Oral contraceptives and breast cancer

Abstract

A population-based case-control study of oral contraceptive use and breast cancer was carried out among young women (less than 43 years of age) at Group Health Cooperative of Puget Sound, Seattle, Washington. Use of oral contraceptives before first pregnancy did not materially differ between cases or controls. The rate ratio estimate of breast cancer incidence in women who had used oral contraceptives before first pregnancy compared to those who had not was 0.9 (95% CI = 0.4, 2.1). There were no meaningful patterns of association between breast cancer and duration of use or formulation of oral contraceptive used before first pregnancy.

PIP: A case-controlled study of the association between oral contraceptive use before pregnancy and breast cancer diagnosed before age 43 was conducted using members of the Group Health Cooperative of Puget Sound, Seattle, Washington. 95 cases were selected from women with tumors diagnosed between July 1978 and December 1983; 136 controls were selected based on duration of membership, and matched for age and duration of membership, and matched for age and duration of membership in the plan. Analytic methods multiple logistic regression; ratio of the rate from coefficients of the logistic regressions; tests for trend were based on comparisons of coefficients to their standard errors using non-categorical models. The occurrence of nulliparity, late age of 1st pregnancy, history of breast lumps, history of maternal cancer and early age of menarche was higher in cases to the extent found in previous studies. Oral contraceptive use was slightly less frequent among cases than controls: 61% vs. 71%, a rate ratio of 0.9 (n.s.) after controlling for age of 1st pregnancy, history of breast lumps, age of menarche and history of maternal cancer. There was no significant difference in total duration of use: the rate ratios were 0.7, 0.7, 1.4 and 1.9 for 1-4 years, 5-9 years, 10+ years and unknown, respectively. The rate ratio for use before 1st pregnancy was 0.9. Age at 1st pregnancy was the strongest confounder in the logistic regression model. Cases and controls did not differ in duration of use before pregnancy: the rate ratios were 0.3, 0.8, 1.3 and 0.3 for 1 year, 1-3 years, 4 or more years or unknown. The test for trend was not significant (p=0.2).