Left ventricular systolic function and mechanical ventilation weaning failure: An updated systematic review and meta-analysis with trial sequential analysis

Objective The impact of left ventricular (LV) systolic dysfunction on weaning failure is unclear. Design Updated meta-analysis assessing the association between LV ejection fraction (LVEF) and weaning failure. Setting A systematic search of MEDLINE and EMBASE for prospective studies reporting weaning according to echocardiographic data measured before starting a spontaneous breathing trial (SBT). Patients or participants Studies included in the meta-analysis that reported on weaning and echocardiographic data. Main variables of interest LVEF, E/e’ ratio, E velocity, deceleration time of the E wave, and e’ velocity. Results Three studies were added in this update ( n = 14, sample 20–30% larger). Lower LVEF ( n = 12) was associated with weaning failure: MD: −4.71 95%CI [−9.18, −0.23]; p = 0.04, I2 = 90%, low certainty of evidence), without subgroups difference according to SBT type or LVEF mean values. Trial sequential analysis showed that results were not robust (information size n = 745/1378). E/e’ ratio (MD: 1.90, 95%CI [0.48,3.32]; p = 0.009, I2 = 84%), E velocity (MD: 8.32, 95%CI [2.59,14.05]; p = 0.004, I2 = 54%) and shorter deceleration time of the E wave (MD: −12.01, 95%CI [−21.13, −2.89]; p = 0.01, I2 = 0%) were associated with weaning failure, whilst e’ velocity was not (MD: -0.59, 95%CI [−1.51,0.32]; p = 0.20, I2 = 64%). Conclusions An association of weaning failure with poorer LV systolic function (LVEF), not detected by the previous meta-analysis, was found. However, such association is highly fragile and with a mean difference below 5%, which seems not clinically relevant. Higher E/e’ ratio and other diastolic parameters confirmed their association with weaning failure, whilst e’ did not.