Perioperative immune dynamics during cardiopulmonary bypass and association with major adverse postoperative events

Purpose To clarify how cardiopulmonary bypass (CPB) affects perioperative immune cell populations and how these changes are associated with postoperative organ dysfunction. Methods We prospectively recruited 60 consecutive patients who underwent CPB as part of elective valvular surgery and/or coronary artery bypass grafting at our center. Peripheral blood samples were collected before surgery, during rewarming, at the end of CPB, and 24 h after surgery. The populations of neutrophils, monocytes, natural killer (NK) cells, T cells, and B cells were analyzed using flow cytometry, and changes were compared between patients with and without major adverse postoperative events (MAEs) within 30 days after surgery. MAEs included acute kidney injury, neurological dysfunction, liver injury, cardiovascular complications, and respiratory dysfunction. Results Of the 60 patients analyzed, 10 (16.7%) developed MAEs. CPB was associated with marked perioperative immune changes, with leukocyte and neutrophil counts peaking at 24 h after surgery, whereas lymphocyte counts declined and reached their nadir at the same time point. During rewarming, patients with MAEs had higher proportions of CD56dimCD314⁺ NK cells (96.74 ± 2.41 vs. 88.41 ± 11.75%, P = 0.035) and CD284⁺ non-classical monocytes (71.36 ± 17.95 vs. 49.13 ± 22.42%, P = 0.006), but lower proportions of CD163⁺ classical monocytes (69.46 ± 27.59 vs. 85.04 ± 15.53%, P = 0.035) and CD45RO⁺ T cells (35.62 ± 6.93 vs. 43.22 ± 11.66%, P = 0.023). At the end of CPB, patients with MAEs had higher proportions of CD4⁺CD38⁺ T cells (59.16 ± 15.16 vs. 41.78 ± 18.60%, P = 0.004) and CD274⁺ unswitched memory B cells (42.69 ± 15.29 vs. 23.80 ± 15.04%, P < 0.001). Exploratory receiver operating characteristic curves (ROC) analyses provided descriptive estimates of within-cohort discrimination for several immune subpopulations. In risk-adjusted Firth penalized logistic regression, higher CD284⁺ non-classical monocytes during rewarming [adjusted odds ratio (OR) = 3.118, 95% confidence interval (CI): 1.253–11.613, P = 0.011] and higher CD274⁺ unswitched memory B cells at the end of CPB (adjusted OR=4.516, 95% CI: 1.630–19.508, P = 0.002) remained statistically associated with MAEs. Conclusion In this single-center study, CPB was associated with dynamic perioperative immune changes, and several immune phenotypes were statistically associated with postoperative MAEs, warranting further validation in larger multicenter studies.