Background: A systematic review and meta-analysis of available studies was performed to investigate the clinical characteristics that can predict COVID-19 disease severity. Materials and Methods: Databases including PubMed, Embase, and Web of Science were searched from December 31, 2019, to May 24, 2020. Random-effects meta-analysis was used for summarizing the Pooled odds ratio (pOR) of individual clinical characteristics to describe their association with severe COVID-19 disease. Results: A total of 3895 articles were identified, and finally, 22 studies comprising 4380 patients were included. Severe disease was more common in males than females (pOR: 1.36, 95% confidence interval [CI]: 1.08-1.70). Clinical features that were associated with significantly higher odds of severe disease were abdominal pain (pOR: 6.58, 95% CI: 1.56-27.67), breathlessness (pOR: 3.94, 95% CI: 2.55-6.07), and hemoptysis (pOR: 3.35, 95% CI: 1.05-10.74). pOR was highest for chronic obstructive pulmonary disease (pOR: 2.92, 95% CI: 1.70-5.02), followed by obesity (pOR: 2.84, 95% CI: 1.19-6.77), malignancy (pOR: 2.38, 95% CI: 1.25-4.52), diabetes (pOR: 2.29, 95% CI: 1.56-3.39), hypertension (pOR: 1.72, 95% CI: 1.23-2.42), cardiovascular disease (pOR: 1.61, 95% CI: 1.31-1.98) and chronic kidney disease (pOR: 1.46, 95% CI: 1.06-2.02), for predicting severe COVID-19. Conclusion: Our analysis describes the association of specific symptoms and comorbidities with severe COVID-19 disease. Knowledge of these clinical determinants will assist the clinicians in the risk-stratification of these patients for better triage and clinical management.