BackgroundVisual hallucinations (VH) are distressing and clinically relevant after cardiac surgery. C-reactive protein–albumin–lymphocyte (CALLY) index reflects systemic inflammation, nutrition, and immune competence. This study evaluated whether preoperative CALLY predicts 7-day postoperative VH.MethodsIn a prospective multicenter cohort (10 cardiac centers; September 2022 to May 2025), adults undergoing coronary or valve surgery were followed through day 7. Preoperative CALLY was calculated from routine laboratory measurements. VH discrimination was assessed using receiver operating characteristic (ROC) analysis, with the optimal cut-off by Youden’s index. Time-to-event analyses employed Cox regression with censoring at day 7, modeling CALLY continuously via restricted cubic splines (with 3 knots at the 10th/50th/90th percentiles) and categorically using a data-driven cut-off. Multivariable models adjusted for age, sex, body mass index (BMI), European System for Cardiac Operative Risk Evaluation II (EuroSCORE II), surgery type, cardiopulmonary bypass (CPB) time, left ventricular ejection fraction (LVEF), and postoperative ventilation hours.ResultsAmong 1,272 patients (150 VH events), the optimal CALLY cut-off was 1.12, separating Low (n=524) and High (n=748) groups. Kaplan–Meier curves diverged early (log-rank p<0.001). High vs. Low CALLY was associated with reduced VH risk (crude HR 0.39, 95% CI 0.28–0.54; adjusted HR 0.38, 95% CI 0.27–0.53; both p<0.001). When modeled continuously, CALLY demonstrated a non-linear association with VH risk (spline terms: HR 0.58 and 2.08; p=0.003 and p=0.023, respectively). Subgroup analyses showed broadly consistent associations, with no statistically significant effect modification by age, sex, diabetes, hyperlipidemia, or surgery type (all p for interaction >0.05).ConclusionsA lower preoperative CALLY index identifies cardiac-surgery patients at higher risk of VH and offers a practical tool for neuropsychiatric risk stratification.