AimsMetformin and acarbose have comparable efficacy as initial therapy for HbA(1c) reduction in Chinese patients with newly diagnosed Type 2 diabetes. However, not all participants achieved glycaemic control. Our aim was to discover a monotherapy predictor for therapeutic response in Type 2 diabetes on the basis of baseline features. MethodsData from the MARCH trial were collected, resulting in 698 individuals being available for longitudinal analyses. All participants were divided into subgroups based on successful and unsuccessful achievement of the glycaemic target according to primary endpoints at week 24 (HbA(1c) < 53 mmol/mol; 7.0%). Logistic regression analysis with stepwise variable selection was performed to assess the independent risk factors for good glycaemic control of monotherapy with metformin or acarbose. ResultsMedian HbA(1c) was 66 1 mmol/mol (8.2 0.07%) in the metformin group at baseline, and 66 1 mmol/mol (8.2 +/- 0.07%) in the acarbose group. After 24 weeks of monotherapy, 79.8% of participants in the metformin group achieved glycaemic targets compared with 78.7% of those in the acarbose group. Multivariate regression analysis showed that BMI and fasting blood glucose were significant independent predictors for the maintenance of good glycaemic control in the metformin group, whereas phase I insulin secretion (Insulin/Glucose at 30 min, I30/G30) and duration of diabetes were associated with good glycaemic control in the acarbose group. ConclusionsFor newly diagnosed Type 2 diabetes, some clinical features and laboratory parameters are important prognostic factors for predicting drug responsiveness. Participants with a higher BMI and lower fasting blood glucose achieved good glycaemic control when metformin was selected as the initial treatment. Acarbose was best for participants with higher phase I insulin secretion (I30/G30) and shorter duration of Type 2 diabetes.