Ability to predict resting energy expenditure with six equations compared to indirect calorimetry in octogenarian men

The accuracy of predictive equations for calculating resting energy expenditure (REE) in elderly people has been questioned. Aging is associated with progressive declines in REE, which partly is explained by loss of fat free mass (FFM). Against this background we aimed to identify the most accurate predictive equation for REE in octogenarian men, taking body composition into account and using indirect calorimetry as reference value. REE was measured in 22 men (mean age 82.6±0.3years) and compared with six predictive equations: two based on FFM and four based on body weight, height and/or age. FFM was derived from Dual-energy X-ray absorptiometry analyses. Spearman's rank correlations showed a moderate to high positive monotonic correlation (r=0.62 to 0.79) between measured and calculated REE (all p<0.005).The mean calculated REE was significantly different from measured REE for all equations except Mifflin-St Jeor. A calculated REE within 10% of measured REE was considered acceptable and the equations of Mifflin-St Jeor, WHO and Harris-Benedict captured 64%, 50% and 45% of the participant, respectively. The Mifflin-St Jeor equation had the lowest root mean square error (138kcal), followed by the equation by Harris-Benedict (189kcal) and WHO (220kcal). The equations from Luhrmann, Henry and Cunningham predicted REE rather poorly in our study subjects, with e.g. <40% of the individuals within 10% of measured REE. Our results indicate that the Mifflin-St Jeor equation (using FFM) is the most accurate equation estimating REE in these octogenarian men. Harris-Benedict or WHO equations are potential alternatives if information on FFM is unavailable, although their accuracy on an individual level is limited.