Implementing rapid pan-microbial metagenomics in paediatric intensive care

BackgroundClinical metagenomics studies have rarely reported rapid results with clinical impact, or deployment of protocols to centres beyond where they were developed. We describe the implementation of a previously validated rapid pan-microbial respiratory metagenomics service to paediatric intensive care, reporting modifications, feasibility, performance, clinical impact, and staff views. MethodsSingle-centre prospective interventional study of a metagenomics service testing respiratory samples from paediatric ICU patients at a London hospital between November 2024 and February 2025. This was based on a well-established published workflow detecting bacteria, fungi, and DNA/RNA viruses, with same-day preliminary results and next-day final results. Performance was assessed against routine microbiology. Clinical impact (change in antimicrobials, diagnosis or infection control) was evaluated through electronic health record review. Clinical staff were surveyed via an anonymous online semi-structured questionnaire. Findings177 samples from 122 patients were tested. 174/177 (98%) passed quality control. Sensitivity at 16 hours sequencing was 89% (95% CI 80-95) for bacteria, 100% (95% CI 54-100) for fungi, and 87% (95% CI 76-95) for viruses, with specificities [≥]99% across all kingdoms. Metagenomics identified 50 additional pathogens in 42/174 (24%) samples, all confirmed by orthogonal PCR testing. Clinical impact occurred in 51/174 (29.3%) samples, consisting of antimicrobial changes in 46/174 (26.4%), including 16/174 (9.2%) cessations and 10/174 (5.7%) initiations, and four samples (2.3%) with infection control implications. Among ICU staff surveyed (n=33), 72% viewed metagenomics positively, 0% viewing it negatively, and 67% perceived additional benefit over routine testing. InterpretationThis clinical pilot of rapid respiratory metagenomics in paediatric intensive care demonstrated high specificity and good sensitivity. Protocols developed for an adult population were adapted to a paediatric population, including for polymicrobial upper respiratory tract samples that are more common for children. However, increased sensitivity and AMR prediction will be required to be truly disruptive. Whilst changes in patient management were observed, particularly for antimicrobial stewardship, delineating the true contribution of mNGS results is difficult in this study. Multi-centre studies over multiple winter seasons, with control arms and rationalising of patient cohorts and samples, are needed to assess impact on patient outcomes and cost-effectiveness. Research in ContextO_ST_ABSEvidence before this studyC_ST_ABSMetagenomic next generation sequencing (mNGS) is increasingly being used in clinical settings for the diagnosis of infection. Technical performance of mNGS versus routine testing is frequently reported, but few studies report clinical impact of mNGS. Even rarer are studies reporting the adaptation of an established protocol from the centre of its development to a new centre with a different patient population (paediatric). We searched PubMed for studies published between 1st Jan 2020 and 1st Jan 2025, using the terms "paediatrics" AND "metagenomics" AND "treatment" AND either "respiratory" OR "pneumonia". This search returned 135 abstracts, of which three were prospective studies using respiratory metagenomic testing in a fashion which impacted patient care. Of these, one study was not configured to detect RNA viruses. The remaining two studies reported 318 paediatric cases undergoing metagenomics. Neither study reported same day reporting of results with the intention of providing actionable results in real time. In summary there is minimal literature on the use of metagenomics as a rapid diagnostic test for respiratory infections in paediatric intensive care, or of adapting adult metagenomic protocols to paediatric populations. Added value of this studyThis observational prospective study is the first to show adaptation of a same day mNGS protocol able to detect bacteria, fungi, DNA and RNA viruses in a paediatric population. We show that modification of the adult protocol to include more upper respiratory and polymicrobial specimens was successful in achieving comparable performance standards. Metagenomics more rapidly identified bacterial pathogens which were also detected by standard-of-care methods as well as additional organisms that were not identified by routine methods. Taken together the data shows impact on antimicrobial prescribing, which could have implications for antimicrobial stewardship and clinical outcomes, although currently the precise contribution of mNGS to these antimicrobial prescribing decisions is unclear. Implications of all the available evidenceThis data suggests a role for respiratory metagenomics for diagnosing respiratory infection in paediatric intensive care. The results show the feasibility of implementing adult protocols in paediatric populations. Randomised control trials and cost-benefit analyses are still needed to assess true impact on length of hospital stay, clinical outcomes, antimicrobial usage and mortality.