This article is based on a presentation at the conference on Women's Health and the Environment: The Next Century--Advances in Uterine Leiomyoma Research held 7-8 October 1999 in Research Triangle Park, North Carolina, USA.
Address correspondence to S.M. Schwartz, 1100 Fairview Ave. N. (MP-381), PO Box 19024, Seattle, WA 98109-1024 USA. Telephone: (206) 667-4660. Fax: (206) 667-5948. E-mail: sschwart@fhcrc.org
This work was supported in part by grants from the National Institute of Child Health and Human Development (HD 29819) and the National Institute of Environmental Health Sciences (NIEHS) (ES 08305), intramural National Institutes of Health support from NIEHS, and the Office of Research on Minority Health. Portions of this work have been adapted from Schwartz SM, Marshall LM. Uterine leiomyomata. In: Women and Health (Goldman MB, Hatch MC, eds). New York:Academic Press, 1999;240-252.
Received 23 February 2000; accepted 2 June 2000.
Uterine leiomyomata are a frequent cause of menstrual abnormalities, pelvic pain, and other symptoms that seriously affect a woman's quality of life. Symptomatic uterine leiomyomata often require major surgery and increased medical utilization. In the United States uterine leiomyomata account for 33% of all hysterectomies (
1). Despite the substantial impact of uterine leiomyomata on gynecologic morbidity, relatively little is known about the etiology of these neoplasms or the characteristics of women who are at increased risk for developing them.
In this review of uterine leiomyomata, we begin by discussing several of the methodologic issues that should be considered when conducting and interpreting research on this condition. We then summarize the epidemiologic literature and conclude by describing several research opportunities for epidemiologists to contribute to an increased understanding of the etiology of uterine leiomyomata.
The major challenge in conducting epidemiologic research on uterine leiomyomata is the large portion of women who develop these tumors but never come to clinical attention. The precise level of undiagnosed disease is uncertain; the evidence comes primarily from pathologic studies of hysterectomy specimens, which represent highly selected patients (
2). Nonetheless, the level of undiagnosed disease is likely to be sufficiently high as to impact strongly the key study design issues of defining cases, identifying appropriate noncases (controls), and collecting data on risk factors present prior to the development of the tumors.
Identifying Cases
Histologic evidence is the gold standard for determining whether a woman has uterine leiomyomata, and the primary public health and clinical impact of uterine leiomyomata derives from the hysterectomies and other surgeries required to treat women when these tumors become symptomatic. Yet, women with histologically confirmed tumors probably represent 10-30% of patients for whom sonographic evidence is available, depending on age and race (3). Epidemiologic studies that include only women with histologically confirmed tumors are challenging to interpret, as such studies may preferentially identify risk factors associated with symptoms (such as uterine bleeding or pelvic pain) and/or large size, rather than the occurrence of leiomyomata (4). Although such risk factors may well provide clues to preventing the development of uterine leiomyomata requiring hysterectomy, they may also reflect characteristics associated with unsuccessful conservative management, patient and/or physician preferences for surgical management, or a combination of these influences. For example, a woman may be more likely to choose hysterectomy as therapy if she has completed childbearing.
The ideal epidemiologic study would identify women with and without uterine leiomyomata prospectively on the basis of regular (e.g., yearly) screening of a defined target population. The preferred method for identifying new cases in such a study, and thus women without uterine leiomyomata as well, would have perfect sensitivity and specificity--an unattainable goal. The best practical screening tool currently available for epidemiologic studies is transvaginal ultrasound, as it would yield a much less selective set of cases than possible if pathologic diagnoses were required. One approach, therefore, might be to identify cases from clinical ultrasound records serving a defined population or on the basis of reported ultrasound diagnoses from a cohort of women being followed over time. However, records from routine clinical ultrasound examinations may well miss many women with leiomyomata because lifestyle, medical, and health insurance factors all may influence which women come to clinical attention. For example, women who have pelvic exams performed in conjunction with prescriptions for hormonal birth control would be expected to have an increased opportunity to have asymptomatic uterine leiomyomata detected and thus be referred for gynecologic ultrasound. Another option would be to recruit a sample of women without clinically detected tumors into a protocol that includes ultrasound examinations to detect asymptomatic, clinically occult, uterine leiomyomata. Such a one-time ultrasound screening of all potential study subjects would only identify prevalent, rather than incident, cases. Regardless of the approach, ultrasound examinations can potentially provide data on tumor characteristics (estimated size, number, and position), and additional data on symptoms could be collected (presence, duration, and severity) to distinguish among the identified uterine leiomyomata on the basis of clinical consequences and other characteristics that may help identify risk factors.
Some uterine leiomyomata will be identified in women incidentally during examinations prompted by symptoms caused by other conditions. Although a researcher may be tempted to classify uterine leiomyomata diagnoses as incidental or not, and exclude the former, often it will not be possible to determine with confidence whether the symptoms are due to the uterine leiomyomata or to other pathologic conditions. Nonetheless, if a characteristic is not a risk factor for uterine leiomyomata but is related to a condition that often coexists with these tumors, a spurious association with uterine leiomyomata could result. To reduce the potential impact of incidental diagnoses, cases should not be selected preferentially from specialty clinics that treat gynecologic, urologic, or infertility conditions. In general, ascertainment of uterine leiomyomata cases from any source in which persons with particular diseases or medical histories are likely to be overrepresented carries the potential to cloud the interpretation of the epidemiologic findings.
Identifying Controls
The challenges in defining a woman's status as a case of uterine leiomyomata have direct implications for establishing the criteria by which the comparison population--the noncases--will be selected. In general, a noncase would be a woman who has an intact uterus and does not have uterine leiomyomata at a defined point in time during which the cases are accrued. Further, controls should be selected to represent the distribution of exposures in the population from which the cases are identified. As with epidemiologic research on other conditions, studies that can specifically define the base population from which the cases derive, and obtain controls from that population, usually will have the strongest potential to provide valid inferences as to uterine leiomyomata risk factors.
The main challenge for control definition is the criterion by which a woman is considered not to have uterine leiomyomata. In published studies conducted to date (see below), controls were women who had not received diagnoses of uterine leiomyomata. However, because not all cases of leiomyomata come to clinical attention, a substantial proportion of controls in each of the studies undoubtedly had leiomyomata. This misclassification can be substantial. In a sample of 393 women 35-49 years of age who had never received a clinical or surgical diagnosis of leiomyomata, between 48 and 73% were found to have this condition when screened using a transvaginal ultrasound administered to identify occult cases (5). Thus, as discussed above, controls without uterine leiomyomata could be identified by recruiting a sample of women without clinically detected tumors into a protocol that includes transvaginal ultrasound examinations to detect asymptomatic, clinically occult, uterine leiomyomata. Such a protocol would be expensive and may be logistically infeasible in some study settings. Whether the reduction in disease misclassification obtained from including uterine leiomyomata-free controls improves the ability to identify risk factors for this condition is likely to depend on the risk factor being studied. For risk factors that reflect the earliest stages of leiomyomata development [e.g., initiation of changes in a myometrial cell that confers growth advantage (6)], the use of ultrasound examinations to identify women without uterine leiomyomata as controls should help clarify relationships. For risk factors that act primarily in the later stages, for example to enhance growth late in reproductive life, a control group comprising women shown by ultrasound to be free of uterine leiomyomata may not enhance the identification of associations. These issues apply in varying degrees to epidemiologic studies of many benign reproductive health conditions, as well as to chronic diseases that are highly prevalent among the elderly (e.g., prostate cancer, carotid atherosclerosis) and thus are not unique to uterine leiomyomata.
Some investigators recommend that in order to exclude the presence of uterine leiomyomata, epidemiologic studies include as controls only women who have had their uteri removed and examined histologically as part of clinical care (2). A similar approach might choose as controls women who have no evidence of uterine leiomyomata based on ultrasound examinations performed as part of clinical care. Because hysterectomy and gynecologic ultrasound are performed to treat or diagnose gynecologic conditions, however, controls identified in such ways may very well yield spurious positive or negative associations with uterine leiomyomata when studying characteristics related to the conditions for which these procedures are performed (7,8).
Collecting Information on Potential Risk Factors
The majority of epidemiologic studies conducted to date have focused on lifestyle and medical characteristics as potential risk factors. Information on these characteristics is typically obtained through interviews with study participants or medical record review. The ideal epidemiologic study would obtain information on such potential risk factors present before disease onset. In the absence of repeated examinations of a cohort of women to identify newly developing uterine leiomyomata among those with no initial evidence of this condition (e.g., using transvaginal ultrasound), it is usually not possible to determine when uterine leiomyomata first occur. In practice, therefore, an investigator typically will attempt to obtain risk factor information that applies to the period prior to initial symptoms. However, this approach is not easily applied to women whose uterine leiomyomata are detected asymptomatically. In addition, for studies that include cases diagnosed histologically, there may be an extended interval between the initial clinical diagnosis of uterine leiomyomata (e.g., by ultrasound) and surgical treatment. In the absence of more definitive data on the natural history of leiomyomata development, there is no clear way to collect risk factor information that pre-dates the actual occurrence of the tumors. Further, extensive attempts to backdate exposure information can be extremely challenging to implement in epidemiologic studies in general (8). As a result, it may be best for an investigator to consider women with uterine leiomyomata as having prevalent disease and to elicit a relatively long-term history of risk factors of interest. This problem with relating risk factors to disease occurrence when the onset is not acute applies in varying degrees to epidemiologic studies of many conditions that have long preclinical phases (e.g., prostate cancer, adenomatous polyps, carotid atherosclerosis) and thus again is not unique to uterine leiomyomata.
Estimated rates of uterine leiomyomata diagnoses vary according to the method of diagnostic confirmation (e.g., clinical examination, ultrasound, or surgical), ranging from 12.8 per 1,000 person-years for all diagnoses to approximately 2.0 per 1,000 person-years for hysterectomy-confirmed cases (
1,3,9,10) (Table 1). Rates increase with age through the reproductive years (
1,3,9,11) and decline in the postmenopausal years (
1,11). Studies have not identified any appreciable change in the rate of uterine leiomyomata over time, although the temporal data available are limited to hysterectomy-confirmed uterine leiomyomata (
1,12,13).
Uterine leiomyomata diagnoses occur 2-3 times as often among black women as white women (1,3,10), whereas Hispanic and Asian women experience rates similar to those among whites (3). The higher rates among blacks are evident at all ages (3,9). Some evidence suggests that uterine leiomyomata rates peak at an earlier age among blacks (35-39 years) than among whites (40-44 years) (3). Several other findings are consistent with the black-white differences in diagnosis rates. Compared to whites, black women are twice as likely to report having physician-diagnosed uterine leiomyomata in the prior 12 months (14). Among women undergoing hysterectomy, 58-65% of blacks have uterine leiomyomata as the primary indication compared to 29-45% among whites (1,15,16). Among women undergoing hysterectomy specifically for uterine leiomyomata, mean age at first diagnosis and mean age at hysterectomy are lower among blacks than whites (14,15). Among women surgically treated for uterine leiomyomata, black women have more uterine leiomyomata, larger tumors, and heavier uterine weight than white women (3,14). The most convincing evidence that black women are at higher risk of leiomyomata comes from a study in which participants from a randomly selected sample of women 35-49 years of age were given ultrasound exams (5). Ultrasound evidence of fibroids was more likely to be found in black women than white women. In the only studies to date that have sought to explain these differences, the excess risk and earlier diagnoses of uterine leiomyomata among blacks was not attributable to differences in the types and severity of symptoms, the use of health care, or putative risk factors (3,14). Additional studies are needed to identify the reason for the difference in the development of uterine leiomyomata between whites and blacks.
The majority of the evidence regarding risk factors for uterine leiomyomata comes from nine formal epidemiologic studies (Table 2). In most studies, cases were women who had undergone hysterectomy or other surgery and whose diagnosis of uterine leiomyomata was confirmed histologically. However, in some studies the designation of case status was based on a lower level of diagnostic certainty. In all studies the comparison population consisted of women in whom uterine leiomyomata had not been clinically diagnosed.
Menstrual and Reproductive History
Studies consistently show that postmenopausal women are at 70-90% reduced risk of uterine leiomyomata (11,17-19). These epidemiologic data are supported by pathologic studies that find a reduction in both the size and number of leiomyomata in postmenopausal compared to premenopausal women (2). Such data provide some of the strongest epidemiologic evidence that a woman's hormonal milieu plays an important role in the occurrence of these tumors. However, the role of specific steroid hormones, such as estrogen and progesterone, cannot be inferred from these findings because the postmenopausal period is characterized by a generalized decline in ovarian function. Mitotic activity in the myometrium appears to be greatest during the luteal phase of the menstrual cycle (20). Thus, if progesterone is an important etiologic factor, it might be expected that a longer history of cycling might be associated with increased risk. No studies have examined summary measures of a woman's cumulative exposure to menstrual cycling as risk factors for uterine leiomyomata. Women with early onset of menses will, on average, have increased lifetime exposure to ovulatory cycles, but early age of menarche has been associated with an increased risk in only some of the studies in which this characteristic has been examined (17-19,21,22).
Studies report the risk of uterine leiomyomata to be 20-50% lower among women who have ever given birth compared to nulliparous women (21-23), and the risk appears to decrease with increasing parity (11,21-23). Although most studies have not found any association between age at first-term birth and risk of surgically confirmed uterine leiomyomata (10,11,21,23), older age at last birth appears to be inversely associated with the risk of surgically confirmed tumors [relative risks (RR) ranging from 0.4 (11) to 0.6 (21) for last birth at age 35 years or older compared to last birth before age 35]. The possible importance of jointly considering the timing of both first and last births was recently examined in a large cohort study (22). Compared to nulliparous women, the risk was reduced similarly (58-65%) regardless of age at first birth (under 25 years of age vs 25 years or older) among women who last gave birth in the prior 3 years. Recent evidence from the Nurse's Health Study (NHS) II indicates that a history of infertility is unlikely to underlie the inverse associations with childbearing and its timing (22). Breastfeeding does not appear to be associated with uterine leiomyomata after accounting for parity, although there has been relatively little research on this topic (19,21).
Several mechanisms may underlie the observed inverse associations with parity. Although endocrine factors change dramatically during pregnancy, the postpregnancy hormonal state also differs from that in nulliparous women (24-26). Further, increasing parity could represent a reduction in a woman's lifetime exposure to normal cycling estrogens and progesterone. Pregnancy also may lead to a reduction in estrogen receptor levels in myometrial tissue (20) that could reduce the sensitivity of these tumors to hormonal stimuli. Alternatively, childbearing may counteract the development of uterine leiomyomata through nonhormonal mechanisms. For example, myometrial hypertrophy during pregnancy could inhibit growth of small clones of transformed myometrial cells, and reductions in collagen content and smooth muscle cell apoptosis during uterine regression could eliminate or reduce the size of minute uterine leiomyomata or selectively remove cells most likely to develop into tumors. The postpartum also could represent an intersection of hormonal and nonhormonal pathways, as loss of smooth muscle cell cytoplasm could diminish estrogen receptor levels and thus, at least temporarily, decrease the estrogen responsiveness of existing uterine leiomyomata.
Obesity and Cigarette Smoking
Several epidemiologic studies (11,21,27), including two prospective investigations (11,27), have found that the risk of uterine leiomyomata increases monotonically with increasing body mass index (BMI). In two of these studies the heaviest women were at 2- to 3-fold greater risk than the leanest women (11,27). Data from other studies, however, show little or no increased risk associated with elevated BMI (17-19). It is unclear whether the lack of consistency among studies can be attributed to methodologic differences in selection of cases and controls, errors in measurement of height and weight, or incomparable analytic approaches. With one exception (18), prior studies have examined BMI and cigarette smoking simultaneously to estimate the independent associations. In the one study in which uterine leiomyomata confirmed by ultrasound were included, the associations tended to be weaker than when analyses were restricted to uterine leiomyomata confirmed by hysterectomy (27). This finding suggests that if obesity is related to uterine leiomyomata, it may be involved in enhancing the growth of uterine leiomyomata and/or promoting the development or severity of symptoms. There is some evidence from one cohort study that women who experience an increase in weight during the reproductive years are at increased risk, whereas weight at 18 years of age is not related to risk (27). Among premenopausal women, obesity or overweight is associated with lower progesterone levels (25,28,29), decreased serum hormone-binding globulin, altered estradiol metabolism, and a physiologic profile that includes hyperandrogenism, hyperinsulinemia, and dyslipidemia. Thus, the possible association between obesity and uterine leiomyomata may reflect associated hormonal changes as well as alterations in metabolic controls that affect myometrial cell signaling through mediators such as insulin receptors, insulinlike growth factors, and peroxisome proliferating activating receptors (30).
Most epidemiologic studies have reported an inverse relation between risk of uterine leiomyomata and cigarette smoking, even after adjustment for BMI (11,18,21,31); the risk ranges from 20 to 50% lower among smokers, depending upon the definition of smoking (e.g., current, ever). The risk among former cigarette smokers does not appear to differ from that of women who have never smoked cigarettes (27,31). Two studies showed an inverse trend associated with the amount of cigarette smoking (11,18), whereas one showed a reduction in risk regardless of amount smoked (31). Cigarette smoking may favorably affect a woman's risk of uterine leiomyomata through hormonal mechanisms, such as by inducing the conversion of estradiol to relatively nontumorigenic catechol metabolites, or by decreasing the length of the menopausal transition by inducing an earlier age at menopause. A long menopausal transition may be a risk factor for leiomyomata because elevated gonadotropins in the presence of estrogen may stimulate growth of myometrial cells (32). As there appears to be little difference in blood or urine estrogen levels between premenopausal smokers and nonsmokers (29,33-36), an inverse relation between cigarette smoking is unlikely to indicate an etiologic role of unopposed or excess estrogen.
Use of Exogenous Steroid Hormones
Studies of the use of steroid contraceptives [combined estrogen-progestin oral contraceptive pills (OCP), and progestin-only injections] and hormone replacement therapy (unopposed estrogen or combined estrogen-progesterone) have clearly identified the procarcinogenic and anticarcinogenic roles of estrogens and progestins, respectively, in the endometrium (37). A number of the epidemiologic studies of uterine leiomyomata have thus examined risk in relation to use of these hormonal medications as an approach to clarifying the specific contributions of steroid hormones in the development of these tumors.
The relation between OCP use and uterine leiomyomata has been studied extensively, but no clear patterns have emerged. The risk of uterine leiomyomata among women who had ever used OCPs has been observed to be reduced (11,21), similar (18,19,38), and increased (39) relative to never users. There is some evidence that current OCP users in particular may be at 20-60% lower risk, whereas past users have the same risk as never users (22,40). Although one study suggests that long-term OCP users are at reduced risk (11), other investigations do not show trends with duration or ages of use (18,22,38), or suggest an increase in risk with increasing years of use of these medications (39). One study found that uterine leiomyomata cases reported more frequent use of OCPs containing progestins that have estrogenic properties (11).
Depot medroxyprogesterone acetate (DMPA) is a progestin-only injectable contraceptive that, although only recently approved for use in the United States, has been used extensively in developing countries. A study in Thailand showed a strong inverse association [RR = 0.4, 95% confidence interval (CI) 0.3-0.5] between a history of DMPA use and the risk of surgically confirmed uterine leiomyomata (21). The risk declined with increasing duration of use, such that use for more than 5 years reduced the risk of uterine leiomyomata by 90%. In addition, the inverse association weakened with increasing time since last use. These data, if confirmed, would support several hypotheses regarding the etiology of uterine leiomyomata, including a direct inhibitory effect of progestins on myometrial tissue proliferation and/or a protective effect of menstrual cycle suppression. Although the latter mechanism is supported by evidence that use of gonadotropin-releasing hormone (GnRH) agonists markedly reduces uterine leiomyomata size, the former mechanism would be inconsistent with the observation that progesterone add-back therapy in patients receiving GnRH experience results in increased growth of their tumors (41).
Patterns of risk associated with different hormone replacement therapy regimens could provide strong evidence regarding the roles of estrogen and progesterone in the development of uterine leiomyomata. If unopposed or excess estrogens are important in the etiology of uterine leiomyomata, users of these medications should be at increased risk, whereas users of combined estrogen and progesterone should be at lower risk or no increased risk. Similarly, if both estrogen and progesterone are playing growth-promoting roles, users of combination estrogen and progesterone replacement therapy should be at increased risk. Unfortunately, few of the published epidemiologic studies of uterine leiomyomata have addressed the relation with hormone replacement therapy. In one study, current estrogen-only users were at 6-fold increased risk of uterine leiomyomata requiring hospitalization compared to never users (39). In a second study, the risk of surgically confirmed uterine leiomyomata was 1.9 times greater (95% CI 1.5-2.5) among women who had ever used estrogen replacement therapy (18). Some evidence of a dose response was observed, with 8 or more years of use associated with a higher risk (RR = 2.3; 95% CI 1.1-9.7) than fewer than 4 years of use (RR = 1.2; 95% CI 0.7-1.9) or for 4-7 years (RR = 1.3; 95% CI 0.5-3.2). Both of these studies were conducted prior to the era in which progestins were commonly added to hormone replacement therapy regimens. As a result, no data are available on associations between combined estrogen-progestin therapy and uterine leiomyomata.
Behavioral and Lifestyle Risk Factors
The associations between uterine leiomyomata and measures of increased body size and obesity reported by some studies (11,21,27) and recent evidence from knock-out mice that high-mobility group genes (overexpressed in a high proportion of uterine leiomyomata due to gene rearrangements) are involved in the development of obesity (42,43) suggest that new etiologic clues and modifiable epidemiologic risk factors might be gained from conducting epidemiologic studies of these tumors in relation to the behavioral and lifestyle determinants and physiologic consequences of increased adiposity. Dietary fat intake and physical activity levels are strong predictors of adiposity, but few data have been reported on the relation between physical activity or diet and uterine leiomyomata. In a study of former college athletes and nonathletes surveyed about their history of benign gynecologic diseases, nonathletes were more likely to report a history of benign uterine tumors (most of which are likely to have been uterine leiomyomata) (44). More recently, a case-control study in which weekly consumption of selected food items was elicited from participants showed increased risks among more frequent consumers of beef and ham and reduced risks among more frequent consumers of green vegetables (45). Data on the relationship between endocrine-active substances from dietary intake or nutritional supplements (such as estrogenic isoflavones in soy products) have not been reported. It will be important to conduct additional studies of these relationships, using prospective designs as well as employing more detailed assessments of dietary and physical activity histories.
Most of the epidemiologic studies conducted to date have focused their attention on risk factors thought to reflect a hormonal etiology. A broader view of endocrine alterations in the etiology of uterine leiomyomata is needed. For example, human studies of possible endocrine-acting environmental factors on these neoplasms should be conducted, given recent evidence from the Eker rat model suggesting possible roles for organochlorine pesticides (
46). In addition, alternative etiologic hypotheses need to be developed and explored in epidemiologic studies. For example, smooth-muscle proliferation may be similar in uterine leiomyomata and atheromas (
47). The extensive research on atherosclerosis can be used to suggest specific avenues to pursue, including obesity-related mechanisms such as hyperinsulinemia. The occurrence of nonuterine leiomyomata in patients immunocompromised due to HIV infection (
48) suggests possible roles for infectious agents and/or immunologic factors. Advances also might result from studies that seek to explain behavioral, social, and biologic factors that underlie the differences in the frequency, size, and number of uterine leiomyomata among women of different racial and ethnic origins. For example, factors involved in the production of keloids, which are benign, collagen-rich skin lesions that occur more frequently among black persons than white persons (
49), may be important in the etiology of uterine leiomyomata.
Although many women receive diagnoses of uterine leiomyomata on the basis of ultrasound or pelvic examination, little is known regarding the rate at which such tumors progress to a point where a woman requires surgical intervention, or which characteristics of the woman or the tumors predict such treatment. In the Maine Women's Health Study, which included only women whose uteri were >= 8 weeks gestational age, 25% of patients had surgery within 1 year of diagnosis (
50). In a study including patients with a broader range of uterine sizes, approximately 26% had surgery within 5 years (
51). The risk of progressing to hysterectomy in that study was higher among women who presented with abnormal bleeding (RR = 1.7; 95% CI 1.1-2.6) or who had a history of surgery for pelvic adhesions (RR = 4.8; 95% CI 2.6-8.9) or gallbladder removal (RR = 2.4; 95% CI 1.1-5.3), and lower among women who had been prescribed oral contraceptives or progestins for relief of symptoms (RR = 0.5; 95% CI 0.3-0.8) (
51). Notably, there was little difference in the risk of progressing to hysterectomy between black and white women, or related to uterine leiomyomata risk factors identified in epidemiologic studies, such as absence of cigarette smoking and nulliparity. Future studies should consider not only medical factors and lifestyle characteristics but also noninvasive measures of uterine leiomyomata characteristics. For example, a recent study suggested that increased tumor vascularity, as measured by Doppler ultrasonography, is strongly associated with increased myoma growth (
52).
Similarly, only two studies have examined the rates and predictors of uterine leiomyomata recurrence following myomectomy (53,54). The risk of recurrence was lower among women with only a single myoma removed and among women who bore at least one child following myomectomy (53,54). As ultrasound-detected uterine leiomyomata are found in nearly one-half of patients within 5 years of surgery (53), studies aimed at identifying risk factors for recurrence could lead to the development of approaches to preventing a considerable number of repeated surgeries and associated costs.
Molecular Epidemiology
Epidemiologic studies of uterine leiomyomata conducted to date have relied on interviews, questionnaires, and medical records for information on potential risk factors. Although the types of data collected through these methods will remain important in future population research, epidemiologic studies that incorporate biologic measures of exposure and disease have considerable potential to test existing and new hypotheses, and thus clarify the etiology of uterine leiomyomata and associated risk factors. For example, blood levels of steroid hormones and other factors (e.g., insulin and insulinlike growth factors, infectious agents, environmental contaminants) could be studied in relation to the future development of uterine leiomyomata and to the recurrence of leiomyomata following myomectomy. It would also be valuable to conduct epidemiologic and clinical studies in which uterine leiomyomata patients are stratified according to ultrasonographic, histopathologic, or molecular characteristics of the tumors. One study employing such molecular fingerprinting found that uterine leiomyomata with abnormal karyotypes tended to be larger than tumors with normal karyotypes (55). By identifying molecularly defined subsets of uterine leiomyomata--such as those expressing high-mobility group proteins (56) or growth factors (57)--that are associated with particular risk factors (e.g., obesity), it may be possible to obtain more specific information as to underlying etiologic mechanisms.
Molecular epidemiologic studies also could add significantly to our understanding of the contribution of inherited genetics in uterine leiomyomata etiology. At present, almost nothing is known regarding genetic factors that affect a woman's risk of developing these tumors. It seems likely that, as with most common adult-onset diseases, genetic influences on uterine leiomyomata will fall into two broad categories: susceptibility genes and disease genes. The former would contribute to uterine leiomyomata by having common variants that confer relatively weak to modest increased risks, possibly in combination with lifestyle or environmental factors or with other such susceptibility genes. One broad class of potential susceptibility genes includes those that encode enzymes and receptors involved in the synthesis, metabolism, and signaling of steroid hormones. For example, a common variant of the cytochrome P450 17 gene has been associated with increased estradiol levels in young women (58). To the extent that carrying this variant is associated with uterine leiomyomata, an etiologic role for excess estrogen would be more plausible. Similarly, studies of the thermolabile genetic variant of the catechol O-methyltransferase gene (59) would allow a test of the hypothesis, raised by laboratory studies (60,61), that catechol metabolites of estradiol contribute to uterine leiomyomata pathogenesis. Disease-causing genes are those for which mutations exist, conferring an extremely high risk of developing uterine leiomyomata, and would contribute (in conjunction with lifestyle and environmental factors) to the clustered occurrence of these tumors within families. The risk of uterine leiomyomata associated with family history of this condition has only been reported by one study (21), and the results suggest strong familial aggregation. However, such findings could be due to reporting bias among women with uterine leiomyomata. Additional studies are needed to confirm familial aggregation of uterine leiomyomata. If confirmed, research will be needed to quantify the relationship, to determine the patterns of inheritance, and to discover the genes and associated mutations that might contribute to uterine leiomyomata occurrence within families.
Current Research
Most of the studies reviewed above were not originally planned to investigate risk factors for uterine leiomyomata. As such, many of the issues regarding case and control definition, ascertainment, and characterization were not addressed. We are aware of four studies that were designed specifically to explore the epidemiology of uterine leiomyomata and that have recently completed recruitment and data collection. These studies are being conducted at the University of Maryland (K. Kjerulff, Principal Investigator), the Johns Hopkins University (E. Faerstein, Principal Investigator), the Fred Hutchinson Cancer Research Center (S. Schwartz, Principal Investigator), and the National Institute of Environmental Health Sciences (D. Baird, Principal Investigator). Each of these studies collected extensive data on potential risk factors (including, in two studies, dietary intake) using in-person or telephone interviews and questionnaires. Several have used vaginal ultrasound exams to identify as controls women who do not have uterine leiomyomata. As the data from these new studies are analyzed and published, the relationship between common exposures such as use of oral contraceptives, smoking, alcohol use, and physical activity should become clarified. All studies collected data on reproductive history as well, so that the potentially protective role of pregnancy can be examined in greater detail. In addition, two studies collected blood samples from many of the participants so that genetic factors and other exposure biomarkers can be examined.
Despite the enormous impact that uterine leiomyomata have on the health of women of reproductive age, surprisingly few studies have been conducted to identify risk factors. Some consistent patterns have emerged in the nearly 15 years since the publication of the first formal epidemiologic study of these tumors: increased risks among black women and decreased risks among postmenopausal women, women who have had live births, and women who smoke cigarettes. Given our lack of knowledge of the mechanisms underlying these relationships, none provide possible avenues toward the primary prevention of uterine leiomyomata. Because most of these studies were not designed specifically to examine risk factors for uterine leiomyomata, methodologic issues limit the interpretation of the findings. Ongoing and future investigations, planned and executed with these methodologic challenges in mind, should clarify previously identified relationships. At present we have insufficient knowledge of the specific physiologic processes underlying the risk factors that have been tentatively identified. Although it is clear that the growth of uterine leiomyomata depends strongly on a woman's steroid hormone milieu, progress identifying modifiable risk factors must involve the development and testing of hypotheses that incorporate other pathologic mechanisms. Recent advances in molecular genetics present opportunities for epidemiologic studies of uterine leiomyomata to incorporate biomarkers of somatic changes found in these tumors and to examine inherited genetic factors related to possible causal physiologic mechanisms. In short, in the new millennium epidemiologic research should yield important advances in our understanding of the etiology of uterine leiomyomata.
REFERENCES AND NOTES
1. Wilcox LS, Koonin LM, Pokras R, Strauss LT, Xia Z, Peterson HB. Hysterectomy in the United States, 1988-1990. Obstet Gynecol 83(4):549-555 (1994).
2. Cramer SF, Patel A. The frequency of uterine leiomyomas. Am J Clin Pathol 94(4):435-438 (1990).
3. Marshall LM, Spiegelman D, Barbieri RL, Goldman MB, Manson JE, Colditz GA, Willett WC, Hunter DJ. Variation in the incidence of uterine leiomyoma among premenopausal women by age and race. Obstet Gynecol 90:967-973 (1997).
4. Cramer DW. Epidemiology of myomas. Semin Reprod Endocrinol 10(4):320-324 (1992).
5. Baird DD, Schectman JM, Dixon D, Sandler DP. African Americans at higher risk than whites for uterine fibroids: ultrasound evidence [Abstract]. Am J Epidemiol 147(11):S90 (1998).
6. Barbieri RL, Andersen J. Uterine leiomyomas: the somatic mutation theory. Semin Reprod Endocrinol 10(4):301-309 (1992).
7. Rothman KJ, Greenland S. Modern Epidemiology. Philadelphia:Lippincott-Raven, 1998.
8. Walker AM. Observation and Inference. Chestnut Hill, MA:Epidemiology Resources Inc., 1991.
9. Velebil P, Wingo PA, Xia Z. Rate of hospitalization for gynecologic disorders among reproductive-age women in the United States. Obstet Gynecol 86:764-769 (1995).
10. Brett KM, Marsh JV, Madans JH. Epidemiology of hysterectomy in the United States: demographic and reproductive factors in a nationally representative sample. J Womens Health 6(3):309-316 (1997).
11. Ross RK, Pike M, Vessey MP, Bull D, Yeates D, Casagrande JT. Risk factors for uterine fibroids: reduced risk associated with oral contraceptives. Br Med J 293(6543):359-362 (1986).
12. Pokras R, Hufnagel VG. Hysterectomy in the United States, 1965-84. Am J Public Health 78:852-853 (1988).
13. Vessey MP, Villard-Mackintosh L, McPherson K, Coulter A, Yeates D. The epidemiology of hysterectomy: findings in a large cohort study. Br J Obstet Gynaecol 99(5):402-407 (1992).
14. Kjerulff KH, Langenberg P, Seidman JD, Stolley PD, Guzinski GM. Uterine leiomyomas. Racial differences in severity, symptoms and age at diagnosis. J Reprod Med 41(7):483-490 (1996).
15. Kjerulff K, Langenberg P, Guzinski G. The socioeconomic correlates of hysterectomies in the United States. Am J Public Health 83(1):106-108 (1993).
16. Meilahn EN, Matthews KA, Egeland G, Kelsey SF. Characteristics of women with hysterectomy. Maturitas 11(4):319-329 (1989).
17. Parazzini F, LaVecchia C, Negri E, Cecchetti G, Fedele L. Epidemiologic characteristics of women with uterine fibroids: a case-control study. Obstet Gynecol 72(6):853-857 (1988).
18. Romieu I, Walker AM, Jick S. Dererminants of uterine fibroids. Post Market Surveil 5:119-133 (1991).
19. Samadi AR, Lee NC, Flanders WD, Boring JR III, Parris EB. Risk factors for self-reported uterine fibroids: a case-control study. Am J Public Health 86(6):858-862 (1996).
20. Kawaguchi K, Fujii S, Konishi I, Iwai T, Nanbu Y, Nonogaki H, Ishikawa Y, Mori T. Immunohistochemical analysis of oestrogen receptors, progesterone receptors and Ki-67 in leiomyoma and myometrium during the menstrual cycle and pregnancy. Virchows Arch A Pathol Anat Histopathol 419(4):309-315 (1991).
21. Lumbiganon P, Rugpao S, Phandhu-fung S, Laopaiboon M, Vudhikamraksa N, Werawatakul Y. Protective effect of depot-medroxyprogesterone acetate on surgically treated uterine leiomyomas: a multicentre case-control study. Br J Obstet Gynaecol 103(9):909-914 (1995).
22. Marshall LM, Spiegelman D, Goldman MB, Manson JE, Colditz GA, Barbieri RL, Stampfer MJ, Hunter DJ. A prospective study of reproductive factors and oral contraceptive use in relation to the risk of uterine leiomyomata. Fertil Steril 70(3):432-439 (1998).
23. Parazzini F, Negri E, La Vecchia C, Chatenoud L, Ricci E, Guarnerio P. Reproductive factors and risk of uterine fibroids. Epidemiology 7(4):440-442 (1996).
24. Bernstein L, Pike MC, Ross RK, Judd HL, Brown JB, Henderson BE. Estrogen and sex hormone-binding globulin levels in nulliparous and parous women. J Natl Cancer Inst 74(4):741-745 (1985).
25. Dorgan JF, Reichman ME, Judd JT, Brown C, Longcope C, Schatzkin A, Campbell WS, Franz C, Kahle L, Taylor PR. Relationships of age and reproductive characteristics with plasma estrogens and androgens in premenopausal women. Cancer Epidemiol Biomark Prev 4(4):381-386 (1995).
26. Musey PI, Collins DC, Bradlow HL, Gould KG, Preedy JR. Effect of diet on oxidation of 17 beta-estradiol in vivo. J Clin Endocrinol Metab 65(4):792-795 (1987).
27. Marshall LM, Spiegelman D, Manson JE, Goldman MB, Barbieri RL, Stampfer MJ, Willett WC, Hunter DJ. Risk of uterine leiomyomata among premenopausal women in relation to body size and cigarette smoking. Epidemiology 9:511-517 (1998).
28. Sherman BM, Korenman SG. Measurement of serum LH, FSH, estradiol and progesterone in disorders of the human menstrual cycle: the inadequate luteal phase. J Clin Endocrinol Metab 39(1):145-149 (1974).
29. Westhoff C, Gentile G, Lee J, Zacur H, Helbig D. Predictors of ovarian steroid secretion in reproductive-age women. Am J Epidemiol 144(4):381-388 (1996).
30. Jump DB, Clarke SD, Thelen A, Liimatta M, Ren B, Badin MV. Dietary fat, genes, and human health. Adv Exp Med Biol 422:167-176 (1997).
31. Parazzini F, Negri E, La Vecchia C, Rabaiotti M, Luchini L, Villa A, Fedele L. Uterine myomas and smoking. Results from an Italian study. J Reprod Med 41(5):316-320 (1996).
32. Kornyei JL, Li X, Lei ZM, Rao CV. Restoration of human chorionic gonadotropin response in human myometrial smooth muscle cells by treatment with follicle-stimulating hormone (FSH): evidence for the presence of FSH receptors in human myometrium. Eur J Endocrinol 134(2):225-231 (1996).
33. MacMahon B, Trichopoulos D, Cole P, Brown J. Cigarette smoking and urinary estrogens. N Engl J Med 307(17):1062-1065 (1982).
34. Longcope C, Johnston CC Jr. Androgen and estrogen dynamics in pre- and postmenopausal women: a comparison between smokers and nonsmokers. J Clin Endocrinol Metab 67(2):379-383 (1988).
35. Zumoff B, Miller L, Levit CD, Miller EH, Heinz U, Kalin M, Denman H, Jandorek R, Rosenfeld RS. The effect of smoking on serum progesterone, estradiol, and luteinizing hormone levels over a menstrual cycle in normal women. Steroids 55(11):507-511 (1990).
36. Daniel M, Martin AD, Faiman C. Sex hormones and adipose tissue distribution in premenopausal cigarette smokers. Int J Obes Relat Metab Disord 16(4):245-254 (1992).
37. Cook LS, Weiss NS. Endometrial Cancer. In: Women and Health (Goldman MB, Hatch MC, eds). San Diego, CA:Academic Press, 1999.
38. Parazzini F, Negri E, LaVecchia C, Fedele L, Rabaiotti M, Luchini L. Oral contraceptive use and risk of uterine fibroids. Obstet Gynecol 79:430-433 (1992).
39. Ramcharan S, Pelligrin FA, Ray R, Hsu J-P. The Walnut Creek Contraceptive Drug Study. A prospective study of the side effects of oral contraceptives. NIH Publ No 81-564. Center Popul Res Monogr 3:69-74 (1981).
40. Royal College of General Practitioners. Oral Contraceptives and Health: An Interim Report from the Oral Contraception Study of the Royal College of General Practitioners. London:Pittman Medical, 1974.
41. Rein MS, Barbieri RL, Friedman AJ. Progesterone: a critical role in the pathogenesis of uterine myomas. Am J Obstet Gynecol 172:14-18 (1995).
42. Van de Ven WJM. Genetic basis of uterine leiomyoma: involvement of high mobility group protein genes. Eur J Obstet Gynecol 81:289-293 (1998).
43. Anand A, Chada K. In vivo modulation of Hmgic reduces obesity. Nat Genet 24:377-380 (2000).
44. Wyshak G, Frisch RE, Albright NL, Albright TE, Schiff I. Lower prevalence of benign diseases of the breast and benign tumours of the reproductive system among former college athletes compared to non-athletes. Br J Cancer 54(5):841-845 (1986).
45. Chiaffarino F, Parazzini F, La Vecchia C, Chatenoud L, Di Cintio E, Marsico S. Diet and uterine myomas. Obstet Gynecol 94:395-398 (1999).
46. Hodges LC, Bergerson JA, Hunter DS, Walker CL. Estrogenic effects of organochlorine pesticides on uterine leiomyoma cells in vitro. Toxicol Sci 54(2):355-364 (2000).
47. Moss NS, Benditt EP. Human atherosclerotic plaque cells and leiomyoma cells. Am J Pathol 78(2):175-187 (1975).
48. Chadwick EG, Connor EJ, Hanson ICG, Joshi VV, Abu-Farsakh H, Yogev R, McSherry G, McClain K, Murphy SB. Tumors of smooth-muscle origin in HIV-infected children. J Am Med Assoc 263(23):3182-3184 (1990).
49. Datubo-Brown DD. Keloids: a review of the literature. Br J Plast Surg 43(1):70-77 (1990).
50. Carlson KJ, Miller BA, Fowler FJ Jr. The Maine Women's Health Study. II: Outcomes of nonsurgical management of leiomyomas, abnormal bleeding, and chronic pelvic pain. Obstet Gynecol 83(4):566-572 (1994).
51. Weber AM, Mitchinson AR, Gidwani GP, Mascha E, Walters MD. Uterine myomas and factors associated with hysterectomy in premenopausal women. Am J Obstet Gynecol 176:1213-1219 (1997).
52. Tsuda H, Kawabata M, Nakamoto O, Yamamoto K. Clinical predictors in the natural history of uterine leiomyoma: preliminary study. J Ultrasound Med 17(1):17-20 (1998).
53. Candiani GB, Fedele L, Parazzini F, Villa L. Risk of recurrence after myomectomy. Br J Obstet Gynecol 98(4):385-389 (1991).
54. Fedele L, Parazzini F, Luchini L, Mezzopane R, Tozzi L, Villa L. Recurrence of fibroids after myomectomy: a transvaginal ultrasonographic study. Hum Reprod 10(7):1795-1796 (1995).
55. Rein MS, Powell WL, Walters FC, Weremowicz S, Cantor RM, Barbieri RL, Morton CC. Cytogenetic abnormalities in uterine myomas are associated with myoma size. Mol Hum Reprod 4(1):83-86 (1998).
56. Morton CC. Warner-Lambert/Parke-Davis Award Lecture. Many tumors and many genes--genetics of uterine leiomyomata. Am J Pathol 153(4):1015-1020 (1998).
57. Andersen J. Growth factors and cytokines in uterine leiomyomas. Semin Reprod Endocrinol 14(3):269-282 (1996).
58. Feigelson HS, Shames LS, Pike MC, Coetzee GA, Stanczyk FZ, Henderson BE. Cytochrome P450c17alpha gene (CYP17) polymorphism is associated with serum estrogen and progesterone concentrations. Cancer Res 58(4):585-587 (1998).
59. Lachman HM, Papolos DF, Saito T, Yu YM, Szumlanski CL, Weinshilboum RM. Human catechol-O-methyltransferase pharmacogenetics: description of a functional polymorphism and its potential application to neuropsychiatric disorders. Pharmacogenetics 6(3):243-250 (1996).
60. Reddy VVR, Hanjani P, Rajan R. Synthesis of catechol estrogens by human uterus and leiomyoma. Steroids 37(3):195-203 (1981).
61. Liehr JG, Ricci MJ, Jefcoate CR, Hannigan EV, Hokanson JA, Zhu BT. 4-Hydroxylation of estradiol by human uterine myometrium and myoma microsomes: implications for the mechanism of uterine tumorigenesis. Proc Natl Acad Sci USA 92(20):9220-9224 (1995).
Last Updated: October 3, 2000
