Sedating critically ill patients often involves prolonged opioid infusions causing opioid tolerance. Naloxone has been hypothesized to limit opioid tolerance by decreasing adenylate cyclase/cyclic adenosine monophosphate activation. The study purpose was to investigate the effect of low-dose naloxone on the maximum cumulative daily fentanyl dose in critically ill children. We conducted a double-blinded, randomized, placebo-control trial from December 2002 through July 2004 in a university PICU. We enrolled 82 children age 1 day to 18 years requiring mechanical ventilation and fentanyl infusions anticipated to last for >4 days were eligible for enrollment. Those receiving additional oral analgesia or sedation, having a history of drug dependence or withdrawal, or having significant neurologic, renal, or hepatic disease were excluded. In addition to fentanyl infusions, patients received low-dose naloxone or placebo infusions. Medications were adjusted using the Modified Motor Activity Assessment Scale. Withdrawal was monitored using the Modified Narcotic Withdrawal Scale. Intervention was a low-dose naloxone infusion (0.25 microg/kg per hour) and the main outcome variable was the maximum cumulative daily fentanyl dose (micrograms per kilogram per day). There was no difference in the maximum cumulative daily fentanyl dose between patients treated with naloxone (N = 37) or those receiving placebo (N = 35). Adjustment for the starting fentanyl dose also failed to reveal group differences. Total fentanyl dose received throughout the study in the naloxone group (360 microg/kg) versus placebo (223 microg/kg) was not statistically different. Placebo patients trended toward fewer rescue midazolam boluses (10.7 vs 17.8), lower total midazolam dose (11.6 mg/kg vs 23.9 mg/kg), and fewer rescue fentanyl boluses (18.5 vs 23.9). We conclude that administration of low-dose naloxone (0.25 microg/kg per hour) does not decrease fentanyl requirements in critically ill, mechanically ventilated children.