Cardiopulmonary bypass time, intraoperative hematocrit and transfusion predict neurocognitive decline in patients undergoing cardiac surgery

BACKGROUND: Neurocognitive decline (NCD) after cardiopulmonary bypass is a well-established phenomenon. While numerous preoperative risk factors have been identified, the influence of intraoperative factors on NCD still needs further exploration. METHODS: We conducted a prospective cohort study to evaluate NCD and associated pre-operative and intra-operative factors. Patients undergoing CABG or valvular surgeries under cardiopulmonary bypass from 2021 to 2023 were recruited. Preoperative characteristics and postop outcomes were extracted from chart review. Intraoperative variables were extracted from anesthesia and perfusionist records. Neurocognitive performance was assessed using the Repeatable Battery for the Assessment of Neuropsychological Status (RBANS) score at baseline, POD4, and 1 month, and NCD is defined as a decrease of ≥ 8 points from baseline. Results were analyzed via multivariable regression models and Pratt score analysis to identify strongest predictors of NCD. RESULTS: Out of the 132 enrolled patients who underwent planned cardiac operations, 94 patients completed neurocognitive assessments on POD4, and 62 patients completed both on POD4 and at 1-month postop. Forty-two (45%) patients had NCD on POD4. NCD in the immediate postop period is more common in patients who are female, with higher BMI, and undergoing valvular surgery compared to CABG (all p < 0.05). Significant intra-op characteristics associated with NCD include lower intraoperative minimum temperature, more transfusion of blood products, and longer CPB and aortic cross-clamp time (all p < 0.05). In multivariable regression models, younger age, lower preoperative creatinine, lower intraoperative hematocrit, longer CPB time and lower volume collected in CellSaver were associated with NCD (all p < 0.05). Pratt score analysis identified CPB time and intraoperative hematocrit as the two strongest intraoperative predictors of NCD on POD4. Patients with NCD had longer ICU and hospital stays, but experienced no significant difference in complication rates compared to the non-NCD cohort. At 1-month follow-up, only 10 (16%) patients experienced persistent NCD, and repeat analyses with multivariable regression models showed only CPB time and cross-clamp time remained statistically significant between cohorts. CONCLUSION: Longer cardiopulmonary bypass time and lower intraoperative hematocrit were strongly associated with early postoperative NCD. Intraoperative strategies to limit hemodilution should be considered as part of neuroprotective management in cardiac surgery.