Background: The survival benefit of statins in nontrial populations of persons with diabetes is unknown.
Objective: : We sought to determine all-cause mortality in fiscal year 2001 (FY01) after statin initiation in FY 99 and/or FY00 in individuals with diabetes in the Veterans Healthcare Administration (VHA).
Methods: Using a retrospective longitudinal cohort analysis, we analyzed 201,102 veterans with diabetes from 104 VHA facilities with medical, pharmacy, and laboratory information from VHA and Medicare databases. Patients with statin exposure, defined as having medication possession coverage >50% of eligible days in FY99 and/or FY00, were characterized as initiators if no statin prescription was found in FY98. Otherwise, they were characterized as continuing users. We defined 4 statin exposure groups: FY99 only, FY00 only, both FY99 and 00, and neither year. All-cause mortality was determined in FY01. Propensity score matched comparisons were used to corroborate results from mixed effects logistic models.
Results: FY01 mortality was no different between FY99-only initiators and the nonexposure group (odds ratio [OR] = 1.08, P = 0.429). In contrast, FY00-only and FY99-00 initiator groups showed odds ratio of 0.75 (P < 0.0001) and 0.71 (P < 0.0001), respectively. There was a similar benefit for continuing users. Propensity analysis demonstrated consistent findings. Increased statin adherence from >50% to >75% was associated with increased benefit (OR = 0.71, P < 0.0001 versus OR = 0.62, P = 0.0002).
Conclusions: One to 2-year statin exposure is associated with a 25% to 29% risk reduction in all-cause mortality of the subsequent year in a high-risk diabetes cohort.