Intraoperative hypotension trajectories and their predictive value for major postoperative complications: a retrospective cohort study.

Background: Intraoperative hypotension (IOH) is a common hemodynamic disturbance during major non-cardiac surgery, yet the prognostic significance of different temporal blood pressure patterns remains unclear. This study aimed to identify distinct IOH trajectories using group-based trajectory modeling (GBTM) and to evaluate their independent and incremental predictive value for major postoperative complications in high-risk surgical patients. Methods: We conducted a retrospective cohort study of 789 adults undergoing elective major abdominal, urologic, or gynecologic surgery between January 2018 and December 2023. Continuous invasive minute-by-minute mean arterial pressure (MAP) recordings were extracted from the anesthesia information management system. IOH was defined as MAP <65 mmHg. GBTM modeled MAP over absolute intraoperative time using polynomial time functions to identify three latent IOH trajectories based on duration and recurrence: transient mild (<10 min), moderate sustained (10-30 min), and prolonged/fluctuating (>30 min or ≥3 episodes). The primary composite outcome included acute kidney injury, postoperative delirium, unplanned ICU admission within 48 h, and 30-day all-cause mortality. Associations were examined using multivariable logistic regression, and predictive performance was evaluated using ROC curves, calibration, bootstrap internal validation, and decision curve analysis. Results: A clear exposure-response relationship was observed across trajectory groups: the primary composite complication occurred in 13.4% of patients in the transient mild group, 20.8% in the moderate sustained group, and 30.7% in the prolonged/fluctuating group (p for trend <0.001). Compared with transient mild hypotension, adjusted odds ratios were 1.58 (95% CI 1.03-2.43) for moderate sustained and 2.42 (95% CI 1.54-3.80) for prolonged/fluctuating trajectories. Incorporating trajectory classification into a clinical model markedly improved discrimination (AUC 0.860 vs. 0.578), calibration, and net clinical benefit compared with conventional IOH measures alone. Conclusion: Distinct intraoperative hypotension trajectories derived from high-resolution arterial pressure data were strongly and independently associated with major postoperative complications and substantially enhanced predictive accuracy beyond standard IOH metrics. Trajectory-based hemodynamic profiling may support individualized blood pressure management and early perioperative risk stratification.