Protective Effect of Sodium-Glucose Cotransporter 2 Inhibitors (SGLT2i) on Liver Function: Results From MELD Analysis in Surgical Revascularization-A Preliminary Study.

BACKGROUND Coronary artery disease is associated with long-term systemic inflammation and perioperative multiorgan vulnerability, including transient hepato-renal dysfunction after cardiac surgery. Sodium-glucose cotransporter 2 inhibitors (SGLT2i) provide cardiorenal protection beyond glucose lowering via anti-inflammatory, metabolic, and endothelial effects, but their effect on perioperative liver-related risk markers, including the Model for End-Stage Liver Disease (MELD) score, in off-pump coronary artery bypass grafting (OPCAB) remains uncertain. MATERIAL AND METHODS We retrospectively studied 111 consecutive patients undergoing elective OPCAB in 2024 (chronic hepatic or renal disease excluded). Patients were stratified by long-term preoperative SGLT2i therapy (≥3 months): no SGLT2i (n = 85) vs SGLT2i (n = 26). MELD score, bilirubin, international normalized ratio , and creatinine were assessed preoperatively and on postoperative days 1 and 6. Between-group comparisons used the Mann-Whitney U test; within-group changes were analyzed using the Friedman test, Kendall's W, and Bonferroni post hoc testing RESULTS Baseline MELD and components were comparable. On postoperative day 1, SGLT2i group exhibited lower bilirubin (P < 0.001), INR (P = 0.04), and MELD score (P = 0.03), but higher creatinine (P = 0.002). By day 6, differences attenuated. Within-group analysis showed significant time effects for MELD, INR, and creatinine in both groups, with large effect sizes (W = 0.48-0.88). Total bilirubin changed over time in the control but not the SGLT2i group. CONCLUSIONS Long-term preoperative SGLT2i therapy was associated with distinct perioperative MELD dynamics in OPCAB, with relative preservation of bilirubin and INR despite creatinine changes, suggesting the potential modulation of hepatic function during perioperative stress. Given the observational design and possible confounding, results are hypothesis-generating and warrant prospective confirmation..