RÉSUMÉ
OBJECTIVE: The objective of this study was to test the hypotheses that reduction of glycine and blocking of the N-methyl-D-aspartate receptor channel complex would be beneficial for both seizure reduction and developmental progress in patients with nonketotic hyperglycinemia. METHODS: We administered benzoate (at doses of 500 to 750 mg/kg/day) and dextromethorphan (at doses of 3.5 to 22.5 mg/kg/day) to four infants with nonketotic hyperglycinemia with follow-up of 3 months to 6 years. RESULTS: Benzoate reduced to normal the glycine concentration in plasma and substantially reduced but did not normalize the glycine concentration in cerebrospinal fluid. Dextromethorphan was a potent anticonvulsant in some but not all patients. There was remarkable interpatient variability in dextromethorphan metabolism. Three patients are living (ages ranging from 4 to 6 years) and are moderately to severely developmentally delayed; two are free of seizures. The third patient, with the slowest development, had intractable seizures for nearly a month before diagnosis, and although seizure-free for 30 months, now has grand-mal seizures. One patient died of intractable seizures at 3 months. CONCLUSIONS: These outcomes suggest that benzoate and dextromethorphan are not uniformly effective in nonketotic hyperglycinemia, but for some patients they improve arousal, decrease or eliminate seizures, and allow for some developmental progress. Trials with additional patients and other receptor channel blockers are warranted.