Decongestion Models and Metrics in Acute Heart Failure: ESCAPE Data in the Age of the Implantable Cardiac Pressure Monitor

Abstract

The United States Food and Drug Administration restricts the use of implantable cardiac pressure monitors to patients with New York Heart Association (NYHA) class III heart failure (HF). We investigated whether single-pressure monitoring could predict survival in HF patients as part of a model constructed using data from the ESCAPE (Evaluation Study of Congestive Heart Failure and Pulmonary Artery Catheterization Effectiveness) trial. We validated survival models in 204 patients, using all-cause 180-day mortality. Two levels of model complexity were tested: 1) a simplified 1-pressure model based on pulmonary artery mean pressure ([PAM]1P) (information obtainable from an implanted intracardiac monitor alone), and 2) a pair of 5-variable risk score models based on right atrial pressure (RAP) + pulmonary capillary wedge pressure (PCWP) ([RAP+PCWP]5V) and on RAP + PAM ([RAP+PAM]5V). The more complex models used 5 dichotomous variables: a congestion index above a certain threshold value, baseline systolic blood pressure of <100 mmHg, baseline blood urea nitrogen level of ≥ 34 mg/dL, need for cardiopulmonary resuscitation or mechanical ventilation, and posttreatment NYHA class IV status. The congestion index was defined as posttreatment RAP+PCWP or posttreatment RAP+PAM, with congestion thresholds of 34 and 42 mmHg, respectively (median pulmonary catheter indwelling time, 1.9 d). The 5-variable models predicted survival with areas under the curve of 0.868 for the (RAP+PCWP)5V model and 0.827 for the (RAP+PAM)5V model, whereas the 1-pressure model predicted survival with an area under the curve of 0.718. We conclude that decongestion as determined by hemodynamic assessment predicts survival in HF patients and that it may be the final pathway for treatment benefit despite improvements in pharmacologic intervention since the ESCAPE trial.

Keywords: Acute heart failure/diagnosis/mortality/therapy; blood pressure monitors; cardiovascular models; hemodynamics; monitoring, physiologic/instrumentation; risk assessment; survival.

Fig. 1

Kaplan-Meier plots show survival for patients in the ESCAPE trial cohort risk-stratified into tertiles for the 5-variable models A) (RAP+PCWP)5V and B) (RAP+PAM)5V, and for C) the 1-pressure risk score model. The plots for the 5-variable models are risk-stratified into tertiles of 67–100 (high), 33–66 (medium), and 0–32 (low). The 1-pressure model used a single pressure measurement, and the highest pressure measured in the cohort was normalized to 100. The corresponding pressure ranges (in mmHg) for PAM in C) are risk-stratified into tertiles of 50–74 (high), 25–49 (medium), and 0–24 (low). H = high mortality risk; L = low mortality risk; M = medium mortality risk; PCWP = pulmonary capillary wedge pressure; PAM = pulmonary artery mean pressure; RAP = right atrial pressure