Electrocardiographic Patterns as Predictors of Mortality in Aluminum Phosphide Poisoning: A Retrospective Cohort Single-Center Study.

Background Aluminum phosphide (ALP) poisoning, prevalent in developing countries, poses a significant public health concern due to its high lethality, especially when ingested. ALP releases phosphine gas, which induces severe cardiotoxic effects, manifesting through various electrocardiographic (ECG) abnormalities. These ECG patterns are critical in predicting mortality among ALP poisoning patients. With mortality rates often exceeding 70%, early identification of these ECG markers can be pivotal in guiding clinical interventions, improving patient outcomes, and reducing the fatality associated with ALP poisoning. This study seeks to develop a prognostic model that uses ECG findings to predict the risk of mortality, providing a vital tool in resource-limited settings for rapid risk stratification and treatment optimization. Objective This study aimed to develop a prognostic model that utilizes ECG findings to predict the risk of mortality in patients with ALP poisoning. Methodology This retrospective cohort study was conducted at the Mardan Medical Complex from December 2022 to June 2024. The study included 64 patients diagnosed with ALP poisoning who met the inclusion criteria, focusing on those with complete medical records and ECG data. Patients were categorized into survivors and non-survivors. Statistical analyses included t-tests for continuous variables, chi-squared tests for categorical variables, and logistic regression to identify significant predictors of mortality. ECG patterns were analyzed using a random forest classifier to determine the most predictive features, and Kaplan-Meier survival curves were used to assess survival probabilities stratified by ECG abnormalities. Interobserver variability in ECG interpretation was resolved by consensus. Results Among the 64 patients studied, 44 (68.75%) were non-survivors, and 20 (31.25%) were survivors. Significant differences were observed in key variables between these groups. Non-survivors had a higher prevalence of atrial fibrillation (70.45% vs. 65.00%; p=0.047) and atrial premature beats (29.55% vs. 0%; p=0.006) and exhibited longer QT intervals (544.32±40.66 ms vs. 420.0±15.89 ms; p<0.001) and QRS durations (137.5±31.1 ms vs. 120.68±41.06 ms; p=0.047). They also had lower Glasgow Coma Scale (GCS) scores (6.61±3.51 vs. 8.2±3.55; p=0.029) and shorter hospital stays (20.39±27.17 hours vs. 75.80±15.00 hours; p=0.032), reflecting rapid clinical decline. Non-survivors ingested a higher number of ALP pills (4.52±1.56 vs. 2.5±0.95; p=0.043). Kaplan-Meier survival curves and logistic regression analyses confirmed that atrial fibrillation, QT prolongation, and wide QRS complexes were the strongest predictors of mortality, highlighting the prognostic significance of these ECG patterns in ALP poisoning. Conclusion This study confirms that the number of ALP pills ingested is a key determinant of the severity of cardiotoxic effects, as reflected in specific ECG patterns such as QT prolongation and atrial fibrillation. Recognizing this dose-dependent relationship can improve clinical outcomes by guiding the intensity of monitoring and treatment strategies based on the quantity of ALP exposure.