Are antibiotics related to oral combination contraceptive failures in the Netherlands? A case-crossover study.

PURPOSE: To investigate whether there is an association between use of antibiotics and breakthrough pregnancy.

METHODS

The study was performed in a population-based prescription database (IADB.nl). We computed case-crossover odds ratios of 397 cases of defined breakthrough pregnancy comparing the use of antibiotics in the exposure window with the use of antibiotics in two control windows. We defined a control group consisting of 29 022 other pregnancies. We computed case-control odds ratios of the use of antibiotics in cases as compared with controls in the different time windows.

RESULTS

The case-crossover odds ratios comparing the use of antibiotics in the exposure window with both control windows were 2.21 (95%CI = 1.03-4.75) and 1.65 (95%CI = 0.78-3.48), respectively. The traditional case-control odds ratios after adjustment for age were 1.71 (95%CI = 1.09-2.66) in the exposure window, 0.81 (95%CI = 0.44-1.47) 2 months before the exposure window, and 1.04 (95%CI = 0.61-1.78) 12 months before the exposure window.

CONCLUSIONS

We did find a relationship between the use of antibiotics and breakthrough pregnancy in a population-based prescription database. The results did not hold for broad-spectrum antibiotics or in a sensitivity analysis. The results are partly not the same as those found in a pharmacoepidemiological study with a similar design using two US pregnancy databases. Both studies can suffer from bias and confounding, but these will be different because of the use of different databases.