Emergency department prediction of heart failure by bedside physical examination

Abstract

Study objectives: A variety of presenting symptoms and features of the bedside examination are all commonly used in the diagnosis and assessment of heart failure patients. We seek to determine the relative performance of presenting symptoms and features of the physical examination, including jugular venous distention (JVD), orthopnea, and computerized detection of abnormal heart sounds, in the emergency department (ED) for the diagnosis of heart failure. Methods: On presentation to the ED, and before treatment, patients with signs of heart failure had standardized reporting of symptoms and bedside examination (JVD, dyspnea, peripheral edema, paroxysmal nocturnal dyspnea, orthopnea), including a standard 12-lead ECG with acoustic signal recording for computerized detection of S3 using the validated AUDICOR algorithm (Inovise Medical, Inc.). Primary diagnosis of heart failure was based on hospital discharge diagnosis or ED diagnosis (if the patient was not admitted). Data were analyzed using logistic regression analysis. Wald values and P values are reported. Results: Of the 138 patients enrolled, 113 received full bedside examination and definitive diagnosis, 46 had a primary diagnosis of heart failure, and 67 had a non–heart failure diagnosis. Within the study population, there were 58.7% male patients in the heart failure group versus 44.8% male patients in the non–heart failure group, 67.4% black patients in the heart failure group versus 49.3% in the non–heart failure group, and the median age category was 66 to 70 years in the heart failure group versus 56 to 60 years in the non–heart failure group. Using a binary logistic regression model, features of bedside evaluation that were significant predictors of a primary diagnosis of heart failure (R2=0.396) were the AUDICOR S3 heart sound (11.599, P=.001) and peripheral edema (9.784, P=.002). JVD (0.981, P=.322), dyspnea (3.001, P=.083), orthopnea (2.857, P=.091), and paroxysmal nocturnal dyspnea (0.208, P=.648) were not significant determinants of primary diagnosis of heart failure. Conclusion: For bedside diagnosis of heart failure, the third heart sound detected by the computerized AUDICOR algorithm and the presence of peripheral edema were the best predictors for a primary diagnosis of heart failure.