BACKGROUND: We evaluated the hypothesized transient effect of oral contraceptive use on the occurrence of gallstone disease in a case-control study. METHODS: The study comprised 82 acute cases, and 249 general population controls, 28 elective cases, and 106 radiodiagnosis controls. Risk-period-specific rate ratios were estimated by logistic regression analysis. RESULTS: During the first 5 years after any year of oral contraceptive use (0- to 5-year risk period) the rate ratio was 1.6 (90% confidence interval, 0.83-2.95), and for the 6- to 10- and 11- to 20-year risk periods, 1.2 (0.67-2.13) and 1.1 (0.69-1.65), respectively. Further evaluation showed substantial confounding by other risk factors. After these were controlled for, the rate ratios were 1.1, 1.8, and 1.0 for the respective risk periods. When oral contraceptive use had been started recently (within 5 years before diagnosis) the rate ratio was 4.8 (0.79-29.7); recent oral contraceptive use was not associated with gallstone disease if preceded by 6 or more years of oral contraceptive use (rate ratio 0.8 (0.15-4.55)). CONCLUSIONS: The results are consistent with a transient effect of oral contraceptive use that does not accumulate beyond 10 years of use. However, the substantial confounding effect in our study raises the question whether the present and previously reported estimates of the effect may be biased by confounding.
BACKGROUND: We evaluated the hypothesized transient effect of oral contraceptive use on the occurrence of gallstone disease in a case-control study. METHODS: The study comprised 82 acute cases, and 249 general population controls, 28 elective cases, and 106 radiodiagnosis controls. Risk-period-specific rate ratios were estimated by logistic regression analysis. RESULTS: During the first 5 years after any year of oral contraceptive use (0- to 5-year risk period) the rate ratio was 1.6 (90% confidence interval, 0.83-2.95), and for the 6- to 10- and 11- to 20-year risk periods, 1.2 (0.67-2.13) and 1.1 (0.69-1.65), respectively. Further evaluation showed substantial confounding by other risk factors. After these were controlled for, the rate ratios were 1.1, 1.8, and 1.0 for the respective risk periods. When oral contraceptive use had been started recently (within 5 years before diagnosis) the rate ratio was 4.8 (0.79-29.7); recent oral contraceptive use was not associated with gallstone disease if preceded by 6 or more years of oral contraceptive use (rate ratio 0.8 (0.15-4.55)). CONCLUSIONS: The results are consistent with a transient effect of oral contraceptive use that does not accumulate beyond 10 years of use. However, the substantial confounding effect in our study raises the question whether the present and previously reported estimates of the effect may be biased by confounding.