Background: Peak systolic global longitudinal strain (GLS) is increasingly used to quantify left ventricular systolic function. The primary objective of this study was to assess whether GLS obtained during intraoperative transesophageal echocardiogram, performed before cardiopulmonary bypass, improves the prediction of postoperative low cardiac output syndrome (LCOS) after adult cardiac surgery.
Methods: GLS from 275 patients undergoing on-pump cardiac surgery was calculated retrospectively using two-dimensional- speckle tracking echocardiography (aCMQ module from Qlab software version 10.5, Philips Medical, Brussels, Belgium). LCOS was defined as the need for inotropic or mechanical circulatory support for >24 hours postoperatively. Patient and procedure characteristics associated with LCOS at the univariable level (P ≤ .05) were entered into a forward stepwise logistic regression to create a first predictive model. A second model was created by adding GLS. The 2 models were compared using the likelihood-ratio test, the area under the receiver operating characteristic (ROC) curve, and the integrated discrimination index. The optimal cutoff value of GLS associated with LCOS was determined by maximizing the Youden index of the ROC curve. Secondary outcomes included time until complete weaning from inotropes, discharge from the intensive care unit and from the hospital, and 30-day mortality.
Results: GLS was significantly associated with LCOS (P < .001) at the univariable level. Predictors of LCOS retained in the first model were cardiopulmonary bypass duration, decreased left ventricular ejection fraction, mitral valve surgery, and New York Heart Association functional class III or IV. Adding the GLS value improved the prediction of LCOS (P = .02). However, the area under the ROC curve did not differ between the 2 models (0.83; 95% confidence interval [CI], 0.77-0.99 vs 0.84; 95% CI, 0.79-0.90; P = .15). The integrated discrimination index associated with addition of GLS was 0.02 (P = .046), meaning that the difference in predicted risk between patients with and without LCOS increased by 2% after adding GLS. A GLS cutoff value of -17% (95% CI, -18.8% to -15.3%) was found to best identify LCOS. After adjusting for covariates included in model 1, a lower GLS value was significantly associated with a lower cumulative probability of weaning from inotropes postoperatively (hazard ratio, 0.90; 95% CI, 0.82-0.97; P = .01). No association was found between GLS and other secondary outcome measures.
Conclusions: GLS is an independent predictor of LCOS after on-pump cardiac surgery. Its incremental value over other established risk factors for postoperative LCOS is, however, limited.
Trial registration: ClinicalTrials.gov NCT02771080.