Background: Prediction equations (PEs) are commonly used in clinical practice to estimate resting energy expenditure (REE) in lieu of measurement with indirect calorimetry (IDC) due to time and cost. Few PEs have been developed for estimating REE in severely obese populations; thus, it is common for a variety of PEs to be used to estimate energy needs in this population.
Objectives: In a sample of severely obese adults: (1) compare the accuracy of common REE PEs versus measured REE with IDC; (2) determine the contributions of FM and FFM to REE; and (3) develop a population specific PE for REE.
Methods: A subset of 501 severely obese adults were analyzed who participated in a previous study that involved measurements of height and weight, calculation of BMI, REE with IDC, and body composition with bioelectrical impedance. Complete data were available on 355 subjects (55 males, 300 females; BMI 46.8 +/- 7.8; age 41.1 +/- 10.9) who also had REE estimated with the following prediction equations: Harris-Benedict (HB), Mifflin, Owen, Bernstein, Robertson-Reid (RR), and Ireton-Jones (IJ). A regression equation was employed to determine contributions of age, gender, weight, and height on REE to develop a PE for REE specifically for severely obese adults. Statistical significance was set at a P < .01.
Results: The HB, Mifflin, Owen, Bernstein, and RR equations all underestimated REE as compared to IDC (P < .0001), whereas the IJ overestimated REE (P < .0001). The HB was the most accurate PE predicting 58% of REE values within +/- 10% of measured REE. Regressions analyses indicated that FFM & FM accounted for 40% and 5% of the variance respectively (R2=0.454). A new prediction equation (Mitchell) was developed that predicted REE within +/-10% of measured REE in 55% of the population. This equation is REE = 6.52 (wt kg) + 13.50 (ht cm) - 4.47 (age years) - 634.06 (males) - 778.70 (females).
Conclusions: Several common PEs for REE poorly predict REE in a population of severely obese adults, which is partly due to the difficulty in accounting for wide variations in FFM and FM. Because the current equation was derived from a large population of severely obese, it may be more appropriate; however, further studies are needed to validate the accuracy. Both the HB REE and the Mitchell PE appear to be reasonable options to estimate REE in severely obese if IDC is not available.