Context: Utilization of the biguanide metformin and a thiazolidinedione (TZD) with new onset diabetes has the benefit of lowering A1cs into the normal range without the problem of severe hypoglycaemia. Objective: To assess the effectiveness of once-daily combined metformin and TZD therapy compared with other therapeutic regimens typically utilized at later stages of type 2 diabetes. Methods: A random chart review of 300 type 2 diabetic patients and extraction of data for body mass index (BMI), duration of diabetes and C-peptide and A1c was performed. In the 210 type 2 diabetic subjects in whom this information was currently available, the data were analysed. Results: Eighty-six patients on once-daily metformin and rosiglitazone had an average A1c of 6.2% (group A), and 58 subjects on triple therapy (metformin, rosiglitazone and glimepiride) (group B) had an average haemoglobin A1c (HbA1c) of 6.9%. The 22 subjects on one injection of insulin per day in addition to triple therapy (group C) had an average HbA1c of 7.6%, and the 44 subjects on more than one insulin injection per day plus metformin and/or rosiglitazone (group D) had an average HbA1c of 8.3%. HbA1cs below 7.0% were found in 91.9% of group A, 21.7% of group B, 36.4% of group C and 56.8% of group D. HbA1cs below 6.5% were found in 78.2% in group A, 15.5% in group B, 22.7% in group C and 31.8% in group D. HbA1cs below 6.0% were found in 41.9% in group A, 6.9% in group B, 9.1% in group C and 13.6% in group D. On univariate analyses, the HbA1c was positively associated with the duration of diabetes and the BMI and negatively associated with random C-peptide levels. Alternatively, on multiple regression analysis, there was no statistical correlation between the duration of diabetes or BMI with the HbA1c. However, there was a strong statistical correlation between the random C-peptide level and the HbA1c (p = 0.002). Conclusion: Early initiation of therapy for type 2 diabetes with a once-daily combination of metformin and rosiglitazone provides the greatest opportunity to achieve A1cs within the normal range. The level of achieved glycaemic control is not dependent on the number or potency of the therapies utilized but is dependent on the level of endogenous insulin production. The use of a TZD as part of initial therapy of type 2 diabetes with its documented ability to preserve or improve β-cell function has the potential to achieve prolonged normoglycaemia in the type 2 diabetic patient. © 2004 Blackwell Publishing Ltd.