Point-of-care echocardiography of the right heart improves acute heart failure risk stratification for low-risk patients: The REED-AHF prospective study

Abstract

Objectives: Validated acute heart failure (AHF) clinical decision instruments (CDI) insufficiently identify low-risk patients meriting consideration of outpatient treatment. While pilot data show that tricuspid annulus plane systolic excursion (TAPSE) is associated with adverse events, no AHF CDI currently incorporates point-of-care echocardiography (POCecho). We evaluated whether TAPSE adds incremental risk stratification value to an existing CDI.

Methods: Prospectively enrolled patients at two urban-academic EDs had POCechos obtained before or <1 h after first intravenous diuresis, positive pressure ventilation, and/or nitroglycerin. STEMI and cardiogenic shock were excluded. AHF diagnosis was adjudicated by double-blind expert review. TAPSE, with an a priori cutoff of ≥17 mm, was our primary measure. Secondary measures included eight additional right heart and six left heart POCecho parameters. STRATIFY is a validated CDI predicting 30-day death/cardiopulmonary resuscitation, mechanical cardiac support, intubation, new/emergent dialysis, and acute myocardial infarction or coronary revascularization in ED AHF patients. Full (STRATIFY + POCecho variable) and reduced (STRATIFY alone) logistic regression models were fit to calculate adjusted odds ratios (aOR), category-free net reclassification index (NRI_cont ), ΔSensitivity (NRI_events ), and ΔSpecificity (NRI_nonevents ). Random forest assessed variable importance. To benchmark risk prediction to standard of care, ΔSensitivity and ΔSpecificity were evaluated at risk thresholds more conservative/lower than the actual outcome rate in discharged patients.

Results: A total of 84/120 enrolled patients met inclusion and diagnostic adjudication criteria. Nineteen percent experiencing the primary outcome had higher STRATIFY scores compared to those event free (233 vs. 212, p = 0.009). Five right heart (TAPSE, TAPSE/PASP, TAPSE/RVDD, RV-FAC, fwRVLS) and no left heart measures improved prediction (p < 0.05) adjusted for STRATIFY. Right heart measures also had higher variable importance. TAPSE ≥ 17 mm plus STRATIFY improved prediction versus STRATIFY alone (aOR 0.24, 95% confidence interval [CI] 0.06-0.91; NRI_cont 0.71, 95% CI 0.22-1.19), and specificity improved by 6%-32% (p < 0.05) at risk thresholds more conservative than the standard-of-care benchmark without missing any additional events.

Conclusions: TAPSE increased detection of low-risk AHF patients, after use of a validated CDI, at risk thresholds more conservative than standard of care.

Keywords: cardiology; echocardiography; emergency department; emergency medicine; heart failure; point-of-care ultrasound; risk stratification.

FIGURE 1

Change in sensitivity (NRI‐Events) and specificity (NRI‐Nonevents) when TAPSE is added to STRATIFY versus STRATIFY alone, for various thresholds of 30‐day risk for serious adverse heart failure events. When added to the STRATIFY risk score, TAPSE ≥ 17 mm increased specificity (TNR/NRI_nonevents, i.e., correct identification of patients without a 30‐day event) without any significant change in sensitivity (i.e., no change in missed events) for risk thresholds benchmarked to the actual 30‐day outcome rate among discharged patients. If the use of TAPSE + STRATIFY were used conservatively (i.e., only at risk thresholds more conservative than the adverse event rate occurring under standard‐of‐care EP disposition decision making) there still would have been a significant increase in the number of low‐risk patients identified without significant change in missed cases. AMI, acute myocardial infarction; CABG, coronary artery bypass graft; NRI, net reclassification index; MCS, mechanical cardiac support; PCI, percutaneous coronary intervention; TNR, true‐negative rate = NRI_nonevents = Δspecificity; TPR, true‐positive rate = NRI_events = Δ sensitivity; TAPSE, tricuspid plane systolic excursion.

FIGURE 2

Relative importance of echocardiographic variables for predicting 30‐day serious adverse heart failure events in conjunction with the STRATIFY decision instrument importance on random forest of each echocardiography variable in predicting the primary outcome of the STRATIFY decision score* after adjusting for the STRATIFY score itself. Eight of the 10 variables which added the most risk prediction to STRATIFY were right heart measures. TAPSE, the primary measure of interest, was the most important echocardiography variable. As a single variable, TAPSE was >70% as important as the entire 13 variable STRATIFY score in predicting STRATIFY's outcome. *30‐day death, CPR, mechanical cardiac support, intubation, new or emergent dialysis, acute myocardial infarction, and/or coronary revascularization. FAC, RV fractional area change; fwRVLS, free‐wall RV longitudinal strain; LV, left ventricle; LVEF, left ventricular ejection fraction; LVGLS, LV global longitudinal strain; PASP, pulmonary artery systolic pressure; PVR, pulmonary vascular resistance; PW, pulsed wave; RV, right ventricle; RVDD, right ventricular diastolic diameter; RVOT, right ventricular outflow tract; TAPSE, tricuspid annular plane systolic excursion.