Obstetrics and Gynecology Clinics
Volume 24 • Number 2 • June 1997
Copyright © 1997 W. B. Saunders Company

 

Dr. Sinan DOĞANTÜRK  

Ankara                                                                              235

ENDOMETRIOSIS

EPIDEMIOLOGY OF ENDOMETRIOSIS

Brenda Eskenazi PhD

Marcella L. Warner PhD

Endometriosis is a puzzling disease with little known about its true prevalence, its distribution in the population, or its risk factors. It is thought to be a relatively common disease, however, with an estimated prevalence among women of reproductive age as high as 10%. ^[90] Given this relatively high prevalence and that women with the disease may have severe pelvic pain and infertility, endometriosis poses a significant public health problem. It is the third leading cause of gynecologic hospitalizations in the United States and a leading cause of hysterectomy. ^[84] In a cost analysis prepared by the US Army, endometriosis cost $2.6 million over a 6-year period for time lost to service because of surgical hospitalization alone. ^[10] This is an underestimate of the cost in the general population because women generally are ineligible for active duty if they have severe symptomatic endometriosis, and time lost from work for unhospitalized sick days could not be accounted for. In a recent analysis ^[45] of women who participated in the US Health Interview Survey, 50% of those reporting having endometriosis had stayed in bed all day because of their condition at some time in the past 12 months, with the average number of bed days being 17.8. Thus, this disease is costly in its effects on women's quality of life, the expense of medical care, and its economic impact in the workplace.

This article summarizes present knowledge about the epidemiology of endometriosis. In this article well-designed and well-controlled studies are emphasized.

From the School of Public Health, University of California, Berkeley, California

This work was supported by Grant Nos. R824761-01 from the Environmental Protection Agency, R01 ES07171-01 from the National Institutes of Health, and EA-M1977 from the Endometriosis Association.

Address reprint requests to
Brenda Eskenazi, PhD
School of Public Health
University of California
140 Warren Hall
Berkeley, CA 94720-7360

236

PREVALENCE OF ENDOMETRIOSIS

Large Studies of Hospitalized or Surgical Cohorts

Endometriosis can be definitively diagnosed only during the course of pelvic surgery, usually laparoscopy or laparotomy. Therefore, most prevalence estimates have been made on the basis of such surgical populations and are therefore highly selective. Table 1 presents the results of a few large studies in which hospital discharge data or surgical records (usually from laparoscopy or laparotomy) were reviewed over a specified period. The earliest of these studies reported on the prevalence of endometriosis in the United States derived from diagnoses of patients discharged from short stay nonfederal hospitals in 1980. ^[67] Among women 15 to 44 years old, 6.3% of the first diagnoses and 6.9% of all diagnoses for hospitalizations for genitourinary problems were because of endometriosis. In a more recent report of all gynecologic diagnoses recorded on over 5 million hospital discharge summaries between 1988 and 1990 from the US National Hospital Discharge Survey, ^[84] 11.2% of the first diagnosis was because of endometriosis. Two large studies reported the prevalence in women who underwent gynecologic surgeries and reported similar rates. Boling et al ^[10] reported that 6.2% of US Army personnel or dependents who had undergone laparoscopy or laparotomy between 1980 and 1985 were diagnosed with endometriosis. Wheeler ^[90] reviewed the surgical and billing records for 21 gynecologic practices in Houston and found that 10.3% of women undergoing all types of gynecologic surgeries, including hysterectomy, were reported to have endometriosis. This overall prevalence included a proportion less than 2% in the subgroup having tubal ligation and 30% in the subgroup having a laparoscopy for other reasons. Wheeler concluded that the prevalence of endometriosis in the general population was close to 10% based on finding 71 cases of histologically confirmed endometriosis among the 858 women who had a vaginal hysterectomy (8.3%) for preoperative diagnoses excluding endometriosis, pelvic pain, and adhesions. Although this estimate may be appropriate, however, it is based on an older patient series in which the entire peritoneal cavity was not visualized.

Table 2 summarizes the average and range in prevalences reported for women undergoing laparoscopy by indication for surgery. In this table, the women who have received laparoscopy for a specified indication and those who were found to have endometriosis presenting with that indication are summed across several studies. We have not included studies in which the specific indication for surgery was unclear or in which surgery was indicated for more than one reason.

Endometriosis was diagnosed in approximately 25% of women (range is 4.5% to 82.0% across studies reviewed) who had a laparoscopy because of pelvic pain, compared with 20% (range 2.1% to 78.0%) of women who presented with infertility. To obtain a closer estimate of the prevalence in the general population, studies of women undergoing

237

tubal ligation have been conducted. It should be noted that these women are probably asymptomatic with proven fertility. Endometriosis was reported in 4.1% of over 10,000 women who received tubal ligation. In those studies in which the staging of the disease was indicated (Acosta, American Fertility Society (AFS), or r-AFS classifications), however, more than 90% of the women with endometriosis on tubal ligation were found to have minimal or mild stage only, compared with approximately two thirds of those presenting with pelvic pain or infertility. Thus, although an incidental diagnosis during sterilization is common (particularly in some studies), the disease tends to be much less progressed than in women who present with symptomatology.

Because of issues of selection bias, the use of surgical patients to estimate disease prevalence in the general population is problematic. There are numerous social, cultural, and medical factors that contribute to or limit a woman presenting with symptomatology or being offered surgery. The degree of pain necessary before surgery is offered may differ by center or by physician. Prevalence in surgical populations may depend on the experience of the surgeon in identifying endometriosis and on referral patterns, especially to centers well known for endometriosis or for evaluation of infertility or pelvic pain. Clearly, estimates based on a population undergoing laparoscopy because of symptomatology will probably exaggerate the general prevalence, whereas estimates derived from asymptomatic women undergoing tubal ligation may be artificially low. Furthermore, the high percentage of mild endometriosis in asymptomatic women undergoing tubal ligation calls into question the definition of the disease; ^{[20] [47]} that is, does the presence of endometrial implants, which is not associated with symptomatology, constitute "disease"?

Requiring symptomatology for disease diagnosis may be justified from a clinical perspective but poses a problem for epidemiologic investigations. For example, if report of pain is required for diagnosis, then certain populations because of cultural and social norms may be more likely to report pain and thus to have surgery and to be diagnosed. Similarly, only women who have tried to reproduce and failed can be considered subfertile, leaving those who have never tried less likely to be diagnosed. In addition, symptomatology has not been well correlated with disease occurrence or severity. For example, among four studies of women referred for laparoscopy because of infertility, ^{[29] [33] [57] [61]} only two of four studies found that women with endometriosis compared with those without were more likely to report dysmenorrhea, and none out of two studies ^{[33] [57]} reported that women with endometriosis were more likely to report pelvic pain or dyspareunia. Furthermore, women with severe symptomatology have been found to be more likely to have severe disease (as determined by AFS stage of disease) in some studies, ^[54] but not in others. ^[30] Others have suggested that symptoms may be related more to the site of the implants (e.g., the association of dyspareunia in women with vaginal lesions), ^[87] the depth of the implants, ^[18] and the number of implants. ^[74] Nevertheless, these studies suggest a biased estimate

238

 

239

240

*2, 9, 18, 26, 35, 38, 39, 46, 48, 49, 51, 55, 72, 85, 86.
2, 3, 7-9, 11, 15, 18, 19, 25-29, 31-33, 38, 39, 46, 49, 51, 52, 57, 61, 66, 79-82, 89, 91.
24, 25, 44, 48, 50, 51, 62-64, 78, 82, 83, 90.

of the prevalence of disease would result if symptomatology were used as a criterion for disease.

Estimates in the General Population

To date, only three studies have attempted to estimate the prevalence of endometriosis in the general population (see Table 1) . The study by Vessey et al ^[88] studied a highly selective segment of the general population: 17,032 women from 17 family planning clinics who were 25 to 39 years old, white, married, and British and had used oral contraceptives, IUD, or diaphragm for at least 5 months. The investigators reviewed medical records up until the end of 1990 (the women then ranged from 41 to 61 years old) for a diagnosis of endometriosis on laparoscopy or laparotomy. A total of 313 cases (1.8%) were identified, including 142 in which endometriosis was the first diagnosis on the surgery report.

The study by Houston et al ^[42] was probably the study to come closest to identifying the frequency of the disease in the general population. This study reviewed the medical records of all white residents of Rochester, Minnesota, between 1970 and 1979 to determine the newly diagnosed cases of endometriosis during this period. Because of biases inherent in estimating incidence from only women who had undergone surgery, they extended their criteria of disease to include symptomatology or positive signs noted on physical examination, such as nodularity or pain. They constructed four diagnostic tiers: (1) histologically confirmed disease, (2) surgically visualized, (3) clinically probable based on report of pain and a positive examination, and (4) clinically possible based only on positive examination. Using census data from 1970 and 1980, they estimated the total and age-specific denominators for the 10-year period. The overall incidence of pelvic endometriosis ranged between 108.8 and 246.9 cases per 100,000 person-years depending on the inclusiveness of the diagnosis. This incidence did not include the estimated

241

loss of 131 cases because of unreviewed medical records. The authors concluded that if the average duration of the disease is 10 years, then the prevalence (incidence × average duration of disease) of endometriosis in the population would have been about 2.5% to 3.3% (including the estimated missed cases). This may be an underestimate of the prevalence, however, in that the average age of diagnosis of endometriosis is in the mid 20s and for most women the disease does not end until menopause: therefore, the duration of disease may be substantially longer than 10 years. For example, if the average duration was 25 years, the prevalence of endometriosis would be 6.2% to 8.2%. Even this prevalence could be an underestimate given that some women may not be examined or report symptoms to their physician.

A recent study ^[45] focused on women's self-report of disease and symptoms rather than on the diagnosis being gathered from medical records. This study included women 18 to 50 years old interviewed between 1984 and 1992 as part of the US Health Interview Survey. Households were sampled by a stratified, multistage probability design. A random subsample of women were then asked whether they had one or more gynecologic conditions presented in a list in the past 12 months. Of 31,617 women surveyed, 211 women reported having received a diagnosis of endometriosis and 1666 reported having menstrual disorders. The estimated annual prevalence rate was 6.9 per 1000 women for endometriosis and 53.0 per 1000 women for menstrual disorders. The authors concluded that this prevalence would probably be an underestimate because some women may find answering questions about gynecologic illness too embarrassing. Also, some women may never have consulted a physician for symptoms or have never received a diagnosis, or if they have, they may not remember the diagnosis or may confuse it with another. It seems likely, therefore, that a proportion of those with unspecified menstrual disorders actually had endometriosis.

All three studies that attempted to estimate prevalence in the general population have in common the limitation that the lack of disease cannot be confirmed in those who have not been examined, undergone laparoscopy, or reported symptomatology.

Caveats in Determining Prevalence

Differences across studies in prevalence depend only partly on whether the population studied is limited to surgical cases or includes the general population. The prevalence may vary because gynecologists may differ in their definition of the disease. This difference in definition may also vary within or across facilities, from country to country, or from decade to decade. Nevertheless, very few of the studies reviewed present their criteria for disease. Some parameters that will alter the prevalence are whether

242

1. The case definition includes adenomyosis.
2. The diagnosis is based on visualization or on histology.
3. Some or all of the implants are excised and examined microscopically.
4. Only typical black powder burn lesions are considered or atypical clear, white, or red lesions are also considered.
5. The cases are defined by laparoscopy, laparotomy, or abdominal or vaginal hysterectomy because the type of surgery will affect the ability of the surgeon to visualize lesions and the amount of material that can be sent to histology.
6. The cases are identified by surgery alone or by symptoms, by signs on pelvic examinations, or by ultrasound. (It is possible that even within the same study, if there was more than one surgeon, more than one criterion for diagnosis may have been used.)

The use of a common definition of disease across studies is imperative because differences in disease definition will result not only in variation in prevalence across studies, but also in which risk factors are identified.

RISK FACTORS

Numerous studies related to the epidemiology of endometriosis have been published. Various risk factors for the disease examined in these studies include (1) sociodemographic characteristics ^{[21] [22] [68]} such as age, ^{[2] [3] [4] [5] [13] [16] [17] [18] [22] [23] [24] [27] [35] [38] [40] [42] [44] [51] [54] [55] [56] [61] [65] [70] [73] [75] [76] [78] [80] [84] [85] [88]} race, ^{[3] [5] [13] [16] [17] [32] [39] [44] [55] [60] [71] [73] [78] [84]} socioeconomic status, ^{[1] [4] [5] [7] [44] [56] [68] [73] [78]} education, ^{[13] [55] [59] [70] [71] [73] [75] [78]} and marital status ^{[13] [22] [61] [68] [70] [71] [73] [78]} ; (2) reproductive history, including menstrual cycle characteristics, ^{[13] [14] [19] [21] [22] [32] [34] [50] [51] [53] [56] [59] [62] [64] [71] [77] [78] [80]} and pregnancy-related factors ^{[6] [11] [12] [21] [26] [52] [61] [62] [63] [64] [78] [88] [89]} such as parity, ^{[2] [5] [13] [23] [24] [38] [39] [44] [56] [59] [62] [68] [70] [71] [76] [78] [80] [88]} gravidity, ^{[22] [23] [24] [44] [56] [59] [68] [78]} length of preceding infertility, ^{[7] [25] [27] [56] [80] [82]} spontaneous or induced abortions, ^{[13] [44] [56] [62] [63] [68] [71]} and age at first pregnancy or birth ^{[13] [51] [62] [63] [71] [78]} ; (3) contraception use ^{[7] [21] [22] [33] [43] [44] [53] [56] [62] [63] [68] [69] [70] [78] [82] [88]} ; (4) personal habits ^{[21] [22]} such as smoking, ^{[7] [19] [22] [56] [75] [78]} alcohol, ^{[36] [73]} caffeine intake, ^[37] exercise, ^{[7] [19]} illicit drug use, ^[7] and number of sexual partners ^{[7] [19]} ; and (5) body characteristics ^[58] such as height, ^{[7] [19] [23]} weight, ^{[7] [19] [23]} body mass, ^{[22] [78] [80]} and hair color. ^[91] Findings reported across studies have been inconsistent, sometimes ranging from positive to protective to no association for the same risk factor. Part of the inconsistency is probably due to the fact that most of these studies are not analytic epidemiologic studies and they may have used different criteria for disease diagnosis or different populations (referred for infertility, pain, etc). Many are either case series, do not have a well-defined comparison group, or have not controlled for potential confounding factors.

For the purposes of this article, the discussion is limited to those studies that are well-designed analytic epidemiologic studies. Criteria considered for inclusion in the discussion include (1) study design is cohort or case-control; (2) cases have a laparoscopy-confirmed diagnosis

243

of endometriosis; (3) criteria for selection of controls are clearly defined; and (4) analysis considers adjustment for potential confounding factors. In this way, it is possible to highlight those risk factors for which there has been some consistency of findings.

Of almost 100 studies reviewed, we have identified a total of six well-designed analytic studies, one cohort, and five case-control studies, for detailed discussion. Within each of these study populations, several risk factors have been examined. The overall findings are summarized in Table 3 . A more detailed discussion by risk factor category examined is presented here.

Sociodemographic Characteristics

The relation of endometriosis and various sociodemographic characteristics, including age, race, and socioeconomic status, has been investigated in all of these studies. Age is the only factor, however, for which a consistent relation has been observed in at least two studies; many other studies could not examine the effects of age because the control group was often frequency matched to cases on age. In general, a positive relationship between endometriosis and age has been observed that is limited to the menstrual or reproductive years. In the cohort study by Vessey et al, ^[88] age group was positively associated with endometriosis, peaking among women aged 40 to 44 years. Compared with women aged 25 to 29 years, the risk of endometriosis was increased among women aged 30 to 34 years (RR=2.1 95%CI 0.6-11.4), 35 to 39 years (RR=4.5 95%CI 1.5-22.9), 40 to 44 years (RR=6.1 95%CI 2.0-30.6). In the case-control study among women undergoing laparoscopic sterilization, Sangi-Haghpeykar and Poindexter ^[78] reported the risk of endometriosis increased with age but was significant only for women older than 36 years (OR=2.0, 95%CI 1.1-3.7), compared with women younger than 25 years.

The association between endometriosis and race has been examined in only one of the six studies. The other studies included only one race: caucasians. In the case-control study among women undergoing laparoscopic sterilization, Sangi-Haghpeykar and Poindexter ^[78] reported a significant association was found for Asian race (OR=8.6, 95%CI 1.4-20.7), but the total number of Asians in the study was only six.

The association between endometriosis and socioeconomic status has been examined by considering several different variables, including education, social class, occupation, and hospital admission class. Although all six studies reviewed examined this relation, only one study reported a positive association. Obermeyer et al ^[68] conducted a case-control study by review of medical records of all patients undergoing laparoscopy at the American University of Beirut Medical Center between 1979 and 1981. A total of 170 laparoscopically confirmed cases were identified and compared with 170 age-matched controls (± 2 years) who were randomly selected from a list of gynecology patients

244

* Dervied from same study population.
Derived from same study population.
Derived from same study population.

245

246

247

who had undergone laparoscopy during the same period. All data were obtained from patient charts and no interviews were conducted; therefore the risk factors that could be investigated were limited to those in the record. After conducting a matched analysis, a significant positive association was found between endometriosis and the highest social class as measured by first-class admission (OR=2.8, 95%CI 1.7-4.4) compared with lower class admissions. No significant differences between cases and controls were found, however, for other measures of social class, such as occupational status or nationality. The authors suggest this may reflect that higher class patients may be receiving laparoscopy for chronic conditions such as endometriosis, whereas lower class patients may receive laparoscopy for more acute conditions. Unfortunately, the authors do not provide data on the reason the controls underwent laparoscopy.

Reproductive Health Factors

The relation of endometriosis and reproductive health factors, including menstrual cycle characteristics and pregnancy-related variables, has been investigated in study populations undergoing laparoscopy for gynecologic conditions including pain and infertility ^{[19] [21] [22] [71]} and tubal sterilization. ^[78] In general, the risk of endometriosis appears to be increased for reproductive health factors that relate to increasing exposure to menstruation (i.e., shorter cycle length, longer duration of flow, reduced parity).

As part of a large multicenter case-control study of primary infertility, Cramer et al ^[19] examined menstrual characteristics and other factors in 286 white women with primary infertility because of endometriosis and 3794 white women who were admitted for delivery of a live-born and who were frequency matched to cases on year of birth, race, and pay status at seven collaborating hospitals from 1981 to 1983. Information on demographic variables, reproductive history, medical and surgery history, contraceptive use, and personal habits was obtained by personal interview. After adjusting for age, collaborating center, education, religion, number of sexual partners, and time between menarche and trying to become pregnant, cases were more likely than controls to have shorter cycle lengths (27 days versus 28-34 days) (O.R.=2.1, 95%C.I. 1.5-2.9) and longer duration of flow (>1 week versus 1 week) (O.R.=2.4, 95%C.I. 1.4-4.0). There was a trend toward increasing risk for endometriosis associated with increasing menstrual pain. Relative to no menstrual pain, the adjusted odds ratios for mild, moderate, and severe pain were 1.7 (95%CI 1.1-2.6), 3.4 (95%CI 2.2-5.2), and 6.7(95%CI 4.4-10.2). Consistent with a pattern of longer, heavier menstrual flow, cases were also more likely to require both napkins and tampons for sanitary protection (OR=1.4, 95%CI 0.9-2.0). No association was found, however, with hygienic practices such as douching after menses (OR=0.9, 95%CI 0.6-1.4). The authors concluded that their findings support the implantation

248

theory of endometriosis as the risk of disease was positively related to menstrual factors that predispose women to greater pelvic contamination by means of retrograde menstruation. One limitation of this study is that controls did not undergo laparoscopy and thus some could have undiagnosed endometriosis. The impact of this, if any, would be to underestimate the risk. In addition, although endometriosis was the primary cause of infertility for the 287 cases, almost 25% of the women had an additional disorder (e.g., ovulatory factor, cervical factor, tubal disease) that may have affected the risk estimates for endometriosis. Unfortunately, the authors did not exclude these women from the analysis so that it is difficult to comment on their impact. Finally, the findings of this study can be generalized only to women with endometriosis manifesting as primary infertility, and not necessarily to women presenting with other signs of endometriosis such as pain.

As part of a case-control study of reproductive factors and risk of endometriosis in women aged 19 to 45 years, Darrow et al ^[22] also examined menstrual cycle characteristics that may affect a woman's exposure to menstruation. From the Reproductive and Endometriosis Center of Western New York, 104 laparoscopically confirmed cases of endometriosis were identified between January 1986 and February 1987 and compared with two groups of controls: (1) 100 friends of cases and (2) 98 patients from the same medical practice with conditions other than endometriosis (42% laparoscopically confirmed). Controls were frequency matched to cases by age. After adjusting for pregnancy, religion, and body mass index, cases under the age of 30 years compared with friend controls of the same age were significantly more likely to have menstrual flow 6 days per month (OR=2.5, 95%CI 1.1-5.9), heavy flow (OR=2.5, 95%CI 1.1-6.3), severe cramps (OR=2.5, 95%CI 1.2-6.0), increasing symptoms (OR=6.6, 95%CI 2.6-16.5), and tampons use for more than 14 years (OR=3.6, 95%CI 1.0-13.5). Adjusted odds ratios for these same variables were elevated but not significantly among women aged 30 years or more. Compared with medical controls under the age of 30 years, cases of the same age were significantly more likely to report menstrual flow 6 days per month (OR=2.3, 95%CI 1.1-5.1), severe cramps (OR=2.3, 95%CI 1.0-5.3), and increasing symptoms (OR=8.3, 95%CI 3.3-21.0). Among those over 30 years of age, there were no differences in menstrual cycle characteristics between cases and controls. Thus, consistent with the findings of Cramer et al, ^[19] the results of this study support the theory that endometriosis is associated with increased exposure to menstruation. The odds ratios for exposure to menstruation were consistently higher using friend controls than medical controls and may be due to the fact that the medical controls had a history of gynecologic problems (e.g., infertility, menstrual problems, tubal abnormalities, history of pelvic inflammatory disease [PID]). If these gynecologic problems share risk factors with endometriosis, the observed associations for medical controls would be expected to underestimate the risk.

Taking a different approach, in a second publication, Darrow et

249

al ^[21] investigated the hypothesis that women who "voluntarily" delay childbirth are at increased risk for endometriosis, perhaps by being exposed to a longer duration of uninterrupted menstrual cycles. Darrow et al ^[21] examined the relationship of endometriosis with sexual and reproductive characteristics, including sexual activity, contraception, age at marriage, and age at first pregnancy. No differences were found between cases and either control group for use of reliable contraception, use of contraception before first pregnancy, age at first intercourse, frequency of intercourse, decisions to delay pregnancy, age at first pregnancy, or body size. Cases were slightly older than controls when they first used contraception (19.8 versus 18.6 years), first began regular intercourse (20.9 versus 19.5 years), and age at first marriage (22.8 versus 21.7 years), but all of the ages were well within the range of a woman's peak fertility. Thus, the authors concluded that the findings did not support the simplistic assumption that endometriosis patients are career women who voluntarily delay pregnancy and thereby leave themselves exposed to more menstrual cycles before their first pregnancy.

Parazzini et al ^{[13] [71]} examined the relationship of reproductive and menstrual characteristics and endometriosis from a case-control study conducted at three obstetrics and gynecology clinics in Lombardy, in northern Italy. The case group included 376 women, age 21 to 48 years, with laparoscopically confirmed endometriosis. The control group included 522 women of similar age who were admitted for acute conditions to the same hospital. After adjusting for age, cases compared to controls were less likely to have irregular menstrual cycles (OR=0.4 95%CI 0.2-0.8), higher parity (OR=0.2 for 1 birth, OR=0.1 for 2 births), and history of induced abortion (OR=0.4, 95%CI 0.2-0.7). No relation was reported for spontaneous abortion, age at first birth, or age of menarche. The authors reported that the findings were similar across the different sites of disease in the pelvis and indications for surgery. Limitations of this study include the fact that controls could have undiagnosed disease as none underwent laparoscopy; nonetheless, the results of this study are consistent with findings of others ^{[19] [22]} that the risk of endometriosis is associated with factors related to increased exposure to menstruation.

Most recently, Sangi-Haghpeykar and Poindexter ^[78] conducted a case-control study to examine the relation between risk factors that included menstrual characteristics, contraception, and reproductive history and endometriosis among women who were undergoing laparoscopic sterilization at Baylor College of Medicine in Houston from September 1987 through December 1993. Of 3384 women, 126 cases of endometriosis were detected and compared with a control group of 504 randomly selected laparoscopic tubal ligation patients. All women, cases and controls, were multiparous. After adjusting for age, race, and the other variables in the model, cases were more likely to have longer cycle lengths (30 versus <30 days) (OR=1.8, 95%CI 1.1-2.7) and long duration of uninterrupted menstrual cycles of 6 years (measured as time since last delivery; OR=2.9, 95%CI 1.3-6.4). There was a trend toward decreasing risk of disease with increasing parity. Compared with

250

women who had two to five live births, the odds ratio was 2.2 for one live birth (95%CI 1.1-4.3). The relation of the above factors to endometriosis varied by location and depth of the lesions. The authors suggested that the inconsistency of findings of this study in relation to others (i.e., increased risk for longer cycle length in this study versus short cycle length in previous studies) may be related to the mild nature of endometriosis in the study population (91.3% Stage I disease). This study included tubal ligation patients, who may be more fertile with fewer gynecologic and reproductive problems than women seeking nonsurgical methods of contraception. The findings may not be applicable to women who have pain, infertility, or severe disease, but rather generalizable to multiparous women.

Contraception Use

The association of endometriosis with contraception is rather complex and has been investigated in a population-based cohort ^[88] as well as hospital-based case-control studies. ^{[69] [71] [78]} A description of each of these studies is presented below.

In the only population-based cohort study conducted to date, Vessey et al ^[88] investigated the relationship between contraception use and endometriosis using the Oxford Family Planning Association Contraceptive Study data. The cases were all laparoscopically confirmed; however, there may have been some disease misclassification in that some who did not receive laparoscopy may have had endometriosis. The authors reported the results of contraceptive use after controlling for age and parity. Compared with never-users, the risk of endometriosis was reduced in current or recent (0 to 12 months previously) oral contraceptive (OC) users (RR=0.4, 95%CI 0.2-0.7), but higher in former (25-48 months previously) OC users (RR=1.8, 95%CI 1.0-3.1). Endometriosis was not associated, however, with duration of OC use. A similar pattern was observed in intrauterine device (IUD) users. Compared with never-users, the risk of endometriosis was reduced in current or recent (0 to 12 months previously) IUD users (RR=0.4, 95%CI 0.2-0.7), but higher in former (49-72 months previously) IUD users (RR=1.4, 95%CI 0.4-3.2). No association was found between diaphragm use and endometriosis.

The results of this study suggest that the risk of endometriosis may be reduced during current and recent OC use but that the risk subsequently increases after use is discontinued. The authors suggested that the finding is not real but due entirely to bias. That is, current OC use masks the symptoms of the disease, which ultimately emerge shortly after OC use is stopped. In women who develop endometriosis while using an IUD, the disease goes unrecognized because the symptoms are similar to symptoms associated with IUD use itself and these women are treated by removal of the device. If this were the case, for both IUD and OC use a risk for disease would be observed only with former users.

To evaluate further the association of contraceptive use and endometriosis,

251

Parazzini et al ^{[69] [70]} analyzed data from the northern Italy case-control study. After adjusting for age, education, marital status, and parity, the odds ratio was 1.6 (95%CI 1.2-2.2) for ever-OC use compared to never-OC use. The association was limited, however, to former OC users (OR=1.9, 95%CI 1.4-2.6). No association was found for current OC use (OR=0.9, 95%CI 0.5-1.9). Consistent with Vessey et al, ^[88] no statistically significant relation emerged between endometriosis and duration of OC use, although the point estimates decreased with number of years of use; compared with never use, the odds ratios for duration of use less than 2 years, 2 to 4 years, and 5 or more years were 2.0 (95%CI 1.3-3.0), 1.5 (95%CI 1.0-2.4), and 1.0 (95%CI 0.6-1.8). In addition, consistent with Vessey et al, ^[88] an increased risk was reported for more recent use. Compared with never-use, the odds ratios for time since last use of less than 5 years, 5 to 9 years, and 10 or more years were 2.0 (95%CI 1.3-3.1), 2.2 (95%CI 1.3-3.6), and 1.5 (95%CI 0.9-3.0). For other forms of contraception considered, compared to never-use, no association was found between endometriosis and ever-IUD use (OR=1.3, 95%CI 0.6-2.8). Unfortunately the authors did not examine the association by current and former IUD use. Again, similar to the findings of Vessey et al, ^[88] there was no association with duration of IUD use less than 2 years (OR=1.7, 95%CI 0.6-4.5) or 2 or more years (OR=0.9, 95%CI 0.5-1.8). No association was found between ever barrier-method-use and endometriosis (OR=0.5, 95%CI 0.3-1.4). Consistent with the interpretation of Vessey, ^[88] the authors suggest a form of selection bias may be responsible for the observed association. Dysmenorrhea is a frequently reported symptom among women with endometriosis and also an important indication for OC use. Thus, women with endometriosis-induced dysmenorrhea would be selectively underrepresented in the never-OC-use group, resulting in a biased estimate of risk for ever-OC use.

Again, consistent with the findings of Vessey et al ^[88] and Parazzini et al, ^{[69] [70]} Sangi-Haghpeykar and Poindexter ^[78] found that current OC use was inversely related to disease (OR=0.5, 95%CI 0.2-0.9), but duration of use was not related. In contrast, there was no association with current IUD use (OR=0.8, 95%CI 0.1-2.3), but compared with never-use, duration of use between 2 to 4 years was associated with an increased risk of disease (OR=3.0, 95%CI 1.1-8.1).

Personal Habits

The relation of endometriosis and personal habits, including smoking, caffeine use, alcohol consumption, and exercise, has been investigated in some studies. In general, the risk of endometriosis appears to be decreased for habits that are related to decreased body estrogen levels.

Cramer et al ^[19] reported an inverse association between endometriosis and smoking, but it was limited to heavy smokers (1 pack/day) who had begun the habit before age 17 years (OR=0.5, 95%CI 0.3-0.9).

252

Using friend controls, Darrow et al ^[22] also found an inverse effect of smoking before age 16 years that was limited to women under age 30 years (OR=0.3, 95%CI 0.1-0.9). Sangi-Haghpeykar and Poindexter ^[78] reported no differences, however, between cases and controls with respect to current or previous smoking. Vessey et al ^[88] also found no association with smoking.

Additional behavioral risk factors have been examined in a subset of the population studied by Cramer et al. ^[19] Grodstein et al ^[37] compared self-reported histories of caffeine consumption among 170 women with primary infertility because of endometriosis (excluding women with an additional disorder) and 3833 women who were admitted for delivery of a live birth at the same center. After adjusting for center, age, number of sexual partners, cigarette smoking, and alcohol intake, a positive association was found between endometriosis and caffeine consumption at the higher levels of intake. Relative to women who consumed 3 g caffeine per month (g/mo) or about 1 cup per day, the odds ratios were 1.1 (95%CI 0.7-1.9), 1.9 (95%CI 1.2-3.0), and 1.7 (95%CI 1.1-2.5) for 3.1 to 5 g/mo, 5.1 to 7 g/mo, and more than 7 g/mo. Although no clear mechanism for this association was given, the authors suggested the most likely explanation for caffeine's influence on endometriosis would be an increase in estrogen levels. It is possible these findings are the result of uncontrolled confounding (menstrual cycle characteristics, see above discussion ^[19] ), however, especially in light of the relatively low exposure levels. For example, the largest effect was observed for consumption of only 3.1 to 5 g/mo, which is equivalent to 1 to 1.7 cups of coffee per day. As with any case-control study, recall bias might explain the findings; infertile cases may have been more likely to report caffeine intake than would controls who had recently completed a successful pregnancy. Also, controls might have underreported caffeine use because they had recently been pregnant, reduced their intake during pregnancy, and reported this intake as their usual consumption. Finally, because approximately one half of the control group was multiparous but were asked about caffeine consumption prior to first-live birth, periods of recall were different for some cases and controls.

Grodstein et al ^[36] also investigated the relationship between moderate alcohol intake and endometriosis in a further subset of the same multicenter study population. Self-reported histories of alcohol consumption were compared among 158 women with primary infertility owing to endometriosis (excluding women with an additional disorder) and 3833 women who were admitted for delivery of a live birth at the same center. Compared with no alcohol use, increased risks of endometriosis were reported for moderate and heavy alcohol use, defined as less than or equal to 1 drink per week (100 g alcohol) and more than 1 drink per week (100 g alcohol), respectively. After controlling for center, age, number of sexual partners, smoking, and caffeine intake, the ORs were 1.7 (95%C.I. 1.2-2.5) for moderate and 1.8 (95%C.I. 1.0-3.2) for heavy alcohol use. The biologic plausibility of this association,

253

especially for the moderate group, is questionable given the extremely low amounts of alcohol consumed. When the analysis was limited to all cases versus only those control subjects who were primiparous (approximately 50%; exact number not given), the ORs were 1.3 (95%CI 0.9-1.9) for moderate and 1.4 (95%CI 0.7-2.6) for heavy alcohol use, suggesting that differential recall may have explained the original finding. This raises further questions about the association with caffeine intake found in the same study population. ^[36]

An inverse association has been reported for exercise, another habit thought to reduce endogenous estrogen levels. In the study by Cramer et al, ^[19] cases compared to controls were found to be less likely to exercise more than 2 hours per week (OR=0.6, 95%CI 0.4-0.8) and were less likely to have begun the habit before age 25 years (OR=0.4, 95%CI 0.3-0.7). Again, the biologic plausibility of this association is not certain because 2 hours of exercise per week is not very highly exposed. Furthermore, it is likely that women with symptomatology of endometriosis such as pelvic pain may reduce exercise, and thus lower levels of exercise may be the effect and not the cause of disease.

Body Characteristics

In the same study population as Darrow et al, ^[22] McCann et al examined the association of body fat distribution with risk of endometriosis in 88 cases and 88 friend controls for whom all the body measurement data (waist, hip, and thigh circumference) were available. ^[58] After adjusting for age, body mass index, ever-pregnant, age at menarche, intensity of menstrual flow, and increasing symptoms, among women less than age 30 years, the risk of endometriosis was inversely associated with waist-to-hip ratio and waist-to-thigh ratio; that is, the risk was greater for women with more peripheral body fat than for women with more centralized fat. For waist-to-hip ratio, relative to women in the highest tertile (ratio of 0.76-1.01), the odds ratios were 3.2 (95%CI 0.6-16.0) and 6.2 (95%CI 1.4-26.7) for the middle (ratio of 0.72-0.76) and lowest (ratio of 0.61-0.72) tertiles, respectively. For waist-to-thigh ratio, a similar but nonsignificant inverse association was observed. Relative to women in the highest tertile (ratio of 1.31-1.75), the odds ratios were 2.7 (95%CI 0.7-9.9) and 2.4 (95%CI 0.64-9.3) for the middle (ratio of 1.22-1.31) and lowest (ratio of 1.03-1.22) tertiles, respectively. This effect was found only for women younger than age 30 years and not for the older women. The authors suggested the lack of findings among older women may be due to a progressive increase in waist circumference with age. Limitations of this study include that there was no direct measurement of body fat, weight and height were based on self-report, the sample size was small, and the controls may have had undiagnosed disease. The authors suggested that greater peripheral body fat may be related to higher estrogen levels in the case group, a hypothesis that is

254

consistent with the notion that the maintenance and activity of the disease depends on estrogen.

Summary of Risk Factors

To date, few well-designed analytic epidemiologic studies of endometriosis have been conducted. In fact, of almost 100 studies reviewed, only 6 studies, 1 cohort and 5 case-control, included a laparoscopy-confirmed case group, provided clear criteria for selection of the control group, and considered potential confounding factors in the analysis. Age is the only sociodemographic characteristic for which a consistent positive relation has been observed. The relation of endometriosis and reproductive health factors, including menstrual cycle characteristics and pregnancy-related variables, has been the primary focus of most studies. In general, the risk of endometriosis appears to be increased for reproductive health factors that relate to increased exposure to menstruation (i.e., shorter cycle length, longer duration of flow, reduced parity). Of other factors examined, an inverse relation with endometriosis has been reported for personal habits that relate to decreased body estrogen levels such as exercise and smoking. Furthermore, endometriosis cases tended to have more peripheral body fat (higher estrogen levels in this group), consistent with the notion that the maintenance and activity of the disease depends on estrogen. Finally, the association of endometriosis with contraception use, in particular OC use, is rather complex but probably due to bias and is not a causal factor of the disease. Overall, given the likely prevalence of endometriosis in the population and its public health significance, surprisingly little is known about risk factors for this disease.

FUTURE DIRECTIONS

Our knowledge about the disease patterns and risk factors of endometriosis is hampered by our inability to diagnose the disease in the general population. To truly understand the epidemiology of endometriosis, population-based cohort studies must become feasible. With the advent of fine-needle laparoscopy, women with less severe symptomatology may be more likely to be diagnosed. Only with the development of a noninvasive biomarker of disease, however, will it be possible to truly study the general population. Currently, numerous laboratories are testing potential biomarkers. Meanwhile, alternative approaches can be developed to study endometriosis. These approaches may include a diagnostic tier approach similar to that used by Houston et al ^[42] or the development of mathematical predictive models.

To understand the underlying mechanism of endometriosis, future studies should study the natural history of the disease, perhaps beginning with premenarchal girls ^[41] , and the association of endometriosis with

255

other diseases such as allergy or autoimmunity. Future epidemiologic investigations should focus on risk factors that are preventable, are biologically plausible, and have high impact on the population.

References

1. Adamson DG: Diagnosis and clinical presentation of endometriosis. Am J Obstet Gynecol 162:568, 1990

2. Ajossa S, Mais V, Guerriero S, et al: The prevalence of endometriosis in premenopausal women undergoing gynecological surgery. Clin Exp Obstet Gynecol 21:195, 1994

3. Arumugam K, Templeton AA: Endometriosis and race. Aust NZ J Obstet Gynaecol 32:164, 1992

4. Arumugam K, Templeton AA: Endometriosis and social class. Med J Aust 153:563, 1990

5. Arumugam K, Welluppilai S: Endometriosis and social class: An Asian experience. Asia-Oceana Journal of Obstetrics and Gynaecology 19:231, 1993

6. Badawy SZA, Cuenca V, Freliech H, et al: Endometrial antibodies in serum and peritoneal fluid of infertile patients with and without endometriosis. Fertil Steril 53:930, 1990

7. Beral V, Rolfs R, Joesoef MR, et al: Primary infertility: Characteristics of women in North America according to pathological findings. J Epidemiol 48:576, 1994

8. Berger MJ, Alper MM: Intractable primary infertility in women exposed to diethylstilbestrol in utero. J Reprod Med 31:231, 1986

9. Blunt A: The place of laparoscopy in "normal" gynaecological practice. Aust NZ J Obstet Gynaecol 12:194, 1972

10. Boling RO, Abbasi R, Ackerman G, et al: Disability from endometriosis in the United States Army. J Reprod Med 33:49, 1988

11. Brincat M, Galea R, Buhagiar A: Polycystic ovaries and endometriosis: A possible connection. Br J Obstet Gynaecol 101:346, 1994

12. Buchan H, Vessey M, Goldacre M, et al: Morbidity following pelvic inflammatory disease. Br J Obstet Gynaecol 100:558, 1993

13. Candiani GB, Danesino V, Gastaldi A, et al: Reproductive and menstrual factors and risk of peritoneal and ovarian endometriosis. Fertil Steril 56:230, 1991

14. Chai S, Wild RA: Basal body temperature and endometriosis. Fertil Steril 54:1028, 1990

15. Chang YS, Lee JY, Moon SY, et al: Diagnostic laparoscopy in gynecologic disorders. Asia-Oceania Journal of Obstetrics and Gynaecology 13:29, 1987

16. Chatman DL: Endometriosis in the black woman. Am J Obstet Gynecol 125:987, 1976

17. Chatman DL, Ward AB: Endometriosis in adolescents. J Reprod Med 27:156, 1982

18. Cornillie FJ, Oosterlynck D, Lauweryns JM, et al: Deeply infiltrating pelvic endometriosis: Histology and clinical significance. Fertil Steril 53:978, 1990

19. Cramer DW, Wilson E, Stillman RJ, et al: The relation of endometriosis to menstrual characteristics, smoking, and exercise. JAMA 255:1904, 1986

20. Crosignani PG, Vercellini P: New clinical guidelines are needed for the treatment of endometriosis. Hum Reprod 9:2205, 1994

21. Darrow SL, Selman S, Batt RE, et al: Sexual activity, contraception, and reproductive factors in predicting endometriosis. Am J Epidemiol 140:500, 1994

22. Darrow SL, Vena JE, Batt RE, et al: Menstrual cycle characteristics and the risk of endometriosis. Epidemiology 4:135, 1993

23. DePriest PD, Banks ER, Powell SE, et al: Endometrioid carcinoma of the ovary and endometriosis: The association in postmenopausal women. Gynecol Oncol 47:71, 1992

24. Dodge ST, Pumphrey RS, Miyazawa K: Peritoneal endometriosis in women requesting reversal of sterilization. Fertil Steril 45:774, 1986

25. Drake TS, Grunert GM: The unsuspected pelvic factor in the infertility investigation. Fertil Steril 34:27, 1980

26. Duignan NM, Jordan JA, Coughlan BM, et al: One thousand consecutive cases of diagnostic laparoscopy. J Obstet Gynaecol Br Commonwealth 79:1016, 1972

256

27. Dunphy BC, Kay R, Barratt CLR, et al: Female age, the length of involuntary infertility prior to investigation and fertility outcome. Hum Reprod 4:527, 1989

28. El-Yahia AW: Laparoscopic evaluation of apparently normal infertile women. Aust NZ J Obstet Gynaecol 34:440, 1994

29. Fedele L, Bianchi S, Di Nola G, et al: Endometriosis and nonobstructive Mullerian anomalies. Obstet Gynecol 79:515, 1992

30. Fedele L, Parazzini F, Bianchi S, et al: Stage and localization of pelvic endometriosis and pain. Fertil Steril 53:155, 1990

31. Federici D, Conti E, Constantini W, et al: Endometriosis and infertility: Our experience over five years. Hum Reprod 3:109, 1988

32. Filer RB, Wu CH: Coitus during menses: Its effect on endometriosis and pelvic inflammatory disease. J Reprod Med 34:887, 1989

33. Forman RG, Robinson JN, Mehta Z, et al: Patient history as a simple predictor of pelvic pathology in subfertile women. Hum Reprod 8:53, 1993

34. Fraser IS: Hysteroscopy and laparoscopy in women with menorrhagia. Am J Obstet Gynecol 162:1264, 1992

35. Goldstein DP, deCholnoky C, Emans SJ, et al: Laparoscopy in diagnosis and management of pelvic pain in adolescents. J Reprod Med 24:251, 1980

36. Grodstein F, Goldman MB, Cramer DW: Infertility in women and moderate alcohol use. Am J Pub Health 84:1429, 1994

37. Grodstein F, Goldman MB, Ryan L, et al: Relation of female infertility to consumption of caffeinated beverages. Am J Epidemiol 137:1353, 1993

38. Gruppo Italiano per lo Studio dell'Endometriosi: Prevalence and anatomical distribution of endometriosis in women with selected gynaecological conditions: results from a multi-centric Italian study. Hum Reprod 9:1158, 1994

39. Hasson HM: Incidence of endometriosis in diagnostic laparoscopy. J Reprod Med 16:135, 1976

40. Hernandez E, Miyazawa K: The pelvic mass: Patients' ages and pathologic findings. J Reprod Med 33:361, 1988

41. Houston DE, Noller K, Melton LJ, et al: The epidemiology of pelvic endometriosis. Clin Obstet Gynecol 31:787, 1988

42. Houston DE, Noller KL, Melton LJ, et al: Incidence of pelvic endometriosis in Rochester, Minnesota, 1970-1979. Am J Epidemiol 125:959, 1987

43. Kirshon B, Poindexter AN: Contraception: a risk factor for endometriosis. Obstet Gynecol 71:829, 1988

44. Kirshon B, Poindexter AN, Fast J: Endometriosis in multiparous women. J Reprod Med 34:215, 1989

45. Kjerulff KH, Erickson BA, Langenberg PW: Chronic gynecological conditions reported by US women: Findings from the National Health Information Survey, 1984 to 1992. Am J Public Health 86:195, 1996

46. Kleppinger RK: One thousand laparoscopies at a community hospital. J Reprod Med 13:13, 1974

47. Koninckx PR: Is mild endometriosis a disease? Hum Reprod 9:2202, 1994

48. Kresch AJ, Seifer DB, Sachs LB, et al: Laparoscopy in 100 women with chronic pelvic pain. Obstet Gynecol 64:672, 1984

49. Liston WA, Bradford WP, Downie J, et al: Laparoscopy in a general gynecologic unit. Am J Obstet Gynecol 113:672, 1972

50. Liu DTY, Hitchcock A: Endometriosis: Its association with retrograde menstruation, dysmenorrhoea and tubal pathology. Br J Obstet Gynaecol 93:859, 1986

51. Mahmood TA, Templeton A: Prevalence and genesis of endometriosis. Hum Reprod 6:544, 1991

52. Mahmood TA, Templeton A: The relationship between endometriosis and semen analysis: A review of 490 consecutive laparoscopies. Hum Reprod 4:782, 1989

53. Mahmood TA, Templeton AA, Thomson L, et al: Menstrual symptoms in women with pelvic endometriosis. Br J Obstet Gynaecol 98:558, 1991

54. Marana R, Muzli L, Caruana S, et al: Evaluation of the correlation between endometriosis extent, age of the patients and associated symptomatology. Acta Eur Fertil 22:209, 1991

257

55. Mathias SD, Kuppermann M, Liberman RF, et al: Chronic pelvic pain: Prevalence, health-related quality of life, and economic correlates. Obstet Gynecol 87:321, 1996

56. Matorras R, Rodiquez F, Pijoan JI, et al: Epidemiology of endometriosis in infertile women. Fertil Steril 63:34, 1995

57. Matorras R, Rodriguez F, Pijoan JI, et al: Are there any clinical signs and symptoms that are related to endometriosis in infertile women? Am J Obstet Gynecol 174:620, 1996

58. McCann SE, Freudenheim JL, Darrow SL, et al: Endometriosis and body fat distribution. Obstet Gynecol 82:545, 1993

59. Meiling H, Lingya P, Baozhen W, et al: A case-control epidemiologic study of endometriosis. Chin Med Sci J 9:114, 1994

60. Miyazawa K: Incidence of endometriosis among Japanese women. Obstet Gynecol 48:407, 1976

61. Moeloek FA, Affandi B, Samil RS: Peritoneal factors in infertile women in Jakarta, Indonesia. J Reprod Med 29:603, 1984

62. Moen MH: Endometriosis in women at interval sterilization. Acta Obstet Gynecol Scand 66, 1987

63. Moen MH: Is a long period without childbirth a risk factor for developing endometriosis? Hum Reprod 6:1404, 1991

64. Moen MH, Muus KM: Endometriosis in pregnant and non-pregnant women at tubal sterilization. Hum Reprod 6:699, 1991

65. Moore JG, Schifrin BS, Erez S: Ovarian tumors in infancy, childhood, and adolescence. Am J Obstet Gynecol 99:913, 1967

66. Musich JR, Behrman SJ: Infertility laparoscopy in perspective: Review of five hundred cases. Am J Obstet Gynecol 143:293, 1982

67. National Center for Health Statistics, McCarthy E: Inpatient utilization of short-stay hospitals by diagnosis: United States, 1980. In Hyattsville, MD, National Center for Health Statistics, 1982

68. Obermeyer CM, Armenian HK, Azoury R: Endometriosis in Lebanon. Am J Epidemiol 124:762, 1986

69. Parazzini F, Ferraroni M: Epidemiology of endometriosis. Br Med J 306:930, 1993

70. Parazzini F, Ferraroni M, Bocciolone L, et al: Contraceptive methods and risk of pelvic endometriosis. Contraception 94:47, 1994

71. Parazzini F, Ferraroni M, Fedele L, et al: Pelvic endometriosis: Reproductive and menstrual risk factors at different stages in Lombardy, northern Italy. J Epidemiol Comm Health 49:61, 1995

72. Pent D: Laparoscopy: Its role in private practice. Am J Obstet Gynecol 113:459, 1972

73. Perper MM, Breitkopf LJ, Brietstein R, et al: MAST scores, alcohol consumption, and gynecological symptoms in endometriosis patients. Alcohol Clin Exp Res 17:272, 1993

74. Perper MM, Nezhat F, Goldstein H, et al: Dysmenorrhea is related to the number of implants in endometriosis patients. Fertil Steril 63:500, 1995

75. Phipps WR, Cramer DW, Schiff I, et al: The association between smoking and female infertility as influenced by cause of the infertility. Fertil Steril 48:377, 1987

76. Redwine DB: The distribution of endometriosis in the pelvis by age groups and fertility. Fertil Steril 47:173, 1987

77. Salamanca A, Beltran E: Subendometrial contractility in menstrual phase visualized by transvaginal sonography in patients with endometriosis. Fertil Steril 64:193, 1995

78. Sangi-Haghpeykar H, Poindexter AN: Epidemiology of endometriosis among parous women. Obstet Gynecol 85:983, 1995

79. Sarram M, Rezazadeh A: Endometriosis in diagnostic laparoscopy in Isfahan, Iran. Int Fed Gynaecol Obstet 24:117, 1986

80. Singh KB, Patel YC, Wortsman J: Coexistence of polycystic ovary syndrome and pelvic endometriosis. Obstet Gynecol 74:650, 1989

81. Stillman RJ, Miller LC: Diethylstilbestrol exposure in utero and endometriosis in infertile females. Fertil Steril 41:369, 1984

82. Strathy JH, Molgaard CA, Coulam CB, et al: Endometriosis and infertility: A laparoscopic study of endometriosis among fertile and infertile women. Fertil Steril 38:667, 1982

83. Trimbos JB, Trimbos-Kemper GCM, Peters AAW, et al: Finding in 200 consecutive

258

asymptomatic women having a laparoscopic sterilization. Arch Gynecol Obstet 247:121, 1990

84. Velebil P, Wingo PH, Xia Z, et al: Rate of hospitalization for gynecologic disorders among reproductive-age women in the United States. Obstet Gynecol 86:764, 1995

85. Vercellini P, Fedele L, Arcaini L, et al: Laparoscopy in the diagnosis of chronic pelvic pain in adolescent women. J Reprod Med 34:827, 1989

86. Vercellini P, Fedele L, Molteni P, et al: Laparoscopy in the diagnosis of gynecologic chronic pelvic pain. Int J Gynecol Obstet 32:261, 1990

87. Vercellini P, Tresidi L, Giorgi OD, et al: Endometriosis and pelvic pain: Relation to disease stage and localization. Fertil Steril 65:299, 1996

88. Vessey MP, Villard-Mackintosh L, Painter R: Epidemiology of endometriosis in women attending family planning clinics. Br Med J 306:182, 1993

89. Waller KG, Lindsay P, Curtis P, et al: The prevalence of endometriosis in women with infertile partners. Eur J Obstet Gynaecol Reprod Biol 48:135, 1993

90. Wheeler JM: Epidemiology of endometriosis-associated infertility. J Reprod Med 34:41, 1989

91. Woodworth SH, Singh M, Yussman MA, et al: A prospective study on the association between red hair color and endometriosis in infertile patients. Fertil Steril 64:651, 1995

 

Ana sayfaya dönmek icin tiklayin http://www.geocities.com/sinandoganturk1.html