Predictive models assessing the impact of intensive care unit admission on the outcomes of acutely poisoned patients

Abstract

Although intensive care unit (ICU) admission is the cornerstone in management of critically acute poisoned patients, the decision of ICU admission is often challenging, especially with limited resources. Hence, our study aimed to assess predictive models of the impact of ICU admission on outcomes of patients with acute poisoning. This retrospective cohort study recruited records of acutely poisoned patients admitted to Tanta University Poison Control Center between 2021 and 2023. Patient demographic and toxicological data, as well as initial clinical and laboratory profiles, were retrieved. Afterward, patients were categorized according to mortality and complicated outcomes. Out of 221 acutely poisoned patients admitted to the ICU, the incidences of mortality and complications in survivors were 54.3% and 57.4%, respectively. Aluminum phosphide (ALP) was the most common cause of poisoning (59%), with a significant association with mortality and predominance in cardiac complications. However, respiratory and neurological complications were evident among illicit substances, cholinesterase inhibitors, and neuropsychiatric drugs. The model anticipating morality included time from presentation to ICU admission, mean arterial pressure (MAP), oxygen saturation, pH, and ALP poisoning. Furthermore, the complication predictive model comprised time from exposure to poison center presentation, time from presentation to ICU admission, and MAP. Both models exhibited good to excellent discrimination performance and consistent calibration. Accordingly, prompt admission of all ALP-poisoned patients to ICU with a highly standardized level of care may alleviate their deleterious outcomes. However, drug categories with reversible courses should be adequately treated with frequent respiratory and hemodynamic monitoring in less-equipped ICUs.