Predicting in-hospital mortality in AECOPD: A comparison of the mSOFA score with DECAF, Ottawa, and GOLD classifications.

BACKGROUND

Acute exacerbations of chronic obstructive pulmonary disease (AECOPD) involve systemic inflammation and organ dysfunction, yet commonly used prognostic tools may rely on subjective elements or focus mainly on respiratory severity. The modified Sequential Organ Failure Assessment (mSOFA) incorporates objective metabolic and perfusion markers, including lactate.

OBJECTIVES

We compared mSOFA with DECAF, the Ottawa COPD Risk Scale, and the GOLD severity classification for predicting in-hospital mortality in emergency department (ED) patients with AECOPD.

METHODS

In this prospective observational cohort study, patients presenting to a university hospital ED with AECOPD (September 2024-May 2025) were enrolled. The primary endpoint was in-hospital mortality. Discrimination was assessed using the area under the receiver operating characteristic curve (AUC) and compared with DeLong's test. Clinical utility was evaluated using decision curve analysis (DCA).

RESULTS

Among 440 patients, in-hospital mortality was 5.9% (n = 26). Non-survivors had higher lactate (2.8 vs. 1.2 mmol/L, p < 0.001) and more severe acidosis (pH 7.28 vs. 7.38, p < 0.001). mSOFA showed the highest discrimination (AUC 0.905; 95% CI 0.860-0.950), followed by Ottawa (0.871), DECAF (0.865), and GOLD classification (0.787). AUCs did not differ significantly among mSOFA, DECAF, and Ottawa (all p > 0.05), while each outperformed GOLD (all p < 0.05). At an optimal cutoff ≥3, mSOFA sensitivity was 88.5% and specificity 81.4%. DCA showed the highest net benefit for mSOFA across threshold probabilities 0.05-0.63.

CONCLUSION

mSOFA demonstrated high discriminative performance and favorable clinical net benefit for predicting in-hospital mortality in ED patients with AECOPD and may be useful for risk stratification.