H. Chen, X. Wang
Background and Objective: To compare clinical features of valproate (VPA)-induced hyperammonemic encephalopathy and asymptomatic hyperammonemia to inform monitoring strategies. Methods: This cross-sectional study at Affiliated Hospital of Jiangsu University included 25 adults with epilepsy hospitalized between January 2024 and March 2026 with serum ammonia >33 umol/L during VPA therapy. Nine met predefined VPA-induced hyperammonemic encephalopathy (VHE) criteria; 16 had asymptomatic hyperammonemia. Clinical, laboratory, electroencephalography (EEG), and neuroimaging parameters were compared. Results: VHE patients had higher ammonia (median 76.0 [IQR 51.2-155.7] vs 42.5 [33.5-66.0] umol/L; P<0.001) and shorter VPA duration (median 0 [0-60] vs 30 [0-120] months; P=0.005) than asymptomatic patients. Crucially, VPA trough concentrations and liver enzymes did not differ between groups. Diffuse EEG slowing occurred exclusively in VHE patients(66.7% vs 0%; P<0.001); prior cerebral infarction was more frequent in VHE (88.9% vs 43.8%; P=0.040). All VHE patients improved clinically within 1-3 days and achieved ammonia normalization within 3-10 days after VPA withdrawal without specific ammonia-lowering therapy. Conclusions: VHE typically emerges early in VPA therapy, is unpredictable by routine therapeutic drug monitoring or liver function tests, and resolves rapidly upon discontinuation. Clinicians should maintain a low threshold for ammonia testing in patients with new neurological symptoms during early VPA treatment, particularly those with prior cerebrovascular disease or diffuse EEG slowing. These findings support integrating ammonia screening into VPA safety protocols beyond standard therapeutic drug monitoring.