Statins and poststroke intracerebral hemorrhage: Concern but increasing reassurance

Post hoc analyses of data from randomized controlled trials can generate useful hypotheses, but need to be considered exploratory. When conducted, such evaluations should adhere to specified criteria and test for a treatment by subgroup interaction for the trial's primary endpoint.(1) The best way to proceed if unexpected but potentially important outcomes occur is less clear. Concern that treatment with HMG-CoA reductase inhibitors (i.e., statins) might increase the risk of poststroke intracerebral hemorrhage (ICH) arose from an unanticipated observation in the Stroke Prevention with Aggressive Reduction in Cholesterol Levels (SPARCL) trial.(2) In SPARCL, patients who had a noncardioembolic ischemic stroke (67%) or TIA (31%) or hemorrhagic stroke (2%) within the prior 1 to 6 months were randomized to high-dose atorvastatin or placebo. SPARCL found an overall reduction in the risk of the primary outcome (fatal or nonfatal stroke; adjusted hazard ratio [HR] 0.84; 95% confidence interval [CI] 0.71-0.99) despite an increase in hemorrhagic stroke (HR 1.66, 95% CI 1.08-2.55) in the atorvastatin group with statistical heterogeneity based on the type of stroke occurring during the trial (ischemic, hemorrhagic, or unclassified stroke). In addition to atorvastatin treatment, subsequent exploratory analyses using multivariable regression revealed higher hemorrhagic stroke risk in those having a hemorrhagic stroke as the entry event (HR 5.65, 95% CI 2.82-11.30), in men (HR 1.79, 95% CI 1.13-2.84), and with increased age (10-year increments, HR 1.42, 95% CI 1.16-1.74). There were no statistical interactions between these factors and treatment.(3) Concern about a statin-related increase in ICH after stroke also arose after a post hoc analysis of the Heart Protection Study showed a nonsignificant increase in hemorrhagic stroke rate in those with a history of cerebrovascular disease (n = 3,280) treated with simvastatin vs placebo (1.3% vs 0.7%; p for heterogeneity = 0.03).(4) Among those treated with atorvastatin in SPARCL, patients with lower low-density lipoprotein (LDL) cholesterol levels had no increased risk of brain hemorrhage.(3) Nonetheless, the fact that at least some observational studies (largely Asian populations) suggested an inverse relationship between cholesterol levels and bleeding risk caused further concern.(5,6 </SUP)