Risk of diabetes associated with statins and fibrates and their association: A disproportionality analysis using the WHO spontaneous reporting database, Vigibase®

Abstract

Relations between hypolipidemic drugs and diabetes are discussed. The aim of our study was to assess the association between diabetes and use of statins or fibrates or their association using the World Health Organization (WHO) global individual case safety reports database, VigiBase®, which records reports from over 130 countries worldwide. We performed a disproportionality analysis including all reports until the 31st December 2017 to measure the risk of reporting “new onset of diabetes” compared with all other reports [as a reporting odds ratio (ROR 95% CI)] for statins, fibrates and their association. The likelihood that diabetes resulted from statin-fibrate pharmacodynamic interaction was also estimated. According to the interaction additive model, a ROR value for coexposure exceeding the sum of the RORs estimated for each individual class of drug supports a potential drug-drug interaction (DDI). To assess the stability of our results, we performed several sensitivity analyses, according to outcome definition and excluding reports with drugs known to induce hyperglycaemia or diabetes. We identified 7633 cases of diabetes (mainly with atorvastatin) and 169,690 non-cases among reports with statins alone, 189 cases (mainly with fenofibrate) and 18,960 non-cases among reports with fibrates alone with 60 cases and 3,187 among reports with the combination statin-fibrates. Statin use alone was associated with an increase in the ROR of diabetes (ROR = 1.75 [1.72–1.78]), but not with fibrate use alone (ROR = 0.76 [0.71–0.82]). The combination of statin + fibrate was associated with an increase in the ROR of diabetes (ROR = 1.46 [1.28–1.67]) but was not significant to support a DDI. Similar results were found in sensitivity analyses. Using the WHO pharmacovigilance database, our study confirmed the signal of diabetes with statins but failed to find any potential signal with fibrates alone and did not suggest a pharmacodynamic DDI with the use of statin + fibrate.