Effect of metformin plus roziglitazone compared with metformin alone on glycaemic control in well-controlled Type 2 diabetes.

Aims To investigate the effect of metformin plus roziglitazione (RSGMET) compared with metformin alone (MET) on glycaemic control in well‐controlled Type 2 diabetes. Methods Subjects (drug naïve or those on glucose‐lowering monotherapy) were randomized (n = 526), following a 4‐week placebo run‐in period, to RSGMET [4 mg rosiglitazone (RSG)/500 mg MET] or MET 500 mg. From weeks 2–18, medication was escalated every 4 weeks (based on gastrointestinal tolerability), then remained at RSGMET 8 mg/2 g or MET 3 g for 14 weeks. Results RSGMET reduced HbA_1c from 7.2 ± 0.6 to 6.7 ± 0.8% at week 32, compared with a reduction from 7.2 ± 0.6 to 6.8 ± 0.9% with MET (treatment difference ?0.13%; P = 0.0357). More subjects achieved an HbA_1c value of ≤ 6.5% at week 32 with RSGMET (51.6 vs. 43.7%), but the treatment difference was not significant (odds ratio 1.37, P = 0.0949). RSGMET produced larger reductions from baseline in mean fasting plasma glucose (adjusted difference ?0.62 mmol/l, P < 0.0001), with the odds ratio of achieving a target of < 7.0 mmol/l being 2.33 (P < 0.0001). Statistically significant differences in favour of RSGMET relative to MET were seen for homeostatic model assessment (HOMA)‐derived estimates of insulin sensitivity and pancreatic B‐cell function, C‐reactive protein (CRP), and systolic blood pressure. Overall rates of gastrointestinal adverse events (relevant to the known profile of MET) were comparable, but with a lower incidence of diarrhoea (8 vs. 18%) with RSGMET. Hypoglycaemia was reported in ≤ 7% subjects per group. Conclusions RSGMET provided similar short‐term glycaemic control to MET with greater improvements in estimates of insulin sensitivity, B‐cell function and CRP, with less diarrhoea and low risk of biochemical hypoglycaemia, suggesting that early use of combination therapy may be appropriate.