Development and validation of a laboratory risk score for the early prediction of COVID-19 severity and in-hospital mortality

Abstract

Background and aims: Coronavirus Disease 2019 is characterized by a spectrum of clinical severity. This study aimed to develop a laboratory score system to identify high-risk individuals, to validate this score in a separate cohort, and to test its accuracy in the prediction of in-hospital mortality.

Methods: In this cohort study, biological data from 330 SARS-CoV-2 infected patients were used to develop a risk score to predict progression toward severity. In a second stage, data from 240 additional COVID-19 patients were used to validate this score. Accuracy of the score was measured by the area under the receiver operating characteristic curve (AUC).

Results: In the development cohort, a step-wise decrease in the average survival duration was noted with the increment of the risk score (p_ANOVA < 0.0001). A similar trend was confirmed when analyzing this association in the validation cohort (p < 0.0001). The AUC was 0.74 [0.66-0.82] and 0.90 [0.87-0.94], p < 0.0001, respectively for severity and mortality prediction.

Conclusion: This study provides a useful risk score based on biological routine parameters assessed at the time of admission, which has proven its effectiveness in predicting both severity and short-term mortality of COVID-19. Improved predictive scores may be generated by including other clinical and radiological features.

Keywords: Biomarkers; COVID-19; Risk score; Severe COVID-19; Short-term mortality.

Fig. 1

Distribution of severe forms incidences, according to the risk score, in the development cohort (a), and in the validation cohort (b), p for trend < 0.0001 respectively.

Fig. 2

a: Distribution of mortality rates, according to the risk score in the development cohort, p _{for trend} < 0.0001. b: Distribution of mortality rates, according to the risk score in the validation cohort, p _{for trend} < 0.0001. c: Average survival duration, according to the risk score in the development cohort, p _ANOVA < 0.0001. d: Average survival duration, according to the risk score in the validation cohort, p _ANOVA < 0.0001.

Fig. 3

a: Receiver operating characteristic (ROC) curve for COVID-19 severity prediction in the development cohort: the area under the ROC curve (AUC) = 0.95 [0.93–0.97], p < 0.0001. b: ROC curve for mortality prediction in the development cohort: AUC = 0.84 [0.78–0.89], p < 0.0001. c: ROC curve for COVID-19 severity prediction in the validation cohort: AUC = 0.74 [0.66–0.82], p < 0.0001. d: ROC curve for mortality prediction in the validation cohort: AUC = 0.90 [0.87–0.94], p < 0.0001.