Abstract
Background There are conflicting results in the literature regarding the association between the antibiotic exposure and breast cancer risk. The aim of this study was to assess this association using a population-based approach.
Methods The source population was the dynamic cohort defined by membership in the Saskatchewan Prescription Drug Plan (Canada) between the years 1981 and 2000. Four matched controls were selected for each case identified by the Saskatchewan Cancer Agency, using incidence density sampling. Detailed drug exposure over a minimum of 15 years before diagnosis allowed studying the respective effects of dosage and timing of antibiotic use on breast cancer risk. Two antibiotic exposure definitions were used: the number of prescriptions and the number of units (tablets, capsules, etc.), which were further categorized into quartiles.
Results A total of 3099 breast cancer cases and 12 396 matched controls were included. The incidence of breast cancer was higher in subjects who had more antibiotic prescriptions during the 1-15 years prior to the index date (RRs = 1.50, 1.63, 1.71 and 1.79 for the four quartiles, respectively, p-trend = 0.0001). Similar results were found when a number of units were considered. We did not find any effect of the timing of antibiotic exposure on breast cancer risk. Similar patterns of increased risk of breast cancer were detected for the specific antibiotic classes.
Conclusions We observed a dose-dependent increase in breast cancer risk in association with the antibiotic exposure up to 15 years in the past. However, the lack of temporal trends and the absence of class-specific effects suggest a non-causal relationship. Copyright (C) 2007 John Wiley & Sons, Ltd.