Intraoperative hypotension burden and blood pressure variability in relation to in-hospital mortality after noncardiac surgery A retrospective cohort study using the MOVER database

Background Intraoperative hypotension burden and blood pressure variability describe different aspects of hemodynamic instability. Their associations with mortality may depend on measurement quality and the severity of the patients included. We examined both measures in invasively monitored adults undergoing noncardiac surgery. Methods This retrospective cohort study used the EPIC component of the Medical Informatics Operating Room Vitals and Events Repository. The first eligible adult inpatient operation per patient was selected. Mean arterial pressure recordings required at least 20 valid measurements and 80% valid-interval coverage during an operating-room stay of at least 30 minutes. Hypotension burden was the time-weighted deficit below 65 mmHg over valid observed intervals; variability was the average absolute difference between consecutive measurements. Logistic models estimated associations with all-cause in-hospital mortality, adjusting for age, sex, American Society of Anesthesiologists physical status, operating-room duration, and procedure category. Results Among 3872 patients, 98 died in hospital (2.53%). In the jointly adjusted model, each interquartile-range increase in log-transformed hypotension burden was associated with higher mortality odds (odds ratio 1.295; 95% confidence interval 1.120–1.498). The variability estimate was imprecise (0.823; 0.636–1.065). Exploratory adjustment for preoperative hospital duration and documented circulatory support yielded estimates of 1.337 (1.157–1.545) for burden and 1.012 (0.776–1.321) for variability. The burden association attenuated after excluding the highest 1% of burden values (1.179; 0.967–1.436) or excluding physical status 5 patients with additional adjustment for preoperative hospital duration (1.166; 0.926–1.468). Stricter numerical quality checks yielded an estimate of 1.359 (1.173–1.573). Conclusions Higher observed hypotension burden was associated with in-hospital mortality in this selected cohort, whereas variability showed no clear conditional association. Attenuation after excluding extreme exposures or the most severely ill patients limits generalization. These findings do not establish a causal treatment effect or a safe blood pressure threshold.