Is Red Bone Marrow a Reliable Matrix for Estimating Blood Ethanol in Decomposed Bodies? A Pilot Study with Focus on the Lipid Fraction

Introduction: Cadaver decomposition can severely compromise toxicological analyses in conventional matrices. Bone marrow (BM), located within the medullary cavity and protected by cortical bone, has emerged as a valuable alternative matrix when blood and soft tissues are unavailable or degraded. BM is classified into red (hematopoietic) and yellow (fatty) types, and their distribution varies by anatomical sites: rib and iliac crest marrow are predominantly hematopoietic, while femoral marrow is primarily fatty. BM lipids may influence xenobiotic distribution according to their physicochemical properties. Evaluating the BM lipid fraction and its changes during decomposition may be crucial for interpreting postmortem toxicological findings. Objectives: This study aimed to: (i) develop and validate a rapid semi-quantitative spectrophotometric method to assess lipid content in red BM (RBM) and psoas muscle (PM); (ii) assess whether lipids affect ethanol concentrations in well-preserved (WP) and decomposed (D) bodies; (iii) explore predictive models to estimate femoral blood (FB) ethanol from RBM. Methods: Lipids was measured in ribs RBM and PM from 63 cadavers (24 WP, 39 D) using a sulfo-phospho-vanillin (SPV) colorimetric assay specifically validated for postmortem matrices. Ethanol (LOQ 0.05 g/L) and ethyl glucuronide (EtG) (LOQ 2.9 ng/mL) were quantified by validated HS-GC-FID and LC-MS/MS methods in FB, RBM, and PM from 166 cadavers (76 WP, 90 D). Results: Lipids were significantly higher in D cases compared to WP cases in both RBM (p=0.010) and PM (p=0.049). However, linear regression analysis showed no significant relationship between lipid and ethanol concentrations in both matrices. Mean and median ethanol ratios were, respectively, 0.99±0.36 and 0.97(FB/PM), 2.74±1.08 and 2.62 (FB/RBM), 2.84±1.18 and 2.77 (PM/RBM), without significant differences between WP and D. Significant linear regression models were found between RBM and both FB (b1=2.3, p«0.001, R²=0.79) and PM (b1=1.97, p«0.001, R²=0.81) ethanol concentrations. Moreover, in approximately 18% of D cases, ethanol was >LOQ in FB and/or PM, while remaining LOQ only in FB and/or RBM but EtG