Limitations of the HOMA-B score for assessment of beta-cell functionality in interventional trials-results from the PIOglim study.

BACKGROUND Drugs with unspecific stimulating effects on beta-cell secretion increase the homeostasis model assessment (HOMA)-B score, indicating improved beta-cell "function." We investigated whether the beta-cell protection provided by adding pioglitazone (PIO) to glimepiride (GLIM) in comparison to up-titrating the GLIM dose alone is reflected by appropriate changes in several measures of beta-cell function, including HOMA-B score. METHODS This double-blind, parallel prospective 6-month study was performed with 82 patients (47 men, 35 women; age, 61 +/- 9 years; duration of disease, 5.3 +/- 4.4 years; body mass index, 32.6 +/- 6.0 kg/m(2); hemoglobin A1c [HbA1c], 7.3 +/- 0.7%) with GLIM monotherapy (1-3 mg). They were randomized to receive a GLIM + PIO combination with up-titration (2 mg + 30 mg/4 mg + 30 mg/4 mg + 4 mg) or to remain on GLIM (up-titration 4/5/6 mg). Observation parameters determined at baseline and end point included HOMA-B, HOMA-IR, HbA1c, glucose, insulin, and intact proinsulin. RESULTS There was a slight increase in the HOMA-B score in the GLIM group but not in the GLIM + PIO arm (baseline/end point: for GLIM, 71 +/- 48/88 +/- 64; for PIO + GLIM, 74 +/- 56/69 +/- 52). Improvements in the other observation parameters were predominantly detected in the PIO + GLIM group (HbA1c, 7.20 +/- 0.61%/6.36 +/- 0.90%; HOMA-IR, 7.0 +/- 4.5/4.1 +/- 2.1; intact proinsulin, 12.4 +/- 10.3/7.6 +/- 4.8 pmol/L [all P < 0.05 vs. baseline]) compared with the GLIM group (HbA1c, 7.45 +/- 0.69%/7.15 +/- 0.97% [P < 0.05]; HOMA-IR, 7.4 +/- 4.5/7.5 +/- 4.3 [not significant]; intact proinsulin, 17.3 +/- 21.6/16.3 +/- 15.5 pmol/L [not significant]). CONCLUSIONS The PIO + GLIM combination led to overall improvement of laboratory biomarkers for beta-cell function, except for HOMA-B. Glimepiride up-titration had no such effects but increased the HOMA-B score. HOMA-B seems to provide misleading results when used as a diagnostic tool in patients treated with sulfonylurea drugs. A corrective term for consideration of proinsulin in the HOMA-B equation may address this limitation.