Statins and Risk of Bleeding: An Analysis to Evaluate Possible Bias Due to Prevalent Users and Healthy User Aspects

Statins are said to protect against a wide range of diseases. We studied to what extent potential bias influences the results of studies on beneficial side effects of statins. We selected 8,188 atrial fibrillation patients who started treatment with anticoagulants at the Leiden Anticoagulation Clinic in the Netherlands between 2003 and 2009 and experienced 1,683 minor and 451 major bleeds during 18,105 person-years of follow-up. Statins were associated with a risk reduction of 9% for bleeds (hazard ratio = 0.91, 95% confidence interval: 0.82, 1.00). Additionally, analyses were stratified by age, incident users (patients who started statins during follow-up, i.e., an inception cohort), and prevalent statin users (statin users at baseline), as restriction to incident users avoids overoptimistic risk estimates. After stratification, the protective associations disappeared or reversed (range of hazard ratios = 0.99-3.22), except for patients aged 75 years or older. This remaining association could be due to another bias as, according to guidelines, in the elderly, statins should be prescribed only to those with a reasonable life expectancy. This could have resulted in a comparison of fit statin users with less fit nonstatin users (healthy user effect). The apparent protective association of statins on bleeds may be due to bias. We recommend stratification by age and incident and prevalent statin use when studying associations of statins with disease outcomes to avoid overoptimistic risk estimates.