ABSTRACT
Introduction Understanding the proportion of patients with COVID-19 who have respiratory bacterial coinfections and the responsible pathogens is important to managing COVID-19 effectively while ensuring responsible antibiotic use. Objectives To estimate the frequency of bacterial coinfection in COVID-19 patients and of antibiotic prescribing and to appraise the use of antibiotic stewardship criteria. Methods Systematic review and meta-analysis was performed using major databases up to May 15, 2020. We included studies that reported a) proportion/prevalence of bacterial coinfection in COVID-19 patients and b) use of antibiotics. Where available, data on duration and type of antibiotics, adverse events, and any information about antibiotic stewardship policies were also collected. Results We included 39 studies with a total of 10 815 patients. The overall prevalence of bacterial coinfection was 10.6% (95% CI 6.8%–14.3%). When only confirmed bacterial coinfections were included the prevalence was 4.6% (95% CI 2.9%–7%). Thirty-five bacterial species were identified, the most frequent being Mycoplasma pneumoniae (n = 12 [34%]), Pseudomonas aeruginosa (n = 4 [11.4%]) and Legionella pneumophila (n = 2 [6%]). The overall antibiotic use was 71.3% (95% CI 63.1%–79.5%). Only one study described criteria for stopping them. All included studies had a moderate to high risk of bias. Conclusions There is currently insufficient evidence to support widespread empirical use of antibiotics in most hospitalized patients with COVID-19, as the overall proportion of bacterial coinfection is low. Furthermore, as the use of antibiotics appears currently to be largely empirical, it is necessary to formulate clinical guidelines to promote and support more targeted administration of antibiotics in patients admitted to hospital with COVID-19.