Multicenter case series of valproic acid ingestion: serum concentrations and toxicity

Abstract

Background: Valproic acid exposures reported to poison centers have increased more than 4-fold over the last 5 years. There are no large case series published on valproic acid ingestion.

Methods: A prospective multicenter case series of all patients reporting an ingestion of valproic acid. Data collected included: age, gender, dose ingested, concomitant medications, symptoms and vital signs, laboratory values, length of hospital stay, and medical outcome. Entrance into the study required a serum valproic acid concentration above the therapeutic threshold of 100 microg/mL. Statistical analysis was by Fisher's exact test.

Results: A total of 335 patients were reported to participating centers of which 186 (55%) had serum valproic acid concentrations greater than 100 microg/mL. Of the 186 cases, 53 were multiple drug exposures leaving 133 cases of sole valproic acid ingestion for evaluation. Age ranged from 2 to 66 years with a mean of 30.1 years +/- 12. Peak serum valproic acid concentrations ranged from 110 microg/mL to 1840 microg/mL with a mean of 378.3 microg/mL +/- 310.2 microg/mL. Time from postingestion to the peak measured valproic acid concentration ranged from 1 to 18 hours, with a mean of 7.4 hours +/- 3.9. Symptoms included lethargy (n = 94), coma (n = 19), tachycardia (n = 24), aspiration (n = 8), metabolic acidosis (n = 8), and hypotension (n = 4). A peak concentration of > 450 microg/mL was more likely to be associated with a moderate or major adverse outcome (p < 0.005). A peak concentration > 850 microg/mL was more likely to be associated with coma (p < 0.005) and acidosis (p < 0.005). Eleven patients experienced transient thrombocytopenia (platelets < 150,000) and all had peak valproic acid concentrations >450 microg/mL. Four patients experienced transient leukopenia (WBC < 3,500). The mean hospital stay for all patients was 42 +/- 33.1 hours. A hospital stay > 48 hours was more likely to be associated with a peak valproic acid concentration > 450 microg/mL (p < 0.05). There were 2 fatalities.

Conclusions: In this case series, patients with peak valproic acid concentrations above 450 microg/mL were more likely to develop significant clinical effects and have longer hospital stays. A peak valproic acid concentration above 850 microg/mL was more likely to be associated with coma, respiratory depression, aspiration, or metabolic acidosis.