A single axial CT image at the third lumbar vertebra (L3) was used to examine body composition parameters (9 (link)). Body composition parameters were quantified using Slice-O-Matic 5.0 software (TomoVision), based on density thresholds in Hounsfield units (HU); 29 to +150 for SM, −190 to −30 for SAT, and −150 to −50 for VAT. Total cross-sectional areas were measured in cm2 for SM, VAT and SAT and mean HU densities were reported for SMD. Incomplete cross-sectional areas of SM and SAT in CT scans (e.g., due to poor positioning or large body size of the patient) were estimated in MATLAB. Cross-sectional areas were normalized for height squared (m2) to obtain SMI (cm2/m2), subcutaneous adipose tissue index (SATI, cm2/m2) and VAT index (VATI, cm2/m2) (10 (link)). All CT scan analyses were performed by one trained researcher (J.S.F.M.) following a standardized protocol. A small subset (n=30) was analyzed by a second researcher to assess interrater reproducibility by calculating intraclass correlation coefficients (0.992 for SMI, 0.999 for VAT and 0.979 for SAT).