Validation of two severity scores as predictors for outcome in Coronavirus Disease 2019 (COVID-19)

BACKGROUND: An established objective and standardized reporting of clinical severity and disease progression in COVID-19 is still not established. We validated and compared the usefulness of two classification systems reported earlier–a severity grading proposed by Siddiqi and a system from the National Australian COVID-19 guideline. Both had not been validated externally and were now tested for their ability to predict complications. METHODS: In this retrospective, single-centre observational study, patients hospitalized with confirmed COVID-19 across all severity stages were enrolled. The clinical severity was graded at admission and during hospitalization. Multivariate Cox regression was used to identify independent risk factors for mortality, a composite primary (mortality, incident acute respiratory distress syndrome, incident mechanical ventilation), a secondary endpoint (mortality, incident acute myocardial injury, incident venous thrombosis, pulmonary embolism or stroke) and progression of severity grades. RESULTS: Of 109 patients 17 died, 31 and 48 developed the primary and secondary endpoint, respectively. Worsening of the severity grade by at least one stage occurred in 27 and 28 patients, respectively. Siddiqi and Australian classification were identified as independent predictors for the primary endpoint (adjusted hazard ratio (aHR) 2.30, p<0.001 and aHR 2.08, p<0.001), for the secondary endpoint (aHR 2.12, p<0.001 and aHR 1.79, p<0.001) and mortality (aHR 2.30, p = 0.071 and aHR 1.98, p = 0.017). Both classification systems showed very good agreement regarding initial grading and good agreement regarding progression of severity stages. CONCLUSIONS: Standardized and objective severity grading is useful to unequivocally stratify patients presenting with COVID-19 for their individual risk of complications.