To the editor,

Recently, in this journal, a study on disease progression in hospitalized Covid-19 patients was reported [1]. We here contribute with data on the risk for hospital-acquired blood stream infection in patients with Covid-19 pneumonia. Little is known about bacterial coinfections in patients with Covid-19. Respiratory viruses confer an increased risk of bacterial pneumonia and bloodstream infection that further is associated with severe disease and increased mortality [2–4]. Knowledge of the aetiology of superinfections may guide preventive measures and choice of empiric antimicrobials. A recent review of eight studies with emphasis on respiratory bacterial and fungal infections of hospitalised patients found an overall coinfection rate of 8% [5]. Studies from the United States and the United Kingdom identified bacteraemia in 1.6% to 5.6% of patients with severe acute respiratory syndrome coronavirus 2 (SARS CoV-2) infection [6–8]. Here, we report the occurrence and type of bloodstream infection as well as risk factors associated with bloodstream infections among patients with Covid-19 pneumonia.

Data included 227 consecutive adults (mean age 66.3 years, SD 17.0 years; females 111 (49%)) admitted to Copenhagen University Hospital, Hvidovre (700 beds, catchment area of population 550.000) from 10th March to 21st May 2020 with confirmed SARS CoV-2 infection by RT-PCR and radiological evidence of pneumonia [9]. Blood culture (BD BACTEC-FX, BD Diagnostics, Denmark) was ordered for 189 (85%) patients. For comparison, we included 2097 consecutive adults admitted for non-Covid-19 reasons with a total of 2800 blood cultures drawn during the same period. Bloodstream infections were categorized as community-acquired or hospital-acquired, if cultures were drawn within or after 48 h of admission. Clinical and microbiological data including the primary focus of infection were prospectively collected by the treating clinical microbiologist, who also assessed the clinical significance of a detected pathogen. Clinical and laboratory data of patients with Covid-19 were retrieved from electronic health records [9]. Groups were compared by parametric or non-parametric tests where appropriate using IBS SPSS Statistics for Windows version 25 (IBM Corp., Armonk, NY., USA). The study was approved by the Danish Patient Safety Authority (31-1521-309) and the Regional Data Protection Centre (P-2020-492).

Clinical findings, procedures during admission and outcomes in the Covid-19 cohort are presented in Table 1. During a median observation time of 6 days (range 0–71 days), we identified a total of 17 bloodstream infections in 12 patients (5.3%) (nine had one bloodstream infection, two had two and one patient had four bloodstream infections) that occurred a median of 16 days after admission (range 5–37 days) (Table 2). All bloodstream infections were hospital-acquired. For 12/17 bloodstream infections, the primary focus of infection was likely an intravascular catheter, for three either a catheter or pulmonary infection and for two it was unknown. Bloodstream infection was associated with the following at admission: lower peripheral oxygen saturation, oxygen supplementation or mechanical ventilation, lower lymphocyte count, and elevated plasma lactate dehydrogenase and C-reactive protein (Table 1). Further, bloodstream infection was associated with admission to intensive care and to renal replacement therapy.

In the non-Covid-19 population, 124 cases of community-acquired bloodstream infection occurred in 122 patients (5.9%) and 32 cases of hospital-acquired bloodstream infection in 30 patients (1.4%). The rate of hospital-acquired bloodstream infection was higher in Covid-19 compared to patients with non-Covid-19 conditions (5.3% vs. 1.4%; p < 0.001). Table 2 lists hospital-acquired pathogens. Enterococcus faecium, Candida species and coagulase-negative staphylococci were the most common pathogens in patients with Covid-19, while Staphylococcus aureus and Enterobacterales were more frequent among cases without Covid-19.

Interestingly, no cases of community-acquired bloodstream infection was detected among the first 227 patients admitted with Covid-19 pneumonia, while hospital-acquired bloodstream infection occurred in 5.3% of patients with Covid-19 compared to 1.4% of patients without Covid-19.

The observed proportion of hospital-acquired bloodstream infection in hospitalized patients with Covid-19 was comparable to results from two previous studies [7,8], but higher than in one study where no clinical information, however, was provided [6]. The discrepancies between studies may be explained by different populations and sample size. One study found in accordance with our results that bloodstream infection occurred late during hospitalization (>5 days after admission for 99.7%), and that the primary focus of infection predominantly was an intravascular catheter [8]. Overall pathogens did not differ between patients with and without Covid-19, although there was a trend towards a higher occurrence of Enterobacterales in the non-Covid-19 population.

More severe disease at presentation (multilobular pulmonary infiltrates, low oxygen saturation and need of oxygen supplementation, higher plasma C-reactive protein and lactate dehydrogenase), stay in intensive care and treatment for organ failure (mechanical ventilation and renal replacement therapy) were risk factors for hospital-acquired bloodstream infection in patients with Covid-19 pneumonia by univariate analysis. The number of cases with bloodstream infection did not permit multivariable analysis. Similarly, another study found that bloodstream infection was more frequent in patients with severe disease requiring mechanical ventilation than not (11.9% vs. 1.8%) [7]. No other studies have reported on risk factors associated with bloodstream infection in patients with Covid-19.

Influenza is associated with bacterial pneumonia and bloodstream infection with Streptococcus pneumonia, S. aureus and S. pyogenes [4,10]. Most occur within the first week of admission [4]. In contrast, the pattern of bloodstream infection associated with Covid-19 was significantly different. All cases were hospital-acquired and occurred late. Median time between bloodstream infection and death was 7.5 days (range 0–12) suggesting that bloodstream infection may have contributed to death but was unlikely the cause.

Our findings support a restrictive use of antimicrobials in patients with Covid-19 upon admission since no bloodstream infections were community-acquired despite initial blood culture in the majority of individuals. This strategy is further supported by previous reports of a low rate of secondary bacterial infection [5,8]. The most common pathogens identified at our institution were staphylococci and E. faecium (50%) followed by Candida species and is comparable to hospital-acquired bloodstream infection in patients without Covid-19.

In conclusion, bloodstream infection associated with Covid-19 were all hospital-acquired and associated with disease severity suggesting that procedures associated with care were more likely causes of bloodstream infection than infection with SARS-CoV-2 per se.