Incremental diagnostic yield of bronchoscopy in non-resolving pneumonia: a prospective observational study from Eastern India

Abstract

Abstract Background Non-resolving pneumonia (NRP) remains a significant diagnostic challenge because of its diverse etiologies, particularly in tuberculosis-endemic regions. Although a structured non-bronchoscopic evaluation establishes the diagnosis in a substantial proportion of patients, many remain without a definitive etiological diagnosis. Bronchoscopy is frequently employed in such patients; however, its incremental diagnostic contribution following an inconclusive non-bronchoscopic evaluation has not been adequately defined. This study aimed to evaluate the incremental diagnostic yield of bronchoscopy following a structured non-bronchoscopic diagnostic evaluation in patients with NRP. Methods This prospective observational study included 130 adult patients with NRP evaluated at a tertiary care centre in Eastern India between June 2024 and December 2025. All patients initially underwent a structured non-bronchoscopic diagnostic evaluation. Patients who remained undiagnosed after the initial evaluation ( n = 78) subsequently underwent flexible bronchoscopy. Bronchoscopic evaluation was performed using a standardized protocol comprising bronchoalveolar lavage (BAL), bronchial brush cytology, and transbronchial lung biopsy (TBLB) during the same bronchoscopy session in all patients. The primary outcome was the incremental diagnostic yield of bronchoscopy, defined as the additional definitive diagnoses established following an inconclusive non-bronchoscopic evaluation. Results The initial non-bronchoscopic evaluation established a definitive diagnosis in 52 of 130 patients (40.0%). The remaining 78 patients underwent bronchoscopic evaluation, which established an additional diagnosis in 34 patients, corresponding to an incremental diagnostic yield of 43.6%. Consequently, the cumulative diagnostic yield increased from 40.0% (52/130) to 66.2% (86/130) following completion of the sequential diagnostic evaluation ( p = 0.040). Bronchoscopic evaluation facilitated the diagnosis of a broad spectrum of infectious and malignant diseases, with tuberculosis (45.1%), malignancy (31.4%), and fungal infections (19.6%) representing the most frequent definitive diagnoses. Conclusion A structured non-bronchoscopic evaluation remains an appropriate first-line diagnostic approach for patients with NRP, particularly in tuberculosis-endemic settings. In patients who remain undiagnosed following the initial evaluation, bronchoscopy provides a significant incremental diagnostic yield and substantially improves the overall diagnostic success of a sequential diagnostic strategy.