Key result
Adding CPB duration and intraoperative transfusion improves postoperative LCOS prediction by ~0.10 AUC.
Why the study?
Most low cardiac output syndrome risk assessment approaches rely primarily on preoperative information, whereas key factors such as cardiopulmonary bypass exposure, bleeding, and transfusion become available only during surgery.
Does updating a preoperative risk prediction model with intraoperative cardiopulmonary bypass duration and allogeneic transfusion (Model B) improve the prediction of postoperative low cardiac output syndrome compared with preoperative characteristics alone (Model A) in older adults undergoing cardiac surgery?
Cohort (n=1,777)
No
Does updating a preoperative risk prediction model with intraoperative cardiopulmonary bypass duration and allogeneic transfusion (Model B) improve the prediction of postoperative low cardiac output syndrome compared with preoperative characteristics alone (Model A) in older adults undergoing cardiac surgery?
Updating preoperative risk estimates at the end of cardiac surgery with two routinely available intraoperative variables (cardiopulmonary bypass duration and allogeneic transfusion) substantially improves discrimination, calibration, and clinical net benefit for predicting postoperative low cardiac output syndrome in older adults.
Answers are grounded in this paper and cite their sources.
May improve LCOS risk stratification with routine intraoperative data; extends preoperative models but leaves open prospective validation and clinical impact.
Background Low cardiac output syndrome (LCOS) is a major complication after cardiac surgery. Most risk assessment approaches rely primarily on preoperative information, whereas key factors such as cardiopulmonary bypass exposure, bleeding, and transfusion become available only during surgery. To align risk assessment with the progressive accumulation of perioperative information, we developed and temporally validated a two-stage strategy that generates an initial risk estimate from preoperative characteristics and then updates that estimate using routinely available intraoperative information. Methods This single-center retrospective cohort study included patients aged ≥ 65 years who underwent eligible cardiac surgery in 2024 as the development cohort. Patients treated from January through June 2025 formed the temporal validation cohort and were not used for model fitting. Model A was based on prespecified preoperative clinical variables, and Model B additionally included cardiopulmonary bypass duration and intraoperative allogeneic transfusion. Missing predictors were handled using 20 multiple imputations, and internal validation used 1,000 bootstrap resamples. Temporal validation assessed the area under the receiver operating characteristic curve (AUC), calibration-in-the-large (CITL), calibration slope, Brier score, and decision-curve net benefit. Paired between-model differences were estimated using 2,000 bootstrap resamples. Results The development cohort included 1,171 patients, of whom 101 developed LCOS (8.6%), and the temporal validation cohort included 606 patients, of whom 50 developed LCOS (8.3%). After 1,000 bootstrap resamples, optimism-corrected development AUCs were 0.699 (95% CI, 0.649–0.753) for Model A and 0.749 (95% CI, 0.705–0.799) for Model B. In the reserved 2025 temporal validation cohort, the AUC of Model B was 0.756 (95% CI, 0.695–0.810), compared with 0.655 (95% CI, 0.571–0.729) for Model A, yielding a paired difference of 0.101 (95% CI, 0.045–0.161; P < 0.001). Model B also had a calibration slope closer to the ideal value (0.858 vs 0.684), with a CITL of − 0.113. Model B showed clear incremental net benefit at risk thresholds of 5% and 10%, and its AUC remained 0.736–0.759 across sensitivity analyses. Brier scores were similar between models (0.0725 vs 0.0739). Conclusions Adding cardiopulmonary bypass duration and intraoperative allogeneic transfusion to the prespecified preoperative model markedly improved LCOS discrimination in temporal validation, produced a calibration slope closer to the ideal value, and yielded incremental net benefit at lower risk thresholds. This two-stage strategy uses routinely available perioperative information to provide interpretable risk updating and may support early postoperative risk stratification, intensified monitoring, and resource allocation.
No takes yet. Share an insight, caveat, or question.
Ning et al. (2026) conducted a cohort in Older adults undergoing cardiac surgery at risk for low cardiac output syndrome (LCOS) (n=1,777). Two-stage perioperative LCOS risk prediction model (Model B: preoperative variables plus cardiopulmonary bypass duration and intraoperative allogeneic transfusion) vs. Preoperative-only risk prediction model (Model A: age, sex, BMI, serum creatinine, serum albumin, LVEF, NYHA class III–IV, non-elective surgery, and previous cardiovascular surgery) was evaluated on Discrimination (AUC) for predicting low cardiac output syndrome (LCOS) within 48 hours after surgery in the temporal validation cohort (AUC difference 0.101, 95% CI 0.045-0.161, p=<0.001). Adding cardiopulmonary bypass duration and intraoperative transfusion to a preoperative model improved temporal-validation AUC for LCOS from 0.655 to 0.756.
Synapse has enriched 8 closely related papers on similar clinical questions. Consider them for comparative context: