Background: Epidemiologic studies of women who use oral contraceptives have reported an increased risk of breast, cervix, and liver cancers, but a reduced risk of other gynecologic neoplasms and (possibly) colorectal cancer. To test the hypothesis that users of oral contraceptives have an overall7 reduction in cancer risk, Hannaford and associates used data from a large population-based study with extended follow-up and the ability to adjust for multiple factors in cancer etiology.

The Study: The British Royal College of General Practitioners performed a study of oral contraception use in primary care from 1968 to 1996. Initially, 1,400 family physicians recruited 23,000 oral contraceptive users and 23,000 “never-users” (control group). The physicians provided data every six months, including type of oral contraceptive used, pregnancies, new illnesses, surgeries, and cancers. Other data gathered included age, smoking status, parity, and social class. Beginning in the 1970s, participants' health records were also flagged, when possible, to enable any new cancer cases and deaths from any cause to be reported to the study from national registries and databases. Cancer incidence in oral contraceptive users was compared with the incidence in never-users. The cancers were grouped into three categories: individual cancers, leading gynecologic neoplasms, and all cancers.

Results: The study covered approximately 744,000 woman-years of observation for oral contraceptive users and 339,000 years for never-users. The median duration of oral contraceptive use was 44 months (range of one to 344 months). Users tended to be younger, smoke more, have a higher initial parity, and come from a lower social class compared with never-users. Users were also more likely to report subsequent use of hormone therapy.

The overall risk of any cancer in oral contraceptive users was significantly reduced by 12 percent, with an adjusted relative risk of 0.88 (95 % confidence interval [CI], 0.83 to 0.94). For specific sites, statistically significant reductions were found for colorectal, ovarian, uterine body, “unknown” and “other” cancers. The combined risk reduction for the main gynecologic cancers was 29 percent in oral contraceptive users. Small and statistically nonsignificant increases in incidence were found for cancers of the lung, cervix, and central nervous system in oral contraceptive users. The groups did not differ in breast cancer incidence.

Overall, cancer risk increased with age and smoking, but oral contraceptive users had lower cancer risks in all age groups except for women younger than 30 years. The difference in risk was statistically significant in women 30 to 39 and 50 to 59 years of age. Users of oral contraceptives also had lower risks (often statistically significant) when users and never-users were compared in subgroups based on smoking, parity, or social class.

Use of oral contraceptives for more than eight years was associated with a significant increase in cancer risk (adjusted relative risk, 1.22; 95 % CI, 1.07 to 1.39), mainly because of an increased incidence of cervical, central nervous system, and pituitary cancers. Conversely, prolonged oral contraceptive use was associated with a significantly reduced risk of ovarian and uterine body cancers. The protective effect of oral contraceptives for ovarian cancer appeared to persist for at least 15 years after discontinuation of use.

Conclusion: The authors estimate a net absolute risk reduction in cancer incidence of 45 per 100,000 woman-years associated with oral contraceptive use. The potential benefit may be greater for specific subgroups, especially older women.

editor's note: How can we use this study's results in everyday practice? The cautious optimists might say that oral contraceptive use appears to reduce rather than increase the risk of cancer overall, despite the issues of large numbers lost to follow-up, differences in oral contraceptive formulations, and other concerns with the study. Being cautious, they will focus on the individual woman and her personal risks for all morbidities, including specific cancers, cardiovascular disease, and unintended pregnancy. Being optimists, they will note that this study confirms other evidence that, in large population-based studies, no increased risk of cancers has been detected in oral contraceptive users.¹ The pessimists may want to reserve judgment until research provides more specific information such as the net effect on morbidity and mortality outcomes, data on more modern formulations and patterns of oral contraceptive use, and outcomes for specific subgroups of women.—a.d.w.