Abstract
BACKGROUND: To evaluate the association of a comprehensive perioperative care strategy on postoperative delirium (POD) incidence and perioperative recovery in elderly patients undergoing surgery for gastrointestinal tumors. This institutional pathway extended routine ERAS care by incorporating a nurse-led, protocolized multicomponent bundle specifically targeting modifiable delirium precipitants across the perioperative period. METHODS: This retrospective study included 260 elderly patients scheduled for elective radical resection for gastric or colorectal cancer between January 2023 and December 2024. Patients were categorized into a conventional care group (control, n = 130) and an enhanced perioperative care group (enhanced care, n = 130). POD incidence and severity within 7 days after surgery were assessed using the Delirium Rating Scale-Revised-98 (DRS-R-98). Secondary outcomes included sleep quality (Pittsburgh Sleep Quality Index, PSQI), melatonin levels, cognitive function (Mini-Cog), activities of daily living (Barthel index), inflammatory and stress biomarkers (CRP, IL-6, cortisol), and recovery parameters [time to first ambulation, first flatus, hospital stay, and Quality of Recovery-15 (QoR-15) scores]. RESULTS: Compared with the control group, the enhanced care group showed significantly lower POD incidence and severity (P < 0.05). Patients managed under the enhanced care protocol also had lower PSQI scores, fewer nocturnal awakenings, higher melatonin levels, better Mini-Cog performance, and higher Barthel index scores (P < 0.001). Postoperative CRP, IL-6, and cortisol levels were also lower in this group (P < 0.05). In addition, time to first ambulation, first flatus, and hospital stay were significantly shorter, whereas QoR-15 scores were significantly higher (P < 0.001). CONCLUSION: In this retrospective cohort, comprehensive perioperative care was associated with lower POD incidence and severity, better sleep, cognitive and functional recovery, and reduced stress and inflammatory markers. Due to the study's retrospective design, findings may be influenced by unmeasured confounders; prospective studies are needed to confirm these results.