Incidence, microbiological and immunological characteristics of ventilator-associated pneumonia assessed by bronchoalveolar lavage and endotracheal aspirate in a prospective cohort of COVID-19 patients: CoV-AP study

Abstract

Abstract Background No univocal recommendation exists for microbiological diagnosis of ventilator-associated pneumonia (VAP). Sampling of either proximal or distal respiratory tract likely impacts on the broad range of VAP incidence between cohorts. Immune biomarkers to rule-in/rule-out VAP diagnosis, although promising, have not yet been validated. COVID-19-induced ARDS made VAP recognition even more challenging, often leading to overdiagnosis and overtreatment. We evaluated the impact of different respiratory samples and laboratory techniques on VAP incidence and microbiological findings in COVID-19 patients. Methods Prospective single-centre cohort study conducted among COVID-19 mechanically ventilated patients in Policlinico Hospital (Milan, Italy) from January 2021 to May 2022. Microbiological confirmation of suspected VAP (sVAP) was based on concomitant endotracheal aspirates (ETA) and bronchoalveolar lavage (BAL). Conventional and fast microbiology (FILMARRAY® Pneumonia Panel plus, BALFAPPP) as well as immunological markers (immune cells and inflammatory cytokines) was analysed. Results Seventy-nine patients were included. Exposure to antibiotics and steroid therapy before ICU admission occurred in 51/79 (64.6%) and 60/79 (65.9%) patients, respectively. Median duration of MV at VAP suspicion was 6 (5–9) days. Incidence rate of microbiologically confirmed VAP was 33.1 (95% CI 22.1–44.0) and 20.1 (95% CI 12.5–27.7) according to ETA and BAL, respectively. Concordance between ETA and BAL was observed in 35/49 (71.4%) cases, concordance between BALFAPPP and BAL in 39/49 (79.6%) cases. With BAL as reference standard, ETA showed 88.9% (95% CI 70.8–97.7) sensitivity and 50.0% (95% CI 28.2–71.8) specificity (Cohen’s Kappa 0.40, 95% CI 0.16–0.65). BALFAPPP showed 95.0% (95% CI 75.1–99.9) sensitivity and 69% (95% CI 49.2–84.7) specificity (Cohen’s Kappa 0.60, 95% CI 0.39–0.81). BAL IL-1β differed significantly between VAP (135 (IQR 11–450) pg/ml) and no-VAP (10 (IQR 2.9–105) pg/ml) patients (P = 0.03). Conclusions In COVID-19 ICU patients, differences in microbial sampling at VAP suspicion could lead to high variability in VAP incidence and microbiological findings. Concordance between ETA and BAL was mainly limited by over 20% of ETA positive and BAL negative samples, while BALFAPPP showed high sensitivity but limited specificity when evaluating in-panel targets only. These factors should be considered when comparing results of cohorts with different sampling. BAL IL-1β showed potential in discriminating microbiologically confirmed VAP. Clinical Trial registration: NCT04766983, registered on February 23, 2021.