Implementation of rapid pan-microbial metagenomics in paediatric intensive care: a single-centre, prospective,...

Few clinical metagenomics studies have reported rapid results with associated clinical impact or described deployment of protocols beyond centres in which they were developed. We aimed to assess the feasibility and clinical impact of implementing a previously validated rapid pan-microbial respiratory metagenomic service in paediatric intensive care, reporting protocol modifications, feasibility, diagnostic performance, clinical impact, and staff perspectives.

This single-centre, prospective, observational study was carried out at a general paediatric intensive care unit (ICU) and a cardiac paediatric ICU in Great Ormond Street Hospital (London, UK) between Nov 4, 2024, and Feb 28, 2025. Respiratory samples were obtained from patients selected by intensive care staff based on clinical suspicion of lower respiratory tract infection, with no exclusion criteria. Samples were analysed by a well-established published metagenomic workflow that simultaneously detects bacteria, fungi, and DNA and RNA viruses and provides same-day preliminary results and next-day final results; separate aliquots were sent for concurrent culture and PCR testing. Clinical impact (change in antimicrobials, diagnosis, or infection control) was evaluated through electronic health record review. Clinical staff were surveyed for their feedback via an anonymous online semistructured questionnaire.

177 samples from 122 patients (median age 3·4 months [IQR 1·2-36·0]; 54% male) were tested. 174 (98%) samples passed quality control. Sensitivity at 16-h sequencing was 89% (95% CI 80-95) for bacteria, 100% (70-100) for fungi, and 87% (76-93) for viruses, with specificities greater than or equal to 99% across all kingdoms. 42 (24%) of 174 samples analysed by metagenomics yielded 50 organisms that were not detected by routine diagnostics, all of which were confirmed by orthogonal PCR testing. Clinical impact was observed in 51 (29%) of 174 samples, comprising changes in antimicrobial therapy in 46 (26%) samples. Of these 46 samples, 15 (33%) resulted in cessation of antibiotics, ten (22%) in initiation of antibiotics, six (13%) in a change of antibiotic therapy, and 15 (33%) provided confidence to continue an empirical antibiotic choice. Of the 33 ICU staff members surveyed, 23 (72%) reported a positive view of metagenomics, none reported negative views, and 22 (67%) perceived additional benefits over routine testing.

This clinical pilot of rapid respiratory metagenomics in paediatric intensive care showed high specificity and good sensitivity. Protocols developed for an adult population were adapted to a paediatric population, including adaptations for polymicrobial upper respiratory tract samples, which are more common in children. However, increased sensitivity and antimicrobial resistance prediction will be required for broader clinical implementation. Although changes in patient management were observed, particularly for antimicrobial stewardship, delineating the true contribution of metagenomic results offered by this study is difficult. Multicentre studies over multiple winter seasons, with control arms and predefined selection of patients and sample types, are needed to assess the impact on patient outcomes and cost-effectiveness.

NHS Respiratory Metagenomics Network (NHS England).