Risk factors for adverse clinical outcomes in patients with COVID-19: A systematic review and meta-analysis

Importance: COVID-19 is a clinically heterogeneous disease of varying severity and prognosis. Clinical characteristics that impact disease course could offer guidance for clinical decision making and future research endeavors and unveil disease pathways. Objective: To examine risk factors associated with adverse clinical outcomes in patients with COVID-19. Data sources: We performed a systematic review in PubMed from January 1 until April 19, 2020. Study selection: Observational studies that examined the association of any clinical characteristic with an adverse clinical outcome were considered eligible. We scrutinized studies for potential overlap. Data extraction and synthesis: Information on the effect of clinical factors on clinical endpoints of patients with COVID-19 was independently extracted by two researchers. When an effect size was not reported, crude odds ratios were calculated based on the available information from the eligible articles. Study-specific effect sizes from non-overlapping studies were synthesized applying the random-effects model. Main outcome and measure: The examined outcomes were severity and progression of disease, admission to ICU, need for mechanical ventilation, mortality, or a composite outcome. Results: We identified 88 eligible articles, and we performed a total of 256 meta-analyses on the association of 98 unique risk factors with five clinical outcomes. Seven meta-analyses presented the strongest epidemiological evidence in terms of statistical significance (P-value <0.005), between-study heterogeneity (I 2 <50%), sample size (more than 1000 COVID-19 patients), 95% prediction interval excluded the null value, and absence of small-study effects. Elevated C-reactive protein (OR, 6.46; 95% CI, 4.85 - 8.60), decreased lymphocyte count (OR, 4.16; 95% CI, 3.17 - 5.45), cerebrovascular disease (OR, 2.84; 95% CI, 1.55 - 5.20), chronic obstructive pulmonary disease (OR, 4.44; 95% CI, 2.46 - 8.02), diabetes mellitus (OR, 2.04; 95% CI, 1.54 - 2.70), hemoptysis (OR, 7.03; 95% CI, 4.57 - 10.81), and male sex (OR, 1.51; 95% CI, 1.30 - 1.75) were associated with risk of severe COVID-19. Conclusions and relevance: Our results highlight factors that could be useful for prognostic model building, help guide patients' selection for randomized clinical trials, as well as provide alternative treatment targets by shedding light to disease pathophysiology.