Random effects meta-analysis of COVID-19/S. aureus partnership in co-infection

Abstract

Aim: To assess the hypothesis that coinfection with SARS-CoV-2 and exacerbates morbidity and mortality among patients, the study aims to report the pooled burden of co-infections in patients hospitalized with COVID-19. Methods: We searched electronic databases and the bibliographies of pertinent papers for articles. We considered studies in which the core result was the number of patients with bacterial () co-infection. We performed random effects meta-analysis (REM) because the studies included were sampled from a universe of different populations and high heterogeneity was anticipated. Using the Cochran’s Q statistic, the observed dispersion (heterogeneity) among effect sizes was assessed. The percentage of total variability in the estimates of the effect size was calculated with the I index. To check for publication bias, the Egger weighted regression, Begg rank correlation and meta-funnel plot were used. We conducted meta-regression analysis to evaluate the variability between our outcomes and the covariates using computational options such as “methods of moments” and then “maximum likelihood” ratio. Results: We included 18 studies and retrieved data for 63,370 patients hospitalized with influenza-like illness, of which about 14,369 (22.67%) tested positive for COVID-19 by rRT-PCR. Of this number, 8,249 (57.4%) patient samples were analyzed. Bacterial, fungal and viral agents were detected in 3,038 (36.8%); in 1,192 (39.2%). Five studies reported MRSA co-infection. Study quality ranged from 6 to 9 (median 7.1) on a JBI scale. From the meta-analysis, 33.1% patients were found to be coinfected (95%, CI 18.0 to 52.6%, Q=3473: df=17, I=99·48%, p=0.00). The rate of /COVID-19 co-infection was 25.6% (95% CI: 15.6 to 39.0, Q=783.4, df=17, I=97.702%, p=0.003).The proportion of COVID-19/ co-infected patients with MRSA was 53.9% (95% CI, 24.5 to 80.9, n=66, 5 studies, Q=29.32, df=4, I=86.369%, p=0.000). With the multivariate meta-regression model, study type (p=0.029), quality (p=0.000) and country (p=0.000) were significantly associated with heterogeneity.Conclusions: The pooled rates of among COVID-19 patients documented in this study support the concern of clinicians about the presence of in co-infections. Improved antibiotic stewardship can be accomplished through rapid diagnosis by longitudinal sampling of patients.

Keywords

• covid-19
• s. aureus
• co-infection
• meta-analysis
• meta-regression