High-density lipoprotein-cholesterol and not HbA1c was directly related to cardiovascular outcome in PROactive

Aim: In PROactive, pioglitazone reduced the incidence of death, myocardial infarction and stroke, and significantly improved HbA1c, systolic blood pressure (SBP), triglycerides and high-density lipoprotein (HDL)-cholesterol relative to placebo. As these glycaemic and lipid parameters are major cardiovascular (CV) risk factors, we assessed their separate contribution to the reduced incidence of CV outcomes. Methods: Patients (n = 5238) with type 2 diabetes and macrovascular disease were randomized to 45 mg pioglitazone or placebo. Relationships among treatment, outcome (time to first event of all-cause mortality, myocardial infarction and stroke) and 10 laboratory measurements and vital signs were investigated using log-linear models. Continuous variable measurements (percent changes from baseline to average of all postbaseline values prior to censoring) were made discrete by categorizing into tertiles. Log-linear models were fitted to multiway tables of discrete data and analysis of deviance used to summarize sources of variation in the data. Results: Although pioglitazone treatment was associated with a decrease in HbA1c and an increase in HDL-cholesterol (HDL-C), only the change from baseline HDL-C predicted the outcome (chi(2) = 28.89, p < 0.0001). No other variables, including HbA1c, triglycerides and systolic blood pressure, showed significant direct associations with outcome. When the analysis was extended to include baseline statin use, this was associated with an improved outcome independently of HDL-C changes. Conclusions: This post hoc analysis suggests that HDL-C, but probably not HbA1c, is a driver of pioglitazone's favourable influence on CV outcome.