Abstract
Breast cancer is the most commonly diagnosed cancer worldwide. Earlier studies have demonstrated that breast cancer patients with particular genomic variants are more susceptible to adverse drug effects (ADEs) when they are receiving endocrine therapy. However, to establish a robust body of evidence with regard to the potential utility and predictive value of these variants, findings from these reports require replication. This study aimed to validate previously reported associations between genomic variants and medically important adverse drug effects (MIADEs) using UK Biobank (UKBB). In 2729 female participants who had received endocrine therapy in the UKBB, no statistically significant genotype-treatment interactions were observed for the outcomes examined after correction for multiple testing. Power was limited for modest interactions involving low-frequency variants and less frequent outcomes, whereas power was high to detect larger interaction effects in common-variant scenarios. Accordingly, the findings do not provide robust evidence to support previously reported pharmacogenomic associations in this dataset, and current evidence does not support the use of pharmacogenomic testing for individualised endocrine therapy selection in clinical practice.