Circulating adiponectin levels are lower in individuals with increased BMI and central adiposity. However, they are paradoxically higher in those with peripheral adiposity. We hypothesized that adiponectin secretion from central and peripheral adipose tissue depots may be associated with adiposity levels and its distribution. A total of 55 subjects (69% women) undergoing elective abdominal surgery (mean age: 53 ± 13 years) were recruited. Health history, anthropometrics, and cardiovascular disease risk factor measurements were obtained. Subcutaneous adipose tissue (SAT) and visceral adipose tissue (VAT) samples were obtained and cultured. Media was collected after 24hr and adiponectin released into the medium was measured using ELISA. We found that mean adiponectin levels from SAT and VAT in all subjects were 17.14±15.27 vs. 15.21±14.28 pg/ml/mg of tissue respectively (p = ns). However, adiponectin secretion from VAT correlated negatively with BMI (r = -0.31, p = 0.01), whereas there was no relationship with SAT (r = 0.08 p = 0.61). Similarly, waist circumference and estimated VAT percentage were both negatively correlated with VAT secretion of adiponectin (r = -0.35, p = 0.01 and r = -0.36, p = 0.02 respectively). These negative correlations were significant only in women on gender-stratified analyses. Adiponectin secretion from VAT decreases with increases in adiposity, while SAT secretion remains unchanged, especially in women. This observation may explain lower circulating adiponectin levels in individuals with central obesity. Further studies are needed to explore the mechanism behind this discrepant adiponectin secretion from SAT and VAT with increases in BMI, particularly among women.

Background Resting metabolic rate (RMR) is a key determinant of daily caloric needs. Respirometry, a form of indirect calorimetry (IC), is considered one of the most accurate methods to measure RMR in clinical and research settings. It is impractical to measure RMR by IC in routine clinical practice; therefore, several formulas are used to predict RMR. In this study, we sought to determine the accuracy of these formulas in determining RMR and assess additional factors that may determine RMR. Methods We measured RMR in 114 subjects (67% female, 30% African American [AA]) using IC. Along with standard anthropometrics, dual-energy X-ray absorptiometry was used to obtain fat-free mass(FFM) and total fat mass. Measured RMR (mRMR) by respirometry was compared with predicted RMR (pRMR) generated by Mifflin–St.Joer, Cunningham, and Harris–Benedict (HB) equations. Linear regression models were used to determine factors affecting mRMR. Results Mean age, BMI, and mRMR of subjects were 46 ± 16 years (mean ± SD), 35 ± 10 kg/m 2 , and 1658 ± 391 kcal/day, respectively. After adjusting for age, gender, and anthropometrics, the two largest predictors of mRMR were race ( p < 0.0001) and FFM ( p < 0.0001). For every kg increase in FFM, RMR increased by 28 kcal/day ( p < 0.0001). AA race was associated with 144 kcal/day ( p < 0.0001) decrease in mRMR. The impact of race on mRMR was mitigated by adding in truncal FFM to the model. When using only clinically measured variables to predict mRMR, we found race, hip circumference, age, gender, and weight to be significant predictors of mRMR ( p < 0.005). Mifflin–St.Joer and HB equations that use just age, gender, height, and weight overestimated kcal expenditure in AA by 138 ± 148 and 242 ± 164 ( p < 0.0001), respectively. Conclusion We found that formulas utilizing height, weight, gender, and age systematically overestimate mRMR and hence predict higher calorie needs among AA. The lower mRMR in AA could be related to truncal fat-free mass representing the activity of metabolically active intraabdominal organs.