Objective: To evaluate the effect of terlipressin on oxygenation, PaO2/FIO2, heart rate, mean arterial pressure, and mortality in children with septic shock refractory to high doses of dopamine/dobutamine and adrenaline. Design and setting: A randomized, nonblind study in the pediatric intensive care unit of a university hospital. Patients and measurements: We studied 58 children with septic shock and refractory hypotension despite fluid loading and high doses of catecholamines, randomly enrolled to terlipressin (TP, n=30) or control (n=28). TP was administered as intravenous bolus doses of 20 mu g/kg every 6 h for a maximum of 96 h. Hemodynamic changes, PaO2/FIO2 rates, length of stay, and mortality rate in PICU were recorded prospectively. Results: Mean arterial pressure and PaO2/FIO2 significantly increased, and heart rate significantly decreased 30 min after each TP treatment, but mortality did not differ from control (67.3% vs. 71.4%). Mean stay in the PICU was shorter in the TP group (13.4 +/- 7.9 vs. 20.2 +/- 9.7 days and was longer among nonsurvivors of the TP group vs. control (10.4 +/- 6.9 vs. 6.2 +/- 3.4 days). Blood urea nitrogen, creatinine, AST, ALT, and urine output of patients in the TP group did not change after terlipressin. Conclusions: Although terlipressin infusion had no effect on mortality, it significantly increases mean arterial pressure, PaO2/FIO2, and survival time in nonsurvivors. Terlipressin seems to cause no adverse effect but warrants further evaluation as a rescue therapy in refractory septic shock.