Background: Sodium nitroprusside (SNP) is often used in postoperative pediatric cardiac surgical patients. Cyanide toxicity may occur with the use of SNP. There is a paucity of literature describing dosing parameters or physical signs and symptoms of toxicity with SNP. Objective: To determine the incidence of cyanide toxicity in postoperative pediatric cardiac surgical patients treated with SNP and identify dosing parameters and physical signs and symptoms that may predict elevated cyanide concentrations. Methods: Medical records of patients who received SNP in the pediatric cardiac intensive care unit from January 2002 through December 2002 were identified and evaluated for cyanide and thiocyanate levels, dosing, and signs and symptoms of toxicity. Patients were included if they had received SNP after cardiac surgery, were 18 years of age or less, and had at least one cyanide or thiocyanate level determined while receiving therapy. Patients were excluded if they had received sodium thiosulfate. The Mann-Whitney U test was used to determine significant differences in mean dose, duration of infusion, renal function, serum lactate, and acid-base status between groups with elevated or nonelevated levels. Logistic regression and receiver operator curve were used to determine variables associated with elevated levels. Relationships between signs and symptoms of toxicity and elevated levels were evaluated with Fisher's exact test. Results: Cyanide concentrations were in the toxic range in 7 of 63 (11%) patients. Patients with elevated concentrations had significantly higher mean dose, cumulative dose, and acid-base excess values. Elevated cyanide levels were independently predicted by mean dose, cumulative dose, and acid-base excess values, and a dose of 1.8 μg/kg/min predicted an elevated cyanide concentration with 89% sensitivity and 88% specificity. Adverse events were not reliable predictors of elevated cyanide levels. Conclusions: Mean dose of SNP is the best predictor of elevated cyanide levels. Adverse events commonly associated with cyanide toxicity may not be reliable indicators of elevated cyanide concentrations.