Concordance between nasopharyngeal swab and bronchoalveolar lavage PCR for respiratory viruses

We aimed to compare the diagnostic performance of nasopharyngeal (NP) swab and bronchoalveolar lavage (BAL) polymerase chain reaction (PCR) assays for respiratory viruses (RVs) in intensive care unit patients with severe pneumonia.

Using prospective cohort data, we analysed paired NP swab and BAL samples from 968 adult intensive care unit patients from January 2010, to December 2023. Agreement between the two sampling methods was assessed using Cohen's κ coefficient. Diagnostic performance metrics, including sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV), for NP swab PCR were calculated using BAL PCR as the reference standard.

Among 968 patients, 318 (32.9%) tested positive for one or more RVs. Concordance between NP swab and BAL PCR was observed in 86.2% of paired samples. Cohen's κ indicated strong agreement (κ = 0.80-0.90) for respiratory syncytial virus, human metapneumovirus, and SARS-CoV-2, and moderate agreement (κ = 0.60-0.79) for rhinovirus, influenza virus, parainfluenza virus, and enterovirus. While specificity (97.4-100%) and NPV (97.8-100%) were consistently high across all viruses, sensitivity (47.4-100%) and PPV (33.3-100%) showed substantial variability by pathogen.

NP swab PCR demonstrated high concordance with BAL PCR and uniformly high NPVs, supporting its utility as a first-line diagnostic tool for severe respiratory viral pneumonia. However, because NP swab PCR provides no information on bacterial or fungal pathogens, BAL remains an essential complementary modality when NP swab results are negative but clinical suspicion persists, or when co-existing or alternative aetiologies are suspected.