BAL in suspected COVID-19 non-critically ill patients.
As recently reported by systematic reviews and metanalysis several retrospective and few prospective observational studies have investigated the role of BAL in suspected or known COVID-19.18,19 To the best of our knowledge, all studies performed in the non-critically ill patients are retrospective. A summary of BAL findings in non-critically ill patients is reported in Table 1 and 2. Between January and February 2020 Chinese scientists reported 5 cases of suspected COVID-19 investigated with BAL showing positivity for SARS-Cov-2 in all cases.20Subsequently, between March and May 2020, Italian virologists confirmed a higher positivity in BAL compared to other specimens (15%, 55/367 positive BAL, compared to 8%, 769/9461 positive NP swabs).21 A small retrospective case series reported a 19% prevalence (3/19 cases) of SARS-Cov-2 infection in BAL performed in patients with negative NP swab.22 During the first COVID-19 wave (March-April 2020), De Clercq et al conducted a retrospective monocentre study in Belgium aimed to evaluate the feasibility of their local diagnostic protocol that included BAL in patients’ diagnostic workup.23 They performed 27 BAL in non-critically ill patients with HCRT changes suspected for COVID-19 and two negative NP swabs and found 26% (7/27) positive BAL for SARS-Cov-2. They also identified one coinfection in SARS-Cov-2 positive (E cloacae) and 63% of other pathogens in negative BAL for SARS-Cov-2 including Mycoplasma pneumoniae, Streptococcus pneumoniae, Haemophilus influenzae, Pneumocysitis jirovecii , and other viruses.23 Another retrospective study conducted in two Belgian centres during the first wave confirmed the utility of BAL in detecting SARS-Cov-2 in 25% (14/55) of non-critically ill patients with negative NP swabs.24 The Authors also underlined the utility of BAL in therapeutic management that was changed after BAL in 60% of cases (33/55), either because other pathogens were identified (one coinfection with Serratia marcescens in SARS-Cov-2 positive cases and 42%, 23/55 of other pathogens in SARS-Cov-2 negative cases including Mycobacterium tuberculosis, Pneumocystis jirovecii, Haemophilus, Serratia, Escherichia coli, virus Influenza type A, Metapneumoviruses, Herpes viruses and Aspergillus fumigatus ) or because an alternative diagnosis was made (18% of cases, 10/55, including rheumatoid arthritis, hypersensitivity pneumonitis, cardiogenic oedema, cryptogenic organizing pneumonia, hepatopulmonary syndrome).24 During the first COVID-19 wave, Mondoni et al. carried out in Italy an observational, retrospective, multicentre cohort study aimed to evaluate the diagnostic yield of bronchoscopy in patients with two negatives NP swabs and suspected COVID-19.25 A total 109 adults, 71% males, age 60 (SD 13.6) years were enrolled, 108 bronchoscopies (99%) performed with flexible scope and 13 with rigid. Two-third of the procedures (N=78) were performed to confirm a COVID-19 diagnosis, and one-third were urgent/life-saving procedures. Only 10% of the procedures were carried out in the ICU setting (8.2% invasive ventilation, 1.8% ECMO). The diagnostic yield of bronchoscopy to detect SARS-Cov-2 in patients with previous negative swabs and a clinical and radiological suspicion of COVID-19 pneumonia was 55.1% (43/78). 1.8% (2/109) patients with both NP swabs and BAL negative for SARS-Cov-2 showed a late NP swabs positivity. Coinfections were detected in 4 cases (3,6% of the total):Haemophilus influenzae, Aspergillus fumigatus, Aspergillus spp. and Candida albicans .25 In the same period (March-April 2020) Patrucco and coworkers conducted a similar Italian observational, retrospective, multicentre cohort study including 131 suspected COVID-19 with two negative NP swabs (male 71%, age 65, range 54-74 years) the majority in Internal Medicine ward (63%), 27.5% in sub-intensive unit and 9% in ICU. SARS-Cov-2 was isolated in 43 (32.8%) BAL.26 Positive patients were younger compared to the negative ones (56 vs. 67, p=0.004) and showed a higher HRCT involvement (ground-glass, peripheral, posterior and multilobar involvement).26 Other microbiological findings were identified in 26 cases (19.8%) and included Herpesviruses, Cytomegalovirus, Staphylococcus aureus, Escherichia coli, Klebsiella pneumoniae, Pseudomonas aeruginosa and fungi . Considering both the identification of COVID-19 and the detection of other causal agent, BAL microbiological analysis was considered clinically useful in 67% of cases.26 Barberi et al in a population of hospitalized patients for suspected COVID-19, negative NP swabs and mild-moderate disease severity (PaO2/FiO2 307, range 254-362), confirmed a BAL positivity of 16% (32/198), 9% (5/54 in patients with negative HRCT).27 Moreover, BAL detected 12.5% (4/32) of coinfections in SARS-Cov-2 positive patients and 33% of other infections in SARS-Cov-2 negative patients. The logistic regression analysis detected two factors predictive of BAL positivity: fever (OR 1.94 per additional °C, 95% CI 1.13-3.33, p=0.016) and HRCT scan involvement grade 2 or more (OR 7.36, 95%CI 2.10-25.77, p=0.002).27 Contrarily to those results, three Italian single centre observational retrospective studies on BAL conducted in the same time period (March-May 2020) in suspected COVID-19 with negative NP swabs (N=81, N=79 and N=28 patients respectively), showed poor BAL performance in detecting SARS-Cov-2 infection with 3/81 (3.7%), 2/79 (2.5%) and 0/28 positive BAL.28-30 In those studies BAL negative for SARS-Cov-2 was still useful to identify other microorganisms (mycobacteria, Pneumocystis, Haemophilus parainfluenzae, Staphylococcus, Pseudomonas, Streptococcus, Enterobacterales, Klebsiella, Candida, and other viruses ).29,30 Two American studies found a 100% concordance between negative NP swabs and BAL conducted in patients that were screened for SARS-Cov-2 before an elective bronchoscopy for suspected diseases other than COVID-19 (obstructive diseases, interstitial lung disease, lung transplant surveillance etc).31,32 In the study conducted by Oberg et al, all but one patient had HRCT non suggestive for COVID-19 (negative HRTC in 58% and indeterminate or atypical in the remaining cases) and none had clinical-laboratory features of COVID-19.31 This study suggests that when the clinical-radiological scenario is not suggestive of COVID-19 and the NP swabs is negative, BAL for COVID-19 is unlikely to be useful, even during a pandemic peak.
Among these small retrospective studies there is a notable variably in the reported utility of BAL for the detection of SARS-Cov-2. This suggests that several factors may influence BAL diagnostic accuracy in detecting SARS-Cov-2, including the heterogeneity of the populations, the variability in BAL technique and sample processing. It is important to mention that BAL diagnostic yield for COVID-19 detection is also influenced by the epidemiological incidence of the disease and may be influenced by the viral variant. With the changing epidemiological scenario and novel omicron variant the BAL diagnostic yield could significantly change. In Figure 2 we present a paradigmatic case in which BAL allowed the diagnosis of COVID-19 (Omicron variant), aspergillosis and transbronchial biopsy detected lung metastasis from melanoma. Prospective studies conducted in larger and more recent populations are needed, particularly considering that the clinical scenario is rapidly changing due to the emergence of the Omicron variant in the vaccinated population.