Clinical Predictors of the Underlying Cause of Instability in the Emergency Department

Abstract

Abnormal vital signs and elevated lactate are used as proxies for physiologic instability in the emergency department (ED) and are associated with ICU admission and worse patient outcomes. Determining the underlying cause of instability can be difficult, and may lead to delay in appropriate ED interventions. This study evaluates the clinical predictors of underlying etiology in ED patients with physiologic instability. We performed a prospective observational study of patients with physiologic instability, defined as lactate ≥4mmol/L, or > 5 continuous minutes of heart rate (HR) ≥ 130, respiratory rate (RR) ≥ 24, shock index ≥1, systolic blood pressure < 90, in an urban, academic ED with 55,000 yearly visits. We excluded patients with isolated seizure, intoxication, psychiatric agitation, or tachycardia due to pain (ie, extremity fracture). We also excluded patients in atrial fibrillation with rapid ventricular response who were discharged once rate control was achieved. Clinical data, including elements of history and physical exam, were collected by chart review. The underlying causes of instability were retrospectively categorized as septic, cardiogenic, hemorrhagic, hypovolemic, and other by an attending Emergency Physician. Clinically significant predictors were determined by logistic regression using cause of instability as the outcome. This study was granted IRB approval with waiver of informed consent. We identified 396 patients who met inclusion criteria, excluded 107, leaving 289 patients for our analysis. The cause of instability was found to be sepsis in 133, cardiogenic in 36, hypovolemic in 34, hemorrhagic in 21, and other in 65 patients. Sepsis was predicted by fever history (OR 3.09, p100.4 (OR 3.28, p < 0.01), exam cellulitis (OR 6.58, p <0.01), SIRS WBC# (OR 2.11, p < 0.01), HIV (OR 6.59, p <0.01), when controlling for all other predictors and missing WBC#. Cardiogenic cause was predicted by history of congestive heart failure (OR 3.08, p <0.01) and shortness of breath (OR 3.37, p <0.01), and exam lower extremity edema (OR 3.58, p <0.01). Hemorrhagic cause was predicted by history of non-GI bleeding (OR 13.73, p = <0.01), and rectal exam showing bright red blood (OR 23.8, p <0.01) or melena (OR 39.67, p <0.01). Hypovolemic cause was predicted by nausea/vomiting (OR 2.53, p < 0.01) and diarrhea (OR 4.32, p = 0.01), while cough was protective (OR 0.34, p= 0.03). Despite the overlap of many clinical covariates between causes of instability, a few showed value in predicting the cause of instability and should be emphasized during the evaluation of these patients. Future studies may integrate these findings into decision tools for predicting the cause of instability.