This study demonstrates that urinary metabolite profiling can accurately distinguish children with suspected community-acquired pneumonia (CAP) from healthy controls. Using magnetic resonance spectrometry and machine learning models, specific metabolites showed high discriminatory power, comparable to combined clinical and metabolite data.
Background
Community-acquired pneumonia (CAP) is a common pediatric illness with diagnostic challenges due to nonspecific clinical signs and variable chest radiograph interpretations. Traditional diagnostic methods often lack reliability and specificity, leading to potential misdiagnosis. Metabolomics, the analysis of small molecules in biological samples, offers a noninvasive approach to identify unique metabolic signatures associated with infections. Prior studies in adults have shown promise, but pediatric data have been limited and small in scale.
Data Highlights
Group
Number of Participants
Age Range (months to years)
Model AUC
CAP Cases
253
3 months to 12 years
Metabolite-only: 0.97
Community Controls
122
3 months to 12 years
Combined clinical + metabolite: 0.99
Key Findings
The metabolite-only model discriminated CAP cases from controls with an AUC of 0.97, nearly matching the combined clinical and metabolite model (AUC 0.99).
Key discriminatory metabolites included 2-aminobutyrate, fumarate, hypoxanthine, acetone, leucine, quinolinate, valine, O-acetylcarnitine, citrate, and trigonelline.
The combined model incorporated clinical factors such as corticosteroid use, fever, cough, rapid breathing, decreased oral intake, difficulty breathing, albuterol use, abnormal sleepiness, vomiting, and wheezing, plus five additional metabolites.
Urine metabolite concentrations provide a noninvasive biomarker panel that may overcome limitations of nonspecific clinical signs and chest radiograph interpretation variability.
The study was sufficiently powered with 253 cases and 122 controls, covering a broad pediatric age range (3 months to 12 years).
Clinical Implications
Urinary metabolite profiling offers a promising, noninvasive diagnostic adjunct for pediatric CAP, potentially improving diagnostic accuracy in ambulatory and emergency settings. Incorporating metabolite biomarkers could reduce reliance on subjective clinical assessments and chest radiographs, guiding more targeted management and antibiotic stewardship.
Conclusion
Urinary metabolomics can accurately discriminate children with suspected CAP from healthy controls, supporting its potential as a reliable diagnostic tool. Further validation and development may enable rapid bedside testing to improve pediatric pneumonia diagnosis.
References
Schubert Research Clinic/Cincinnati Children’s Hospital Medical Center/2024 -- Urinary Metabolites Associated with Suspected Community-Acquired Pneumonia
by Lilliam Ambroggio, Todd A Florin, Kayla Williamson, Grace Bosma, Brandie D Wagner, Larisa Yeomans, Jae Hyun Kim, Heidi Sucharew, Maurizio Macaluso, Richard M Ruddy, Kathleen A Stringer, Samir S Shah
A large English cohort study found influenza hospitalization more than doubled the short-term risk of new-onset diabetes, with prediabetes and critical care admission among the strongest predictors.
