Context: GLP-1 analogues have recently been promoted as anti-hyperglycemic agents in critically ill patients with systemic inflammation, but the effects of TNF-α on glucose metabolism during GLP-1 administration are unknown. Objective: To determine whether infusion of TNF-α at high physiological levels impairs GLP-1's effects on glucose metabolism. Design: Randomized, controlled, cross-over trial. Setting: Hospital clinical research laboratory. Participants: Twelve healthy males (age 24±3 y; BMI 22.9±1.3 kg/m(2)). Interventions: Following an overnight fast, either saline (0.9%) or recombinant human TNF-α (1000 ng/m(2)/h) was infused from t = 0-6 hours. At t = 2 hours, GLP-1 infusion (0.5 pmol/kg/min) began. From t = 4-6 hours, the GLP-1 infusion rate was increased to 1.2 pmol/kg/min. Plasma glucose was clamped at 5 mmol/L throughout via a variable-rate 20% dextrose infusion. Trials were 7-14 days apart. Main outcome measures: Endogenous glucose production (EGP) was measured by [6,6-(2)H2]glucose isotope tracer dilution method. Results: GLP-1 infusion suppressed plasma glucagon (P<0.01), elevated plasma insulin and C-peptide (P<0.01), and suppressed EGP (P<0.001) during saline infusion. In contrast, infusion of TNF-α increased plasma TNF-α and IL-6, elevated body temperature, and blunted the GLP-1-induced suppression of EGP during high dose GLP-1 infusion (all P<0.05, TNF-α vs. saline). However, TNF-α infusion lowered plasma GLP-1 during high dose GLP-1 infusion (P<0.001). Conclusions: TNF-α induces systemic inflammation and reduces plasma GLP-1, thereby reducing the suppression of EGP during GLP-1 infusion. This may have clinical relevance if GLP-1 analogue drugs are used for the treatment of hyperglycemia in critically ill patients.