OBJECTIVE To test a method to predict the end-diastolic pressure-volume relationship (EDPVR) from a single beat in patients with heart failure. RESULTS Patients (New York Heart Association class III-IV) scheduled for mitral annuloplasty (n=9) or ventricular restoration (n=10) and patients with normal left ventricular function undergoing coronary artery bypass grafting (n=12) were instrumented with pressure-conductance catheters to measure pressure-volume loops before and after surgery. Data obtained during vena cava occlusion provided directly measured EDPVRs. Baseline end-diastolic pressure (P(m)) and volume (V(m)) were used for single-beat prediction of EDPVRs. Root-mean-squared error (RMSE) between measured and predicted EDPVRs, was 2.79+/-0.21 mm Hg. Measured versus predicted end-diastolic volumes at pressure levels 5, 10, 15 and 20 mm Hg showed tight correlations (R(2)=0.69-0.97). Bland-Altman analyses indicated overestimation at 5 mm Hg (bias: pre-surgery 44 ml (95% CI 29 to 58 ml); post-surgery 35 ml (23 to 47 ml)) and underestimation at 20 mm Hg (bias: pre-surgery -57 ml (-80 to -34 ml); post-surgery -13 ml (-20 to -7.0 ml)). End-diastolic volumes were significantly different between groups and between conditions, but these differences were not dependent on the method (ie, measured versus predicted). RMSEs were not different between groups or conditions, nor dependent on V(m) or P(m), indicating that EDPVR prediction was equally accurate over a wide volume range. CONCLUSIONS Single-beat EDPVRs obtained from hearts spanning a wide range of sizes and conditions accurately predicted directly measured EDPVRs with low RMSE. Single-beat EDPVR indices correlated well with directly measured values, but systematic biases were present at low and high pressures. The single-beat method facilitates less invasive EDPVR estimation, particularly when coupled with emerging non-invasive techniques to measure pressures and volumes.