Uterine leiomyoma (UL) occurs in 20-40% of women, mainly in their late reproductive years; it is widely quoted that UL frequency is much higher among blacks. Symptomatic cases comprise up to 50% of the total; abnormal vaginal bleeding, pelvic pain, infertility and adverse pregnancy events may occur. An unknown proportion of the cases is estimated to undergo malignant transformation into leiomyosarcomas. Although conservative management is now more common, UL is still the leading cause of hysterectomy in the United States. Clinical observations support the idea that unopposed estrogenic stimulation of the myometrium is associated with UL. However, only three epidemiologic studies have been published; consistent findings were a positive association with nulliparity, and a negative association with menopause and smoking. Conflicting data emerged regarding the role of oral contraceptives, replacement estrogens, obesity, and duration of menstrual life. These studies were potentially subject to selection bias, as they were based only on surgical cases of a predominantly asymptomatic or mildly symptomatic condition. A practice-based case-control study will be conducted among Baltimore area women under the care of a group of gynecologists affiliated with The Johns Hopkins Medical Institutions. The primary aim is to determine whether UL is associated with several markers of increased uterine exposure to unopposed endogenous and exogenous estrogens. Other factors to be studied include race and socioeconomic status; atherogenic risk factors (self-reported obesity, hypertension, and diabetes, and smoking history); family history of UL; and possible sources of uterine irritation (pelvic inflammatory disease, use of intrauterine contraceptive device, vaginal douching, and perineal use of talc). Cases of UL diagnosed by these practitioners, and histologically or sonographically confirmed, will be ascertained. Eligible cases will be area residents, premenopausal, with no history or presence of invasive breast or gynecologic cancers. Controls will be frequency matched to cases on age within 5-year groups, randomly selected among women visiting the same physician for routine gynecologic evaluation during the calendar month when the case's tumor was first clinically suspected. Controls will be excluded if they are postmenopausal, hysterectomized, have clinical UL, or history or presence of invasive breast or gynecologic cancers. About 350 cases and 350 controls are expected to be ascertained in nine months. Trained female interviewers will utilize a structured questionnaire in telephone interviews; in addition, selected information will be abstracted from medical records. Data on medical monitoring and health behavior indicators will also be collected, to allow control for confounding. Bivariate and stratified analyses will be conducted, and logistic regression models will be fitted; agreement between interview and medical records data will be studied. Study findings are expected to provide currently undefined demographic and epidemiologic information for preventive strategies, and for identification of high risk groups in need of monitoring.
