Comparison of pioglitazone and metformin efficacy using homeostasis model assessment.

aimsTo compare clinical efficacy of two different insulin sensitizers, pioglitazone and metformin, and to reveal factors that influence the clinical efficacy.

methodsSeventy-eight Japanese subjects with Type 2 diabetes mellitus poorly controlled with sulphonylureas [38 men and 40 women, aged 57 +/- 9 years, body mass index 25.2 +/- 1.4 kg/m2, and HbA1c 8.3 +/- 0.6% (means +/- SD)] were randomly assigned to groups for the addition of either pioglitazone or metformin and followed up for 4 months. A decrease in HbA1c levels was compared with baseline factors including homeostasis model assessment of insulin sensitivity (HOMA-R) and beta-cell function (HOMA-beta) with 71 subjects who completed the study.

resultsThe overall decrease in HbA1c levels was similar for the pioglitazone (-1.2 +/- 0.2%) and metformin (-1.3 +/- 0.1%) groups. In the pioglitazone group, the decrease in HbA1c levels was negatively correlated with baseline HOMA-R (r=-0.698, P<0.0001) and HOMA-beta (r=-0.680, P<0.0001). In contrast, the decrease was positively correlated with baseline HOMA-beta (r=0.556, P=0.0004) in the metformin group. Multivariate analysis revealed that either HOMA-R or HOMA-beta was a main determinant of the decrease in HbA1c levels in the pioglitazone group. In the metformin group, baseline levels of fasting glucose were also included as an independent determinant in addition to HOMA-beta. The subjects with greater HOMA-R (> or =4.0) or HOMA-beta (> or =40%) displayed better response to pioglitazone than to metformin, and vice versa.

conclusionsIn Type 2 diabetic subjects poorly controlled with sulphonylureas, addition of pioglitazone or metformin resulted in a comparable reduction in HbA1c levels. Subjects with greater insulin resistance or preserved beta-cell function displayed better response to pioglitazone, whereas subjects with reduced beta-cell function displayed better response to metformin.