Pulmonary superinfection diagnosed with bronchoalveolar lavage at intubation in COVID patients: A Swedish single‐centre study

Patients with severe coronavirus disease 2019 (COVID) pneumonia and acute respiratory distress syndrome (C‐ARDS) on invasive mechanical ventilation (IMV) have been found to be prone to having other microbial findings than severe acute respiratory syndrome coronavirus 2 (SARS‐2)‐CoV‐19 in the bronchoalveolar lavage (BAL) fluid at intubation causing a superinfection. These BAL results could guide empirical antibiotic treatment in complex clinical situations. However, there are limited data on the relationship between microbial findings in the initial BAL at intubation and later ventilator‐associated pneumonia (VAP) diagnoses.

Routine serum inflammatory markers could not be used to identify this complication.
Microorganisms located in BAL at intubation were rarely associated with later VAP development.

K E Y W O R D S
adult respiratory distress syndrome, bronchoalveolar lavage, COVID-19, pneumonia, ventilatorassociated pneumonia

Editorial Comment
This article highlights the importance of broncholaveolar lavage (BAL) for the assessment of pulmonary superinfections.Although performed in a specific population of COVID-acute respiratory distress syndrome patients, the study conveys the general message that microorganisms observed in BAL at intubation may be not associated with the later VAP development, as has been often seen in COVID patients.

| INTRODUCTION
Pneumonia and acute respiratory distress syndrome (ARDS) due to COVID-19 have imposed several difficult challenges for critical care physicians worldwide. 1Although several aspects of COVID-19-related ARDS (C-ARDS) are similar to pneumonias of other aetiologies, others are new and different. 2,3Some of these patients are in a highly inflammatory state when they arrive at the intensive care unit (ICU).Hence, recognising and diagnosing patients with a simultaneous pulmonary superinfection without using bronchoalveolar lavage (BAL) has been challenging. 4,59][10] Several reports have indicated that ventilator-associated pneumonia (VAP) is more common in C-ARDS patients than in those with other causes of ARDS [11][12][13][14] ; however, the presence of superinfection at intubation and the possible relationship to the later development of VAP remains unclear.
This study aimed to analyse the incidence of a simultaneous pulmonary superinfection at the time of intubation of C-ARDS patients, and to determine if these microorganisms could be associated with the later development of VAP.We further hypothesised that elevated inflammatory markers would be indistinguishable between C-ARDS patients with or without superinfections at intubation.

| Study design
This retrospective cohort study of prospectively collected data analysed all patients at their first intubation and then treated with invasive mechanical ventilation (IMV) for C-ARDS in our ICU between 14 March 2020 and 13 March 2021.Patients who arrived in the ICU already intubated were excluded.

| Settings
The seven-bed mixed ICU admits 500 patients per year.Baseline data were summarized using descriptive statistics.Categorical variables are reported as frequencies and continuous variables as medians and interquartile ranges.Differences between groups were analysed using a χ 2 test, Fisher's test, or Mann-Whitney U test, as appropriate.All statistical analyses were performed using IBM SPSS Statistics for Windows version 27 (IBM Corp., Armonk, NY, USA).

| Definitions
ARDS was defined using the criteria of the Swedish Intensive Care Registry adopted from the Berlin definition. 17Comorbidities were defined using the International Classification of Diseases and Related Problems, Tenth Edition (ICD-10) codes registered prior to admission.
Immunosuppression was defined according to the Simplified Acute Physiology Score 3 (SAPS 3). 18COVID-19 infection was confirmed by a positive severe acute The cut-off value for a significant BAL culture was ≥10 4 colonyforming units (CFU) per mL.Pulmonary superinfection was defined as a significant microbial analysis from BAL fluid within 24 h after intubation.VAP was defined according to the European Centre for Disease Prevention and Control in patients who were intubated for ≥48 h with two modifications (Supporting information S1).4,19 First, due to the severity of the lung disease, classical BAL with 150 mL of 0.9% saline was not performed.Instead, lavage was performed using approximately 20 mL of 0.9% saline.Candida spp. ad coagulase-negative staphylococci were considered as colonising microorganisms and not responsible for superinfection or VAP.Second, due to the extensive and dynamic infiltrates in all lungs of C-ARDS patients, the criterion of an x-ray or computed tomography scan with new or progressive infiltrate was not included in VAP diagnosis.

| Microbiological and other laboratory analyses
BAL was routinely performed for microbiological analysis within 24 h after intubation with a single-use bronchoscope (Ambu ® aScope™, Ambu A/s, Ballerup, Denmark).The rationale for routine BAL was diagnosing or excluding a potential bacterial or fungal superinfection, and to be able to discontinue or de-escalate antibiotics whenever appropriate.Blood samples were collected simultaneously to analyse inflammatory markers, such as C-reactive protein (CRP), procalcitonin (PCT), neutrophil granulocytes, and lymphocytes.
BAL was also performed whenever there was suspicion of a new infection or deterioration in ventilatory parameters.Blind techniques were not used to obtain lower respiratory secretions for microbiological analyses, nor were tracheal secretions analysed.
All microorganisms were isolated from cultures with standard methods, except for atypical microorganisms, such as Mycoplasma pneumoniae, Pneumocystis jirovecii, Chlamydophila psittaci, Chlamydophila pneumoniae, and Legionella pneumophila, which were diagnosed using the PCR technique (Supporting information S1).

| Ethics
The Swedish Ethical Review Authority approved the study protocol on 26 August 2020, (DNR 2020-02758).The Declaration of Helsinki and its revisions were followed.The need for informed consent was waived due to the retrospective design of the study.

| RESULTS
A total of 134 patients were treated for C-ARDS during the study period.Approximately 114 patients (85%) were intubated and subsequently put on IMV (Figure 1).

| BAL within 24 h of intubation
Among all of the patients, only two had not a BAL performed with microbial analysis within 24 h after intubation (one patient was immediately transferred to another hospital and one culture was lost in the laboratory).
Hence, data from 112 (98%) patients were available for analysis.
The 31 (28%) patients with a significant BAL analysis and 81 (72%) with a non-significant BAL analysis within 24 h after intubation are presented in Table 1.
There were no significant differences in baseline characteristics, including antibiotic use at intubation, or inflammatory markers (CRP, PCT, neutrophil granulocytes, or lymphocytes), between patients with or without a superinfection.Seven of the 31 (23%) patients with positive microbiological findings had more than one species isolated.All isolates are listed in Table 2. Five isolates were identified as MRSA, but none as ESBL, ESBL-carba, or VRE.
No significant differences in 30-day mortality were observed between the group with or without superinfection at intubation (6.5% and 9.9%), and superinfection was not clinically identified as a direct cause of death.

| BAL after 48 h of IMV
Fourteen patients were transferred to other units due to a shortage of ICU beds within 48 h after intubation.Hence, these patients could not be evaluated for VAP.Ninety-eight (88%) of the 112 patients were on IMV for more than 48 h in our ICU and could develop VAP.
The median number of BAL sampling sessions (including at intubation) for microbiological analysis, which were conducted for each of the 98 patients on IMV for more than 48 h, was two (range: 1-20).
Among the patients who underwent microbiological analysis, a single episode of VAP was diagnosed in 36 patients (37%).One patient experienced two VAP episodes.
The microorganisms responsible for VAP are listed in

Streptococcus agalactiae 1
Note: Only one patient had an indistinguishable bacterial strain responsible for a superinfection and ventilator-associated pneumonia.Abbreviations: ESBL, extended beta lactamase; MRSA, methicillin-resistant Staphylococcus aureus; VAP, ventilator associated pneumonia.
absence of a superinfection.The microorganism observed in BAL at intubation was rarely associated with later VAP development.
The microorganisms from BAL at intubation were predominantly of upper airway, oral, and pharyngeal flora.This aligns with other studies and indicates the colonisation or micro-aspiration of microorganisms prior to or during intubation. 5,21However, our study observed a higher incidence (28%) of superinfection when compared with two recently published studies evaluating BAL at intubation in patients with C-ARDS (21% and 10%, respectively).It is challenging to determine whether there is a pulmonary superinfection in C-ARDS patients prior to intubation without a BAL analysis.Severe C-ARDS is often associated with increased levels of inflammatory markers, especially an elevated PCT, which has been proposed as a marker of superinfection and leads to a subsequent need for antibiotics.8][9] Our findings could not support the use of CRP, PCT, neutrophil granulocytes, or lymphocytes as markers for the presence of superinfection intubation.This finding is similar to that of another study that also used BAL as a diagnostic tool for diagnosing pulmonary superinfections. 5A recently published multicentre study from Spain analysed CRP and PCT as markers for all bacterial superinfections in early COVID-19 patients treated in the ICU.Their findings did not confirm that these markers could be used at a single time point to identify bacterial superinfections. 10veral studies have shown that pandemics caused by different strains of influenza viruses are prone to be superinfected by different bacteria.Streptococcus pneumoniae was the predominant bacterial pathogen during the pandemics of 1918 and 1968 but during 1957 S. aureus accounted for 44% of the deaths. 23During the 2009 pandemic, S. pneumoniae accounted for the majority of secondary bacterial infections in hospitalized patients but the majority of deaths were caused by Streptococcus pyogenes, Haemophilus influenzae, and Gram-negative rods. 23,24The underlying mechanisms for this variation are likely multifactorial and encompass various aspects, such as the varying ability among viruses to damage respiratory tract epithelial cells, an influence on the local and systemic immune system, facilitation of the release of bacteria from biofilms through viral mediation, and induction alterations in normal microbial flora. 23Based on our findings and those of other studies conducted in ICUs, it is not apparent that a specific microorganism contributes to superinfections in C-ARDS patients.
Currently, our routine involves antibiotics at intubation for patients with severe C-ARDS, irrespective of the levels of inflammatory markers.
Empirical antibiotic treatment is suspended if the mandatory microbiological BAL analysis at intubation is negative.For patients who have less severe cases of C-ARDS, antibiotics are not routinely administered unless there is a specific indication.If the initial BAL result indicates a superinfection, antibiotics will be administered after consultation with an infectious disease specialist.To counteract antibiotic overuse, several complementary techniques (BAL analysis of inflammatory cells or inflammatory markers) to distinguish colonisation from infection have been developed, although they are not currently established for clinical use, especially not in C-ARDS patients. 5,25 the beginning of the pandemic, a substantial proportion of intubated patients were treated with antibiotics, as recommended by the WHO guidelines. 6This most likely resulted in the overuse of antimicrobial drugs.Gradually, the use of antibiotics has decreased, which seems adequate but studies still indicate an overuse. 26The overuse of antibiotics is a well-known risk factor for several complications, such as the development of multi-resistant bacteria, increased risk of healthcare-related infections such as Clostridioides difficile infections, and an overall increase in antibiotic resistance. 26Performing BAL at intubation and repeatedly during ventilator treatment of patients with C-ARDS may be a way to optimize antibiotic treatment.
8][29][30] However, at the start of the pandemic, it became evident that VAP was a clinical problem in our setting, which is supported by other studies. 12,13,31Our VAP incidence of 37% falls in the middle range of other published studies (21%-64% [pooled mean 50%]). 11The first Swedish study, analysing VAP in COVID-19 patients was based on registry data and utilized ventilator-associated respiratory tract infection as an endpoint, reported an incidence of 30%. 32The second study, primarily analysing the incidence and risk factors for ventilator-associated lower respiratory tract infections, revealed an incidence of 29%. 33The variation in incidence could be explained by many factors related to the COVID-19 disease, such as patient case mix, treatment, diagnostic routines, and logistics.Furthermore, the relatively high incidence of VAP could depend on numerous other factors such as the immunological effects of SARS-CoV-2, mucus formation, prone position, over-crowded ICUs, lack of equipment, and understaffed ICUs.This has been described in other studies. 11,34veral of the species observed in BAL at intubation were also commonly observed in VAP BAL.Surprisingly, we observed only one patient with an indistinguishable strain in the initial superinfection and later VAP development.This is consistent with previous studies and could be attributed to the administration of antibiotics targeting the initial microorganism. 5,21Thus, VAP was caused by microorganisms acquired during IMV and highlights the importance of preventive strategies in order to minimize VAP incidence.
The microorganisms responsible for superinfection and VAP in our study were similar to those reported in other studies evaluating C-ARDS patients. 12,13,31The overall incidence of multi-resistant strains in Swedish ICUs is low. 35We identified MRSA from BAL cultures in five (16%) patients.This is a surprisingly high incidence from a Swedish perspective.The reason for this is unclear but all patients except one were of non-Scandinavian origin or had lived abroad for a long time.Genetic analysis of these five strains did not support the notion that the strains were transmitted from one patient to another.The incidence of other multi-resistant strains, such as ESBL, ESBL-carba, and VRE, was low in our study.This reflects the overall low incidence of bacterial strains with these resistant mechanisms in the Swedish ICU population. 35cause of the difficulties in diagnosing a superinfection, we believe it is highly recommended to utilise microbial BAL analysis in this clinical setting.Our BAL routines for patients with invasive mechanical ventilation are well established and robust.A low frequency of BAL for microbiological culture will most likely result in a false low VAP incidence.The median number of BAL samples for microbiological analyses performed in our patients was two.Given the median time of 14 days with regards to IMV, we believe this to be a reasonable BAL frequency.
Neither the presence of a secondary microorganism observed in BAL at intubation nor the occurrence of VAP were found to be associated with mortality in this small study.However, there are conflicting results from other studies.Some studies have not reported the impact of VAP on mortality, while others have demonstrated an increased risk. 5,12- 14,36It is worth noting that the overall mortality rate among C-ARDS patients in our study was relatively low (12%), which makes it challenging to evaluate superinfection and VAP as risk factors for death. 16e study has several limitations.First, the retrospective singlecentre design reduces generalizability.Second, we were unable to analyse adherence to basic hygiene or VAP preventive routines, which could have influenced VAP incidence.However, a team from the Department of Infection Control and Prevention performed several inspections and conducted educational sessions throughout the study period to increase the knowledge about healthcare-related infections.
Third, BAL for the diagnosis of microorganisms was performed with 20 mL of 0.9% saline.BAL with larger volumes of 0.9% saline or a protected specimen brush might have increased the incidence of the microbial findings.The use of mini-BAL is supported by a recent analytical review. 22Fourth, due to the small sample sizes in subgroups regression analysis could not be performed.Finally, galactomannan data analysis was not part of the clinical routine during the first 2 months of the study, which could have resulted in an underestimation of the number of CAPA.
This study has several strengths.First, the routine for BAL diagnosis at intubation and on suspected VAP was well established, both in terms of the indications for BAL and in the microbial analysis with culture, PCR, and galactomannan.8][29][30] Data from our COVID-ICU have indicated low incidences of catheter-related bloodstream infections and C. difficile-related diarrhoea (unpublished data).
In conclusion, one fourth of the patients with C-ARDS had a pulmonary superinfection at intubation.The significance of this finding needs to be further explored.Routine serum inflammatory markers could not be used to identify a superinfection at intubation.In light of these findings, we think that BAL can be considered at intubation to gain insights into what is frequently an intricate scenario.

16 2. 3 |
All physicians working in the ICU are either in training, or board-certified specialists in anaesthesia and intensive care medicine.All of the nurses are registered as ICU nurses.Due to the sudden influx of COVID-19 patients requiring ICU treatment, a separate 12 bed ventilatorequipped COVID-ICU opened in a post-operative unit.During the pandemic, there was a shortage of ICU-registered nurses; therefore, nurse anaesthetists were recruited.Care routines and general outcome results for these ICU patients with COVID-19 have been previously published elsewhere. 15,Data collection and statistical analyses All baseline data were collected from two computer medical record systems (Metavision™ and Cambio Cosmic™) and validated by two specially trained ICU-registered nurses.The first author extracted all microbial data.

20 PaƟents 28 F
The isolates classified as multi-resistant were methicillin-resistant Staphylococcus aureus (MRSA), vancomycin-resistant enterococci (VRE), and extended-spectrum beta-lactamase (ESBL) producing Gram negative bacteria, including ESBL carbapenemase activity (ESBL-carba).Isolates were considered indistinguishable if they shared the same phenotype and antibiogram.COVID-19-associated pulmonary Aspergillosis (CAPA) was categorized as proven, probable, or possible according to the European Confederation of Medical Mycology/International Society for Human and Animal Mycology.I G U R E 1 Flowchart representing the study.BAL, bronchoalveolar lavage; ICU, intensive care unit; VAP, ventilator-associated pneumonia.*All transferred to another intensive care unit due to limited resources.

Table 2 .
The predominant species were S. aureus, Klebsiella spp., Escherichia coli, Microorganism responsible for superinfection and VAP.
4 | DISCUSSIONThis study revealed that approximately one fourth of C-ARDS patients had a pulmonary superinfection in the lungs after being diagnosed with BAL within 24 h after intubation.Routine inflammatory markers could not be utilised as reliable indicators to predict the presence orT A B L E 2