In Reply Dr Finkelstein argues that the increase in odds ratios (ORs) of musculoskeletal diseases in statin users may be attributable to imperfect propensity score matching. In addition to the propensity score–matched cohort, we examined the risk of outcomes in different cohorts (Table 2 of our article)1 including a whole population cohort, a no-comorbidity index cohort, a musculoskeletal incident cohort, and a 2-year statin cohort. Each of these cohorts included different populations who varied in baseline characteristics. We also used different definitions for musculoskeletal outcomes. Our results remained consistent throughout.
Mansi I, Mortensen EM, Frei CR. Statins and Musculoskeletal Adverse Events—Reply. JAMA Intern Med. 2014;174(2):303-304. doi:10.1001/jamainternmed.2013.12683