Clinical Characteristics Associated with Bacterial Bloodstream Coinfection in COVID-19

Inappropriate antibiotic use in COVID-19 is often due to treatment of presumed bacterial coinfection. Predictive factors to distinguish COVID-19 from COVID-19 with bacterial coinfection or bloodstream infection are limited.We conducted a retrospective cohort study of 595 COVID-19 patients admitted between March 8, 2020, and April 4, 2020, to describe factors associated with a bacterial bloodstream coinfection (BSI). The primary outcome was any characteristic associated with BSI in COVID-19, with secondary outcomes including 30-day mortality and days of antibiotic therapy (DOT) by antibiotic consumption (DOT/1000 patient-days). Variables of interest were compared between true BSI (n = 25) and all other COVID-19 cases (n = 570). A secondary comparison was performed between positive blood cultures with true BSI (n = 25) and contaminants (n = 33) on antibiotic use.Fever ( 38 °C) (as a COVID-19 symptom) was not different between true BSI (n = 25) and all other COVID-19 patients (n = 570) (p = 0.93), although it was different as a reason for emergency department (ED) admission (p = 0.01). Neurological symptoms (ED reason or COVID-19 symptom) were significantly higher in the true BSI group (p 0.01, p 0.01) and were independently associated with true BSI (ED reason: OR = 3.27, p 0.01; COVID-19 symptom: OR = 2.69, p = 0.03) on multivariate logistic regression. High (15-19.9 × 10True BSI in COVID-19 was associated with neurological symptoms and nonsignificant higher WBC, and led to overall higher 30-day mortality and worse patient outcomes.