Added Diagnostic Utility of Clinical Metagenomics for the Diagnosis of Pneumonia in Immunocompromised Adults

Background: In the evaluation of community-acquired pneumonia, 30% to 60% of cases remain undiagnosed, despite extensive conventional microbiologic testing (CMT). Clinical metagenomics (CM) is an unbiased pathogen detection method that can increase diagnostic yield.

Research question: Does adding clinical metagenomics to conventional microbiologic testing improve the diagnostic yield for pneumonia in immunocompromised adults?

Study design and methods: We performed a noninterventional prospective study of immunocompromised adults with pneumonia who underwent bronchoscopy and BAL over 2 years. CMT was performed per standard of care. A commercial CM test was performed on residual BAL fluid. Final microbiologic diagnoses were based on CMT vs CMT + CM. Final clinical diagnoses for CMT and CMT + CM were made based on laboratory results in conjunction with clinical and radiologic findings. Hypothetical impact of CMT + CM on management and antimicrobial stewardship was also assessed.

Results: A total of 30 immunocompromised adult patients (31 episodes of pneumonia) were included. Final microbiologic diagnoses were made in 11 cases (35%) with the use of CMT and in 18 cases (58%) with the use of CMT + CM. Bacterial pneumonia was diagnosed in five cases (16%) by CMT and in 13 cases (42%) by CMT + CM; fungal pneumonia was diagnosed in six cases (19%) by CMT and in seven cases (23%) by CMT + CM, and viral pneumonia was diagnosed in two cases (6%) by CMT and in five cases (16%) by CMT + CM. The hypothetical impact of CMT + CM on management was deemed probable in one case, possible in eight cases, and unlikely in two cases, whereas the impact on antimicrobial stewardship was possible in 13 cases and unlikely in seven cases. Final clinical diagnoses were made in 20 of 31 cases (65%) based on CMT and in 23 of 31 cases (74%) based on CMT + CM.

Interpretation: CMT + CM increased diagnostic yield in immunocompromised adults with pneumonia from 35% to 58%, mostly by the detection of additional bacterial causes but was less useful for fungal pneumonia.