MSACL 2026 Abstract

Urinary Tract Infection Produces an Acute Urinary Metabolomic Signature of Host Muscle-Protein Catabolism in Pregnancy

Thomas Horvath (1), Cindy Shope (2), Maxim Seferovic (2)
(1) Department of Pathology, Texas Children’s Hospital and Baylor College of Medicine, Houston, TX, United States (2) Department of Obstetrics & Gynecology, Texas Children’s Hospital and Baylor College of Medicine, Houston, TX, United States

Maxim Seferovic, PhD (Presenter) Baylor College of Medicine

Relevant Financial Disclosures (within past 24 months, reported on Apr 22, 2026) No relevant financial relationship(s) to disclose.

INTRODUCTION:
Targeted urinary metabolomics captures microbial products (e.g. histamine, GABA) alongside host protein-turnover markers: 3-methylhistidine (3-MeHis), released from actin/myosin during muscle proteolysis, and 1-methylhistidine (1-MeHis). Here we set out to profile eleven metabolites in a longitudinal pregnancy cohort to assess whether we could associate these markers to pregnancy complications including UTIs which are common in pregnant women and associated with poor outcomes like preterm birth.

METHODS:
Eleven urinary metabolites were profiled by targeted LC-MS/MS in 92 pregnant subjects (37 UTI-positive, 54 controls) across four pregnancy visits, delivery, and postnatal. PQN normalization corrected GFR-driven dilution drift. Dated UTI episodes enabled pre/concurrent/post-infection timing. Linear mixed-effects models, within-UTI-subject contrasts, outlier refits, bootstrap antibiotic mediation, and UTI × DM interaction were fit.

RESULTS:
UTI was associated with elevated 3-MeHis (β = +0.30 SD, p = 0.007) and 1-MeHis (β = +0.25, p = 0.039). Visits within 2 weeks of diagnosis showed β = +0.52 SD for 3-MeHis (p = 0.003), attenuating to +0.16 at >2 wk post-infection (p = 0.50) and absent pre-infection (β = −0.05, p = 0.89) demonstrating a temporal effect to the infection. A within-UTI-subject analysis (concurrent vs same-woman pre+post visits, 28 subjects) was β = +0.67 SD (p = 0.058), confirming acute induction of metabolite levels. Pre-infection samples 7–9 wk before UTI showed z = −0.77 and −0.41 (below control mean), which excludes baseline predisposition. UTI subjects received more antibiotics (3.07 fold courses, p < 0.002), but total courses did not mediate the elevation (20.5%, ns). The 3-MeHis increase in UTI remained positive after antibiotic adjustment (c' = +0.18) in mediation analysis indicating independent effect. We observed a significant synergistic interaction between UTI and maternal diabetes. While a UTI independently increased 3-MeHis levels, the presence of diabetes potentiated this effect, with an additional increase of 0.54 in the interaction model and tracked UTI status regardless of preterm-birth outcome.

CONCLUSIONS:
UTI in pregnancy produces an acute, temporally anchored urinary methylhistidine signature, likely explained as host muscle-actin/myosin catabolism from infection-induced catabolic stress rather than microbial production. The signal peaks at infection, is absent pre-infection in the same women, is not explained by antibiotics, and is amplified by maternal diabetes. The finding position 3-MeHis as a candidate noninvasive readout of infection-induced host proteolysis in perinatal medicine.