Bloodstream Infections After Cardiac Surgery: Pathogen Distribution, Antimicrobial Resistance, and a Nomogram Prediction Model

Background: Postoperative bloodstream infection (BSI) is a severe complication following cardiac surgery, associated with significant mortality. This study aims to characterize pathogen profiles, antimicrobial resistance and independent risk factors for BSI in a Chinese cardiac surgery cohort. Methods: We conducted a retrospective cohort study of postoperative BSI in patients who underwent cardiac surgery at the First Affiliated Hospital of Nanjing Medical University from January 2019 to December 2024. Patients who developed postoperative BSI were identified and a 1:1 propensity score matched analysis were performed to adjust for baseline characteristics of patients. Univariate and multivariate analysis were applied to identify independent risk factors for postoperative BSI. R software was used to construct a prediction nomogram, followed by internal validation. Results: During the study, a total of 10,007 patients underwent heart surgery, and 164 pairs of patients were successfully matched using propensity score matching. Gram-negative bacteria were the most common pathogens (64.1%), followed by Fungi (18.2%) and Gram-positive bacteria (17.6%). Among Enterobacterales , 33.8% were carbapenem-resistant, with carbapenem-resistant Klebsiella pneumoniae (CRKP) accounting for 63.6% of K. pneumoniae isolates. Meanwhile, 85.0% (17/20) of A. baumannii isolates were multidrug-resistant A. baumannii (MDRAB). The 90-day mortality rate in the BSI group was 39.0%, significantly higher than that in the non-BSI group (4.9%, P < 0.001). Multivariate logistic regression identified impaired cardiac function (New York Heart Association II–IV, NYHA II–IV) (OR=5.583, 95% CI:2.681– 11.627), pulmonary infection (OR =4.053, 95% CI:1.692– 9.710), emergency surgery (OR=9.629, 95% CI:2.962– 31.301), coronary artery bypass grafting (CABG) (OR=2.796, 95% CI:1.096– 7.134), continuous renal replacement therapy (CRRT Postoperative ICU stay ≥ 14 d) (OR=4.758, 95% CI:1.854– 12.211) and postoperative ICU stay ≥ 14 d (OR=14.616, 95% CI:6.765– 31.577) as independent risk factors for postoperative BSI. The nomogram showed good discrimination and calibration after internal validation, with an optimism-corrected C-index of 0.913 and a mean absolute calibration error of 0.026. Conclusion: Postoperative BSI in cardiac surgery patients is associated with high mortality. A nomogram based on impaired cardiac function (NYHA II–IV), pulmonary infection, emergency surgery, CABG, CRRT and postoperative ICU stay ≥ 14 d displayed high accuracy and consistency in predicting the risk of postoperative BSI. These results facilitate early risk stratification and guide preventive strategies in high-risk cardiac surgical patients.