An epidemiologic study of epithelial ovarian malignancies : with a focus on hormone-related factors

Abstract

The main purpose of this thesis was to examine estrogen and progestin effects of hormone replacement therapy, in relation to the risk for epithelial ovarian neoplasias. Other intentions were to assess reproductive factors, oral contraceptive use, previous gynecologic surgeries, and family history of ovarian cancer according to the risk of ovarian neoplasias, and to connect these findings to different hypotheses regarding ovarian carcinogenesis. Finally, part of the work investigated whether preventive life-style modifications exist against these malignancies.

All results were based on a case-control study conducted in all of Sweden between October 1st, 1993 and December 31st, 1995. Subjects were 828 women with newly diagnosed, histologically confirmed epithelial ovarian neoplasias and 3899 population controls, all 50-74 years of age. For all epithelial ovarian cancer analyses (EOC, n=655) we used the histology classification reported by field pathologists, while borderline ovarian tumor data (BOT, n=193) were analyzed according to the reviewed histologies. Data were collected through mailed questionnaires. Odds ratios (OR) and 95% confidence intervals (CI) were estimated by the use of unconditional logistic regression.

Our key findings were elevated risks of EOC among ever users compared with never users of both unopposed estrogens (OR= 1.43, 95% Cl 1.02-2.00), and estrogens supplemented with sequential progestins (OR=1.54, 95% Cl 1.15-2.05), while no excess risk appeared with the use of estrogens continuously combined with progestins (OR=1.02, 95% Cl 0.73-1.43). Increased risks were seen for serous, mucinous, and endometrioid EOC, and for all EOC the greatest risk increase appeared with hormone use exceeding 10 years. Parity reduced EOC risk (OR=0.61 95% Cl 0.46-0.8 1) for uniparous compared with nulliparous women, and multiparity decreased EOC risk further. Ever compared with never use of oral contraceptives protected from EOC overall (OR=0.73 95% Cl 0.59-0.90), and for serous, endometrioid, and clear-cell, but not mucinous EOC.

The protection increased with duration of oral contraceptive use, and persisted up to 25 years after cessation of use. The risk of EOC appeared to decrease with incomplete pregnancies, early menopausal age, late age at first birth, unilateral oophorectomy, and coffee consumption; increase with ovarian cancer family history; and was unrelated to menarcheal age, lactation, irregular menses, menopausal symptoms, and alcohol use. Most epidemiologic findings on EOC risk seem consistent with the retrograde transportation hypothesis, which may operate through retrograde menstrual bleeding. A life-style including leisure time physical activity and a normal body mass index (BMI) may reduce EOC risk.

We observed no protection from serous (OR=1.40 95% Cl 0.87-2.26) and mucinous BOT (OR=1.04 95% Cl 0.61-1.79), among ever compared with never users of oral contraceptives, and duration of use was unrelated to BOT risk. An excess risk of serous BOT appeared among ever compared with never users of unopposed estrogens (OR=2.07 95% CI 1.08-3.95), but not for HRT including sequential or continuous progestins. No association between HRT and mucinous BOT was seen. The risk of BOT seemed to decrease with parity, lactation, incomplete pregnancies, tubal ligation, and leisure time physical activity; increase with irregular menstrual cycles, ovarian cancer family history and alcohol use; and was unrelated to age at first and last birth, and menopausal symptoms. The OR of serous BOT for the highest compared with the lowest BMI category was 6.47 (95% Cl 3.09-13.5). We suggest that hormonal situations where estrogens are not counteracted by progestins may increase the risk of serous BOT.