Left and right ventricular diastolic dysfunction as predictors of difficult separation from cardiopulmonary bypass

Abstract

As the evaluation of diastolic function can be complex in the setting of a busy cardiac operating room, its assessment may benefit from an algorithmic approach using transesophageal echocardiography. We developed a diagnostic algorithm which was then applied in a series of cardiac surgery patients to determine whether moderate to severe left ventricular diastolic dysfunction (LVDD) and right ventricular diastolic dysfunction (RVDD) can predict difficult separation from cardiopulmonary bypass (DSB). An algorithm using pulsed-wave Doppler interrogation of the mitral and tricuspid valve, the pulmonary and hepatic venous flow, and tissue Doppler interrogation of the mitral and tricuspid annulus was developed. The study was divided in two phases involving two groups of patients undergoing cardiac surgery. In phase I, echocardiographic evaluations of patients (n = 74) were used to test the reproducibility of the algorithm and to evaluate inter-observer variability using Cohen's kappa values which were calculated in three specific periods. In phase II, the algorithm was applied to a second group of patients (validation group, n = 179) to explore its prognostic significance. The primary end-point in phase II was DSB. In phase I, the kappa coefficients for LVDD and RVDD algorithms were 0.77 and 0.82, respectively. In phase II, moderate or severe degrees of LVDD were observed in 29 patients (16%) and moderate to severe RVDD was observed in 18 patients (10%) before cardiac surgery. Both moderate and severe LVDD (P = 0.017) and RVDD (P = 0.049) before surgery were observed more frequently in patients with DSB. Moderate and severe LVDD and RVDD can be identified with very good reproducibility, and both degrees of diastolic dysfunction are associated with DSB.