We aimed to quantify for the first time the relationship between statin adherence and ischemic stroke (IS) in patients with diabetes.

Using Finnish health registers, we assembled a cohort of 52 868 statin initiators with diabetes in 1995–2006. We conducted a nested case-control analysis matching cases with IS with up to four controls for age, sex, date of statin initiation and follow-up duration. Adjusted rate ratios for IS were estimated with conditional logistic regression. Additional potential confounders were considered with inverse probability weighting and the role of unmeasured confounding using external adjustment. Statin adherence was measured as the proportion of days covered (PDC).

Among 1703 cases and 6799 controls, good adherence to statins (PDC ≥ 80%) was associated with a 23% decreased incidence of IS (95%CI 14–32%) compared with poor adherence (PDC < 80%). This association remained broadly unchanged when stratified by sex, age, history of atherosclerotic cardiovascular disease or IS. There was a dose–response relationship between adherence level and the risk of IS (RR 0.63 [0.53–0.75] for PDC ≥ 80% versus PDC < 20%, P for trend <0.0001). Among patients with good adherence, those initiating with low intensity statin therapy had a 15% lower incidence (95%CI 2–27%) and those initiating with moderate intensity therapy a 29% lower incidence (16–41%) of IS compared with those with poor adherence who initiated with low intensity therapy. Our sensitivity analyses supported the robustness of the results.

In diabetes, poor statin adherence may considerably increase the risk of IS both in primary and secondary prevention of IS