Pioglitazone slows progression of atherosclerosis in prediabetes independent of changes in cardiovascular risk factors

Abstract

Objective: To determine whether changes in standard and novel risk factors during the Actos Now for Prevention of Diabetes trial explained the slower rate of carotid intima media thickness (CIMT) progression with pioglitazone treatment in persons with prediabetes.

Methods and results: CIMT was measured in 382 participants at the beginning and up to 3 additional times during follow-up of the Actos Now for Prevention of Diabetes trial. During an average follow-up of 2.3 years, the mean unadjusted annual rate of CIMT progression was significantly (P=0.01) lower with pioglitazone treatment (4.76×10(-3) mm/year; 95% CI: 2.39×10(-3)-7.14×10(-3) mm/year) compared with placebo (9.69×10(-3) mm/year; 95% CI: 7.24×10(-3)-12.15×10(-3) mm/year). High-density lipoprotein cholesterol, fasting and 2-hour glucose, HbA(1c), fasting insulin, Matsuda insulin sensitivity index, adiponectin, and plasminogen activator inhibitor-1 levels improved significantly with pioglitazone treatment compared with placebo (P<0.001). However, the effect of pioglitazone on CIMT progression was not attenuated by multiple methods of adjustment for traditional, metabolic, and inflammatory risk factors and concomitant medications, and was independent of changes in risk factors during pioglitazone treatment.

Conclusions: Pioglitazone slowed progression of CIMT, independent of improvement in hyperglycemia, insulin resistance, dyslipidemia, and systemic inflammation in prediabetes. These results suggest a possible direct vascular benefit of pioglitazone.

Trial registration: ClinicalTrials.gov NCT00220961.

Figures

Figure 1

Annual CIMT progression rates, mm/year × (103) by intention to treat groups. White bars represent those randomized to placebo. Black bars represent those randomized to piogitazone treatment. Error bars represent 95% CI. All models include random subject effects representing individual differences from the group mean for baseline CIMT and rate of CIMT progression. The difference in progression rate between treatment groups was significant in all models (P < 0.01). Model 1: Unadjusted Model 2: Adjusted for age, study center, gender, race/ethnicity (non-Hispanic Whites vs. others) and prior CVD Model 3: Model 2 plus HbA1c, systolic and diastolic blood pressure, BMI, triglyceride/HDL-C ratio and Matsuda index; P < 0.01. The results were not different when adjusted for other lipid values such as total cholesterol/HDL-C ratio, or total cholesterol, triglycerides, LDL-C and HDL-C individually or in combination. Model 4: Model 3 plus adiponectin, leptin, CRP, TNF-α, IL-6, MCP-1, PAI-1 Model 5: Model 4 plus concomitant medications (lipid lowering agents, anti-hypertensives, and drugs containing salicylic acid, all parameterized as yes versus no) Model 6: Parsimonious model fitted including all covariates from Model 5 that were P1c, and total cholesterol/HDL-C ratio)