59% between 45 to 64 and to 71% among the women aged over 64. The sample shows no statistical differences to the age distribution of the Andorran population (Table 2).
UI prevalence distributed in age groups and types
Age-related sample and population distribution
Sample distribution (%)
Population distribution (%)
15–24
10,2
14,2
25–44
50,1
45,6
45–64
24,4
25,1
>64
10,2
15,1
Missing
5,1
0
There are not significant statistical differences.
If we distribute the women affected by UI according to their level of education, the percentage of women reporting some kind of UI slightly decreases with the level of education (Table 3) but without any statically significant differences. Interestingly, university graduates were the ones who report less UUI.
UI in relation to the level of education
UPS
PS
S/PS
UG
CONTINENT
52.6 %
60.5 %
66.8 %
66.9 %
UI
47.4 %
39.5 %
33.2 %
333.1 %
UPS – Unfinished primary school. PS – Primary school. S/PS – Secondary /Professional school. UG – University graduate
Risk factors related to UI
The risk of Sporadic UI is greater in parous than in non-parous women. The risk is 1.67 higher (C.I. 95% 0.96 – 2.92) with one child, 1.86 (C.I. 95% 1.07 – 3.22) with two children and 2.76 (C.I. 95% 1.46 – 5.20) with more than two children. Taking any drugs is associated with a higher risk of Sporadic UI 1.90 (C.I. 95% 1.30 – 2.77).
The risk of self-reported RUI is also greater in parous than in non-parous women. The risk is 1.47 higher (C.I. 95% 0.64 – 3.34) with one child, twofold for women who have two children (OR = 2.71 C.I. 95% 1.25 – 5.87) and threefold for three and more children (OR = 3.18 C.I. 95% 1.35 – 7.47). Taking any drugs (OR = 1,73 C.I. 95% 1.08 – 2.77) or having any health problem (OR = 1,96 C.I. 95% 1.22 – 3.13) is also associated with a higher risk of RUI.
RUI and social status
For women who clearly manifested RUI, we tried to determine the existing relation between social status and the following factors: the degree of concern (Fig. 2), the limitations it causes (Fig. 3), medical visits due to this problem (Fig. 4), and the treatment followed for UI (Fig. 5).
Degree of concern in relation to social status
Women with limitations due to RUI distributed by social status
Women that consulted a health professional about RUI distributed by social status
Women following a treatment for RUI distributed by social status
Women in the middle class group appear more functionally limited and more concerned by UI. Curiously, this group is less inclined to seek medical advice on UI, but does follow the recommended treatment more consistently.
IU effects risk factors
After an adjustment for age and social class, if we compare women with Sporadic UI, with women with RUI, the latter are more concerned by UI (risk 6.5: CI 95% 2.07 – 20.41), are more limited (risk 5.8: CI 95% 3.08 – 10.90), have a greater tendency to ask for medical advice (risk 3.78:CI 95% 1.96 – 7.28), and are more inclined to follow treatment consistently (risk 6.68: CI 95% 2.60 – 17.16).
After an adjustment for age and UI level, middle class women are more limited by UI (risk 2.9: CI 1.34 – 6.31) than upper class women.
Finally, after an adjustment for UI level and social class, women aged 15 to 29 are more limited by UI (risk 5.26: CI 95% 1.51 – 18.12) than women aged 60 years and over. Women aged 45 to 59 are more inclined to follow treatment (risk 10.82: CI 1.28 – 91.30) than women aged 15 to 29.
Discussion
Very often, the records or analysis of the studies of a community's health problems do not consider differences in the gender, socio-economical and cultural status, or the ethnic origins of the subjects of the study.
Little is therefore known about the effect of those factors on women's health, the use that women make of the health services, or the reasons why they do or do not consult on a health problem that greatly affects their quality of life.
This study is a report on the prevalence of UI in the female population of Andorra, as well as the socio-economical characteristics of the affected women and their approach when consulting the health services on this problem. The results raise a series of questions needing further examination.
Despite the bias of a study based purely on the female population attending any of the health services during the surveyed period, the sample can be considered representative of the country's female population since the data are comprehensive in terms of the current census.
Our sample is representative of the women attending health services during one year, and not of the whole Andorran population. This methodological option, chosen for economical and operative reasons, may cause a bias. However, the prevalence observed is similar to the data found in other published studies [3,7-12]. Similarly, prevalence distribution also shows an increase of UI with age [13,21]. It is important though to underline that only RUI increases, whereas the proportion of women manifesting Sporadic UI decreases with age. The data do not give information as to whether women suffering episodes of SporadicUI when younger show a greater predisposition to RUI in the later stages of life. In this survey, the subject herself contributes her own perception of UI. Consequently, one may ask if younger women involved in a full working, social, and family life are apt to perceive Sporadic UI as a minor, unconsciously masked, problem.
Women with a higher level of education manifest a higher UI prevalence. Possibly, a higher level of education leads to a higher degree of awareness of the presence of UI, while at lower levels of education UI may be perceived as an unavoidable problem.
The extent of female concern about UI when related to social status also produced a remarkable outcome. Women in the three social levels reported being "somewhat bothered" by UI, whereas this is often a highly limiting and uncomfortable problem.
Middle-class women show a higher UI prevalence but consult health services less frequently about this problem. However, when they do ask for assistance, women in this category follow UI treatment more consistently than the rest of the sample.
Another widely known factor is that UI is usually a hidden problem, either because the patient considers the problem as normal or is embarrassed to point it out to the health professional, from whom she frequently gets only a technical-medical answer [6,22]-. Moreover, approaching this problem without a multidisciplinary perspective is associated with a higher relapse rate, something that contributes to the perception that it is a problem without solution, which is manifestly untrue [23-25].
These factors could explain the paradoxical behaviour of women affected by such an impairing condition [3-5,12,26]. Even when assuming the presence of UI, less than 50% ask for medical advice in search of a solution or an improvement to this problem.
Conclusion
The results of this study can be considered reliable, as well as a starting point to identify UI features in women of our country. Furthermore, the study itself could have a sensitizing effect on the participants (professionals and patients) and result in a better knowledge of the illness and its solutions. Some reports show consensus and indicate a need for an increased degree of awareness, both in the female population and among health professionals, as to the actual relevance of UI.
Competing interests
None declared.
Pre-publication history
The pre-publication history for this paper can be accessed here:
Chronic pelvic pain is a common condition with a major impact on health-related quality of life, work productivity and health care utilisation. The cause of the pain is not always obvious as no pathology is seen in 40–60% of the cases. In the absence of pathology there is no established treatment. The Lee-Frankenhauser sensory nerve plexuses and parasympathetic ganglia in the uterosacral ligaments carry pain from the uterus, cervix and other pelvic structures. Interruption of these nerve trunks by laparoscopic uterosacral nerve ablation (LUNA) may alleviate pain. However, the balance of benefits and risks of this intervention have not been reliably assessed. LUNA has, nevertheless, been introduced into practice, although there remains controversy regarding indications for LUNA. Hence, there is an urgent need for a randomised controlled trial to confirm, or refute, any worthwhile effectiveness. The principal hypothesis is that, in women with chronic pelvic pain in whom diagnostic laparoscopy reveals either no pathology or mild endometriosis (AFS score ≤ 5) LUNA alleviates pain and improves life quality at 12 months.
Methods/Design
The principal objective is to test the hypothesis that in women with chronic pelvic pain in whom diagnostic laparoscopy reveals either no pathology or mild endometriosis (AFS score ≤ 5) LUNA alleviates pain and improves life quality at 12 months. A multi-centre, prospective, randomised-controlled-trial will be carried out with blind assessment of outcomes in eligible consenting patients randomised at diagnostic laparoscopy to LUNA (experimental group) or to no pelvic denervation (control group). Postal questionnaires including visual analogue scale for pain (primary outcome), an index of sexual satisfaction and the EuroQoL 5D-EQ instrument (secondary outcomes) will be administered at 3, 6 and 12 months. The primary assessment of the effectiveness of LUNA will be from comparison of outcomes at the one-year follow-up, although the medium-term and longer-term risks and benefits of LUNA will also be evaluated.
The sample size for this trial has been estimated as 420 patients in total using the hypothesis that LUNA will alleviate pain symptoms (i.e. reduce pain scores on a VAS) more than no intervention at one-year following diagnostic laparoscopy and taking into consideration 20% loss to follow-up. The intention to treat analysis to address the principal research questions will be conducted using the one-year follow-up data.
Background
Pelvic pain remains the single most common indication for referral to a Gynaecology clinic accounting for 20% of all outpatient appointments [1,2]. Five percent of all new appointments are for chronic pelvic pain[3]. It continues to be one of the most difficult and perplexing problems encountered in Gynaecology. Pelvic pain has a major impact on health-related quality of life, work productivity and health care utilisation. It is also a major cause of workplace absenteeism [4]. An estimated 158 million pounds are spent annually on the management of this condition in the health service [5]. In primary care, the annual prevalence is 38/1000 in women aged 15–73, a rate comparable to that of asthma (37/1000) and chronic back pain (41/1000)[6].
An initial troublesome problem in addressing research issues related to pelvic pain is the difficulty in making a definite diagnosis. The following definition is commonly used and will be adopted for our research proposal: Constant or intermittent, cyclic or acyclic pain, that persists for 6 months or more and includes dysmenorrhoea, deep dyspareunia and intermenstrual pain [7].
An effective treatment for pelvic pain has evaded gynaecologists for centuries. Even today only 20–25% patients respond to conservative management [8]. When such treatment fails, a diagnostic laparoscopy is usually performed [2,3,9]. Chronic pelvic pain accounts for up to 40–60% of all diagnostic laparoscopies [2]. The cause of the pain is not always obvious as no pathology is seen in 40–60% of the cases. When definite pathology is found the likely causes are endometriosis (25%) and adhesions (25%) [2]. There is evidence that negative findings at laparoscopy and follow up with ultrasound will provide reassurance and relief to some patients. For patients without obvious pathology there is no established treatment.
In recent years, operative laparoscopy has developed rapidly and now makes laparoscopic pelvic denervation a practicable option in the management of chronic pelvic pain. The nerve plexuses and parasympathetic ganglia in the uterosacral ligaments, first described in the last century [10], and subsequently confirmed [11,12], carry "pain" from the uterus, cervix and other pelvic structures. The surgical approach to the management of pelvic pain is based on interruption of these nerve trunks that can be achieved by transection of uterosacral ligaments.
The transection of the uterosacral ligaments and the nerve plexuses it contains is the simplest of the surgical procedures for pelvic pain. The original work by Doyle described vaginal and abdominal approaches to divide the attachments of the uterosacral ligaments to the cervix [13,14]. With the wider use of minimal access therapy there is a renewed interest in the division of the Frankenhauser nerve plexus in the uterosacral ligaments laparoscopically using lasers or electro-diathermy. In an attempt to relieve patients' symptoms clinicians frequently perform laparoscopic uterosacral nerve ablation, or LUNA as it is sometimes called. However the efficacy of this procedure has not been assessed objectively using methodologically sound research.
The Need for the LUNA Trial
Health technology assessment in surgical interventions requires an initial evaluation of the safety and stability of new interventions followed by randomised trials [15]. The initial evaluative evidence alone is not sufficient to assess the clinical effectiveness of LUNA for which randomised research remains the gold standard.
Evidence from the medical literature: In the past, pelvic denervation has been proposed for treatment of chronic pelvic pain [13] but it has not been widely adopted due to the need for open invasive surgery. However, recent developments in minimal access surgery make LUNA a practicable treatment option. Our systematic review [16] has shown that the currently available research evidence on LUNA is inconclusive. Therefore further research is required to generate effectiveness evidence in the form of a high quality randomised controlled trial. This need is also supported by the recent Cochrane Reviews [17,18] that recommend rigorous research to assess surgical interventions in chronic pelvic pain.
Professional consensus
We conducted a survey of UK gynaecologists associated with the British Society of Gynaecological Endoscopy in 1998 to determine the extent to which LUNA was being used in practice [19]. Our survey indicated that despite the lack of definitive evidence, many gynaecologists familiar with the technique were using LUNA as a therapeutic option. Of the 247 respondents, 108 (44%) offered LUNA to patients: 78% for chronic pelvic pain, 66% endometriosis and 18% for dyspareunia/other reasons. Crucially, this survey indicated that 93 of 108 (86%) gynaecologists currently performing LUNA were willing to recruit patients in a randomised trial of LUNA. In this situation equipoise applies (i.e. the technique has been introduced without definite evidence but opinion regarding its use is not yet solidified) making the need for a trial even more urgent.
Pilot study
We have undertaken a pilot of the LUNA trial with the objective of assessing its feasibility. It has shown acceptability to patients. It has also established trial management procedures, piloted questionnaires, measured compliance and standardised operating procedures. A confidential interim analysis was recently reviewed by an independent data monitoring committee when the first 60 patients had completed 6 months follow-up. The committee recommend that a larger study is needed for adequate statistical power in the trial to evaluate LUNA reliably.
Research ProposalObjectives
1. To test the hypothesis that in women with chronic pelvic pain in whom diagnostic laparoscopy reveals either no pathology or mild endometriosis (AFS score ≤ 5) LUNA alleviates pain and improves life quality at 12 months (principal objective).
2. To test the hypothesis that response to LUNA differs according to the site and cause of the pain by two secondary analyses: (i) Women with central pain, (ii) women with no visible pathology.
3. To explore the variation in LUNA's effectiveness and side effects at different periods of follow-up (3, 6, months and 1, 2, 3, 5 and 10 years).
Design and setting
A multi-centre, prospective, randomised-controlled-trial involving centres in the UK and elsewhere will be carried out with single blind assessment of outcomes in eligible consenting patients randomised at diagnostic laparoscopy to LUNA (experimental group) or to no pelvic denervation (control group). Postal questionnaires including visual analogue scale for pain (primary outcome), an index of sexual satisfaction and the EuroQol 5D-EQ instrument (secondary outcomes) will be administered at 3, 6 and 12 months. The primary assessment of the effectiveness of LUNA will be from comparison of outcomes at the one-year follow-up, although the medium-term and long-term risks and benefits of LUNA will also be evaluated postal questionnaires to the women at 2, 3, 5 and 10 years after laparoscopy.
Eligibility
All new patients presenting to the Gynaecology outpatient clinic with pelvic pain (cyclical or noncyclical) and/or dyspareunia, and requiring diagnostic laparoscopy for evaluation of these conditions, will be invited to participate (Protocol, patient information leaflet and the relevant forms and questionnaire available on the LUNA trial website)[20]. They will be provided with information leaflet-Appendix A1 [see Additional file 1]. When they consent to participate (for consent form see Appendix A2- additional file 1), they are registered prior to randomisation (Appendix B- additional file 1) and an On study form is filled (Appendix C-see additional file 1)
Inclusion criteria
• Pelvic pain of longer than 6-month duration.
• Pain located within the true pelvis or between and below the anterior iliac crests.
• Associated functional disability.
• Lack of response to medical treatment.
• Diagnostic laparoscopy planned.
Exclusion criteria
• Previous LUNA.
• Mild, moderate and severe endometriosis (American Fertility Society score >5).
• Previous surgery for endometriosis.
• Previous surgery for pelvic inflammatory disease.
• Previous hysterectomy.
• Adnaexal pathology.
Interventions
Diagnostic laparoscopy plus uterosacral nerve ablation (experimental group) or laparoscopy without pelvic denervation (control group).
Laparoscopic uterosacral nerve ablation will be carried out in a uniform manner by named surgeons in each of the participating centres following a common protocol as described in the standard surgical text [21]. Routine preparation will be made for a diagnostic laparoscopy with the patient under general anaesthesia. Following pneumoperitoneum, a laparoscope will be used to visualise the pelvis. Before embarking on operative laparoscopy an anatomical pelvic assessment will be performed to identify pelvic structures and pathology. At this stage patients with pathology outlined in the exclusion criteria will be excluded. It is expected that around 30% of women will be unsuitable for LUNA at operation and will be 'registered only' cases. Eligible patients will be randomised by a telephone call to the Birmingham University Clinical Trials Unit.
Clear identification of the uterosacral ligaments is a prerequisite to treatment with lasers or electro-diathermy. The posterior leaf of the broad ligament will be carefully inspected to identify the course of the ureters, which on rare occasions could be particularly close to the uterosacral ligaments. Care will also be taken to note thin walled pelvic veins, which often lie lateral to the uterosacral ligaments. If accidentally punctured, they may cause troublesome bleeding requiring further endoscopic endocoagulation. The uterosacral ligaments will be identified by manipulation of the uterus in the right and left lateral planes. The ligaments will then be ablated with laser or micropoint electro-diathermy or endocoagulation depending upon the surgeons' preference. The ablation will start as close to the posterior aspect of the cervix as possible and continue for a minimum of 1 cm posterolaterally on either side. The aim of the procedure is to destroy the sensory nerve fibres and the secondary ganglia as they leave the uterus and come to lie within the uterosacral ligaments.
The safe conduct of operative laparoscopy for LUNA requires the use of two ports, one for delivery of the energy source (laser or diathermy) and another for manipulation. These are in addition to the umbilical port used for the laparoscope itself. In contrast, diagnostic laparoscopy in women with no pathology requires only one port in addition to the umbilical laparoscopic port. This difference in number of ports has potential for introducing bias by compromising patient blinding to group allocation. A sham incision (see Blinding and Ethical Considerations) in the control group is used to overcome this problem.
Randomisation
Consenting eligible patients will be randomised to diagnostic laparoscopy plus uterosacral nerve ablation (experimental group) or to no pelvic denervation at the time of diagnostic laparoscopy (control group). Important issues related to randomisation are considered below:
Allocation Sequence
The subjects will be allocated to groups using a chance procedure, blocking and stratification [22]. Stratified block randomisation will be employed to ensure that there will be nearly equal numbers of patients in the two groups within the prognostic subgroups, even if the study ends prematurely. Variable block size will be used to avoid any possibility of foreknowledge. Stratified allocation is used so that chance imbalances in the stratification variable do not have an effect on the outcome. Since response to treatment may depend on the surgeons' technique for laparoscopic uterosacral nerve ablation, analyses will be retrospectively stratified according to the surgeons participating in the trial.
Intraoperative randomisation
Allocation concealment is a crucial factor in avoiding bias in randomised trials [23]. Although it is not possible to blind the surgeon, it is essential to keep the surgeon blind to the group allocation until after the irrevocable decision to enter the woman into the trial has been made. So, in LUNA, treatment allocation will be issued at diagnostic laparoscopy, after the surgeon has inspected the pelvis and ensured that the patient fulfils all of the inclusion criteria and she does not have any of the exclusion criteria. Women may be randomised or registered into the study by telephoning the toll free Randomisation Line on 0800 953 0274 (+44 121 687 2319 from outside the UK) or by Internet randomisation at and clicking on the randomisation button. Passwords for Internet randomisation will only be allocated to centres with ethical approval.
Following surgery, the surgeon fills in operation details on a post-surgery form (Appendix D- see additional file 1)
Blinding
In the LUNA study, patients will be kept blind to their treatment allocation until the follow-up in the trial is complete. Patients may also show a placebo effect if they know they have received the active treatment. The magnitude of placebo effects should not be underestimated. There is clear evidence that inadequacies of blinding in randomisation lead to exaggeration of treatment effect in randomised trials [24].
However, there is a potential problem in the maintenance of blinding in the LUNA trial. As mentioned earlier, patients allocated to have LUNA will have the standard operative laparoscopy with three ports (one 10 mm umbilical port and two 5 mm lateral ports), whereas patients allocated to the control group under normal circumstances would have standard diagnostic laparoscopy with two ports (one 10 mm umbilical port and one 5 mm lateral or midline port). By noting the different number of incisions some patients might become aware of their group allocation and this might alter their response. In order to maintain patient blinding, a sham 5 mm skin incision is made superficially in a lateral port site. This approach in avoiding bias due to lack of blinding has been used in a previous trial of laparoscopic nerve ablation [25] and has also received ethical approval in this trial.
A second purpose of blinding is to prevent differences in other aspects of patient management introducing biases affecting the results. The patient's GP will therefore be kept blind to treatment allocation. Double blinding is not possible in LUNA, however, as the surgeons performing the surgical intervention on the patients will be aware of the group allocation. However, the likelihood that this will lead to bias in outcome assessment is low as the patient outcome assessments in this study will be conducted by self-administered questionnaires, avoiding any possible bias from surgeons' knowledge of group allocation.
EndpointsPrimary Outcome measure: Visual analogue scale for assessment of pain
Pain is difficult to measure, partly because it is accompanied by other sensations and partly because the reaction component affects the judgement of the pain regardless of the intensity of the stimulus. A measure of pain is nevertheless essential to the outcome of this clinical trial. Visual analogue scales (VAS) originally devised as measures of well being,[26] have been successfully adapted to measure pain [27,28]. This technique involves use of a 10 cm line on a piece of paper representing a continuum of the patients' opinion of the degree of pain. It is explained to the patient that the one extreme of the line represents "no pain at all" while the other represents "as much pain as she can possibly imagine". The subject rates the degree of pain by placing a mark on the line and scale values are obtained by measuring the distance from zero to that mark. Visual analogue scales (VAS) in the assessment of pain have been established to be reproducible and accurate [28]. VAS has commonly been used in measurement of chronic pain [29]. All the studies of LUNA included in systematic reviews so far have used this measure, or a variation on it, for assessing outcome. Individual pain scores have sufficient psychometric strengths to be used in chronic pain research involving group comparison designs[30].
Secondary Outcome measuresAssessment of sexual function
Sexual function is an important aspect of life quality in patients with pelvic pain. Pain itself is an anti-aphrodisiac, and together with discomfort and altered self image, it impacts upon sexual function [31]. Its assessment in an objective manner is therefore an important part of the LUNA trial. There are several sexual function instruments, with high levels of reliability, validity and responsiveness, which yield comparable results across occasions and individuals, making them suitable for monitoring therapeutic progress in randomised research [31].
The Sexual Activity Questionnaire (SAQ)[32] will replace the Brief Index of Sexual Satisfaction (ISS) [33] for the assessment sexual function. The outcome measure has been changed because of poor acceptability and compliance with the ISS. The SAQ has excellent internal consistency and test retest reliability. It also has excellent concurrent and construct validity and has been shown to be acceptable to women in other clinical trials [34] In the questionnaire it will be clearly stated that the measure of sexual function covers material that is sensitive and personal. Participants will be reassured that their responses will be kept completely confidential and that if they do not wish to answer any questions, they will be allowed to leave the questionnaire blank.
Health-related quality of life Instrument
Health-related quality of life (HRQL) instruments are becoming powerful tools for outcome assessments in randomised trials. Quality of life instruments assess aspects of patient's health status usually not grasped by conventional clinical indices; hence, they can be applied as complementary assessments together with VAS and ISS. Quality of life has to be defined clearly and patient's perception of normal performance serves a pivotal role in this context. HRQL instruments are administered with questionnaires assessing a number of different domains, i.e. areas of behaviour or experience that the instrument is attempting to measure [35].
Resource usage outcome measures
LUNA is a quick, safe and inexpensive procedure for women already undergoing diagnostic laparoscopy. Our hypothesis is about a clinically important effect without an excess of complications. If the hypothesis is confirmed, then any benefits will essentially be "dominant" outweighing the relatively small costs of intervention. Therefore, we do not plan a formal economic evaluation at this stage. However, data on health resource use will be collected partly as efficacy outcome measures (i.e. less need for medical care for pelvic pain indicates greater efficacy), and partly to allow an economic evaluation to be carried out, should significant complications occur. Other measures will include analgesic use, consultations at general practice and hospital, and time off work. Again these are both economic outcomes and indicators of residual pain.
Follow-up
Postal Questionnaires to assess pain and sexual function will be administered at enrolment in the trial and then at 3, 6, 12, 24 and 36 months after laparoscopy. The outcomes at 12 months will be used to address the primary research question. This time interval is chosen because laparoscopy alone has a placebo effect for up to 3–6 months in some patients [36,37]. The 24 and 36-month follow up will be used to monitor medium-term effects of the intervention. Existing participants in the trial who consented to 3 years of follow-up will be asked to consent to long-term follow-up (10 years) once they have reached the 3 year follow up time-point. Participants randomised after June 2003 are asked to consent to 10 years of follow-up at entry for long term follow up.
Ineligible Patients – follow-up only
Women who are ineligible for the LUNA trial because of moderate to severe endometriosis, significant adhesions, significant pelvic inflammatory disease, other significant pathology or those for whom LUNA is not technically feasible should be registered with the Trial Office for follow-up only. The woman should be told that she was not eligible for the trial randomisation, and the reasons why, and asked if she would agree to complete the follow-up questionnaires at 6 and 12 months.
This non-random cohort will provide comparative data on the natural history of patients with chronic pelvic pain with significant pathology.
Side-effects quantification
The centres have been advised to fill an " adverse event form" in case of immediate and delayed complications if any are associated with the procedure.
Sample Size and Power Considerations
The sample size for this trial has been estimated using the hypothesis that LUNA will alleviate pain symptoms (i.e. reduce pain scores on a VAS) more often than no intervention at one-year following diagnostic laparoscopy. Cohen describes 'effect sizes' of 0.2 and 0.5 standard deviations (SD) as 'small' and 'medium' [38]. Interim analyses of the pilot study indicate that the SD of the difference in change in VAS scales between LUNA and no pelvic denervation groups will be about 4.0. This corresponds to small and medium effect sizes on VAS of 0.8 and 2.0 respectively and is consistent with other studies of chronic pelvic pain, where clinically important symptom alleviation has been defined as a reduction in pain score of 2 or more [39]. To confirm or refute a small to medium effect of LUNA (0.3 SD difference), based on ά = 0.05 and β = 0.2 (80% power), 175 patients in each group (i.e. 350 patients in total) will be required. Considering a 20% loss to follow-up, the sample size is inflated to 210 patients in each group (i.e. 420 patients in total).
Stratification Variables
Randomisation will be conducted using minimisation, stratified by the four variables:
a. Presence or absence of some minimal pathology (minimal endometriosis ± ablation; adhesions requiring adhesiolysis only; minimal pelvic inflammatory disease)
b. Site of pain (presence of central pain or not)
c. Parity of the woman (nulliparous or parous)
d. Whether the woman is sexually active or not
The first two variables form the prespecified subgroup analyses, and the other two variables are included as having impact on dysmenorrhoea and dyspareunia respectively.
Data AnalysisType of analysis
Intention to treat
The main analysis to address the principal research questions will be conducted using the one-year follow-up data. The mean differences in VAS pain scores; sexual satisfaction and life quality scores in the two groups will be compared using a two-sample t-test. The rates of women with clinically significant (2 VAS point) alleviation of pain symptoms will also be compared producing a relative risk estimate with 95% confidence intervals (Mantel-Haenzel test). Baseline characteristics of the patients enrolled in the two groups will be compared to ensure that randomisation has produced comparable groups of patients. The use of additional treatment (co-intervention) for pelvic pain following LUNA or no pelvic denervation will be assessed for any systematic difference between the two groups.
Subgroup analyses are limited by statistical power and can produce spurious results particularly if many are undertaken. Our literature review [16] and consultation with gynaecologists [19] suggests that the effectiveness of LUNA may be greater for central compared to non-central pain and if there is no associated pathology (i.e. no endometriosis). Therefore, we have chosen to limit secondary analyses to these subgroups only. The LUNA trial is powered to detect a small to medium overall difference and if a larger treatment benefit is found then other subgroup analyses will be undertaken, appropriately cautiously. The LUNA trial is powered overall at 80% to detect a 0.3 SD difference in effect. Our pilot study shows that 60% patients have mainly central pain and 70% have no pathology. Hence, in the subgroup with central pain the power will be 80% to detect a 0.4 SD treatment effect. In the subgroup with no pathology the power will be 80% to detect a 0.35 SD treatment effect.
Frequency of analyses
Interim analysis at completion of 6 months follow-up of the first 60 patients has already been completed as pilot study.
Bi-annual analyses of recruitment, compliance and loss to follow-up are being carried out for LUNA Trial Management Committee.
An annual interim analysis of effectiveness is done for confidential review by independent Data Monitoring Committee to determine whether the principal question has been answered and to monitor adverse events.
Main analyses of effectiveness of LUNA will be done at completion of one year follow up of the total study sample. Additional analysis of long-term effects will be done at two and three-year follow-up points.
Ethical considerations
When there is uncertainty about the appropriate therapy, scientific clinical trials are the best scientifically ethical way to resolve uncertainty and thereby benefit both the individual patients and all others concerned in their care [40]. The need for a "sham" incision in this trial is the main ethical issue and it is required because without it the patients cannot be kept blinded. The purpose of blinding is to prevent various biases from affecting the results. The need for blinding in surgical trials has been emphasised in the medical literature [41] and there is empirical evidence that inadequacies of blinding in randomisation lead to exaggeration of therapeutic efficacy in randomised trials [42]. Blinding of patients in surgical trials is clearly indicated when the intervention primarily treats symptoms and when the outcomes are based on patients' own assessment. LUNA is an intervention for treating chronic pelvic pain (a diagnosis based on symptoms) and the outcome assessment is based on patients' responses on a visual analogue pain scale and a quality of life instruments. Hence, the use of a "sham" incision is justified if bias is to be avoided in the LUNA trial and this approach has been used in a previous trial of LUNA [43]. Ethical approval for the LUNA trial procedures has already been obtained from the Multicentre Research Ethics Committee (MREC).
Conduct of the Study and Data Management
The trial will be managed from the Birmingham Clinical Trials Unit. Each investigating centre will carry out the study in accordance with the study protocol and to the MRC guidelines on Good Clinical Practice in Clinical Research (1998). Patients will be invited to participate if they fulfil all the inclusion criteria and do not have any exclusion criteria (See Appendix B Eligibility Checklist and Randomisation Form). They will be provided with a laparoscopy and LUNA trial information leaflet (Appendix A1) and signed consent obtained prior to laparoscopy (Appendix A2). Consenting patients should be asked to complete the Enrolment Questionnaire and On Study Form. The final decision to enrol patient in the trial will depend on the findings at laparoscopy when the surgeon will perform LUNA or not after determining eligibility as shown in Eligibility Checklist and Randomisation Form (see Appendix B). At the end of the procedure the surgeon will complete the Post Surgery Form (Appendix D). All the three forms for each patient enrolled will be photocopied to keep a record at the participating centre and the originals will be sent to the Birmingham Clinical Trials Unit, which will act as the co-ordinating centre. At 3, 6 and 12 months after enrolment, the Follow-up Questionnaire will be mailed to the patients with a pre-paid self-addressed envelope. Recruitment is expected to take 12 months and follow up for the main endpoints another 12 months. At completion of the main study further Follow-up questionnaires will be mailed out at 24 and 36 months.
Finishing date
September 2005; expected reporting date: December 2006
List of abbreviations used
AFS- American Fertility Society
5D EQ- 5 dimensional European Quality of Life Questionnaire
GP- general practitioner
LUNA- laparoscopic uterosacral nerve ablation
RR- relative risk
95% CI- 95% confidence intervals
VAS- visual analogue scale
HSG-health service guidelines
SAQ- Sexual Activity Questionnaire
HRQL- Health-related quality of life
Competing interests
None declared.
Authors' contributions
The LUNA Trial Collaboration consists of a management group and a collaborative group.
The management group consists of Pallavi Latthe (PL), Khalid S Khan (KSK) Janesh K Gupta (JKG) from Department of Obstetrics and Gynaecology, Jane Daniels (JD), Robert K Hills (RKH) and Richard Gray (RG) from Birmingham Clinical Trials Unit, and Richard Lilford (RJL) from Department of Public Health and Epidemiology, all from University of Birmingham, UK. KSK and RG were involved in developing the original idea of the trial. KSK, JKG, RJL and RG were co-applicants on the successful funding proposal and are responsible for the overall direction of the project. PL, JD and RKH have contributed to adaptations from the original design. PL co-ordinates clinical work within various study centres and recruits patients along with other collaborators. RG and RKH provide statistical advice. All of the above authors are members of the trial management group and have approved the final manuscript.
The collaborative group consists of clinicians and centres that recruit patients in this trial. At the time of submitting this protocol, the following clinicians (centres) have recruited patients: P Chein (Ninewells Hospital, Dundee, UK); K Chin (Staffordshire General Hospital, UK); M O'Connor, A Baxter, T Farrell (Royal Hallamshire Hospital, Sheffield, UK); LS Dwarakanath (City Hospital, Birmingham, UK); S Irani (Birmingham Heartlands and Solihull Hospital NHS Trust, UK); KSK, PL, JKG, C Mann, TJ Clarke (Birmingham Women's Hospital, Birmingham, UK); V Kay (Forth Park Hospital, Kirkcaldy, UK); D Phillips (Perth Royal Infirmary, UK); JS Samra, C Cox (New Cross Hospital, Wolverhampton, UK); S Keay (Walsgrave Hospital, Coventry, UK); E Shaxted (Northampton General Hospital, UK) and J Thornton (City Hospital, Nottingham, UK) The following centres have obtained ethics approval and await recruitment: A Ali (Darlington Memorial Hospital, UK); C Chandler (Billinge Hospital, Wigan, UK); E Downes (Chasefarm Hospital, Enfield, UK); A Hextall (St Albans City Hospital, UK);J Kabukoba (Sandwell Hospital, West Bromwich, UK); M Katesmark (Epsom General Hospital, UK); P. Kirwan (Leicester General Hospital, UK); K Louden (Royal Hampshire Hospital, UK);J Moohan (Altnagelvin Area Hospital, UK); S Najja (Queen's Park Hospital, Blackburn, UK); K Olah (Warwickshire General Hospital, UK); M Padwick (Watford General Hospital, UK); P Owen (Stobhill Hospital, Glasgow, UK) and A Tapp (Royal Shrewsbury Hospital, UK).
Pelvic sensory nerve pathways and site for luna
The laparoscopic uterosacral nerve ablation (luna) trial schema
Study Flow Chart
Clinic
Theatre
Follow up (months/ years)
3
6
12
2
3
5
10
Patient information and consent (App A)
x
Eligibility Checklist and Randomisation form (App B)
x
x
Enrolment Questionnaire (Appendix C1)
x
Surgery Form (Appendix D)
x
Letter to GP (Appendix E)
x
Follow-up Questionnaire (Appendix C2, C3)
x
x
x
x
x
x
x
Pre-publication history
The pre-publication history for this paper can be accessed here:
Supplementary Material
Additional File 1
Appendix A1: Patient information leaflet. Appendix A2: Consent form. Appendix B: Registration and randomisation form. Appendix C: On study form. Appendix D: Post surgery form.
Click here for file
Acknowledgements
Funding: WellBeing, UK (Charity grant no. CF/371)
Data Monitoring Committee: P Brocklehurst, D Braunholtz and JA Jordan
Acute pancreatitis rarely complicates pregnancy. Although most pregnant women with acute pancreatitis have associated gallstones, less common causes such as drugs have been reported.
Case presentation
We report the case of a 34-year-old woman who underwent medical abortion with mifepristone and gemeprost and received codeine as pain-relief during the induction of abortion. She developed a severe acute necrotizing pancreatitis which required 14 days of intensive care. Other possible etiological factors, i.e. gallstone, alcohol intake and hyperlipidemia, were excluded.
Conclusions
The reported case of acute pancreatitis was most likely drug-induced.
Background
Acute pancreatitis rarely complicates pregnancy. Acute pancreatitis complicated 1 in 3300 pregnancies at a large public hospital in Dallas, Texas [1], whereas in southern California 1 in 1500 women were affected [2]. Other published reports cite incidences ranging widely from 1 in 1000 to 1 in 12,000 live births [3,4]. Although most pregnant women with acute pancreatitis have associated gallstones, less common causes such as trauma, drugs, and ethanol ingestion have also been reported [5,6]. We report the case of a woman who developed severe acute necrotizing pancreatitis that required 14 days of intensive care following medical abortion.
Case presentation
The patient was a pregnant gravida IV, para I, 34-year-old woman who due to severe malformation of the fetus underwent medical abortion. She was healthy, was not on any medication and had no known allergies. There was no history of alcohol or any other substance abuse. Previous interrupted pregnancies had been due to miscarriages in the first trimester and on one occasion medical abortion due to a diagnosed Turner's syndrome of the fetus.
A chorion villi sample during the present pregnancy had shown a normal female chromosomal constitution. However, repeated ultrasound examinations revealed poor fetal growth and a complete lack of amniotic fluid. There was no sign of fluid leakage. Ultrasound examination at week 17 showed a cystic formation in the fetal abdomen, most likely representing a serious renal malformation not compatible with life. The patient decided to terminate pregnancy and was admitted to Uppsala University Hospital for a medical abortion. Treatment was initiated at week 18 of pregnancy with 600 mg mifepristone given orally after which the patient experienced mild gastric pain, which is commonly seen after mifepristone administration [7].
The next day, four vaginal pessaries of 1 mg gemeprost were administered every third hour. At the same time as treatment with gemeprost was initiated, she also received oral doses of a codeine 30 mg/paracetamol 500 mg combination due to her above mentioned mild gastric pain. Abortion was induced the same day and the patient underwent exceresis. The fetus exhibited skeletal malformations. However, six hours after administration of the first doses of gemeprost and of codeine/paracetamol, the patient developed severe epigastric pain. She had at this point received a total of 2 mg of gemeprost, 60 mg of codeine, and 1 g of paracetamol, apart from the 600 mg of mifepristone given the day before. Her pain was at first judged as gastritis, but omeprazole had little effect. Visual analogue scale (VAS) score was at this stage between 8 and 9 out of 10, where 10 indicates maximum subjective pain. Repeated doses of 2 mg morphine i.v., and at one occasion 5 mg ketobemidon i.m., were necessary to relieve the pain. An acute CT of the abdomen revealed a swollen pancreas as seen with pancreatitis, and plasma amylase levels were elevated (up to 23 μkat/L, reference interval 0.2–0.8 μkat/L). Although a CT scan is not optimal for imaging gallstones, there were no signs of cholelithiasis or dilated biliary ducts. The patient's pancreatitis was classified as severe acute necrotizing pancreatitis, a condition with a high mortality rate [8,9]. When the diagnosis was made, she was immediately transferred to the intensive care unit. Her pain was later controlled through epidural anaesthetics. Results from repeated blood samples are shown in table 1. The patient was febrile and had an increased leukocyte blood count and plasma CRP as well as decreased levels of antithrombin and elevated levels of fibrinogen and fibrin d-dimer, indicating activation of the coagulation system. Apart from a decreased plasma albumin level, all liver parameters were normal. The values slowly returned towards normal during the following 14 days of intensive care. During this time, antibiotic treatment with cefuroxime and metronidazole was given. At follow up 1 month later, her plasma amylase level had not yet returned to normal (2.0 μkat/L) and was still elevated after 2 months (1.1 μkat/L).
Blood monitoring during the 14 days of hospitalization.
Day 1 (day)
Day 1 (evening)
Day 2
Day 3
Day 4 (morning)
Day 4 (evening)
Day 5
Day 6
Day 8
Day 10
Day 14
B-leukocyte count (109/L)
21.0
16.3
18.7
19.5
17.0
17.6
14.9
13.6
-
12.3
7.4
B-hemoglobin (g/L)
115
100
86
81
80
84
82
112
115
110
125
B-platelet count (109/L)
269
217
208
200
226
216
249
-
-
373
-
P-CRP (mg/L)
126
147
199
329
290
210
190
173
122
75
<5
P-sodium (mmol/L)
137
-
-
-
-
-
-
136
140
141
-
P-potassium (mmol/L)
3.4
-
-
-
-
-
-
3.6
4.0
4.3
-
P-creatinine (μmol/L)
80
64
62
78
66
62
67
75
74
81
-
P-glucose (mmol/L)
-
-
-
-
-
8.3
-
-
-
-
-
P-urea (mmol/L)
-
1.6
1.8
2.1
-
-
1.4
-
-
-
-
P-cytatin C (mg/L)
-
-
-
0.80
-
-
-
-
-
-
-
GFR. calculated (ml/min)
-
-
-
100
-
-
-
-
-
-
-
P-bilirubin (μmol/L)
11
12
10
5
8
-
-
-
-
-
-
P-albumin (g/L)
-
26
21
22
22
-
-
-
29
28
-
P-triglycerides. fasting (mmol/L)
-
-
-
-
-
-
-
-
-
2.33
-
P-calcium (mmol/L)
-
-
-
-
1.75
-
-
-
-
-
-
P-LD (μkat/L)
6.5
4.7
4.3
5.5
5.7
-
-
-
-
-
-
P-ALP (μkat/L)
2.8
2.1
2.1
3.1
3.5
-
-
-
-
-
-
P-amylase (μkat/L)
23
16
11
4.1
2.4
-
2.3
3.0
-
-
-
P-ASAT (μkat/L)
0.46
0.32
0.37
0.30
0.26
-
-
-
-
-
-
P-ALAT (μkat/L)
-
-
0.12
0.25
0.11
-
-
-
-
-
-
INR
-
1.2
1.2
1.2
1.1
1.1
1.1
-
1.0
-
-
P-antithrombin (%)
-
75
64
-
65
-
30
-
-
-
-
P-APT time (s)
-
28
28
30
31
29
-
-
-
-
-
P-fibrinogen (g/L)
-
3.7
3.7
-
4.2
-
-
-
-
-
-
P-fibrin d-dimer (mg/L)
-
0.5
0.5
-
0.8
-
-
-
-
-
-
Examination of the fetus revealed a likely Potter's syndrome with malformations of the hands and feet, poor development of the skeleton in the lower extremities, 13 ribs and a fused lumbar vertebrae. The kidneys were absent, there was a cloacal malformation, and only two vessels were seen in the umbilical cord.
Drug-induced pancreatitis is considered to be rare [10]. However, the exclusion of other possible etiological factors combined with the coincidence in time with the intake of both codeine/paracetamol and mifepristone and gemeprost make these drugs suspect as causative agents. On her previous medical abortion she received only two vaginal pessaries of 1 mg gemeprost, and no mifeprisone, without complications. On this occasion, she also received a combination of codeine 30 mg/paracetamol 500 mg. These facts would point out mifepristone or possibly gemeprost as the likely causative agents. However, there have been a handful of published cases of pancreatitis complicating treatment with codeine [10-12]. A common feature of these reported cases are previous cholecystectomies [11,12]. The patient reported here had not been cholecystectomized. Also, three cases of possible codeine-precipitated pancreatitis have been reported since 1965 to the Swedish Drug Information System (SWEDIS) handling reports on suspected adverse reactions to drugs used in Sweden [13]. Codeine is known to cause rapid but transient spasm of the sphincter of Oddi [11]. Laboratory studies have shown that codeine may cause a mild, transient hyperamylasemia [11]. Pancreatitis following paracetamol overdose have been previously reported [11], but the doses taken in the present case are unlikely to have been causative. Gemeprost is a synthetic prostaglandin E1 (PGE1) analogue. Studies indicate that PGE1 is a modulator of pancreatic blood flow and protein production. For instance, PGE1 stimulated the production and secretion of alpha-amylase from minces of porcine pancreas in vitro [14], and enzyme output in dogs in vivo [15]. PGE1 further increases mesenteric and pancreatic blood flow [15]. Thus, PGE1 is not devoid of actions on the pancreas. Progesterone has also been shown to exert modulating action on the pancreas. In rats, progesterone stimulated pancreatic cell proliferation in vivo [16]. Progesterone receptors have also been shown to be present in human pancreatic tissue [17]. However, the effect of a progesterone receptor antagonist such as mifepristone on the pancreas has not been studied. Evidence thus exists that both PGE1 and progesterone exert modulatory action on the pancreas, but to our knowledge, no reports on pancreatitis following treatment with gemeprost or mifepristone have been published.
It is worth noting that a small number of pregnant women with acute pancreatitis have an associated hyperlipidemia, usually hypertriglyceridemia. In the reported case, the p-triglyceride level 10 days after diagnosis was elevated to 2.33 mmol/L (about 75 mg/dl). However, mild to moderate hyperlipidemia may be secondary to acute pancreatitis [18], and it is generally believed that a triglyceride level of >1,000 mg/dl is needed to precipitate an episode of acute pancreatitis [19]. Furthermore, a follow-up measurement of p-triglycerides six months later was 1.22 mmol/L.
Conclusions
In conclusion, due to the coincidence in time and exclusion of other possible etiological factors such as cholelithiasis, alcohol and hyperlipidemia, the reported case of severe acute necrotizing pancreatitis was most likely drug-induced. Since codeine, although uncommonly, has previously been reported to precipitate pancreatitis, this drug must be the first to suspect. However, we cannot exclude gemeprost or mifepristone as the causative agents.
Competing interests
None declared.
Authors' contributions
PH coordinated the collection of information on the patient and wrote the manuscript. HA was the patient's treating physician and provided all of the information on the patient, as well as reported the case to the local pharmacovigilance agency. EH first received the report, made a preliminary summary of the case, and searched SWEDIS for previous reports on drug-induced pancreatitis. All authors participated in the final adjustments of the manuscript.
Pre-publication history
The pre-publication history for this paper can be accessed here:
It has been suggested that bone loss and cognitive decline are co-occurring conditions, possibly due to their relationship with estrogen. Cognitive decline has been associated with various nutritional deficiencies as well. The purpose of this study was to determine if cognitive function is related to bone mineral density of various skeletal sites as well as to various dietary components.
Methods
Cross-sectional study with 97 healthy, Caucasian, postmenopausal women (59.4–85.0 years) enrolled in a larger longitudinal study, investigating the effects of sodium on bone mass. The subjects were divided into two groups based on cognition scores. Group 1 represented lower and Group 2 higher scores on cognitive function. Bone mineral density from the whole body, lumbar spine, femur and forearm were measured with the Lunar DPX-MD instrument. Anthropometry was measured by standard methods. Cognition was assessed using the Mini Mental State Examination. Cumulative (over 2 years) dietary intake from 3-day records was analyzed by Food Processor® (ESHA Research, Salem, OR) and cumulative physical activity was assessed using Allied Dunbar National Fitness Survey for older adults.
Results
Subjects' cognition scores ranged from 22–30 (normal, 27–30), indicating all subjects had either mild or no cognitive impairment. Multiple Analysis of Covariance adjusted for age, height, weight, physical activity, alcohol, calcium, sodium and energy intake, showed a statistically significant association between cognition and bone mineral density of all measurable sites (η2 = 0.21, P < 0.01). However, after Analysis of Covariance follow-up tests and Bonferroni correction, the differences for individual bone sites diminished, though Group 2 had higher adjusted means for all sites except for the femoral neck, Ward's triangle and trochanter. There was a positive significant association between cognition score and carbohydrate and potassium intake (η2 = 0.07, P = 0.050). Group 2 did have a significantly higher potassium intake (P = 0.023). In multiple regression, saturated fat had a significant negative relationship with cognitive function.
Conclusions
It appears mild degree of cognitive impairment may be a marker for lower bone mineral density as well as for a diet lower in carbohydrate and potassium intake, and higher in saturated fat. Consequently, older women with cognitive impairment may benefit of being screened for potential bone loss and poor nutrition.
Background
It is well known that a decline in estrogen is related to bone loss and possibly a loss in cognitive function as well. It has been suggested that bone loss and cognitive decline are co-occurring conditions, likely due to their relationship with estrogen [1,2]. Recent studies suggest estrogen replacement therapy may benefit cognitive function in postmenopausal women without memory impairment [3]. Other studies report estrogen therapy may benefit women with Alzheimer's disease (AD) [4] or those at risk for AD [5]. Conversely, studies also report estrogen deficiency is not associated with poor cognitive function [6]. These conflicting results could be due to various reasons, including the type of cognitive assessment tool, inclusion (or not) of confounders, degree of cognitive decline, age of the subjects, to name just a few.
Few studies have examined the relationship between cognitive function and bone mineral density (BMD), both as possible markers of cumulative estrogen exposure. Two studies found a positive relationship with total hip BMD [1,2] and femoral neck BMD [1] and cognitive function in older adults presumably due to higher cumulative exposure to estrogen [1,2]. Women with osteoporosis were found to have poorer cognitive function (presumably due to lower estrogen exposure) [2]. However, both of these studies used elderly subjects with significant cognitive impairment and/or history of osteoporotic fractures and one of them [1] did not control for physical activity, a significant predictor of both BMD [7] and cognitive function [8].
Both cognitive function and BMD also depend on various nutrients in food. Studies have found elderly people who performed better on cognitive tests had higher intakes of energy, fruits, vegetables, fiber, carbohydrate, protein, B-vitamins, vitamins A and C, calcium, phosphorous and iron [9-11], and overall better healthy diet score [12]. In general, it appears a more balanced diet that contains more vitamins and minerals is associated with better cognitive function. Substantial research in the past decade has focused on the relationship between antioxidants from diet and/or supplements and cognitive function. However, no clear consensus has been reached as to whether taking antioxidant supplements will help preserve cognitive function and prevent the development of dementia. Some studies suggest vitamin E [13,14], vitamin C [15] and beta-carotene [15,16] positively benefit cognitive function, while other studies have not found a relationship between vitamin E [16,17], vitamin C [13,14,16-18] or beta-carotene [13,17,18] and cognitive function. There is also controversy as to whether particular fats in the diet are beneficial or potentially harmful to cognitive function due to their relation to the development of atherosclerosis, thrombosis, and inflammation of arterial walls [17,19].
Due to the above inconsistencies and uncertainties in the literature we investigated cognitive function and its relationship with bone, as well as with nutrients from both diet and supplements, in healthy postmenopausal women with either no or mild cognitive impairment. Our research questions were as follows: 1) Does bone mass in various skeletal sites differ by cognitive function among postmenopausal women when controlling for confounders known to impact bone, such as age, height, weight, hours of total activity, alcohol consumption and total energy, sodium and calcium intake? 2) How do various nutrients in the diet relate to cognitive function in postmenopausal women when controlling for the above confounders?
MethodsSubjects
This cross-sectional evaluation included 97 currently non-smoking, Caucasian, postmenopausal women free of chronic diseases (including severe osteoporosis) and medications (including estrogen) known to affect bone. All subjects reported at least 12 years of education (high school graduate), with many having additional college education. Subjects were part of a larger longitudinal study investigating the effects of a reduced sodium intake on bone, as described previously [20] and were randomly asked to participate in this study. Subjects were instructed to maintain a calcium intake of 1200 mg a day, the current recommended adequate intake [21], and were given calcium citrate supplements, if necessary, to assure adequate calcium intake throughout the study (as per the protocol of a larger study). All data for this study were collected by one person, a registered dietitian (RAB). The Institutional Human Subjects Review Board approved study protocol and subjects signed informed consent.
Anthropometry and bone densitometry
Weight and height were measured by standard procedures in indoor clothes without shoes. BMD (g/cm2) was measured by dual X-ray absorptiometry with a Lunar DPX-MD instrument (GE Medical Systems, Madison, WI, USA) using specialized software for whole body, lumbar spine, femur (neck, trochanter, Ward's triangle and shaft) and forearm (including radius and ulna), as described previously [22]. The BMD and anthropometries, were measured every 6 months throughout the study, but only those obtained at the time when cognition test was given (usually in a second year of the study) were used in analyses and in relation to cognitive status. Quality assurance of our densitometer was performed daily and coefficients of variation and precision of the instrument were reported previously [22].
Cognitive assessment
Cognition was assessed using the Mini Mental State Examination (MMSE) [23]. The MMSE is a brief, commonly used screening tool to assess cognitive status. It consists of various graded questions and tasks generating a maximum of 30 possible points. Each subject was asked a series of questions from eleven categories; orientation to time, orientation to place, registration, attention and calculation, recall, naming, repetition, comprehension, reading, writing and drawing. Scores are classified as: normal cognitive function with a score of 27–30; mild cognitive impairment with a score of 21–26; moderate cognitive impairment with a score of 11–20; and severe cognitive impairment with a score of 0–10 [24]. Cognitive function for each subject was assessed one time, usually in the second year of the study, and evaluated with the corresponding bone and anthropometric measurements conducted at that time, or with the cumulative average of other variables of interest.
Dietary and alcohol consumption assessment
Subjects were instructed by a registered dietitian to record 3 days of dietary intake (2 week and 1 weekend day) every 6 months during the study. Food models and pictures were used for instruction. After the records were completed, the same dietitian followed-up and rechecked every record with each participant, particularly about additional food or snacks consumed, portion sizes, and ways of preparation. The same dietitian analyzed nutrient intake from the records using Food Processor® (ESHA Research, Salem, OR). Mean daily intake, including total energy and all other macro- and micronutrients (vitamins, minerals and fatty acids) was calculated. Supplement use was recorded during each visit, as well. The detailed description of diet assessment is reported earlier [20]. Cumulative average intake of each considered nutrient, to the point of the cognition assessment (usually in a second year), was used to assess its relationship with cognitive functioning. Alcohol consumption was assessed using questionnaire designed to determine long-term (at least for a last year and extending back to several years), frequency, amount, and source of intake, with the help of the same dietitian. It was expressed as drinks per day, from which g/day of alcohol was calculated, as described previously [25].
Physical activity
Physical activity (PA) was assessed using the modified Allied Dunbar National Fitness Survey for older adults [26] and was collected every 6 months. A measure of total activity was assessed based on minutes per week engaged in weight bearing activities of a moderate intensity defined as an activity of at least 4 kcal/min, such as recreational activities (tennis, hiking, weight lifting), walking, heavy housework, gardening and do-it-yourself activities. Data collected included frequency and duration of each activity and were expressed as number of hours per week engaged in the above activities. The questionnaire was filled with the participants on the site and with the help of a dietitian. We used this questionnaire in our previous studies in postmenopausal women and found it reliable and easy to complete [27]. Cumulative average activity score to the point of the cognitive assessment was used in the subsequent analyses.
Data analysis
Data analysis was conducted with SPSS statistical software (version 8.0). Pearson's r was calculated among all variables, as a preliminary analysis. To determine the relationship between cognitive function and BMD, subjects were divided into two groups based on the mean (27.9) for the MMSE score. Group 1 comprised of subjects below and group 2 of subjects above the mean. One-way Analysis of Variance (ANOVA) was conducted for various bone sites and covariates to determine if there were significant group differences. Multiple Analysis of Covariance (MANCOVA) with univariate follow-up tests and Bonferroni corrections to the alpha level were used to check for group differences in BMD of the whole body, all sites of the hip (neck, trochanter, Ward's triangle and shaft), lumbar spine and forearm. Subjects' age, height, weight, hours of total activity, alcohol consumption and total energy, calcium and sodium intake were included in the analysis as confounders. The use of these confounders was based on scientifically proven and theoretically presumed evidence for their possible effects on bones and memory.
To determine the relationship between cognitive function and diet, subjects were divided into two groups based on MMSE criteria for normal cognitive function (score of ≥27). Group 1 (<27), was considered to have mild cognitive impairment while Group 2 (≥27), was considered to have normal cognitive function. Multiple regression models were calculated to find the best model using various nutrients to predict cognitive function. ANOVA was conducted for all nutrients and covariates. MANCOVA with confounders (described above) and univariate follow-up tests were used to check for group differences in nutrient intake. Significance level was set at P ≤ 0.05.
Declaration of sources of funding
This work was funded in part by the NRI/USDA 2001-00836, Donaghue Medical Research Foundation DF98-056, University of Connecticut Office for Sponsored Programs and Mission Pharmacal®, San Antonio, TX, USA. The financial sponsors played no role in the design, execution, analysis and interpretation of data, or in writing the study.
ResultsCognitive function and BMD
All bone sites, whole body, hip (neck, trochanter, Ward's triangle and shaft), spine and forearm, were sufficiently highly correlated with each other, as expected (r ranged from 0.43–0.98, P = 0.0001), to justify the use of MANCOVA for this data set. Subjects' mean MMSE score was 27.9, with a range of 22–30, indicating all subjects had either mild or no cognitive impairment (Table 1). Descriptive statistics for each group and all subjects combined are presented in Table 1. ANOVA was conducted to assess group differences for age, height, weight, total activity, alcohol consumption, and energy, sodium and calcium intake. Group 2 had significantly higher body weight (F1,95 = 8.57, P = 0.004) than Group 1, while other variables of interest were not significantly different.
Mean ± SD for descriptive characteristics of all subjects and for each group1
Variable
Group 1 n = 34
Group 2 n = 63
All Subjects n = 97
Age
69.7 ± 6.9
69.7 ± 6.7
69.7 ± 6.7
Weight (kg)2
63.9 ± 9.0
70.7 ± 11.9
68.3 ± 11.4
Height (cm)
160.6 ± 6.1
161.6 ± 7.3
161.2 ± 6.9
MMSE score3
25.4 ± 1.5
29.2 ± 0.8
27.9 ± 2.1
Alcohol (drinks/day)
0.4 ± 0.6
0.3 ± 0.5
0.3 ± 0.5
Energy (kcal/day)
1596 ± 252
1645 ± 345
1628 ± 315
Total calcium (food and supplements) (mg/day)
1426 ± 350
1438 ± 279
1434 ± 304
Sodium (mg/day)
2114 ± 661
2322 ± 955
2249 ± 866
Total Activity (hr/wk)
5.4 ± 3.7
6.3 ± 6.0
6.0 ± 5.3
1Groups were stratified to below (Group 1) and above (Group 2) the mean for cognitive function score of 27.9 2statistically significant difference between groups, P < 0.01 3 Mini Mental State Examination score
MANCOVA was conducted for the above groups to determine the relationship between cognitive function and BMD for whole body, all sites of the hip, lumbar spine and forearm. Box-M test was not significant indicating the observed covariance matrices of the dependent variables were equal across both groups: Box-M = F(28,15625) = 30.11, P = 0.502. Levene test revealed the error variance of the dependent variables was equal across both groups, P ranged from 0.407 to 0.948. MANCOVA results revealed statistically significant differences between the two groups of subjects, Wilks' Λ = 0.79, F(7,77) = 2.96, P = 0.009, multivariate η2 = 0.21. Several of the covariates significantly influenced the combined dependent variables: total activity, Wilks' Λ = 0.79, F(7,77) = 2.93, P = 0.009, multivariate η 2 = 0.21; weight, Wilks' Λ = 0.73, F(7,77) = 4.11, P = 0.001, multivariate η2 = 0.27; and age, Wilks' Λ = 0.72, F(7,77) = 4.28, P = 0.001, multivariate η2 = 0.30. ANCOVA with Bonferroni correction to the alpha level (P, 0.05/7 variables = 0.007) was conducted on each dependent variable as a follow-up test to MANCOVA. The significance between groups slightly diminished, but the trend remained. Table 2 lists the adjusted and unadjusted means of BMD for each group at each bone site. Group 2 had higher unadjusted BMD for all sites except for the Ward's triangle, and higher adjusted means for all sites except for the femoral neck, Ward's triangle, and trochanter. Groups were significantly different using Bonferroni correction to the alpha level for several of the covariates: weight was significantly different for all sites of the hip BMD; total activity was significantly different for the femoral neck and Ward's triangle BMD; and age was significantly different for forearm BMD.
Adjusted1 (in bold) and unadjusted means ± SD, with corresponding P values of bone mineral density (BMD) of each bone site for two groups2 of subjects
Bone Site (BMD) (g/cm2)
Group 1
Group 2
P-values
Femoral Neck
0.825
0.822
0.883
0.795 ± 0.09
0.838 ± 0.12
0.093
Ward's Triangle
0.731
0.678
0.042
0.700 ± 0.11
0.691 ± 0.14
0.820
Trochanter
0.758
0.742
0.776
0.719 ± 0.12
0.761 ± 0.13
0.093
Femoral Shaft
1.036
1.041
0.838
0.999 ± 0.13
1.057 ± 0.17
0.062
Lumbar Spine3
1.082
1.099
0.659
1.057 ± 0.16
1.113 ± 0.21
0.237
Forearm4
0.431
0.445
0.265
0.426 ± 0.05
0.448 ± 0.67
0.173
Whole Body
1.083
1.090
0.672
1.067 ± 0.08
1.098 ± 0.92
0.101
1 Adjusted for age, height, weight, total hours of current activity a week, alcohol consumption, and total energy, calcium and sodium intake, 2Groups were stratified to below (Group 1) and above (Group 2) the mean for cognitive function score of 27.9 3BMD of the second, third and fourth lumbar vertebrae 4BMD of the ulna and radius
Cognitive function and diet
To explore the relationship between diet and cognitive function we calculated Pearson's r for all variables. Both carbohydrate and total potassium intake (food and supplements) were significantly correlated to MMSE score, r = 0.22, P = 0.029 and r = 0.20, P = 0.048, respectively. In a multiple regression model containing age, height, total activity, alcohol consumption, and total vitamins B6, B12, C, A, and E intake, the latter implied in some studies to having influence on cognition) weight and energy were positive predictors and saturated fat intake (mean ± SD = 18.8 g ± 8.0 g) was a negative predictor of MMSE, with multiple R2adjusted = 0.071.
Subjects were divided into two groups based on MMSE criteria for normal and impaired cognition, with Group 1 classified as mild cognitive impairment (MMSE score <27), and Group 2 classified as normal cognitive impairment (MMSE score ≥27). ANOVA was conducted to assess group differences for age, height, weight, total activity, alcohol consumption and energy intake. Body weight was significantly higher in group 2, (F1,95 = 8.57, P = 0.004), whereas other variables of interest were not different between groups, P ranged from 0.235 to 0.980.
MANCOVA for the above groups was conducted to determine the relationship between cognitive function and carbohydrate and total potassium intake, as combined dependent variables. MANCOVA results revealed a statistically significant difference among the two groups of subjects, Wilks' Λ = 0.93, F(2,86) = 3.11, P = 0.050, multivariate η2 = 0.07. Energy intake was significantly related to carbohydrate and total potassium intake, Wilks' Λ = 0.38, F(2,86) = 69.61, P = 0.0001, multivariate η2 = 0.62. ANCOVA was conducted for potassium and carbohydrate as a follow-up test to MANCOVA. After Bonferroni correction to the alpha level (P, 0.05/2 variables = 0.025), total potassium intake was significantly different between groups, P = 0.023, with multivariate η2 = 0.058. Table 3 lists the adjusted and unadjusted means of carbohydrate and total potassium intake for each group. Group 2 had higher unadjusted and adjusted means for both variables. Groups were significantly different using Bonferroni correction to the alpha level in energy intake for both carbohydrate and total potassium intake, P = 0.0001.
Adjusted1 (in bold) and unadjusted means ± SD for average intake of carbohydrates and total potassium for two groups2 of subjects
Nutrient
Group 1
Group 2
Carbohydrate (g/day)
206
216
200 ± 36
218 ± 44
Total Potassium (mg/day)3
2527
2831
2478 ± 522
2846 ± 553
1Adjusted for age, height, weight, total hours of activity/week, alcohol consumption and total energy intake 2Groups were stratified to below (Group 1) and above (Group 2) the normal cognitive function score of 27 3significantly different between groups, P = 0.02
Discussion
Our results suggest subjects with lower MMSE scores had overall lower bone mass than subjects with higher MMSE scores. MANCOVA results indicated a moderate effect (η2 = 0.21), or 21% of generalized variability in BMD from the whole body, all sites of hip, spine and forearm is accounted for by differences in MMSE scores after correcting for subjects' age, height, weight, total activity, alcohol consumption, and total energy, calcium and sodium intake. Unadjusted group means were lower in Group 1 (MMSE score <27.9) for all bone sites except the Ward's triangle (Table 2). However, with the ANCOVA follow-up tests to MANCOVA the group differences for individual bone sites diminished. Several of the confounders had a significant influence on various bone sites, which might have accounted for the diminished significance with ANCOVA. Additionally, stronger statistical significance may have been reached in other bone sites with more subjects in each group.
Our results based on unadjusted means, agree with Yaffe et al, who also found older women with poor cognitive function to have lower total hip BMD [2]. Zhang et al, found elderly men and women with moderate verbal memory impairment (a measure of cognitive function) to also have lower femoral neck BMD [1]. Lui at al [28] in a large prospective cohort study of over 4000 women found those with more rapid total hip bone loss were more likely to have a decline in cognitive function than those who had lower rates of bone loss. It therefore appears based on the above evidence lower levels of cognitive function are associated with lower hip BMD. The above researchers did not measure the whole body, lumbar spine or forearm to assess their relationship with cognitive function. In our study, these bone areas were significantly lower in subjects with lower MMSE scores although the statistical significance slightly diminished after the univariate follow-up tests.
We attribute the diminished difference between the groups to several factors. First, the average score on the MMSE in our subjects was 27.9, which is considered normal cognitive function [24], indicating on average our subjects were not cognitively impaired. Group 1 (MMSE score of <27.9) had only 34 subjects, and 10 of them had a score of 27, which is considered normal. Therefore out of 97 subjects only 24 had a MMSE score between 22–26 (mild cognitive impairment), and 73 were in the normal range for cognition. If there were more subjects with mild cognitive impairment, stronger significant group differences may have been reached even after the correction tests. Second, subjects in Group 1 had significantly lower weight than those in Group 2. Lower weight is associated with lower bone mass [27,29]. In this data set weight was a stronger predictor for BMD than MMSE score, and when entered as a co-variate in MANCOVA, statistical significance was diminished in univariate follow-up tests. Third, the MMSE only measures overall cognitive function. Measuring verbal memory specifically, similarly to Zhang et al [1], may have shown a stronger positive relationship with BMD.
Regarding diet, the main finding was that total potassium intake was significantly higher in subjects with normal cognitive function compared to subjects with mild cognitive impairment. When comparing adjusted means, it appears about 300 mg of additional potassium a day may positively benefit cognitive function. MANCOVA results indicated a small effect (η2 = 0.07), or about 7% of generalized variability in both potassium and carbohydrate intake was accounted for by differences in MMSE scores after correcting for subjects age, height, weight, total activity, average daily alcohol consumption and vitamins intake. Adjusted means for both potassium and carbohydrate intake were higher in Group 2 (normal cognitive function). It is possible that both carbohydrate and total potassium intake may have served as markers for a diet higher in fruits and vegetables (excellent sources of both carbohydrates and potassium). Such kind of diet may benefit cognitive function in older women. This research supports the findings of other studies, which found older people who consume more of the above foods to have better cognitive function than people who consume less [9-12]. None of the other nutrients in our study examined by MANCOVA were significant predictors of cognitive function.
Saturated fat was a significant negative predictor of MMSE score in multiple regression analysis, which agrees with the findings of Ortega et al [9]. They found elderly people with scores of 28 or higher on the MMSE to have lower intakes of saturated fat. The Rotterdam study found no significant relationship with any type of fat in the diet and risk of dementia or its subtypes in a prospective evaluation in over 5,000 elderly subjects [19]. These subjects had a normal cognition at baseline and the diet was compared with the incidence of a diagnosis of dementia after a mean follow-up of 6 years. However, the researchers did not measure cognitive function in their subjects so it is not known whether milder cases of cognitive decline such as in our study may be related negatively to saturated fat intake.
There are a few limitations to our study. Although some variables were expressed as cumulative averages collected within 2-year period, this is a cross-sectional evaluation with respect of cognitive functioning that was evaluated only once. Therefore all findings have to be taken within that context and within the limitations of cross-sectional study. Education level was not controlled for, which has been found to affect individual scores on the MMSE [24]. However, this study population did report having at least a high school education and many reported having a college education as well. None of the subjects were currently using estrogen but 14 reported past estrogen use for a short time (up to 1 year). The past-estrogen use was not controlled for, however, in our previous analyses, it did not influence BMD or MMSE scores for those individuals [20,27]. We used two criteria for determining the cut-off levels for MMSE scores on which to stratify subjects into groups. One was the internal criterion (mean for MMSE score of 27.9) and was used for the evaluation of the relationship between BMD and cognition. Another was an external criterion distinguishing normal from cognitively impaired subjects (MMSE score of 27) and was used for the evaluation of the relationship between diet and cognition. We performed all analyses with each cut-off level, and although the results were similar, those based on above described stratification showed slightly higher significance and therefore they are reported.
Our study also has several strengths and is distinct with respect to a comprehensive approach in measuring and evaluating BMD of various skeletal sites, precise identification and inclusion of covariates and analysis of their independent effect on bone and/or cognition, as well as the comprehensive dietary and activity assessment. By using MANCOVA we were able to simultaneously evaluate all measurable skeletal sites and their relationship with MMSE scores and consequently strengthen the overall conclusion. An important issue in this context is possible interaction among age, BMD, cognitive function and dietary intake and physical activity. For example, age alone may be responsible for both the decline in BMD and cognitive function, hence the association between the two. A similar case could be made for dietary intake and physical activity. In order to overcome these effects, all our analyses were performed with taking into account the variables that might interact, be co-linear, and/or independently influence either BMD or cognition. The dietary and activity assessment was a result of a cumulative average over a 2-year period. This provides additional strength to the data and the assurance that both assessments are fairly accurate and representative of subjects' long term intake and activity, not necessarily a case in some other studies.
Conclusions
It appears based on the findings of this study and others, mild degrees of cognitive impairment may be associated with lower levels of BMD specifically in the hip. Lumbar spine and forearm BMD seems to be slightly lower (2–3%) in subjects with lower MMSE scores, based on adjusted means in our study, but clearly more research is needed to determine these relationships. Our findings also indicate a diet containing carbohydrate sources that are also high in potassium such as fruits and vegetables may benefit cognitive function, while saturated fat may influence it negatively. A prudent action (if practical and possible) may be the evaluation of bone status and dietary intake in older women presenting with cognitive impairment, for both potential bone loss and poor diet lacking in good sources of potassium and containing excess amounts of saturated fat.
Competing interests
None declared.
Authors' contributions
RAB carried out the nutritional assessment and analysis, bone and anthropometric measurements and cognitive and physical activity surveys, as well as wrote the draft of the manuscript. JZI conceived and designed the study, obtained the funding, and wrote the final version. Both authors contributed in the statistical analysis and read and approved the final manuscript
Pre-publication history
The pre-publication history for this paper can be accessed here:
Acknowledgements
This work was funded in part by the NRI/USDA 2001-00836, Donaghue Medical Research Foundation DF98-056, University of Connecticut Office for Sponsored Programs and Mission Pharmacal®, San Antonio, TX, USA. The financial sponsors played no role in the design, execution, analysis and interpretation of data, or in writing the study.
Common vaginal infections that manifest in women are usually easily diagnosed. However, Chlamydia infection is often asymptomatic, leading to infertility before it is detected. If it occurs in pregnancy, it could lead to significant neonatal morbidity. It may also play a role with other viral infections for e.g. Human Papilloma Virus in the development of cervical cancer. The objective of this study was to determine the prevalence of Chlamydia infection in women undergoing screening for cervical abnormalities as a part of a research project in primary and secondary care institutions in the United Arab Emirates.
Methods
In this cross sectional study married women attending primary and secondary care participating in a large nationwide cervical abnormalities screening survey were offered Chlamydia testing using a commercially available test kit. This kit uses a rapid immunoassay for the direct detection of Chlamydia trachomatis antigen in endocervical swab specimens. As this study was performed in a traditional Islamic country, unmarried women were excluded from testing, as the management of any positive cases would create legal and social problems. All married women consenting to take part in the study were included irrespective of age.
Results
Of 1039 women approached over a period of eight months 919 (88.5%) agreed to participate. The number of women in the 16 to 19 years was small (0.01%) and 30% were aged over 40 years. The prevalence of Chlamydia infection in this study was 2.6% (95% confidence interval 1.2–3.3%), which was marginally higher in women screened in secondary care (p = 0.05).
Conclusion
This is one of the few reports on the prevalence of Chlamydia infection in women from the Middle East. Due to cultural and social constraints this study excluded a large proportion of women aged less than 19 years of age. Hence no direct comparisons on prevalence could be made with studies from the West, which all included younger women at high risk of Chlamydia. However this study emphasizes the importance of cultural factors while interpreting results of studies from different cultures and communities.
Background
United Arab Emirates (UAE), situated in the Arabian Gulf Peninsula, is a young country formed in 1971. Due to the judicious use of its newfound natural resources, the country has seen tremendous growth in its infrastructure including healthcare provision. According to the World Bank statistics on UAE [1], in the year 2002, the life expectancy at birth was 75.4 years, infant mortality rate 8 per 1000 and childhood immunisation rates reached 95%. The gross national income per capita in 1998 was US$ 19,550. Although economic growth has produced a dramatic improvement in the living conditions of its people, it has brought its own problems. There has been an epidemiological transition in terms of morbidity and mortality in the Emirates. According to the Ministry of Health [2], cardiovascular disorders (23.9%), accidents (16.8%) and cancer (8.3%) accounted for 48.9% of the total deaths during 1998.
Although UAE is a traditional Islamic society, recent globalisation and rapid economic development have had their own impact on the health and lifestyle of the population. There are anecdotal reports of the practice of promiscuity appearing in certain sections of society. Holiday travel has also increased over the last decade in the community. Due to these reasons there is a possibility that sexually transmitted infections (STI) might pose a significant public health threat in the future. There is no population-based data on STI in UAE. The first two authors of this report undertook a population-based study to determine the prevalence of cervical abnormalities in UAE community. As in many previous reports, [3,4] this provided an ideal opportunity to establish the prevalence of Chlamydia trachomatis (CT) infection as this information is currently lacking in UAE.
CT is the most common curable sexually transmitted infection in the UK [5]. The complications of CT infection are distressing and expensive, including infertility, ectopic pregnancy, and chronic pelvic pain [6]. In most countries, selective screening, in high-risk groups, is recommended largely because testing for CT is expensive and requires considerable technical expertise. But it has been argued [7] that in high prevalence populations, selective screening should not be used and all women should be screened.
The prevalence of CT infection varies extensively depending on the health care setting. It is low in primary care but very much higher among women attending STI clinics. Ideally, to establish the prevalence in any community, screening should be undertaken at a population level. This may not always be feasible due to logistic difficulties and the delicate nature of screening for STI in healthy women. The alternative is to screen women attending primary and secondary care self-presenting themselves for a variety of reasons. This study adopted this approach and attempted to establish the prevalence of Chlamydia trachomatis infection in women in Abu Dhabi Emirate, a province of UAE to determine the magnitude of this STI in this community.
Methods
UAE is formed of seven emirates (provinces) and the largest emirate is Abu Dhabi, which covers over 3/4th of the landmass where nearly 40% of the population live. The country attracts a large number of expatriates and their number is equal or in some emirates exceeds the number of UAE citizens. Due to resource constraints, only women attending primary and secondary care in Abu Dhabi Emirate were targeted for the Chlamydia prevalence survey. Both citizens and expatriates were targeted in the survey. UAE terrain is predominantly desert in nature, hence the settlements are concentrated in urban areas and a network of primary health centres serves the population. This survey was a part of a major study on prevalence of cervical abnormalities among women in UAE.
UAE is still largely a traditional Islamic country and premarital sex is apparently very rare due to cultural, legal and societal taboos. In the only state medical school in the country there is still gender segregation and men and women medical students are taught separately. Unmarried women were excluded from testing in this study, as the management of any positive cases would create legal and social problems. Premarital sex is illegal in UAE. "According to the UAE Penal Code, he who indulges in sexual intercourse with a woman to whom he is not married even if it is with her consent, will be sentenced to no less than a year in jail" [8]. Also women attending the clinic seeking treatment specifically for vaginal discharge were excluded from the study as earlier reports show the prevalence of CT to be higher in these women [9,10]. Twenty-seven out of a total of 42 health centres were included in the study. These primary health centres were selected by stratified random sampling, and a ratio of 4:1 was adhered to for urban and rural areas i.e. for every rural centre four urban centres were selected. The strata were based on the mean monthly attendance. Chlamydia swabs were offered opportunistically to women attending gynaecology clinics during 2000 in five maternity hospitals serving the population of Abu Dhabi Emirate and the two largest major government hospitals were targeted and included in the study. An endocervical swab was taken which was analyzed using one of the commercially available kits (Clearview) according to standard protocols described by the manufacturer [11]. This is a rapid method and provides a simple direct detection assay for Chlamydia antigen in endocervical swab specimens, for either single testing or batch use. In this test Chlamydia antigen is extracted from the specimen by heating the swabs at 80°C with the Extraction Buffer. This detected all 15 seroversions of Chlamydia.
All positive results were reported back immediately to the health care providers who took over the responsibility of prescribing treatment and further management. UAE University medical school's research and ethics committees approved the study and administrative approval was obtained from His Excellency the Undersecretary, Ministry of Health and from the Director of Primary care, Government of UAE.
Statistical analysis
Data entry and analysis were performed using the statistical package SPSS version 10 [12]. The quality of the entry process was checked by re-entering a random sample of 10% of cases and running frequencies to check for extreme values. In the analysis, appropriate frequencies were generated. The difference between means of continuous variables was tested by "t" tests and associations between categorical variables by chi-squared and Fishers's exact test as appropriate. The prevalence of Chlamydia infection was analysed by certain sociodemographic and clinical variables.
Results
Of 1039 women approached, 919 (88.5%) agreed to participate and the demographic characteristics of responders and non-responders are presented in Table 1. One in 4 women reported living in a polygamous marriage. Women attending primary care contributed to over 63% of the sample. The overall prevalence of vaginal Chlamydia infection among women was 2.6% and the 95% confidence interval was 1.2 to 3.3%. Prevalence of infection was lower in women of polygamous marriage (1.9%) compared to women of monogamous marriage (2.3%) although this difference was not statistically significant. Prevalence by source of swabs, participants' age, nationality, education and presence of gynaecological complaints are presented in Table 2.
Chlamydia test refusal by socio demographic variables
Characteristics
Chlamydia Swab#
Accepted
Refused N (%)
Chi2
p value
Age
<=Median
487
49 (9.1)
6.5
0.011
>Median
428
71 (14.2)
Education
Illiterate
191
14 (6.8)
2.3
0.133
Educated
652
75 (10.3)
Nationality
U.A.E.
443
37 (7.7)
13.0
0.000
Non-U.A.E
468
82 (14.9)
Gynecological complaint
Yes
714
88 (11.0)
1.1
0.284
No
198
31 (13.5)
# Figures do not always add up to 919 due to missing information
Prevalence of Chlamydia by socio demographic characteristics, source of smear and presence of gynaecological complaints
Characteristic
Prevalence of Chlamydia
Positive N (%)
Negative N (%)
Source of smear*
PHC
9 (1.6)
565 (98.4)
Hospital
12 (3.6)+
326 (96.4)
Age group (years)&
16–19
1 (8.3)
11 (91.7)
20–29
8 (2.7)
290 (97.3)
30–39
9 (2.8)
315 (97.2)
40–49
1 (0.5)
221 (99.5)
50–59
0 (0)
46 (100)
>=60
1 (16.7)
5 (83.3)
Nationality ||
UAE
10 (2.3)
433 (97.7)
Gulf countries
2 (2.7)
72 (97.3)
African Arab
3 (3.4)
86 (96.6)
Middle East
2 (1.7)
116 (98.3)
Indian subcontinent
1 (1.0)
96 (99.0)
Others
3 (3.7)
79 (96.3)
Education level**
Illiterate
10 (2.3)
433 (97.7)
Primary school
2 (2.7)
72 (97.3)
Secondary school
3 (3.4)
86 (96.6)
College & Higher
2 (1.7)
116 (98.3)
Gynaecological Symptom ++
Positive
16 (2.26) &&
691 (97.94)
Negative
4 (2.0)
194 (98.)0
* Source missing in 7 subjects, + Chi2 3.7, p = 0.054, &Age missing in 11 subjects || Nationality missing in 15 subjects, ** Education level missing in 83 subjects ++ Data on symptoms missing in 14 subjects, &&Symptoms include lower abdominal pain and vaginal discharge
Discussion
To our knowledge, there have been few studies on the prevalence of genital Chlamydia infection from the Middle East [13,14]. A hospital-based study [13] from Jordan reported a prevalence of 4.6% among symptomatic patients with urethritis using a cryptic plasmid- based PCR technique. In our study non-responders differed significantly from responders (Table 1). Contrary to a study from Sweden [15] where younger women refused participation more often than older women the reverse was true in our study. This could be due to the different cultural setting in which these studies were performed. In addition, non-UAE nationals were more likely to refuse testing. It is possible that non-nationals were concerned about being labelled as carriers of sexually transmitted infection; hence there was a higher non-response in this group. Over 60% of women had attended primary care and the prevalence of Chlamydia infection was marginally higher in women attending secondary care (Table 2). As we wanted to study the prevalence of CT in apparently healthy women, we excluded women who specifically sought care for vaginal discharge. As these women might have a higher prevalence of CT infection it is possible that their exclusion could have led to some underestimation of prevalence in this population. Although women seeking treatment for vaginal discharge were excluded, it is possible that women attending secondary care had pre-existing genital infections thus the marginally higher prevalence. The prevalence was highest in the sexually highly active age group of 16–29 year, i.e. 16–29:2.89% vs 30+: 1.87%.
In comparing prevalence across different studies, one needs to take in to account the background risk of included women and the testing technique employed, as this would influence the prevalence rates. One consideration, which confounds the results of this survey being compared with studies from other countries, is the age group of women included in this study. This study only included married women and excluded unmarried women aged less than 19 years and over represented women aged 40 years and over. It is likely that the population being studied is at low risk of Chlamydia. However, in UAE society premarital sex is totally forbidden and any such activity would lead to punishment by law [8] and other sanctions by society. Although no figures are available on premarital sex from this community anecdotal evidence suggests that even if it exists it is likely to be very rare. In UAE, clinicians do not often encounter young unmarried women with STI (Dr. Neriger, Consultant in Infectious disease, Tawam hospital, Al-Ain, UAE, personal communication).
The Chlamydia prevalence in this study of 2.6% was similar to studies performed in primary care in Sweden (2.7%) [15] and Greece (3.2%) [16] but was much lower than studies performed in other populations; Pennsylvania family planning clinics (11.2%) [17], antenatal clinics in Bangkok (8.6%) [18], pregnant women attending secondary care in India (17%) [19], and in a US college gynaecology clinic (8.1%) [20]. As described above the lower prevalence in our study could be due to the large number of low risk women included and exclusion of younger women who could potentially be at a higher risk of STI.
The prevalence of polygamy was much higher in UAE (26%) as compared to neighbouring Kuwait [21] (12.5%). It is possible that the lower prevalence of infection in women of polygamous marriages could be due to their husbands having a much lower prevalence of sexually transmitted infections because they live in a stable relationship within the marriage setting. Prevalence of infection did not vary by the presence of any gynaecological complaints.
In this survey, Clearview test was employed to determine the prevalence of infection. This test is easier to perform with minimum expertise as test kits can be used off the shelf. The performance of the test varies depending on the gold standard used to compare the results. The sensitivity of the test was only 50% when amplification technologies, the AMP CT (Gen-Probe) and LCx (Abbott) assays were used as a comparison [22]. However, the sensitivity rose to 76% when a DAKO IDEIA Chlamydia test was used as the gold standard [23]. The advantages and disadvantages of using Clearview for rapid detection have been discussed elsewhere [24]. According to this decision analysis in one off testing situations, rapid tests contribute significantly to the detection and treatment of Chlamydia infections among women. As ours was a survey of women attending health care institutions including primary health care centres and as the resources, including expertise were limited, we used the Clearview to obtain results at point of care. In low prevalence situations, tests with low sensitivity such as Clearview may not be ideal as it has been reported that even very good tests have low positive predictive value while testing low prevalence population [25]. Hence, it is possible that the prevalence in this community might be higher and this needs further studies using other tests for Chlamydia, which have a higher sensitivity.
The US preventive task force [26] has recommended that all sexually active women aged 25 and younger and other asymptomatic women at increased risk of infection should be routinely screened. Due to the selective nature of women included in our study and the type of test employed it is hard to arrive at any definite conclusions on the prevalence of genital Chlamydia based on this study and hence to put forward any firm recommendations. However it is worth mentioning here that in our population, the prevalence of Chlamydia is much lower than other populations even though 40% were recruited from secondary care.
Conclusions
In this study performed among a conservative Middle Eastern society the acceptance of CT testing was high (89%). The prevalence of CT infection in this population was 2.6%. The prevalence was marginally higher in women attending secondary care although it was not different if women had any gynaecological complaints or not. Due to the exclusion of young women and the type of test employed no firm conclusions can be drawn on the prevalence in the population. Further studies are warranted including younger women using more sensitive tests to screen for Chlamydia. However it remains to be seen how researchers in future could overcome the legal, social and logistics constraints we faced while planning to screen young unmarried women in this traditional Islamic community. Till such studies become available we propose selective Chlamydia testing of the population based on clinical grounds. Our study also emphasises the importance of cultural factors while interpreting results of studies from different cultures and communities. Further epidemiological studies are warranted in this population to determine the correlates of Chlamydia positivity to plan preventive activities.
Competing interests
None declared.
Authors' contributions
SGA is the Principle Investigator and grant holder and was involved in planning, coordinating the research, sample collection and handling, supervising data entry and analysis, writing and editing of the manuscript. PB is the Principle co-investigator and was involved in planning the epidemiological aspects of the research, coordinating and supervising data entry and analysis, and writing and editing of the manuscript. NO supervised the sample and data collection and entering data into database and helped with the analysis of data and editing the manuscript. SAK analysed the samples and edited the manuscript. SM was involved in the planning and coordination of the research at the primary health care level, and supervised data entry and analysis and edited the manuscript. IS was involved in planning and coordination of the research at the secondary care level, data analysis, and editing the manuscript.
Pre-publication history
The pre-publication history for this paper can be accessed here:
Acknowledgements
This study was funded by grant No: NP/99/25 from the Research Grants Committee of the Faculty of Medicine & Health Sciences, UAE University, Al-Ain, UAE.
The authors would like to thank all the three reviewers for their constructive comments and useful suggestions and Professor F. Branicki, Chair, Department of Surgery, Faculty of Medicine & Health Sciences, UAE University for his comments on the revised manuscript.
United Arab Emirates data profileMinistry of Health, Statistics, UAE, StatisticsMeijerCJCalameJJde WindtEJRisseEKBlekerOPKenemansPQuintWGMeddensMJPrevalence of Chlamydia trachomatis infection in a population of asymptomatic women in a screening program for cervical cancer19898127302498093SmithJRMurdochJCarringtonDFrewCEDougallAJMacKinnonHBaillieDByfordDMForrestCADavisJAPrevalence of Chlamydia trachomatis infection in women having cervical smear tests19913028241995120Chlamydia trachomatisSummary and conclusions of CMO's expert advisory group1998Clinical guidelines and standards for genital chlamydia infection1997WeinstockHSBolanGAKohnRBalladaresCBackAOlivaGChlamydia trachomatis infection in women: a need for universal screening in high prevalence populations?19921354171736659Gulf News on line edition. Ask the lawMarch 15, 2002WilliamsHTabriziSNLeeWKovacsGTGarlandSAdolescence and other risk factors for Chlamydia trachomatis genitourinary infection in women in Melbourne, Australia2003793141257661110.1136/sti.79.1.31VishwanathSTalwarVPrasadRCoyajiKEliasCJde ZoysaISyndromic management of vaginal discharge among women in a reproductive health clinic in India20007630361102688910.1136/sti.76.4.303CLEARVIEW for health care providersSPSS Base for WindowsAwwadZMAl-AmaratAAShehabiAAPrevalence of genital chlamydial infection in symptomatic and asymptomatic Jordanian patients2003720691456322410.1016/S1201-9712(03)90053-1MawajdehSMAl-QutobRSchmidtAMeasuring reproductive morbidity: a community-based approach, Jordan2003246354914627210BrannstromMJosefssonGBCederbergALiljestrandJPrevalence of genital Chlamydia trachomatis infection among women in a Swedish primary health care area1992244161589724NelsonMEPrevalence of Chlamydia trachomatis infection among women in a multiphysician primary care practice199282983021419130WinterLGoldyASBaerCPrevalence and epidemiological correlates of Chalamydia trachomatis in rural and urban populations19901730362305334NiamsanitSNunthapisudPLimpongasanurakSPrevalence of Chlamydia trachomatis among women attending an antenatal clinic in Bangkok199819609133266370PaulVKSinghMGuptaUBucksheeKBhargavaVLTakkarDNagVLBhanMKDeorariAKChlamydia Trachomatis infection among pregnant women: prevalence and parental importance199912114SwinkerMLYoungSACleavengerRLNeelyJLPalmerJEPrevalence of Chlamydia trachomatis cervical infection in a college gynecology clinic: relationship to other infections and clinical features19881513363067388ChalebyKWomen of polygamous marriage in an inpatient psychiatric service in Kuwait19851735683965613LauderdaleTLLandersLThorneycroftIChapinKComparison of the PACE 2 assay, two amplification assays, and Clearview EIA for detection of Chlamydia trachomatis in female endocervical and urine specimens1999372223910364589WoolleyPDPumphreyJApplication of 'Clearview Chlamydia' for the rapid detection of cervical chlamydial antigen199782578914715910.1258/0956462971919868GiftTLPateMSHookEW3rdKasslerWJThe rapid test paradox: when fewer cases detected lead to more cases treated: a decision analysis of tests for Chlamydia trachomatis1999262324010225593GrimesDASchulzKFUses and abuses of screening tests200235988141189730410.1016/S0140-6736(02)07948-5Screening Chlamydial infection. US Preventive Services Task Forceoai%3Apubmedcentral.nih.gov%3A459232!!!pmc!bmcwh