Association of Documented Active Warming with Prolonged Postoperative Hospitalization After Endoscopic Spine Surgery: A Retrospective Cohort Study

Intraoperative temperature loss may occur during endoscopic spinal surgery because of general anesthesia, operative exposure, and irrigation. This study examined whether documented active warming was associated with intraoperative thermal outcomes and prolonged postoperative hospitalization after endoscopic spinal surgery. This single-center retrospective cohort included adults who underwent elective single-level endoscopic spinal surgery under general anesthesia from 2020 through 2023. All 115 patients in the active warming group (AWG) received documented forced-air warming, whereas none of the 135 patients in the conventional warming group (CWG) received active warming. Warmed intravenous fluids were additionally used in 86 AWG patients. The primary outcome was postoperative length of stay (LOS) >3 days. The primary logistic regression model was adjusted for age, sex, BMI, ASA physical status ≥ III, diabetes mellitus, surgical approach, surgical duration, and calendar year. Expanded-model, stabilized inverse probability of treatment weighting (IPTW), and categorical-calendar-year sensitivity analyses were also performed. The AWG had a higher median nadir temperature (36.4 °C vs. 35.7 °C), a smaller median temperature fluctuation (0.5 °C vs. 1.0 °C), and a lower incidence of intraoperative hypothermia (19/115 [16.5%] vs. 71/135 [52.6%]); all p < 0.001. Postoperative LOS >3 days occurred in 28 of 115 AWG patients (24.3%) and 59 of 135 CWG patients (43.7%). Documented active warming was associated with lower adjusted odds of prolonged postoperative hospitalization (OR 0.52, 95% CI 0.30-0.89; p = 0.018), with similar estimates in the expanded model (OR 0.54, 95% CI 0.31-0.94; p = 0.029), stabilized-IPTW analysis (OR 0.50, 95% CI 0.30-0.84; p = 0.008), and categorical-calendar-year sensitivity analysis (OR 0.53, 95% CI 0.30-0.92; p = 0.024). Documented active warming was associated with more favorable intraoperative thermal profiles and lower adjusted odds of postoperative LOS >3 days. The retrospective, nonrandomized design precludes causal inference and subgroup-specific treatment conclusions.