American Journal of Epidemiology Vol. 153, No. 1 : 27-29
Copyright © 2001 by The Johns Hopkins University School of Hygiene and Public Health
ORIGINAL CONTRIBUTIONS |
Invited Commentary: Studying the Epidemiology of Uterine Leiomyomata—Past, Present, and Future
1 Program in Epidemiology, Division of Public Health Sciences, Fred Hutchinson Cancer Research Center, Seattle, WA.
2 Department of Epidemiology, School of Public Health and Community Medicine, University of Washington, Seattle, WA.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONNew approaches and findings...The future of UL...REFERENCES |
|---|
Uterine leiomyomata have a substantial impact on women's reproductive health, but epidemiologists have focused relatively little energy on identifying risk factors for this condition. Only a handful of studies, most of which were not designed to address methodological challenges posed by these tumors, have been conducted. These studies focused almost exclusively on reproductive and hormonal characteristics as possible risk factors, but consistent relations have not emerged. Three new reports (from two studies) in this issue of the Journal target the paucity of information on uterine leiomyomata risk factors by testing novel hypotheses, by employing designs that incorporate subclinical tumors or account for variable management of clinically recognized disease, or by using a combination of these approaches. The success of these strategies and the contributions of the new findings are discussed. Recommendations are made for a program of research that eventually could improve our knowledge of uterine leiomyomata etiology and yield clues to the prevention of associated morbidity.
leiomyoma; uterine diseases
Abbreviations: UL, uterine leiomyomata
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONNew approaches and findings...The future of UL...REFERENCES |
|---|
Since 1986, when the first formal epidemiologic investigation of uterine leiomyomata (UL) was published (1
), reports have appeared from only seven studies (2, 3). Given the large number of US women affected by these "benign" tumors—they are the primary indication for 200,000–300,000 hysterectomies performed each year among premenopausal women (4)—it is striking that epidemiologists have committed so little energy to understanding the etiology of this condition. The paucity of population research on UL is all the more disappointing given that few previous studies were designed with the specific goal of identifying risk factors for these tumors. The extant studies have largely examined risk factors for hysterectomy-treated UL, focusing on the hypothesis that the development of these tumors is strongly tied to a woman's hormonal milieu (5, 6). Thus, characteristics directly or indirectly reflecting a woman's exposure to estrogens and progestogens (e.g., menstrual history, childbearing, exogenous hormone use, cigarette smoking, obesity) have been assessed using questionnaires or medical records. From these investigations, it seems fairly certain that postmenopausal women are at 70–90 percent reduced risk of UL, while having had a live birth is probably associated with a more modest reduction in risk (20–50 percent) (2). Unfortunately, there is little consistency in results for other markers of hormonal status, and studies of alternative etiologic mechanisms are virtually nonexistent. Perhaps the most noteworthy epidemiologic feature of UL is the far greater risk among Black women as compared with White women (7–9).
Two features of UL pose key challenges for epidemiologic research. First, these tumors develop in a high proportion (>45 percent) of women by the fifth decade of life without coming to clinical attention (10). To the extent that efforts are not made to identify subclinical tumors, either associations will be biased toward the null if misclassification is nondifferential with respect to risk factors, or spuriously positive or negative associations will result if differential misclassification is present (e.g., if disease ascertainment varies with health-care-seeking behavior or lifestyle factors). Second, among those tumors that come to clinical attention, only a portion of affected women progress to hysterectomy. The decision for surgical intervention primarily reflects the severity of a woman's symptoms (bleeding or pelvic pain) and her desire to maintain childbearing potential (11). It is not difficult to imagine how studies restricted to hysterectomy-treated cases could yield inconsistent results, given that determinants of hysterectomy may be related to risk factors under investigation and that measurement of risk factors during the etiologically relevant periods could be compromised by inclusion of characteristics related to nonsurgical management following an initial clinical diagnosis. The increasing availability of myomectomy (surgical removal of individual leiomyomata) further limits the strength of inferences that should be drawn from studies of hysterectomy-managed cases.
|
New approaches and findings in UL epidemiology |
|---|
|
TOPABSTRACTINTRODUCTIONNew approaches and findings...The future of UL...REFERENCES |
|---|
The three reports (from two studies) appearing in this issue of the Journal (12
–14) are noteworthy because they attempt to address these methodological challenges and to answer questions about risk factors in ways that eventually could improve our knowledge of UL epidemiology. In the study by Chen et al. (12), all women had received a laparoscopic examination, which provided an opportunity for the investigators to reduce bias due to disease misclassification. The sensitivity of laparoscopy for detecting UL is likely to be low, however, because: 1) in contrast to transvaginal ultrasound, tumors located completely within the myometrium or extending into the endometrium are likely to be missed; and 2) the main tubal sterilization study apparently did not specifically include a protocol for determining the presence or absence of uterine pathologies. Thus, potential gains in validity may not have been realized. Chen et al. (12) also are the first to have examined possible differences in UL risk factors between Black and White women. Unfortunately, the conclusions that can be drawn from these analyses are extremely limited, because Chen et al. used an automated backwards elimination modeling technique for scientific inference and reporting (see table 4 of their paper). Thus, rather than provide a complete set of multivariate-adjusted associations that can be compared between Black and White women, Chen et al. have included only some of the many factors under investigation in table 4, and there are blank spaces where certain odds ratio estimates are missing for either Black or White women.
Chen et al. acknowledge the challenge in tying exposure histories to the onset of UL, given the apparently long preclinical phase and paucity of knowledge about the natural history of these tumors. However, the problem is compounded in their investigation by the inclusion of a substantial number of cases with a history of UL, as well as several exposures (birth control methods, smoking) on which no historical details were available. Among the cases with a history of UL, many of the lifestyle, reproductive, or menstrual characteristics ascertained as of entry into the original tubal sterilization study clearly postdated the development and clinical diagnosis of the UL, and may well have changed in the interim. Interestingly, the authors state that their results are unaffected by the inclusion of these women (12). Rather than feel reassured, however, one might be concerned that, with the possible exception of number of living children, the exposure ascertainment for most characteristics missed the etiologically relevant period regardless of the case definition. The work by Chen et al. reminds us, ultimately, that epidemiologic research on UL has relied for too long on studies that were not designed to address the key methodological challenges presented by this disease.
The two reports from a study designed de novo to investigate risk factors for UL (13, 14) are therefore particularly welcome. As one would hope with such a study, the definition of cases and controls, the breadth and detail of data on potential risk factors, and the recording of clinical characteristics of the cases' diagnoses allowed more rigorous testing of existing hypotheses, as well as investigation of novel etiologic mechanisms. For example, by defining as cases women recently ascertained to have a first diagnosis of UL, Faerstein et al. (13, 14) minimized biases arising from factors that influence progression to surgical intervention. In addition, through collection of data on health-care-seeking behavior, they were able to address issues of ascertainment bias in the analysis.
Collection of data on each case's clinical characteristics strengthened Faerstein et al.'s analysis in at least two ways. First, associations were examined after cases were stratified according to the presence or absence of symptoms, and the similarity of the results supports the absence of ascertainment bias due to inclusion of asymptomatic but clinically detected cases. Second, the authors used information from diagnostic ultrasound examinations to address the question, Is a history of infertility a cause or a consequence of UL? In finding that the association with infertility in their data was limited to cases with submucous UL (those that protrude into the endometrial cavity and thus could cause infertility by interfering with implantation), Faerstein et al. have provided a provocative refutation of previous studies (15, 16). While this novel analysis was not without limitations (data on tumor site were available only for a small subset of the cases, and there were no formal tests of heterogeneity by tumor site), it demonstrates the benefits of focused data collection in testing existing hypotheses regarding UL etiology.
Faerstein et al. push the field forward even further in their second paper (14), with the first systematic investigation of etiologic relations that extend beyond hypotheses focused on steroid hormones. The authors report moderate independent associations with clinically recognized high blood pressure, which they view as a marker of a "pro-atherogenic" state. While Faerstein et al. base their interpretation of the atherogenic hypothesis on prior studies of smooth muscle cell physiology, there is conflicting information as to the relative roles of smooth muscle cell proliferation versus extra cellular matrix restructuring in driving the growth of UL (17–19). The hypothesis that local irritants and/or infectious agents could be etiologically relevant to UL has much less direct support from nonepidemiologic studies. As such, not only do the reported associations with pelvic inflammatory disease, talc use, and possibly Chlamydia trachomatis infection provide leads for epidemiologists to follow (some of which are outlined by Faerstein et al.), but the findings also present important opportunities for new investigations in animal models, such as the Eker rat (20).
|
The future of UL research |
|---|
|
TOPABSTRACTINTRODUCTIONNew approaches and findings...The future of UL...REFERENCES |
|---|
A variety of research strategies should be pursued to advance our understanding of the epidemiology and pathogenesis of UL. To mitigate the impact of disease misclassification, investigators could use transvaginal ultrasound examinations within a defined population to achieve more complete ascertainment of asymptomatic, clinically silent UL. This approach should increase the sensitivity for detecting and quantifying associations with risk factors and should permit simultaneous analyses in relation to clinically apparent UL, clinically silent UL detected by ultrasound, and UL-free controls that may allow for more powerful tests if associations with risk factors are graded with respect to disease severity. Differences in risk factors between clinically apparent and ultrasound-detected silent UL may also help distinguish among factors that are involved in the initiation or early development of these tumors and those factors promoting growth in the later stages. In settings where the use of transvaginal ultrasound examinations is not feasible, restricting the study population to relatively young women may enhance identification of risk factors if the prevalence of clinically silent UL is found to be fairly low in this group. Focusing on younger women might be advantageous for additional reasons. First, young women stand to suffer the most "years of reproductive potential lost" from hysterectomy. Second, patients with early-onset UL may be those most likely to exhibit familial aggregation of this condition (21
) and thus potentially show stronger associations with possible genetic risk factors, such as polymorphisms of genes involved in steroid hormone synthesis, metabolism, and signaling. While rigorously designed case-control studies will continue to be important tools for identifying risk factors, it will be essential to assemble cohorts of women among whom the development and/or growth of UL can be assessed longitudinally in relation to lifestyle, medical, genetic, and physiologic characteristics. Finally, important contributions to the etiology of these tumors could come from studies examining the relation between a woman's risk factor status and molecular characteristics of UL that she develops. Such "molecular fingerprinting" studies, for example, might be able to clarify inconsistent epidemiologic findings regarding the role of obesity, since somatic changes in the high mobility group protein gene family are common in UL (22), and model experimental systems suggest that high mobility group proteins are involved in mammalian growth (23).
UL are a high research priority in women's health (24). In October 1999, the first scientific conference dedicated to emerging research on UL was held (25), and the proceedings featured many examples of how interdisciplinary endeavors may yield progress in understanding the pathogenesis of these tumors. As with other diseases, epidemiologists should take the lead in the next generation of studies that will identify risk factors for UL and determine approaches to reducing the burden of these tumors.
|
ACKNOWLEDGMENTS |
|---|
This work was supported in part by National Institutes of Health grants HD-29819 (National Institute of Child Health and Human Development) and ES-08305 (National Institute of Environmental Health Sciences).
The author is especially appreciative of the rewarding discussions held on this topic with Drs. Donna Baird, Lynn Marshall, and Kristen Kjerulff. The author thanks Dr. Mary Anne Rossing for providing feedback on an earlier version of this commentary.
|
NOTES |
|---|
Reprint requests to Dr. Stephen M. Schwartz, Fred Hutchinson Cancer Research Center, 1100 Fairview Avenue North, MP 381, Seattle, WA 98109-1024 (e-mail: sschwart{at}fhcrc.org).
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONNew approaches and findings...The future of UL...REFERENCES |
|---|
- Ross RK, Pike MC, Vessey MP, et al. Risk factors for uterine fibroids: reduced risk associated with oral contraceptives. Br Med J (Clin Res Ed) 1986;293:359–62.
- Schwartz SM, Marshall LM. Uterine leiomyomata. In: Hatch M, Goldman M, eds. Women and health. New York, NY: Academic Press, Inc, 1999:240–52.
- Schwartz SM, Marshall LM, Baird DD. Epidemiologic contributions to understanding the etiology of uterine leiomyomata. Environ Health Perspect 2000;108:821–7.
- Wilcox LS, Koonin LM, Pokras R, et al. Hysterectomy in the United States, 1988–1990. Obstet Gynecol 1994;83:549–55.[Web of Science][Medline]
- Rein MS, Barbieri RL, Friedman AJ. Progesterone: a critical role in the pathogenesis of uterine myomas. Am J Obstet Gynecol 1995;172:14–18.[Web of Science][Medline]
- Nowak RA. Fibroids: pathophysiology and current medical treatment. Baillieres Best Pract Res Clin Obstet Gynaecol 1999;13:223–38.[Medline]
- Kjerulff KH, Langenberg P, Seidman JD, et al. Uterine leiomyomas: racial differences in severity, symptoms and age at diagnosis. J Reprod Med 1996;41:483–90.[Web of Science][Medline]
- Marshall LM, Spiegelman D, Barbieri RL, et al. Variation in the incidence of uterine leiomyoma among premenopausal women by age and race. Obstet Gynecol 1997;90:967–73.[Web of Science][Medline]
- 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 1997;6:309–16.[Web of Science][Medline]
- Baird DD, Schectman JM, Dixon D, et al. African Americans at higher risk than whites for uterine fibroids: ultrasound evidence. (Abstract). Am J Epidemiol 1998;147(suppl):S90.
- Davies A, Magos AL. Indications and alternatives to hysterectomy. Baillieres Clin Obstet Gynaecol 1997;11:61–75.[Web of Science][Medline]
-
Chen C-R, Buck GM, Courey NG, et al. Risk factors for uterine fibroids among women undergoing tubal sterilization. Am J Epidemiol 2001;153:20–6.
[Abstract/Free Full Text] -
Faerstein E, Szklo M, Rosenshein N. Risk factors for uterine leiomyoma: a practice-based case-control study. I. African-American heritage, reproductive history, body size, and smoking. Am J Epidemiol 2001;153:1–10.
[Abstract/Free Full Text] -
Faerstein E, Szklo M, Rosenshein N. Risk factors for uterine leiomyoma: a practice-based case-control study. II. Athero- genic risk factors and potential sources of uterine irritation. Am J Epidemiol 2001;153:11–19.
[Abstract/Free Full Text] - Parazzini F, Negri E, La Vecchia C, et al. Reproductive factors and risk of uterine fibroids. Epidemiology 1996;7:440–2.[Web of Science][Medline]
- Marshall LM, Spiegelman D, Goldman MB, et al. A prospective study of reproductive factors and oral contraceptive use in relation to the risk of uterine leiomyomata. Fertil Steril 1998;70:432–9.[Web of Science][Medline]
- Rein MS, Barbieri RL, Welch W, et al. The concentrations of collagen-associated amino acids are higher in GnRH agonist-treated uterine myomas. Obstet Gynecol 1993;82:901–5.[Web of Science][Medline]
- Kawamura N, Ito F, Ichimura T, et al. Correlation between shrinkage of uterine leiomyoma treated with buserelin acetate and histopathologic findings of biopsy specimen before treatment. Fertil Steril 1997;68:632–6.[Web of Science][Medline]
- Upadhyaya NB, Doody MC, Googe PB. Histopathological changes in leiomyomata treated with leuprolide acetate. Fertil Steril 1990;54:811–14.[Web of Science][Medline]
- Howe SR, Everitt JL, Gottardis MM, et al. Rodent model of reproductive tract leiomyomata: characterization and use in preclinical therapeutic studies. Prog Clin Biol Res 1997;396:205–15.[Web of Science][Medline]
- Schwartz S, Voigt L, Tickman E, et al. Familial aggregation of uterine leiomyomata. (Abstract). Am J Epidemiol 2000;151(suppl):S10.
- van de Ven WJ. Genetic basis of uterine leiomyoma: involvement of high mobility group protein genes. Eur J Obstet Gynecol Reprod Biol 1998;81:289–93.[Web of Science][Medline]
- Anand A, Chada K. In vivo modulation of Hmgic reduces obesity. Nat Genet 2000;24:377–80.[Web of Science][Medline]
- National Institutes of Health. Agenda for research on women's health for the 21st century: a report of the Task Force on the NIH Women's Health Research Agenda for the 21st Century. Executive summary. Bethesda, MD: National Institutes of Health, 1999.
- McBride G. Benign but no longer forgotten—fibroids get their own conference. Lancet 1999;354:1450.[Web of Science][Medline]
CiteULike 
Connotea 
Del.icio.us What's this?
Related articles in Am. J. Epidemiol.:
- Risk Factors for Uterine Leiomyoma: A Practice-based Case-Control Study. I. African-American Heritage, Reproductive History, Body Size, and Smoking
- Eduardo Faerstein, Moyses Szklo, and Neil Rosenshein
Am. J. Epidemiol. 2001 153: 1-10.[Abstract] [FREE Full Text] - Risk Factors for Uterine Leiomyoma: A Practice-based Case-Control Study. II. Atherogenic Risk Factors and Potential Sources of Uterine Irritation
- Eduardo Faerstein, Moyses Szklo, and Neil B. Rosenshein
Am. J. Epidemiol. 2001 153: 11-19.[Abstract] [FREE Full Text] - Risk Factors for Uterine Fibroids among Women Undergoing Tubal Sterilization
- Chao-Ru Chen, Germaine M. Buck, Norman G. Courey, Kimberly M. Perez, and Jean Wactawski-Wende
Am. J. Epidemiol. 2001 153: 20-26.[Abstract] [FREE Full Text]
This article has been cited by other articles:
|
|
 |
|
  A.R.W. Williams, H.O.D. Critchley, J. Osei, S. Ingamells, I.T. Cameron, C. Han, and K. Chwalisz The effects of the selective progesterone receptor modulator asoprisnil on the morphology of uterine tissues after 3 months treatment in patients with symptomatic uterine leiomyomata Hum. Reprod., June 1, 2007; 22(6): 1696 - 1704. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 D. D. Baird, D. B. Dunson, M. C. Hill, D. Cousins, and J. M. Schectman Association of Physical Activity with Development of Uterine Leiomyoma Am. J. Epidemiol., January 15, 2007; 165(2): 157 - 163. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
R. Boynton-Jarrett, J. Rich-Edwards, S. Malspeis, S. A. Missmer, and R. Wright A Prospective Study of Hypertension and Risk of Uterine Leiomyomata Am. J. Epidemiol., April 1, 2005; 161(7): 628 - 638. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
H. Klip, F. E. van Leeuwen, R. Schats, C.W. Burger, and for the OMEGA project group Risk of benign gynaecological diseases and hormonal disorders according to responsiveness to ovarian stimulation in IVF: a follow-up study of 8714 women Hum. Reprod., September 1, 2003; 18(9): 1951 - 1958. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
C.-R. Chen, G. M. Buck, J. Wactawski-Wende, and K. M. Perez Chen et al. Respond to "Studying the Epidemiology of Uterine Leiomyomata" by Schwartz Am. J. Epidemiol., January 1, 2001; 153(1): 30 - 30. [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||