DOI: 10.1111/acem.70429 ISSN: 1069-6563

Association of Initial Peripheral Venous pO 2 With 48‐Hour Clinical Deterioration in Emergency Department Patients Undergo

Jill Corbo, Peter Gruber, Nicholas Acosta, Aunaly Palmer, Benjamin W. Friedman

ABSTRACT

Background

Peripheral venous pO 2 is routinely reported with venous blood gas (VBG) results despite uncertainty regarding its prognostic significance. We evaluated whether initial peripheral venous pO 2 was associated with 48‐h clinical deterioration and provided prognostic information beyond age and lactate.

Methods

We conducted an ambispective observational cohort study of adults undergoing VBG testing at a single urban academic emergency department, with prospective enrollment and index‐data collection and retrospective ascertainment of 48‐h outcomes. The primary outcome was a composite of ICU admission, death, vasopressor use, or positive‐pressure ventilation within 48 h. The primary analysis modeled venous pO 2 continuously and used multivariable logistic regression adjusted for age and lactate. Incremental performance was evaluated exploratorily by comparing an age‐and‐lactate model with a model additionally containing venous pO 2 .

Results

Among 973 patients, 103 (10.6%) experienced the composite outcome. Median venous pO 2 was 33 mmHg (IQR 25–44). Venous pO 2 was not associated with clinical deterioration after adjustment for age and lactate (adjusted odds ratio [aOR] 1.054 per 10‐mmHg increase; 95% CI 0.957–1.161; p  = 0.289). The AUROC was 0.649 for the age‐and‐lactate model and 0.654 after adding venous pO 2 (ΔAUROC 0.005; DeLong p  = 0.390).

Conclusion

In this single‐center cohort of adult ED patients undergoing VBG testing, no association was observed between peripheral venous pO 2 and 48‐h clinical deterioration. Adding venous pO 2 did not improve apparent prognostic performance beyond age and lactate. We cannot exclude smaller associations or potential utility in selected disease‐specific populations. Confirmation in larger multicenter cohorts using consecutive enrollment and external validation is needed.

More from our Archive