Assessment of left ventricular diastolic function by transesophageal echocardiography before cardiopulmonary bypass : clinical implications of a restrictive profile

Résumé 0

Objective Left ventricular (LV) diastolic function can be assessed by transesophageal echocardiography before cardiopulmonary bypass in the setting of cardiac surgery. The objective of this study was to determine whether the assessment of LV diastolic dysfunction (LVDD) improves mortality risk prediction. Design Retrospective single-center cohort study. Setting Single tertiary cardiac surgery center. Participants Data from patients undergoing cardiac surgery using cardiopulmonary bypass (CPB) and for which an evaluation for LVDD was performed before CPB between February 1999 and November 2015. Interventions Cases were reviewed retrospectively from a transesophageal echocardiography hemodynamic database. LV diastolic function was graded as normal, impaired relaxation (grade 1), pseudo-normalization (grade 2), or restrictive (grade 3) determined by mitral inflow waves, tissue Doppler imaging of the mitral annulus, and pulmonary venous flow. The main outcome was in-hospital mortality. Measurements and Main Results A total of 760 patients were included, 144 (18.9%) patients with normal diastolic function, 331 (43.6%) patients with grade 1 LVDD, 218 (28.7%) patients with grade 2 LVDD, and 67 (8.8%) patients with grade 3 LVDD. In-hospital mortality occurred in 31 patients (4.1%). The presence of grade 3 LVDD was associated with an increased likelihood of in-hospital mortality (odds ratio [OR]: 19.39, confidence interval [CI]: 2.37-158.48, p = 0.006). In contrast, LV systolic dysfunction was not independently associated with increased mortality. When added to the Parsonnet score, the addition of diastolic function resulted in a net reclassification improvement of in-hospital mortality (NRI: 0.419 CI: 0.049-0.759, p = 0.02), and in integrated discrimination improvement (IDI: 0.0179 CI: 0.0049-0.031, p = 0.007). Difficult separation from CPB was observed more frequently in patients with grade 3 LVDD (62.9% v 36.1%, p = 0.01). Conclusions In contrast to LV systolic dysfunction, restrictive LVDD is associated with an increased risk of in-hospital mortality in cardiac surgical patients. Further studies should explore how this information may be used by the attending anesthesiologist to tailor perioperative management.