Association between cumulative intraoperative hypotension and postoperative ICU admission after gastrointestinal lumen cancer surgery: a retrospective cohort study using the INSPIRE database
Abstract
Background Intraoperative hypotension is associated with postoperative organ injury, but its relationship with postoperative intensive care unit (ICU) admission after gastrointestinal lumen cancer surgery remains uncertain. Methods We conducted a retrospective cohort study using the INSPIRE perioperative database. Adults undergoing their first gastrointestinal lumen excision or resection, identified by ICD-10-PCS 0DB*/0DT* codes and a temporally aligned ICD-10-CM C15–C21 diagnosis, were included. The primary exposure was cumulative duration of mean arterial pressure (MAP) <65 mmHg, categorized as 0, >0 to <10, 10 to <30, and ≥30 min. Multivariable logistic regression adjusted for demographic, clinical, laboratory, and surgical factors. Secondary analyses characterized the empirical MAP measurement intervals, time-weighted area below MAP 65 mmHg, and hypotension burden normalized to the monitored, strict operation, and strict anesthesia windows. Results Among 7,405 patients, 1,157 (15.62%) were admitted to the ICU. ICU admission rates were 7.15, 10.91, 15.25, and 32.01% across increasing exposure categories. In the complete-case expanded model (n = 6,045; 966 events), adjusted odds ratios were 1.02 (95% CI, 0.75–1.40), 1.29 (95% CI, 1.00–1.66), and 1.78 (95% CI, 1.36–2.33), respectively, compared with no exposure. Of 319,661 consecutive MAP intervals, 99.43% were 5 min. Each 50-mmHg·min increase in the area below MAP 65 mmHg was associated with an OR of 1.07 (95% CI, 1.04–1.10), and each 5-min/h increase in MAP <65 mmHg duration within the strict operation window with an OR of 1.14 (95% CI, 1.07–1.21); both p < 0.001. Conclusion Sustained intraoperative MAP <65 mmHg, particularly for ≥30 min, was associated with higher odds of postoperative ICU admission after adjustment for measured clinical and surgical factors. Findings remained consistent when hypotension depth and time-normalized burden were considered, but do not establish causality or an intervention threshold.