Julinha M Thelen, A G Noud Buenen, Marjan van Apeldoorn, Heiman F Wertheim, Mirjam H A Hermans, Peter C Wever
BACKGROUND: During the coronavirus disease 2019 (COVID-19) pandemic in the Netherlands it was noticed that very few blood cultures from COVID-19 patients turned positive with clinically relevant bacteria. This was particularly evident in comparison to the number of positive blood cultures during previous seasonal epidemics of influenza. This observation raised questions about the occurrence and causative microorganisms of bacteraemia in COVID-19 patients, especially in the perspective of the widely reported overuse of antibiotics and the rising rate of antibiotic resistance.
METHODS: We conducted a retrospective cohort study on blood culture results in influenza A, influenza B and COVID-19 patients presenting to two hospitals in the Netherlands. Our main outcome consisted of the percentage of positive blood cultures. The percentage of clinically relevant blood cultures, isolated bacteria and 30-day all-cause mortality served as our secondary outcomes.
RESULTS: A total of 1331 viral episodes were analysed in 1324 patients. There was no statistically significant difference (p = 0.47) in overall occurrence of blood culture positivity in COVID-19 patients (9.0, 95% CI 6.8-11.1) in comparison to influenza A (11.4, 95% CI 7.9-14.8) and influenza B patients (10.4, 95% CI 7.1-13.7,). After correcting for the high rate of contamination, the occurrence of clinically relevant bacteraemia in COVID-19 patients amounted to 1.0% (95% CI 0.3-1.8), which was statistically significantly lower (p = 0.04) compared to influenza A patients (4.0, 95% CI 1.9-6.1) and influenza B patients (3.0, 95% CI 1.2-4.9). The most frequently identified bacterial isolates in COVID-19 patients were Escherichia coli (n = 2) and Streptococcus pneumoniae (n = 2). The overall 30-day all-cause mortality for COVID-19 patients was 28.3% (95% CI 24.9-31.7), which was statistically significantly higher (p = <.001) when compared to patients with influenza A (7.1, 95% CI 4.3-9.9) and patients with influenza B (6.4, 95% CI 3.8-9.1).
CONCLUSIONS: We report a very low occurrence of community-acquired bacteraemia amongst COVID-19 patients in comparison to influenza patients. These results reinforce current clinical guidelines on antibiotic management in COVID-19, which only advise utilization of antibiotics when a bacterial co-infection is suspected.
