Medicine (Baltimore). 2020 Oct 2;99(40):e22439. doi: 10.1097/MD.0000000000022439.
BACKGROUND: The mortality rate associated with Covid-19 varies considerably among studies and determinants of this variability are not well characterized.
METHODS: A systematic review of peer-reviewed literature published through March 31, 2020 was performed to estimate the mortality rate among hospitalized patients in China with a confirmed diagnosis of Covid-19. Hospital mortality rates were estimated using an inverse variance-weighted random-effects meta-analysis model. Funnel plot symmetry was evaluated for small-study effects, a one-study removed sensitivity analysis assessed the influence of individual studies on the pooled mortality rate, and metaregression assessed the association of potential confounding variables with mortality rates.
RESULTS: The review included 16 observational studies involving 1832 hospitalized patients with a diagnosis of Covid-19. The surveillance period among studies ranged from December 16, 2019 to February 23, 2020. The median patient age was 53 years and 53% were males. A total of 38.5% of patients presented with at least 1 comorbidity, most commonly hypertension (24.0%), cardiac disease (15.1%), and diabetes mellitus (14.4%). Fever and cough, reported in 84.8% and 61.7% of patients respectively, were the most common patient symptoms. The pooled mortality rate was 9.9% (95% confidence interval 6.1% to 14.5%). Funnel plot asymmetry was not observed and the meta-analysis results were not substantially influenced by any single study since the pooled mortality rate ranged from 8.9% to 11.1% following iterative removal of one study at a time. Substantial heterogeneity in the mortality rate was identified among studies (I = 87%; P < .001). In a metaregression that included demographics, patient risk factors, and presenting symptoms, only a higher prevalence of diabetes mellitus was associated with a higher mortality rate (P = .03).
CONCLUSIONS: In a meta-analysis of hospitalized patients in China with a diagnosis of Covid-19, the mortality rate was 9.9% and a higher diabetes mellitus prevalence was independently associated with a worse prognosis. The independent influence of diabetes mellitus with Covid-19 mortality should be viewed as hypothesis-generating and warrants further study.