Purpose: To assess the association between serum ammonia level upon admission during the initial intensive care unit (ICU) stay and mortality.
Materials and methods: This retrospective cohort study included 2703 adult patients in eICU Collaborative Research Database. The ICU mortality within ammonia deciles were assessed. Logistic regression analyses were performed to analyze the relationship between ammonia and mortality.
Results: We defined three ammonia categories: <47, 47-111, and ≥111 μg/dL, corresponding to low, intermediate, and high ICU mortality. Increased ammonia was significantly associated with increased ICU mortality (per 10 μg/dL increase: odds ratio, 1.070 [95% confidence intervals, 1.05-1.09]; intermediate vs. low: 1.90 [1.41-2.56]; high vs. low: 4.38 [2.99-6.41]) and in-hospital mortality (1.06 [1.04-1.08]; 1.45 [1.13-1.87]; 3.41 [2.43-4.79]). Adding ammonia to the Acute Physiology and Chronic Health Evaluation (APACHE) IV score improved the area under the curve from 0.826 to 0.839 (P < 0.001) and from 0.806 to 0.813 (P = 0.001) for ICU and in-hospital mortality, respectively. Interaction and subgroup analyses demonstrated consistent results in patients with different APACHE IV scores, with or without hepatic diseases.
Conclusions: Elevated serum ammonia level in critically ill patients upon admission was an early risk factor for higher ICU and in-hospital mortality.
Keywords: Ammonia; Intensive care unit; Mortality; eICU collaborative research database.
Copyright © 2021. Published by Elsevier Inc.