Delayed valproic acid toxicity: A retrospective case series

Study Objective: We estimate how often an initially undetectable or nontoxic valproic acid (VPA) concentration misrepresents the potential toxicity of an acute VPA overdose. Methods: All patients followed up by the San Diego and San Francisco Divisions of the California Poison Control System between January 1997 and June 1999 with a documented serum VPA concentration of 120 μg/mL or greater at any time were identified through a search of poison center databases. Cases involving patients who had serial VPA concentrations and an initial VPA concentration that was unmeasurable or less than 100 μg/mL were also identified. The methods of decontamination, clinical outcomes, highest measured concentration, and times after ingestion of the initial and highest measured concentration of VPA were determined by means of chart review for these cases. Results: Of the 173 patients identified with VPA concentrations of 120 μg/mL or greater, 21 (12.1%) had a detectable initial concentration not in the toxic range, and 5 (2.9%) had initial concentrations of less than the limit of detection. These 26 cases involved acute ingestion of VPA preparations. Four of these 26 patients were discharged to home or to a psychiatric facility after the initial results were obtained. Conclusion: Delayed toxicity might occur in patients after an acute overdose of enteric-coated VPA. Initial serum VPA concentrations might be misleading. Serial measurements documenting declining VPA concentrations or prolonged observation are recommended to determine whether a patient is medically safe for discharge or psychiatric placement. [Ingels M, Beauchamp J, Clark RF, Williams SR. Delayed valproic acid toxicity: a retrospective case series. Ann Emerg Med. June 2002;39:616-621.]