AIM: Dipeptidyl peptidase-4 (DPP-4) inhibitors are used in the treatment of type 2 diabetes. Available sub-group analysis of clinical trials does not allow a clear identification of predictors of therapeutic response to these drugs. The aim of this study is the assessment of predictors of response to DPP-4 inhibitors. MATERIALS AND METHODS: A meta-analysis was performed, exploring correlation between 24-week effects on HbA(1c) of maximal doses of DPP-4 inhibitors, compared either with placebo or with other active drugs, matches to baseline characteristics of patients enrolled in 63 randomized clinical trials, either published or unpublished but disclosed on different websites were studied. RESULTS: DPP-4 inhibitors significantly reduce HbA(1c) at 24 weeks [by 0.6 (0.5-0.7)%] when compared with placebo; no difference in HbA(1c) was observed in comparisons with thiazolidinediones and α-glucosidase inhibitors, whereas sulfonylureas and metformin produced a greater reduction of HbA(1c) , at least in the short term. DPP-4 inhibitors produced a smaller weight gain than thiazolidinediones, and showed a lower hypoglycaemia risk than sulfonylureas. The placebo-subtracted effect of DPP-4 inhibitors on HbA(1c) was greater in older patients and in those with lower fasting plasma glucose at baseline. Similar results were obtained in comparisons with thiazolidinediones and metformin. CONCLUSIONS: Although drugs for type 2 diabetes are studied in heterogeneous samples of patients, their efficacy can be predicted by some clinical parameters. DPP-4 inhibitors appear to be more effective in older patients with mild/moderate fasting hyperglycaemia. These data could be useful for a better definition of the profile of patients who are likely to benefit most from these drugs.