Delayed gadolinium-enhanced MRI of cartilage (dGEMRIC) reflects cartilage glycosaminoglycan (GAG) distribution. The technique assumes that the plasma levels of the contrast agent Gd-DTPA(2-) are the same across individuals after intravenous (IV) injection, when dosing by weight. However, adipose tissue has lower extracellular water (ECW) than lean tissue. The aims of this study were to measure (1) plasma Gd-DTPA(2-) levels vs body mass index (BMI), and (2) dGEMRIC vs BMI after correcting for the dose-BMI effect.(1) Plasma Gd-DTPA(2-) levels were analyzed at 3-90 min after IV injection per body weight in 24 individuals with BMI between 21.5 and 46.5. (2) dGEMRIC was compared with BMI in 19 asymptomatic volunteers and 23 with osteoarthritis (OA).(1) Plasma Gd-DTPA(2-) kinetics were similar in obese and non-obese groups, however, overall concentration was higher in the obese group. A very obese subject (BMI 45) would have 1.4 times higher Gd-DTPA(2-) concentration than a lean subject (BMI 20), which translates into a bias in dGEMRIC of up to 20%. (2) With dose bias taken into account, dGEMRIC showed no correlation with BMI in asymptomatic knees. In OA knees, unnarrowed femoral compartments demonstrated a negative correlation between dGEMRIC and BMI (R=0.57, P=0.004). No correlation was seen in radiographically narrowed compartments.BMI can be a source of dosing bias in dGEMRIC and a correction factor should be considered in cross-sectional studies with a large range of BMI. There is no correlation between dGEMRIC and BMI in asymptomatic knees, but a negative correlation in OA knees.