OBJECTIVE: To determine the utility of the metabolic autopsy in the hospitalized pediatric patient. STUDY DESIGN: This was a retrospective review of all metabolic autopsies performed at a large pediatric hospital over a 5-year period. Premortem clinical diagnoses were correlated with autopsy findings and results of postmortem testing. RESULTS: Of the 23 metabolic autopsies performed, a metabolic disorder was diagnosed before death in 4 and after death by extensive studies initiated before death in 2. In the remaining 17 cases, postmortem samples were inadequate for subsequent enzymatic analysis in 2, a nonmetabolic explanation for symptoms was identified in 4, and no unifying diagnosis could be defined in 8. A metabolic disorder was diagnosed by postmortem tissue study in 3 of the 17 cases; in all 3 of these cases the patient died in the neonatal period after limited premortem investigation for primary lactic acidosis. For the 8 subjects who had undergone an extensive laboratory workup before death, in each case metabolic autopsy failed to establish a diagnosis. CONCLUSIONS: In a small but significant percentage of cases (18%), the metabolic autopsy successfully identified an undiagnosed metabolic disease. However, metabolic autopsy following an extensive nondiagnostic clinical workup is unlikely to yield a specific metabolic diagnosis.