Abstract Background Risk prediction scores are important clinical decision-making tools for management of patients with COVID-19. In the UK, the ISARIC 4C score was developed to predict in-hospital mortality from readily available clinical characteristics capturing disease severity and internally validated on a large cohort within the UK but requires external validity to confirm its generalizability and to what extent the same risk factors affect prognosis in COVID-19 patients with cardiovascular disease (CVD) is less clear. Purpose To assess the prognostic validity of ISARIC 4C score in predicting in-hospital mortality in individuals hospitalized with COVID-19 and CVD across WHO regions. Methods We conducted an external validation of ISARIC 4C score to predict mortality using data from the WHF global COVID-19 and CVD cohort (5313 patients recruited between August 2020 – September 2021), compared to the UK cohort of 1721 patients recruited from March – June 2020. We assessed the validity of the ISARIC 4C score to prognosticate in-hospital mortality. The variables used for 4C score were age, sex, number of comorbidities, respiratory rate, peripheral oxygen saturation on room air, urea and C-reactive protein. Each value was then assigned a score according to the ISARIC 4C guidelines and the total score was calculated. 4C score ranging from 0-3, 4-8, 9-14, and 15-19 was classified as "low", "intermediate", "high" and "very high", respectively. The discriminatory ability of the 4C score for in-hospital mortality were tested for the WHF cohort by WHO region compared to the UK cohort using area under the curve (AUC) for receiver operating characteristic (ROC) curves. Results The study included 7,052 individuals, of whom 1,268 (17.9%) died in-hospital. Mean age was 61 years, males: 59% and pre-existing hypertension: 46.1%, diabetes 33.8%, coronary artery disease 11.4%, heart failure: 4.1% and stroke 3.5%. Mean ISARIC 4C score among survivor’s vs non-survivors was 9.0 and 12.0, (P 0.001). Overall mortality rates across the pre-defined risk groups were 3.4% (Low), 6.3% (Intermediate), 21.2% (High), and 43.4% (Very High) (Figure 1). The AUC for discrimination of in-hospital mortality by ISARIC 4C was 0.77 (95% CI: 0.75 – 0.80) in the UK and 0.75 (95% CI: 0.69 – 0.79) in WHF cohort (Figure 2). Diagnostic accuracy measures for cut-offs at 3, 8, 11, and 14 can be found in Table 1. The cut-off value of 11 had sensitivity of 61.4%; specificity, 73.3.0%, positive predictive value: 33.5% and negative predictive value: 89.6%. Conclusion This study represents the first global validation of a simple 4C risk score and confirms the prognostic validity of ISARIC 4C mortality risk score in patients hospitalized with COVID-19 and CVD across WHO regions with varying population characteristics and different ethnic groups, and can be used to identify and prioritize care for patients at greatest risk of death.
Singh et al. (Sat,) studied this question.