Stroke prediction and stroke prevention with atorvastatin in the Collaborative Atorvastatin Diabetes Study (CARDS)

Abstract

Patients with Type 2 diabetes have an elevated risk of stroke. The role of lipid levels and diabetes-specific factors in risk prediction of stroke is unclear, and estimates of efficacy of lipid-lowering therapy vary between trials. We examined predictors of stroke and the effect of atorvastatin on specific stroke subtypes in Type 2 diabetes in the Collaborative Atorvastatin Diabetes Study (CARDS) [a trial of 2838 participants with mean low-density lipoprotein cholesterol < 4.14 mmol/l, no history of macrovascular disease and randomized to atorvastatin 10 mg daily or placebo]. Median follow-up was 3.9 years. Cox regression models were used to estimate the effect of atorvastatin on stroke rate and risk of stroke associated with baseline risk factors. Risk factors that predicted stroke in univariate models were examined in a multivariable model. Independent risk factors predicting stroke were age [10-year increments; hazard ratio (HR) 2.3, P < 0.001], microalbuminuria (albumin : creatinine ratio > 2.5 mg/mmol; HR 2.0, P = 0.007) and glycaemic control (HbA(1c) > 10%; HR 2.7, P = 0.007). Women were at lower risk of stroke (HR 0.3, P = 0.004). Lipids did not predict stroke. Of 60 first strokes, 47 were non-haemorrhagic, 13 were indeterminate and none was definitely haemorrhagic. Atorvastatin treatment was associated with 50% reduction in non-haemorrhagic stroke (95% confidence interval 9%-72%P = 0.024), similar to the 48% reduction (11%-69%) for all strokes combined. Diabetes-specific risk factors are important predictors of stroke in Type 2 diabetes. Despite the lack of association between baseline lipids and first stroke, there was a reduction of 50% of non-haemorrhagic strokes associated with atorvastatin treatment in the CARDS population.