Comparative epidemiology of intraoperative hypotension and hypoxemia across three large retrospective cohorts.

BACKGROUND: Intraoperative hypotension is associated with morbidity and mortality, but comparable evidence for intraoperative hypoxemia remains sparse. Their joint epidemiology remains unclear. METHODS: We analyzed three retrospective noncardiac surgical cohorts (INSPIRE, MOVER, MSHS) totaling 165,646 cases. We defined binary hypotension as mean arterial pressure (MAP) <65 mmHg for >10 minutes, and hypoxemia as SpO2 <90% for ≥2 minutes (MOVER and MSHS) or a 5-minute median of 85% (INSPIRE). We calculated time-weighted average (TWA)-MAP burden in all cohorts, and TWA-SpO2 in MOVER and MSHS. Using logistic regression, we estimated associations of preoperative characteristics with each event, and of binary and TWA burden measures with acute kidney injury (AKI), non-routine ICU admission, and in-hospital mortality. AKI analyses included all cohorts; ICU and mortality analyses included INSPIRE and MSHS. RESULTS: Hypotension prevalence (12.7-28.8%) exceeded hypoxemia prevalence (2.3-4.9%); they co-occurred in <1% of cases. Hypotension clustered near induction, whereas hypoxemia had bimodal onset near induction and emergence. The two events had largely distinct preoperative associations. After mutual adjustment, binary hypoxemia was associated with all three outcomes, and binary hypotension with ICU admission and mortality but not AKI. TWA-MAP and TWA-SpO2 each showed graded associations with all three outcomes, although TWA-SpO2 estimates for ICU admission and mortality were from MSHS alone. Among event-positive cases, post-induction hypotension (64.8-88.1%) and mid-case hypoxemia (14.9-51.2%) were common but not associated with outcomes beyond TWA burden. CONCLUSIONS: Hypotension and hypoxemia rarely co-occurred and had distinct preoperative associations, but each was independently associated with outcomes. Further research should evaluate hypoxemia burden.