Valproate metabolites and hepatotoxicity in an epileptic population

Idiosyncratic hepatotoxicity, although rare, is of major concern when one is treating patients with valproate (VPA). Several clinical criteria are associated with an increased risk of developing this complication, but more specific predictors are needed. It has been postulated that 4-en-VPA or one of its further metabolites may be responsible for the hepatic toxicity and that under certain conditions the metabolism of VPA is shifted to this product. We postulated that measurement of serum concentrations of 4-en-VPA or another metabolite might be a simple technique that would be predictive of risk for developing idiosyncratic hepatotoxicity. Because this complication is rare, we chose to analyze our data by a multiple linear regression model, exploring associations between VPA or three of its metabolites and clinical risk factors for hepatotoxicity. 4-en-VPA correlated with older age and absence of encephalopathy. 4-en-VPA was only seen in patients receiving polytherapy; all patients were also receiving CBZ. 2-en-VPA correlated with poor nutritional status. We conclude that routine measurement of serum 4-en-VPA is unlikely to be a useful predictor of risk for developing fatal hepatotoxicity. Serum concentrations of 4-en-VPA may not reflect presence or effects in the liver as it may be metabolized to further intermediates or be bound to tissue. Thus, serum levels of 4-en-VPA do not reflect its important role in the pathogenesis of hepatotoxicity. This metabolite was detected only in patients receiving polytherapy, a potent risk factor for developing this rare complication.