Electrocardiographic predictors of adverse cardiovascular events in suspected poisoning

Poisoning is the second leading cause of injury-related fatality in the USA and the leading cause of cardiac arrest in victims under 40 years of age. The study objective was to define the electrocardiographic (ECG) predictors of adverse cardiovascular events (ACVE) complicating suspected acute poisoning (SAP). This was a case-control study in adults at three tertiary-care hospitals and one regional Poison Control Center. We compared 34 cases of SAP complicated by ACVE to 101 consecutive control patients with uncomplicated SAP. The initial ECG was analyzed for rhythm, intervals, QT dispersion, ischemia, and infarction. ECGs were interpreted by a cardiologist, blinded to study hypothesis and case data. Subjects were 48% male, with mean age 42 +/- 19 years. In addition to clinical suspicion of poisoning in 100% of patients, routine toxicology screens were positive in 77%, most commonly for benzodiazepines, opioids, and/or acetaminophen. Neither the ventricular rate, the QRS duration, nor the presence of infarction predicted the risk of ACVE. However, the rhythm, QTc, QT dispersion, and presence of ischemia correlated with the risk of ACVE. Independent predictors of ACVE based on multivariable logistic regression were prolonged QTc, any non-sinus rhythm, ventricular ectopy, and ischemia. Recursive partitioning analysis identified very low risk criteria (94.1% sensitivity, 96.2% NPV) and high risk criteria (95% specificity). Among patients with SAP, the presence of QTc prolongation, QT dispersion, ventricular ectopy, any non-sinus rhythm, and evidence of ischemia on the initial ECG are strongly associated with ACVE.