Venous Congestion, Not Cardiac Index is Associated with Diuretic Resistance

Abstract

Background Diuretic resistance is associated with worse outcomes in acute decompensated heart failure (ADHF). Traditional dogma has held that low cardiac index (CI) is a driver of diuretic resistance by reducing renal perfusion. Recently, accumulating evidence has shown that venous congestion has a more potent effect on cardiorenal interactions than low CI. However, venous congestion also likely identifies patients with increased volume overload and the potential for a greater diuretic response. Hypothesis Right atrial pressure (RAP) will be associated with diuretic response while CI will have little to no association. Methods We analyzed a subset of patients from the Evaluation Study of Congestive Heart Failure and Pulmonary Artery Catheterization Effectiveness (ESCAPE) dataset (n=190). Diuretic efficiency (DE) was defined as daily urine output (UOP) per doubling of loop diuretic dose, and compared with baseline, final (day of pulmonary artery catheter (PAC) removal) and change in PAC variables. Daily UOP was additionally examined with time-matched PAC measurements. Results DE was weakly positively correlated with baseline RAP (r=0.16, p=0.03) but negatively correlated with final RAP (r=-0.21, p=0.009) and change in RAP from baseline (r=-0.27, p=0.001). There was no correlation between DE and the remaining PAC variables, including CI. In high vs. low DE groups dichotomized by median value, the low DE group had significantly higher final RAPs (9.0, IQR 6.0-15.5 mmHg vs. 8.0, IQR 4.8-12.0 mmHg, p=0.04) and smaller reductions in RAP (-2.0 mmHg, IQR-5.0-2.0 mmHg vs. -5.0, IQR=-10.0-0.0 mmHg, p=0.03). There were no DE group differences for baseline CI (p=0.78), final CI (p=0.42), or change in CI (p=0.92). There were no significant differences in DE in subjects with CI 2.2 L/min/m2 (p=0.95). Time-matched PAC and UOP data showed a weak positive correlation between RAP and UOP adjusted for diuretic dose (r=0.13, p=0.03), but no correlation between UOP and CI (p=0.18). Conclusions We were unable to identify any association between CI and DE. Higher baseline RAP predicted better DE, but later in the hospital course higher RAP predicted worse DE, likely indicating that elevated RAP is a marker rather than a mediator of diuretic response. Overall, these results suggest that PACs provide limited value in the workup of diuretic resistance.