Prognostic Stratification of Clinically Stable Patients with Heart Failure by Echocardiographic Pressure/Volume Loop Model

Abstract

Pressure/volume (P/V) loops provide useful information on left ventricular performance and prognosis in patients with heart failure (HF) but do not lend themselves to routine clinical practice. The authors developed a noninvasive method to compute individualized P/V loops to predict adverse clinical outcomes in patients with stable HF, which the authors believe can be used clinically. A derivation cohort (n= 443 patients) was used to develop an echocardiography P/V loop model, using brachial arterial pressure and trans-thoracic two-dimensional Doppler echocardiographic data. Each patient's P/V loop was depicted as an irregular pentagon, and a centroid was derived for each loop. The centroid distance (CD) from a reference centroid (derived from 101 healthy control subjects) was computed. This model was prospectively applied to 435 patients who constituted the validation cohort. The study end point was a composite of cardiac death or hospitalization for HF among study patients. In the derivation cohort, CD was threefold greater among patients who experienced adverse events than those who did not. During a follow-up period of 30 months (15-45 months), event rates were 35% (72 of 206 patients) and 12% (29 of 237 patients P<.001), respectively, among patients with CD>33mL/mm Hg and those with CD≤33mL/mm Hg (prognostic cutoff derived by receiver operating characteristic analysis). Multivariate Cox analysis identified CD as an independent predictor of adverse outcome (hazard ratio, 1.61; 95% CI, 1.03-2.50) independently of left ventricular end-diastolic volume, pulmonary capillary wedge pressure, and left ventricular ejection fraction. These conclusions were confirmed in the validation cohort. The authors propose a method to create a noninvasive P/V loop and its centroid. These data provide useful pathophysiologic and prognostic information in patients with HF.