Difference of Ammonia Level as Predictor of Delayed Neurologic Complication in Patients with Glufosinate Ammonium Herbicide Poisoning Presented with Alert Mentality

ABSTRACT

PURPOSE: The serum ammonia level was postulated as a surrogate marker for severe neurotoxicity in glufosinate ammonium (GLA) poisoning. The aim of this study is to evaluate whether the level of serum ammonia can predict delayed neurologic complications in patients with GLA poisoning presented with alert mentality. METHODS: Thirty-six GLA-poisoned patients presented to our emergency department with alert mentality initially were analyzed retrospectively. The baseline characteristics, laboratory findings, ammonia level (initial and second ammonia level, frequency of hyperammonemia, and difference of ammonia level), and clinical outcomes were compared between non-neurologic (n=16) and neurologic complication groups (n=20). RESULTS: Neurologic complications occurred in 20 patients (55.6%) with 14.3 hours (median) of latent period. The initial ammonia level and frequency of initial hyperammonemia did not show any difference between the two groups. However, the difference of ammonia level between the 2nd and 1st samples was an independent predictor of delayed neurologic complication (adjusted odds ratio; 1.184 (95% confidence interval (CI); 1.01-1.387, p=0.037)). The area under the curve and cut-off point of the difference of ammonia level for the prediction of delayed neurologic complication was 0.936 (95% CI; 0.756-0.992) and 15.4 umol/L respectively. CONCLUSION: The difference of ammonia level rather than the initial ammonia level could be used to predict delayed neurologic complication in GLA-poisoned patients presented with alert mentality.