Oral contraceptive use by formulation and breast cancer risk by subtype in the Nurses' Health Study II: a prospective cohort study

Abstract

Oral contraceptive use has been associated with a higher breast cancer risk; however, evidence for the associations between different oral contraceptive formulations and breast cancer risk, especially by disease subtype, is limited. This study aimed to evaluate the associations between oral contraceptive use by formulation and breast cancer risk by disease subtype. This prospective cohort study included 113,187 women from the Nurses' Health Study II with recalled information on oral contraceptive usage from 13 years of age to baseline (1989) and updated data on usage until 2009 collected via biennial questionnaires. A total of 5799 breast cancer cases were identified until the end of 2017. Multivariable Cox proportional hazards models estimated hazard ratios and 95% confidence intervals for the associations between oral contraceptive use and breast cancer risk overall and by estrogen and progesterone receptor and human epidermal growth factor receptor 2 status. Oral contraceptive use was evaluated by status of use (current, former, and never), duration of and time since last use independently and cross-classified, and formulation (ie, estrogen and progestin type). Current oral contraceptive use was associated with a higher risk for invasive breast cancer (hazard ratio, 1.31; 95% confidence interval, 1.09-1.58) when compared with never use, with stronger associations observed for longer durations of current use (>5 years: hazard ratio, 1.56; 95% confidence interval, 1.23-1.99; ≤5 years: hazard ratio, 1.19; 95% confidence interval, 0.95-1.49). Among former users with >5 years since cessation, the risk was similar to that of never users (eg, >5 to 10 years since cessation: hazard ratio, 0.99; 95% confidence interval, 0.88-1.11). Associations did not differ significantly by tumor subtype. In analyses by formulation, current use of formulations containing levonorgestrel in triphasic (hazard ratio, 2.83; 95% confidence interval, 1.98-4.03) and extended cycle regimens (hazard ratio, 3.49; 95% confidence interval, 1.28-9.53) and norgestrel in monophasic regimens (hazard ratio, 1.91; 95% confidence interval, 1.19-3.06), all combined with ethinyl estradiol, was associated with a higher breast cancer risk when compared with never oral contraceptive use. No association was observed for current use of the other progestin types evaluated (norethindrone, norethindrone acetate, ethynodiol diacetate, desogestrel, norgestimate, and drospirenone), however, sample sizes were relatively small for some of the subgroups, limiting these analyses. Current oral contraceptive use was associated with a higher risk for invasive breast cancer regardless of disease subtype, however, the risk in former users was comparable with never users 5 years after cessation. In analyses by progestin type, associations were observed for select formulations containing levonorgestrel and norgestrel. Assessment of the associations for newer progestin types (desogestrel, norgestimate, drospirenone) was limited by sample size, and further research on more recently introduced progestins is warranted.