Clinical and Historical Features Associated with Severe COVID-19 Infection: A Systematic Review and Meta-Analysis

Background: In low-resource settings (LRS), a COVID-19 severity scoring tool could ensure timely access to appropriate care and guide use of critically limited resources. We systematically evaluated existing literature for associations between COVID-19 illness severity, and historical characteristics, clinical presentations, and investigations measurable in LRS. Methods: A systematic review of four databases from December 01, 2019 to June 01, 2020 identified all studies assessing potential associations between clinical characteristics and investigations, and illness severity. Data for all variables that were statistically analysed in relation to COVID-19 disease severity were extracted. A meta-analysis was conducted to generate pooled estimates of odds ratios (pORs). This review was registered with PROSPERO (CRD42020178098). Findings: Of 8402 relevant articles, 79 were eligible for inclusion, analysing 27713 patients with laboratory-confirmed COVID-19. A total of 202 features were analysed in relation to COVID-19 severity. Of these, 81 were deemed feasible for assessment in LRS: appropriate data were available for meta-analysis of 44 (54·3%). Meta-analysis identified 19 significant predictors of severe COVID-19, including: stroke (pOR: 3.08 (95% CI [1.95, 4.88])), shortness of breath (pOR: 2·78 (95% CI [2·24-3·46])), and chronic kidney disease (pOR: 2.55 (95% CI [1.52-4.29])). I 2 testing suggested substantial heterogeneity across samples. Interpretation: We identified multiple variables predictive of severe COVID-19. Due to the novelty of the disease, there is currently no prognostication tool purpose-designed and validated for LRS. Findings may inform such a tool that can offer guidance on clinical treatment and disposition, ultimately reducing morbidity and mortality due to the pandemic. Funding: No funding was received for this study. Declaration of Interests: The authors declare no competing interests.