A simple urine test could one day tell doctors which children with a blocked kidney need surgery and which can safely be watched, thanks to a new study that mapped the molecular fingerprints of one of the most common birth defects of the urinary tract. Researchers in Brazil have used an untargeted metabolomics strategy, combining two powerful analytical platforms, to reveal a set of small molecules in urine that change in a highly characteristic way as congenital blockage at the ureteropelvic junction progresses from a harmless finding to a genuine threat to kidney function. The work, carried out at the Hospital das Clínicas of the University of São Paulo Medical School, offers the most comprehensive molecular portrait to date of ureteropelvic junction obstruction, or UPJO, in children.
UPJO is no rarity. Fetal hydronephrosis, the swelling of the kidney caused by urine accumulating behind a blockage, is detected in roughly 0.25 to 1 percent of pregnancies by routine antenatal ultrasound, and UPJO is the most commonly diagnosed congenital urinary tract anomaly, affecting about one in 1,500 live births. It remains the leading cause of kidney failure in infants and children. Yet the condition is notoriously difficult to manage because it behaves like a spectrum rather than a single disease. Some children harbor an obstruction that silently destroys the kidney, while others carry a dilation that never causes harm. Approximately 20 percent of affected children will eventually need surgery, but deciding which ones has long been a source of controversy among pediatric urologists.
Current diagnostic practice relies on ultrasound measurements of the renal pelvis, nuclear medicine scans that assess differential kidney function and urinary drainage, and the appearance of symptoms such as pain, infection or blood in the urine. Severe obstruction demands surgical repair, typically the Anderson-Hynes dismembered pyeloplasty, to prevent irreversible loss of renal function, whereas children with non-obstructive hydronephrosis can be followed conservatively for years. The problem is that these imaging-based reference standards are unsatisfactory at predicting which asymptomatic patients will deteriorate. Several candidate protein biomarkers, including retinol-binding protein-4, NGAL, KIM-1, TGF-beta 1, MCP-1, endothelin-1 and epidermal growth factor, have been screened over the years, but none has entered clinical practice due to modest performance and limited availability.
The Brazilian team took a different route. Instead of examining proteins or genes, they turned to metabolomics, the comparative study of the complete set of low-molecular-weight metabolites, the metabolome, expressed under given conditions. Because metabolites sit at the downstream end of biology, their patterns can integrate the effects of genes, proteins and environment, making them sensitive reporters of tissue stress and injury. The researchers reasoned that a blocked kidney under pressure, and one that is merely dilated but functioning normally, should leave distinguishable chemical traces in the urine. To maximize metabolome coverage, they employed two orthogonal analytical techniques: gas chromatography coupled to mass spectrometry, ideal for volatile organic acids and derivatized small molecules, and high-performance liquid chromatography coupled to high-resolution quadrupole time-of-flight mass spectrometry with electrospray ionization, which captures a broader range of polar and semi-polar compounds.
The prospective observational study, approved by the local ethics committee and conducted between September 2016 and June 2019, enrolled 37 children aged 2 to 12 years in three age-matched groups. Ten children had established obstructive hydronephrosis and underwent pyeloplasty based on accepted guideline criteria, including symptomatic obstruction, differential renal function below 40 percent, a drop of more than 10 percent in renal function on serial scans, or poor drainage after diuretic challenge. Fifteen asymptomatic children with non-obstructive hydronephrosis were managed conservatively and remained clinically and radiologically stable throughout follow-up. Twelve healthy children without any urinary tract problems served as controls. All patients had normal estimated glomerular filtration rates above 90 ml/min/1.73 m2, underscoring that the molecular changes detected reflect early, subclinical perturbation rather than overt renal failure. Urine samples were collected aseptically, centrifuged, stored at minus 80 degrees Celsius, and screened for infection before analysis.
The analytical workflow was exacting. For gas chromatography, urine was treated with urease to remove urea, proteins were precipitated with cold isopropanol, and dried extracts underwent methoximation with methoxyamine hydrochloride and silylation with BSTFA before injection onto a 5-percent-phenyl-methylpolysiloxane capillary column. Liquid chromatography separated compounds on a pentafluorophenyl column with a formic acid-methanol gradient, and mass accuracy was maintained with continuous lock-mass infusion. Quality control samples, pools of all study urines, were injected after every five regular samples, and only features with relative standard deviations below 30 percent in these pools proceeded to identification. This dual-platform design proved crucial: in one comparison, the supervised multivariate model failed to separate groups using GC-MS data alone, yet the univariate approach on the same data revealed significant metabolites, and the HPLC-MS platform compensated where GC-MS was blind.
Statistically, the study was rigorous for a pilot of this size. Univariate analyses using ANOVA with Tukey post-hoc testing and Benjamini-Hochberg false discovery rate correction, together with multivariate partial least-squares discriminant analysis validated by 7-fold cross-validation, 200-permutation testing and CV-ANOVA, produced strong performance metrics: adjusted p-values below 0.05, q-values between 0.04 and 0.1, and areas under the receiver operating characteristic curve ranging from 0.662 to 0.882 for univariate findings, and from 0.923 to 0.995 for multivariate models, with Q-squared prediction values between 0.714 and 0.859. Out of 94 discriminatory metabolites identified across all group comparisons, 17 achieved particular significance because they were confirmed jointly by both univariate and multivariate methods, the strongest statistical evidence the design could provide.
The chemistry told a coherent biological story. Comparing non-obstructive hydronephrosis with controls, three metabolites, 1-methylhistidine and 3-methylhistidine, 1-methylhistamine and 3-methylhistamine, and the dipeptide glutamyl-hydroxyproline, emerged as markers of the earliest metabolic perturbation implanted by the obstruction. Nine metabolites, including mandelic acid, pantothenic acid, furoylglycine, gluconic acid, 3-hydroxyphenylacetate, 4-hydroxyphenylacetate, 3-hydroxy-3-methylglutaric acid, pentose alcohols and tagalose, distinguished obstructive cases from controls, reflecting the long-term metabolic impact of established blockage. Cystine and methionine sulfoxide showed the largest fold increases of all, at 156 and 152 percent respectively, in obstructive versus control children, hinting at profound oxidative stress and protein oxidation. Most clinically compelling, five metabolites, ascorbic acid, furoylglycine, gluconic acid, pentoses and threitol, separated the children who required surgery from those who did not, and therefore constitute the leading candidates for a decision-support biomarker panel. In pathway analyses, beta-alanine metabolism was the only route significantly altered at early stages, consistent with a developing antioxidant defense, and its progression implicated tyrosine, phenylalanine, alanine, aspartate, glutamate, and amino sugar and nucleotide sugar metabolisms as the disease advanced.
Some of the individual findings carry mechanistic weight. Beta-alanine was decreased by 36 percent in the obstructive versus non-obstructive comparison, and pantothenic acid, a coenzyme A precursor with membrane-protective properties, fell by 49 and 59 percent against controls, reinforcing the link between obstruction and mounting oxidative burden on the podocytes of the glomerular filter. 4-Hydroxyphenylacetic acid, a key metabolite of the tyrosine and phenylalanine pathways, rose by 112 percent in obstructive versus non-obstructive children and 84 percent against controls, and had already been associated with declining renal function and podocyte injury elsewhere; recent work also suggests it can protect renal tubular cells from necroptosis during sepsis. Disturbances in the amino sugar pathway, with alpha-glucosamine 1-phosphate elevated and N-acetyl-D-mannosamine, the precursor of sialic acids essential to the glomerular filtration barrier, depleted in the surgical group, connect the metabolite signature directly to the structural integrity of the kidney filter.
The authors are candid that this exploratory study was not designed to deliver a definitive clinical biomarker, and that the modest cohort size and age range limit generalizability. Even so, the convergence of two analytical platforms, validated statistics and biologically plausible pathways lends the 94-metabolite signature considerable credibility as a diagnostic and prognostic scaffold. The team emphasizes that larger cohorts for external validation, targeted quantitative assays and biological testing are the necessary next steps before any of these molecules reaches the clinic. If that validation succeeds, the dream of a non-invasive urine test that spares asymptomatic children unnecessary surgery while protecting those with true obstruction from silent kidney loss may move a decisive step closer to reality.
Subject of Research: Urinary metabolomics profiling of children with ureteropelvic junction obstruction to identify biomarkers of obstructive hydronephrosis
Article Title: Urinary metabolomics profile in children with ureteropelvic junction obstruction
Article References: da Silva, L. H. F., Mello, M. F., Kataoka, F. G. H., Amsei Filho, U. F., Farah, J. P. S., Yamaguchi, L. F., Kato, M. J., de Oliveira, M. A. L., Faccio, A. T., Andrade, L., Dénes, F. T., Lopes, R. I., & Tavares, M. F. M. (2026). Urinary metabolomics profile in children with ureteropelvic junction obstruction. Metabolomics, 22(5), Article 154. https://doi.org/10.1007/s11306-026-02527-0
Image Credits: AI Generated
DOI: 10.1007/s11306-026-02527-0
Keywords: metabolomics, ureteropelvic junction obstruction, hydronephrosis, biomarkers, pediatric urology, GC-MS, HPLC-MS, urine test, kidney obstruction, pyeloplasty, oxidative stress, beta-alanine metabolism
Cite Scienmag News
APA MLA Chicago
Copy citation Download RIS
Tags: beta-alanine metabolismBiomarkersbiomarkers predicting kidney function decline in childrenclinical management of ureteropelvic junction obstructioncongenital urinary tract anomalies diagnostic toolsearly detection of fetal hydronephrosisGC–MSHPLC-MShydronephrosiskidney failure risk assessment in childrenkidney obstructionMetabolomicsmetabolomics in pediatric nephrologymolecular fingerprinting of ureteropelvic junction obstructionnon-invasive urine testing for kidney surgery needsOxidative stresspediatric urologypyeloplastyuntargeted metabolomics in pediatric kidney diseasesureteropelvic junction obstructionurinary molecular signatures forurine biomarkers for pediatric kidney obstructionurine testurine-based diagnostics for congenital kidney conditions
