Standardized preanalytical sample handling is a cornerstone of reliable molecular analytics, particularly where subtle multimolecular differences can drive biological interpretation. In different clinical studies, blood-based samples are often processed using varying workflows, and the influence of individual technical parameters remains poorly quantified. In particular, preanalytical differences can limit study comparability and complicate the robust evaluation of the approach for different applications and specific medical questions. Here, we systematically investigate the effects of major preanalytical variables on infrared molecular fingerprinting (IMF) of blood serum and plasma, performed via Fourier transform infrared (FTIR) spectroscopy. We examine the impacts of venous blood draw tube types, tube fill volumes, delay times prior to centrifugation, centrifugation conditions, freezing and thawing procedures, and delay times between final sample preparation and infrared spectroscopy. We found that the type of blood collection tube, the fill volume of the tube, a delay of more than 4 h before centrifugation, more than two freeze-thaw cycles, and delays prior to measurement can result in systematic changes in infrared fingerprints that are detectable through multivariate analysis and machine learning. Importantly, for most conditions, the magnitude of parameter-related spectral variation remains smaller than the intrinsic biological variability across individuals, supporting the feasibility of IMF for realistic clinical settings. The presented results provide actionable guidance and a basis for harmonizing preanalytical workflows, reducing preanalytical confounding in IMF-based biofluid analysis, and improving cross-study comparability, with relevance that may generalize to other molecular profiling technologies, clinical studies, and biobanking.
Dietmann et al. (Wed,) studied this question.