Predictors and outcomes of respiratory bacterial coinfections in patients with COVID‐19 admitted to hospital: An observational prospective study

No abstract available.

Predictors and outcomes of respiratory bacterial coinfections in patients with COVID-19 admitted to hospital: An observational prospective study To the Editors: Bacterial coinfections are common complications of viral respiratory infections. In patients with influenza, bacterial coinfections range from 2% to 65%. 1 In patients with coronavirus disease 2019 (COVID-19) coinfections are less common (<10%). 2,3 The public health organizations warned against the indiscriminate use of antibiotics in COVID-19 to avoid selection of antimicrobial resistance and reduce the risk of bacterial superinfections. 4 Unlike coinfections, superinfections occurring during hospital course are common and associated with inappropriate antibiotic use on admission. 5,6 Unfortunately, antibiotic prescriptions in COVID-19 patients remain unacceptably high. 7 We performed a prospective, observational study to identify predictors of respiratory coinfections in COVID-19. Patients with COVID-19 pneumonia consecutively admitted to the University Hospital of Pisa, Italy (September 2020-September 2021) were included. All patients were prospectively followed-up from Emergency Department (ED) admission to hospital discharge or death. Both microbiological and laboratory exams were performed according to standards of practice and attending physicians judgement. Microbiological examinations were performed on sputum, urine and blood during the first 24 h after admission. Results of tracheobronchial aspirates, and bronchoalveolar lavage fluid (BAL), when performed, were considered as well as serologic tests for Chlamydophila pneumoniae, Legionella pneumophila or Mycoplasma pneumoniae. The aetiology was considered definite if 1 of the following criteria was met: positive urinary antigen for Legionella pneumophila or Streptococcus pneumoniae; bacterial yield in cultures from valid sputum (>25 polymorphonuclear cells and 10 epithelial cells per high-power field) of 10 6 CFU/mL, from tracheobronchial aspirates of 10 5 CFU/mL, from BAL fluid of 10 4 CFU/mL or from protected specimen brush cultures of 10 3 CFU/mL; or occurrence of seroconversion (a four-fold increase in IgG titers for C. pneumoniae [1:512] or an increase in IgM titers for C. pneumoniae [1:32] or M. pneumoniae [any titer]). 8 Charlson Comorbidity index was calculated for each patient. 9 Respiratory coinfections were defined as infections diagnosed within the first 24 h from hospital admission. 2 Superinfections were infections diagnosed more than 48 h after admission. 2 The primary objective was to identify factors independently associated with bacterial respiratory coinfections. The secondary objective was to explore the impact of coinfections on a composite endpoint of death or need for invasive mechanical ventilation within 30 days from admission.
Multivariable regression analyses were performed to identify predictors of coinfection and predictors of the composite endpoint, using a forward stepwise procedure, and entering all variables with univariate p < 0.05 and those of clinical relevance. For the composite endpoint, the time at risk was defined as the days between hospital admission and the occurrence of endotracheal intubation or death. Statistical significance was established at p < 0.05. The analysis was performed using a commercially available statistical software package (SPSS 27.0; IBM, Armonk, NY).
This research was previously presented at the 32nd ECCMID 2022.
A total of 71 (13%) patients met the composite endpoint. On Cox regression analysis, respiratory coinfections, Charlson Comorbidity Index, age and ferritin value on admission were predictors of death or endotracheal intubation (Table 1).
In this study, we found that respiratory coinfections occur in 8% of hospitalized patients with COVID-19 and are associated with high risk of progression to severe COVID-19. The proportion of coinfections changed from the beginning of the pandemic. The study by Garcia-Vidal et al. conducted during the first wave reported a prevalence of 3.1%. 2 A more recent observational study by Moreno-Garcia et al. (February 2020-February 2021) reported that coinfections occurred in 9% of patients. 10 Although in this study both respiratory coinfections and bacteraemia caused by E. coli, S. aureus, P. aeruginosa and H. influenzae were included, the results are similar to our findings. During the course of the pandemic, a wide use of steroids-frequently administered also before hospital admission-may have contributed to an increase in the proportion of bacterial coinfections. Moreover, since disease severity has decreased during time according to vaccination and spread of new variants (e.g., omicron), factors such as coinfections might more strongly contribute to disease severity causing hospital admission. The Spanish colleagues found that patients with higher PCT values had more risk to be affected by coinfections than those with lower levels. 10 We found that the trend in PCT values (increase ≥50% in the first 72 h from admission) instead of the absolute value is associated with increased risk of coinfections, highlighting the need of PCT monitoring during the first 3 days from hospital admission.
Our findings are also in line with the study by Russell CD 7 . However, in this study S. aureus and H. influenzae were more frequently identified compared to our study. This may be due to different local epidemiology, more strict microbiological surveillance and differences in the patient population. In fact, Russell included a higher proportion of immunocompromized patients (29.3%) and chronic obstructive pulmonary disease (COPD; 28.7%) 7 compared to our study (16.6% and 11%, respectively).
We found an association between coinfections and severe COVID-19. This relationship may be dual. First, patients with coinfections may have an increased risk of disease progression; thus, identification of coinfected patients may allow physicians to select patients who need antibiotic therapy. Second, it is also possible that patients with more severe disease had a high risk of presenting with bacterial coinfections on admission. However, the indiscriminate use of antibiotics in all patients with severe COVID-19 should be avoided and not encouraged because of the risk of multidrug-resistant organisms selection. Thus, implementing the diagnosis of coinfections on ED and correlating microbiological data with radiological, clinical and laboratory features may guide the start of antibiotic therapy only in patients who really need antibiotics. The start of antibiotic therapy on admission should be guide by several factors, such as disease severity and increase in PCT values. Subjects with mild to moderate COVID-19 who need oxygen therapy without an increase in the PCT levels during the first 3 days from hospital admission, empiric antibiotic therapy may be avoided.  Our study has some limitations: (i) it is an observational single-centre study and the sample size is relatively low; (ii) the association between coinfections and poor outcome might be related to higher disease severity of patients with coinfections; (iii) we did not consider the impact of treatments on the outcome of patients, although in our hospital we used a standardized therapeutic approach based on an internal guide for the management of COVID-19, updated on regularly basis. 11 In conclusion, the identification of COVID-19 patients with a respiratory coinfection on hospital admission is crucial to reduce inappropriate antibiotic use and to select those subjects at higher risk of mortality who may beneficiate from antibiotic therapy. Our findings may help clinicians to avoid inappropriate antibiotic therapy in patients with COVID-19 admitted to the hospital.