# Association Between ST-Segment Deviation in Electrocardiography and 30-Day Mortality in Non-Cardiac Critically Ill Patients: A Retrospective Single-Center Study

## TL;DR

This study finds that ST-segment deviation on ECG is linked to higher 30-day mortality in non-cardiac ICU patients, even after adjusting for severity of illness.

## Contribution

The study demonstrates that ST-segment deviation is an independent predictor of mortality in non-cardiac ICU patients beyond the SOFA score.

## Key findings

- ST-segment deviation was present in 41% of ICU patients and more common in non-survivors.
- ST-segment deviation independently predicted 30-day mortality (HR = 1.534) after adjusting for SOFA score.
- Adding ST-segment deviation to the SOFA score modestly improved model discrimination.

## Abstract

Background/Objectives: ST-segment deviation (STD) on electrocardiography (ECG) may reflect myocardial injury in critically ill patients. However, its prognostic significance in non-cardiac intensive care unit (ICU) populations remains unclear. We aimed to assess the association between STD on ICU admission and 30-day mortality and to evaluate its incremental prognostic value beyond the SOFA score. Methods: In this retrospective single-center study, we included 307 consecutive ICU patients (median age: 64.0 years; 65.5% men). Patients with acute cardiac conditions were excluded. STD was defined as ≥1 mm ST elevation or depression in any lead on standard 12-lead ECG performed on admission. The primary outcome was 30-day all-cause mortality. Prognostic associations were assessed using multivariable Cox regression adjusted for SOFA score. Discriminative performance was evaluated by comparing ROC curves for models with and without STD, with bootstrap-based testing (1000 iterations) to assess significance. Results: STD was present in 126 patients (41.0%) and occurred more frequently in non-survivors (47.6% vs. 36.5%, p = 0.033. In Cox regression, STD was independently associated with 30-day mortality (HR = 1.534; 95% CI: 1.081–2.177; p = 0.017), even after adjustment for SOFA score. This association remained statistically robust in bootstrap validation. The addition of STD amplitude to the SOFA score modestly improved model discrimination with a borderline significant difference between the areas under the curve (ΔAUC = 0.005, p = 0.0581). Conclusions: ST-segment deviation on the admission ECG is an independent predictor of 30-day mortality in non-cardiac critically ill patients and may enhance risk stratification beyond the SOFA score.