This study examined the effect of training on physical activity and substrate utilization in the elderly. Before the start, in week 6 and week 12 (T0, T6 and T12) data on physical fitness, physical activity and substrate utilization were collected in the exercise (11 males, 11 females; 63 +/- 8 yrs) and control group (6 males, 5 females; 59 +/- 4yrs). Physical activity was registered with a tri-axial accelerometer and substrate utilization was calculated from resting respiratory exchange ratio (RER) by indirect calorimetry. At T6 physical activity on training days was significantly higher than on non-training days (33.4 +/- 10.3 vs. 26.5 +/- 7.8 counts min(-1); p <0.001). At T12, after adjusting for training activity, physical activity on training days was significantly lower than on non-training days (23.7 +/- 8.4 vs. 28.2 +/- 9.3 counts min(-1); p <0.01). RER decreased significantly (0.02 +/- 0.03; p <0.05), indicating a relatively larger fat oxidation. Changes in RER were negatively correlated with pre-training RER. In conclusion, in elderly an increase in structured training (exercise) is compensated for by a corresponding decrease in non-training physical activity. Training increased relative fat utilization in elderly with a high pre-training RER, whereas elderly with a low pre-training RER decreased their relative fat utilization.