Assessing right ventricular systolic function in a population of unselected patients before cardiac surgery: A multiparametric approach is mandatory

Background and aim: According to recent American guidelines, RV dysfunction can be diagnosed on a single parameter: 1) Peak systolic velocity during ejection period (S') <10 cm/sec 2) RV fractional area change (RVFAC) <35%. The aim of our study was to assess these recent recommendations in a large non-selected cohort of patients awaiting cardiac surgery and to evaluate less validated RV function criteria. Methods: Four hundred and thirteen patients (means values of age and LVEF respectively: 70.3±10.3 years and 62±13%) were enrolled of which 63% were awaiting valve surgery, 49% coronary artery bypass grafting and 3% others. To evaluate RV function, the following parameters were obtained: parameters derived from pulsed tissue Doppler at the tricuspid annulus (S', right myocardial performance index (RMPI), isovolumic acceleration (IVA)), RVFAC, and 2D Strain of the basal lateral wall. Indices of RV preload (IVC collapse index and right atrium RA area) and afterload (systolic pulmonary artery pressure, pulmonary vascular resistance) were also recorded. Results: Out of 413 patients, 320 patients (77.5%) had normal RV function (group 1, defined by S'>10cm/s and RVFAC> 35%,). RV dysfunction was probable in 54 patients (group 2, defined by S'<10cm/s OR RVFAC< 35%). RV dysfunction was highly probable in 39 patients (group 3, defined by S'<10cm/s AND RVFAC< 35%). Using group 1 and 3, other less-validated parameters were evaluated: IVA ≤1.8 m/s2 and basal 2D-Strain ≥-17% had both the best diagnostic value to detect RV dysfunction with a sensitivity of 86% and a specificity > 80%. A cut-off value of RMPI30.60 allowed an acceptable discrimination (AUC 0.79) to predict RV dysfunction with a specificity of 80%. Moreover, IVA was found to be the least load-dependent while basal 2D-Strain and RMPI appeared to be afterload and preload dependent. Conclusions: In patients with a suspicion of RV dysfunction, S' and RVFAC are frequently discordant (58% of the patients). Using a group of highly probable RV dysfunction (S'<10cm/s AND RVFAC< 35%), we found that IVA and basal 2D-Strain have both a good diagnostic value. Moreover, contrary to 2D-strain, IVA was not influenced by loading conditions, adding to its diagnostic value. Our results underline the need of a multiparametric approach to diagnose RV dysfunction, and for this setting both IVA and 2D strain could help.