Interactive Effect between On-Scene Hypoxia and Hypotension on Hospital Mortality and Disability in Severe Trauma

Abstract

Background: It is unclear whether effect size of the hypoxia is different on in-hospital mortality and disability according to hypotension status in the field. Methods: Adult severe trauma (ST) patients during 2012–13 who were treated by emergency medical services (EMS) and had abnormal revised trauma scores in the field or who had positive trauma triage criteria were analyzed. Exposure was hypoxia (<94%) measured by EMS. End points were hospital mortality and disability defined as a Glasgow Outcome Scale that decreased by 2 points or more. Multivariable logistic regression with interaction model between hypoxia and hypotension was used for outcomes to calculate the adjusted odds ratios (AOR) with 95% confidence intervals (95% CIs) after adjusting for potential confounders. Results: A total of 17,406 EMS-ST patients were analyzed. Of those, 2,598 (14.9%) died, and 3,292 (21.5%) were considered disabled at discharge. The total hypoxia group showed higher mortality and disability indices (35.7 and 51.2%) than the non-hypoxia group (10.7 and 15.9%) (each p-value < 0.0001). The AOR of hypoxia was 2.15 (1.92–2.40) for mortality and was 1.97 (1.75–2.21) for disability. In the interaction model, AORs for mortality by hypoxia in the hypotensive and non-hypotensive groups were 2.66 (2.32–3.04) and 1.74 (1.61–1.87), respectively (P < 0.0001 for interaction). The AORs for disability in the hypotensive and non-hypotensive groups were 2.17 (1.87–2.53) and 1.55 (1.42–1.69), respectively (P < 0.0001 for interaction). Conclusions: The effect of hypoxia was much greater in the hypotensive group than in the non-hypotensive group both in terms of mortality and disability.