Carnitine deficiency and hyperammonemia in children receiving valproic acid with and without other anticonvulsant drugs.

Plasma ammonia and total and free carnitine were measured in 84 children requiring anticonvulsant drugs: 32 patients (group A) on valproic acid alone, 28 children (group B) on polytherapy including valproic acid, and 24 patients (group C) on polytherapy without valproic acid. The other anticonvulsant drugs used in groups B and C were carbamazepine and phenobarbital. Plasma ammonia concentrations were elevated in both group A and B compared with controls. Group B patients showed significantly higher hyperammonemia than group A (59.9 +/- 16.3 micrograms/dl vs. 36.7 +/- 12.4 micrograms/dl; P < 0.05). Group C patients had plasma ammonia levels similar to those of controls (31.1 +/- 14.7 micrograms/dl vs. 29.7 +/- 12.1 micrograms/dl; NS). In both group A and group B patients, plasma ammonia levels were correlated with the valproic acid dosage (r = 0.32, P < 0.01) and with serum concentrations of valproic acid (r = 0.41, P < 0.001). Moreover, a significant correlation between plasma ammonia and duration of valproic acid therapy was found in the patients as a whole (r = 0.31, P < 0.01). Plasma total and free carnitine concentrations were significantly reduced in groups A and B (total carnitine 36.9 +/- 6.9 mumol/l vs. 32.9 +/- 9.7 mumol/l; free carnitine 28.9 +/- 5.1 mumol/l vs. 25.7 +/- 4.3 mumol/l, respectively) compared with group C patients who did not receive valproic acid and in whom values were similar to controls (total carnitine 46.1 +/- 9.0 mumol/l vs. 47.7 +/- 10.1 mumol/l; free carnitine 40.1 +/- 7.1 mumol/l vs. 42.9 +/- 8.0 mumol/l, respectively). Twenty-eight patients (18 of group A and 10 of group B) were re-evaluated and showed a complete normalization of plasma ammonia, and total and free carnitine levels which were similar to controls. Our data suggest that hyperammonemia is an important problem in patients receiving valproic acid, particularly in association with other anticonvulsant drugs. This increase of plasma ammonia and the concomitant reduction of carnitine seem to be transient and completely reversible.