OBJECTIVES: Type 2 diabetes is a progressive disease and most patients experience deterioration in glycemic control over time necessitating the use of combination therapy. We sought to determine the comparative efficacy and safety of third-line antidiabetic agents in patients with type 2 diabetes after failure of metformin and TZD therapy. METHODS: We performed a literature search of MEDLINE and CENTRAL through May 2014 and included randomized controlled trials (RCTs) of ≥12-weeks duration evaluating the addition of a noninsulin agent in patients with type 2 diabetes inadequately controlled on stable, optimized metformin and TZD therapy (≥1500mg metformin and ≥50% maximum TZD dose for ≥4 weeks). Network meta-analysis was performed on identified trials. Endpoints of interest were changes from baseline in HbA1c, body weight, systolic blood pressure (SBP); and the risk of hypoglycemia, urinary (UTI) and genital tract infection (GTI). RESULTS: Eleven RCTs evaluating dipeptidyl peptidase-4 (DPP-4) inhibitors (linagliptin, sitagliptin), sulfonylureas (SUs), (glibenclamide, glimepiride), glucagon-like peptide-1 (GLP-1) analogs (exenatide, liraglutide, dulaglutide, taspoglutide) and sodium glucose cotransporter-2 (SGLT2) inhibitors (canagliflozin, empagliflozin) were included. All agents analyzed reduced HbA1c vs placebo (range, 0.55-1.17%). All SUs were associated with significant weight gain (range, 3.31-7.29 kg), while significant weight loss was seen with all GLP-1 analogs (range, 1.53-2.2 kg) and SGLT-2 inhibitors (range, 2.08-2.95 kg). A reduction in SBP was seen in canagliflozin and GLP-1 analogs (range, 2.39-5.05 mmHg). The relative risk (RR) of confirmed hypoglycemia was significantly increased with dulaglutide, exenatide and glimepiride vs. placebo (RR range, 2.65-6.17); and was higher (RR>1.0) than all other agents except linagliptin. No agent reported an increased risk for UTI or GTI vs. placebo. CONCLUSIONS: In conclusion, when added to stable, optimized metformin and TZD therapy, all noninsulin antidiabetic agents reduced HbA1c, but differed in their effects on body weight, hypoglycemia and SBP.