To investigate energy requirements in healthy elderly subjects, we assessed the association of total energy expenditure (TEE) with resting metabolic rate (RMR), physical activity, body composition, and energy intake in 13 individuals (aged 56 to 78 years, six women and seven men). Free-living TEE was measured using doubly labeled water, RMR was measured by respiratory gas analysis, and energy expenditure of physical activity (EEPA) was derived from the difference between TEE and RMR, assuming the thermic response to feeding contributes 10% of TEE. Fat mass (FM) and fat-free mass (FFM) were obtained from underwater weighing, VO2max was determined from a bicycle test to exhaustion, energy intake was obtained from a 3-day food diary, and leisure time activity (LTA) was determined by structured interview. TEE was 2,406 +/- 438 kcal/d (range, 1,856 to 3,200 kcal/d, or 1.25 to 2.11 times RMR) and was related to VO2max (r = .79, P = .001), LTA (r = .74, P = .004), FFM (r = .69, P = .009), and FM (r = -.64, P = .018). The association between TEE and VO2max persisted after adjustment for FFM (partial r = .58, P = .036). EEPA was related to LTA (r = .83, P less than .0001) and FM (r = -.58, P = .039). Energy intake underestimated TEE by 31% +/- 18% in women and by 12% +/- 11% in men. Using stepwise regression, TEE was best predicted by VO2max and LTA (total adjusted r2 = .86). We conclude the following: (1) TEE varies greatly within healthy elderly subjects due to variations in physical activity; (2) VO2max has an important role in predicting energy requirements in older individuals; and (3) healthy older individuals underreport energy intake.