Preoperative platelet-to-lymphocyte ratio and post-hepatectomy hemorrhage risk after elective hepatectomy: a single-center cohort study with exploratory calendar-time assessment of a candidate cutoff

Post-hepatectomy hemorrhage (PHH) is clinically important after liver resection, yet simple preoperative bleeding risk markers remain limited. We assessed whether preoperative platelet-to-lymphocyte ratio (PLR) was associated with PHH and derived an ROC-based threshold. In this analysis of a prospectively maintained single-center cohort, 4,570 consecutive hepatectomy cases were screened from 2022 to 2024 and 4,386 adults undergoing elective hepatectomy were analyzed. PLR was calculated from the complete blood count closest to surgery within 7 preoperative days. The primary endpoint was ISGLS-defined PHH within 30 days, adjudicated from a fixed postoperative hemoglobin baseline and documented postoperative bleeding. A 2022–2023 cohort was used for exploratory threshold derivation, and a 2024 calendar cohort was used for internal temporal testing. PHH occurred in 233 of 4,386 patients (5.3%), including 121 grade B/C events (2.8%). In derivation, continuous PLR yielded an AUC of 0.664, and the exploratory candidate threshold was 124.8 (sensitivity 58.4%; specificity 67.1%). Across the full cohort, each 25-unit lower PLR was independently associated with PHH (adjusted OR 1.10, 95% CI 1.06–1.14; p < 0.001). PLR ≤125 was also independently associated with any PHH (adjusted OR 1.88, 95% CI 1.44–2.46; p < 0.001) and grade B/C PHH (adjusted OR 1.76, 95% CI 1.24–2.50; p = 0.002). In the internal temporal test cohort, continuous PLR yielded an AUC of 0.647 and the fixed threshold ≤125 separated groups with higher versus lower PHH incidence (8.9% vs. 3.8%). At the derivation-defined threshold, the positive predictive value remained low (9.0% in derivation and 8.9% in the internal temporal test cohort). Lower preoperative PLR was associated with higher post-hepatectomy hemorrhage risk and with a small increase in model discrimination when added to standard clinical variables. A candidate threshold near 125 separated higher- and lower-risk groups in an internal temporal test; however, its low positive predictive value does not support stand-alone clinical use, and the threshold remains exploratory pending external multicenter validation.