This study was performed to evaluate carnitine deficiency in a large series of epilepsy children and adolescents treated with old and new antiepileptic drugs with or without ketogenic diet. Plasma free carnitine was determined in 164 epilepsy patients aged between 7 months and 30 years (mean 10.8 years) treated for a mean period of 7.5 years (range 1 month-26 years) with old and new antiepileptic drugs as mono or add-on therapy. In 16 patients on topiramate or lamotrigine and in 11 on ketogenic diet, plasma free carnitine was prospectively evaluated before starting treatment and after 3 and 12 months, respectively. Overall, low plasma levels of free carnitine were found in 41 patients (25%); by single subgroups, 32 out of 84 patients (38%) taking valproic acid and 13 of 54 (24%) on carbamazepine, both as monotherapy or in combination, showed low free carnitine levels. A higher though not statistically significant risk of hypocarnitinemia resulted to be linked to polytherapy (31.5%) versus monotherapy (17.3%) (P=.0573). Female sex, psychomotor or mental retardation and abnormal neurological examination appeared to be significantly related with hypocarnitinemia, as well. As to monotherapy, valproic acid was associated with a higher risk of hypocarnitinemia (27.3%) compared with carbamazepine group (14.3%). Neither one of the patients on topiramate (10), lamotrigine (5) or ketogenic diet (11) developed hypocarnitinemia during the first 12 months of treatment. Carnitine deficiency is not uncommon among epilepsy children and adolescents and is mainly linked to valproate therapy; further studies are needed to better understand the clinical significance of serum carnitine decline.