The effect of adherence to statin therapy on the hazard of cardiovascular mortality in the Netherlands

Background: The biological efficacy of statin therapy has been demonstrated in various clinical trials. However, end users may differ from trial participants in relevant ways. Therefore, observational studies are needed to assess clinical effectiveness. Our objective was to assess the clinical effectiveness of adherence to statin therapy in reducing cardiovascular mortality in the Netherlands. Objectives: The aim of this study was to assess the clinical effectiveness of adherence to statin therapy in reducing cardiovascular mortality in the Netherlands. Methods: Individual-level mortality information from Statistics Netherlands was linked to pharmacy dispensing data that came from the representative database IADB.nl. We used extended Cox models with adherence to statin therapy as the primary exposure and time to cardiovascular mortality as the primary outcome. We adjusted for age, sex, birth cohort, socio-economic status, diabetic status, and the utilization of various cardiovascular drugs. Covariates were allowed to vary over time. We achieved population-averaged effect estimates through implementation of the parametric G-formula. We also performed a subset analysis by calendar period corresponding to periods of particular cardiovascular prescribing guidelines and a subset analysis by dispensing background to assess the influence of healthy adherer bias. Results: The conditional estimate was that being fully adherent to statins reduced the hazard of cardiovascular mortality by about 47% (HR: 0.53; 95%CI: 0.46 to 0.61), compared with being fully non-adherent to statins. The population-averaged estimate was of similar magnitude. In addition, we found evidence that estimates of clinical effect approached estimates of trials just after the introduction of statins in the population but became potentially more confounded in later calendar years. Conclusions: The study provides evidence of the clinical effectiveness of statins, although the final estimates may still be affected by healthy adherer bias