Differences among canine serum, plasma and urine metabolite profiles from samples that were collected at the same time

Background Biospecimen choice can substantially influence metabolite measurements, yet matched comparisons of serum, EDTA plasma, lithium-heparin plasma, and urine remain limited in canine targeted metabolomics. Objectives To evaluate cross-matrix comparability of targeted small polar metabolite profiles in dogs and to assess how different biospecimens capture diet- and health-associated variation. Animals Client-owned dogs from a previously reported feeding trial. Methods Targeted ultra-performance liquid chromatography–tandem mass spectrometry (UPLC-MS/MS) was performed on matched biospecimens from a baseline blood-matrix cohort (serum, EDTA plasma, lithium-heparin plasma) and an end-of-trial serum-urine cohort. Metabolites with >20% missing values within a matrix and samples with >20% missing metabolite values within a biofluid were excluded. Cross-matrix agreement was assessed using Pearson correlation and mean paired log 10 concentration differences. PCA and PERMANOVA were used to compare global matrix effects and diet- and health-associated separation. Results After filtering, 88 metabolites remained for serum, 85 for EDTA plasma, and 88 for lithium-heparin plasma in the blood-matrix cohort. Globally, EDTA plasma separated clearly from both serum and lithium-heparin plasma, whereas serum and lithium-heparin plasma overlapped substantially. At the metabolite level, 10/84 (11.9%) serum–EDTA plasma, 16/88 (18.2%) serum–lithium-heparin plasma, 20/84 (23.8%) lithium-heparin plasma–EDTA plasma, and 4/75 (5.3%) serum–urine comparisons were comparable to varying degrees. Serum showed the clearest diet-associated separation at baseline (PERMANOVA pseudo- R 2 = 0.13, p = 0.013) and end-of-trial (pseudo- R 2 = 0.22, p = 0.021). Health-associated separation was weak and exploratory. Conclusions and Clinical Importance Biospecimen choice materially affects targeted metabolite profiling in dogs. Serum and lithium-heparin plasma were globally more similar to each other than either was to EDTA plasma, whereas urine was a distinct complementary matrix rather than a blood surrogate. Serum was the most robust matrix for detecting diet-associated variation.