Peripheral venous oxygen saturation as an auxiliary indicator for improving the prediction of early in-hospital mortality in the emergency department

Peripheral venous oxygen saturation (SpvO₂), derived from a non-invasive peripheral venous blood gas (VBG) sample, reflects the balance between systemic oxygen delivery and consumption. Despite its physiological relevance, its prognostic utility for early in-hospital mortality in the emergency department (ED) has not been systematically evaluated. This single center retrospective observational study enrolled adult patients (≥ 18 years) who presented to the ED of a tertiary academic teaching hospital for medical reasons and underwent VBG analysis within the first hour of arrival between January 2020 and December 2024. Patients with cardiac arrest on arrival, trauma or external-cause presentations, SpO₂ <94%, or SpO₂–SpvO₂ gap < 20% were excluded. The primary outcome was 24-hour in-hospital mortality. Univariable and multivariable logistic regression, receiver operating characteristic (ROC) analysis, and Youden-index-derived optimal cutoff analysis were performed. Pre-specified subgroup analyses were conducted stratified by illness severity (NEWS ≤ 4 vs. > 4) and supplemental oxygen use. Of 69,827 screened patients, 20,967 met inclusion criteria; 119 (0.58%) died within 24 h. Mean SpvO₂ was significantly lower in non-survivors than survivors (40.3 ± 23.1% vs. 47.2 ± 17.9%; p < 0.001). On adjusted multivariable logistic regression, SpvO₂ remained an independent predictor of 24-hour mortality (adjusted OR 0.965 per 1% increase, 95% CI 0.944–0.986; p = 0.001). The Youden-optimal SpvO₂ cutoff was 20.8%, yielding sensitivity 31.9%, specificity 91.9%, and negative predictive value 99.6%. The SpvO₂ + Lactate + NEWS combination model achieved an AUC of 0.924, significantly outperforming NEWS alone. The independent association of SpvO₂ with mortality was consistent across all pre-specified subgroups for illness severity (NEWS ≤ 4 vs. > 4) and supplemental oxygen use. SpvO₂ is an independent prognostic marker of 24 h in-hospital mortality in ED patients across a broad range of illness severities. Although its standalone discriminative performance is modest, combining SpvO₂ with lactate and NEWS substantially improves risk stratification.