ORAL-CONTRACEPTIVE USE AND BREAST-CANCER RISK - A METAANALYSIS OF VARIATIONS WITH AGE AT DIAGNOSIS, PARITY AND TOTAL DURATION OF ORAL-CONTRACEPTIVE USE

Objective To review published reports of risk of breast cancer with oral contraceptive use. Design Meta-analysis of study results, using regression techniques to explore the inter-study heterogeneity. Setting Twenty-seven epidemiological studies of breast cancer risk and oral contraceptive use published 1980-1989. Main outcome measures Relative risk of breast cancer with oral contraceptive use; variations by age at diagnosis, parity, and total duration of use, and with study design characteristics. Results The overall relative risk estimate was 1.16 (95% CI 1.07-1.25) for the < 45 years group, 1.21 (95% CI 0.99-1.47) for the nulliparous subgroup, and 1.27 (95% CI 1.12-1.44) for durations of use of more than 8 years. Source of subjects, their marital status, the type of study, data collection methods, matching for geographical area, the number of adjustment factors used in the study analysis, the country of study and the calendar period in which the study ended were all important explanatory factors for inter-study variation, but a substantial proportion of inter-study variation remained unexplained. Conclusions These meta-analyses suggest that risk of breast cancer may be raised by around 20% in younger, nulliparous and long use duration subgroups of oral contraceptive users. Risk estimates appear to be influenced by several study design factors, which should be considered carefully in designing and reviewing future studies.