Comparative effectiveness of statin plus fibrate combination therapy and statin monotherapy in patients with type 2 diabetes: use of propensity-score and instrumental variable methods to adjust for treatment-selection bias
Pharmacoepidemiology and Drug Safety, 05/03/2012
Clinical Article
Suh HS et al. – The authors did not find a difference in effectiveness regarding cardiovascular outcomes between the statin plus fibrate combination therapy and the statin monotherapy after controlling for hidden bias.
Methods- A retrospective database analysis was performed using a large managed care claims database of patients identified with type 2 diabetes based on diagnosis codes from January 2002 through December 2003 and continuously enrolled for the entire study period, 5.5 years.
- A statin plus fibrate combination therapy group (patients who used statins less than 6 months and augmented with fibrates for more than 6 months) and a statin monotherapy group (patients who used statins persistently) among patients with type 2 diabetes were followed for 3 years to examine the relationship between the intervention and cardiovascular events using a multivariable logistic regression model, propensity score method, and instrumental variable approach.
- The statin plus fibrate combination therapy group of 318 and the statin monotherapy group of 9928 were identified from 75515 diabetics.
- After adjusting for factors that can impact cardiovascular outcomes, the combination therapy group did not significantly experience a reduction in cardiovascular disease, as compared with the statin monotherapy group (OR=0.77; p=0.083).
- The statin plus fibrate combination therapy group was significantly associated with a reduction in cardiovascular events after propensity matching (OR=0.53; p=0.002).
- Using the physician prescribing preference instrument to adjust for unmeasured confounding, they did not find evidence that subjects in the statin plus fibrate combination therapy group versus stain monotherapy group experienced a significant reduction in cardiovascular events (p=0.124).
