Pre-analytical variation is a major challenge in metabolomics, yet most stability studies have focused on healthy volunteers and overlooked the impact of disease and medication. To address this gap, we conducted a pilot study in systemic lupus erythematosus (SLE) to assess serum metabolite stability under delayed centrifugation. Peripheral blood from 10 SLE patients and 5 healthy controls (HC) was stored at room temperature for 1–24 h before processing and analyzed by untargeted LC–MS-based metabolomics. This design enabled direct evaluation of pre-analytical delay in the context of clinical heterogeneity. Principal component trajectories showed reproducible temporal shifts in HC but dispersed patterns in SLE, underscoring disease- and treatment-related influences. Linear mixed-effects models identified metabolites with condition-specific kinetics, including glucose, choline, glycerophosphocholine, and pyroglutamic acid. Mycophenolate intake was further associated with distinct AMP dynamics. These findings demonstrate that both disease state and medication reshape apparent metabolite stability, highlighting the need for strictly controlled preparation and well-characterized cohorts in clinical metabolomics.

Background/Objectives: Pre-analytical variation is a major challenge in metabolomics, yet most stability studies have focused on healthy volunteers and have overlooked the impact of disease and medication. To address this gap, we conducted a pilot study in systemic lupus erythematosus (SLE) to assess serum metabolite stability under delayed centrifugation. Methods: Peripheral blood from 10 SLE patients and 5 healthy controls (HC) was stored at room temperature for 1–24 h before processing and analyzed by untargeted LC-MS-based metabolomics. This design enabled direct evaluation of the effect of pre-analytical delay within the context of clinical heterogeneity. Results: Principal component trajectories showed reproducible temporal shifts in HC but dispersed patterns in SLE, indicating disease- and treatment-related influences. Linear mixed-effects models identified metabolites with condition-specific kinetics, including glucose, choline, glycerophosphocholine, and pyroglutamic acid. Mycophenolate intake was further associated with distinct AMP dynamics. Conclusions: These findings demonstrate that both disease state and medication reshape apparent metabolite stability, highlighting the need for strictly controlled sample handling and well-characterized clinical cohorts in metabolomics studies.