Introduction: There is increasing evidence of a causal association between statin use and type II diabetes mellitus (T2DM) in adults. However the risk of incident T2DM in children treated with statins is unknown, and a recent study demonstrating a protective effect of familial hypercholesterolemia suggests there may be heterogeneity of effect by the presence of genetic dyslipdemia. We estimated the association between statin use and the risk of T2DM in a commercially insured population ages 8 to 20 from 2004 to 2013 with and without a diagnosis of dyslipidemia.
Methods: New users of statins with a minimum 75 percent of days covered (PDC) in the first year of use were matched to up to 10 non-users using year-specific propensity scores within a 0.01 caliper. The outcome was defined as incident T2DM more than 12 months after the match date to address possible confounding by indication and allow for a biologically plausible level of cumulative exposure. Cox proportional hazard models were used to estimate the relative hazards for T2DM among statin users and non-users. Time was defined from the end of the 12 month latency period to either the outcome or censoring. Sensitivity analyses varied the width of this exclusionary window, as well as PDC to test the robustness of our results to study assumptions.
Results: There were a total of 13,443,478 patients who met eligibility criteria, 755 (0.1%) of whom were started on statins and met the minimum PDC. Of these, 748 (98%) were matched with no significant difference in the year-specific match rate. Statin use was associated with an increased risk of T2DM in children without dyslipidemia diagnosis (HR 2.75, 95%CI 1.20-6.30, p=0.01), but not in children with a recorded diagnosis of dyslipidemia (HR 1.01, 95%CI 0.59-1.70, p=0.98). The results were robust to variations in the latency period and PDC (figure).
Conclusion: Statin use was associated with increased risk of new onset T2DM in children without a dyslipidemia diagnosis, but not in children with a diagnosis of dyslipidemia.