Challenges of indirect calorimetry in mice.

TO THE EDITOR: the recent paper by Longo et al. (5) provides an excellent description of the challenges of performing indirect calorimetry in mice. Mice that gain weight or are weight stable in their home cages often lose weight during the indirect calorimetry measurements, indicating that the procedure alters the usual behavior of the animals. To tackle this difficulty, the authors proposed a simple statistical adjustment of their data based on the measured body weight changes. While we applaud the spirit of this attempt to address a typically ignored difficulty with indirect calorimetry in mice, we have several questions and comments about their results and interpretation. The reported unadjusted energy intake and expenditure values cannot explain the observed weight changes. For example, the mice that were provided the high-fat diet for the first time during indirect calorimetry (i.e., the STDIO group) gained 2.57 g over 4 days and had a measured 2.11 kcal/day positive energy balance (see Table 2 in Ref. 5). Assuming that the weight gain was primarily due to increased fat mass (2), the measured energy imbalance would lead to 2.11 kcal/day 4 days 9.4 kcal/g 1 g of weight gain. Alternatively, the energy density of the deposited tissue would have to have been 2.11 kcal/day 4 days 2.57 g 3.3 kcal/g, implying that 80% of the mass change was due to fat-free mass (4), which seems highly unlikely. Incommensurate measurements of energy imbalance and weight change are widespread in this field, and we implore practitioners of indirect calorimetry to address this issue. Our most important criticism is that the proposed statistical adjustment of the data to correct for the observed changes of body weight was incomplete. If the adjusted data had completely corrected for the observed weight changes, as intended, then the adjusted data should demonstrate a state of energy balance. However, the reported adjusted energy intake and energy expenditure values are not equal (see Table 2 in Ref. 5) in violation of the energy balance principle. This demonstrates that the statistical adjustment was incomplete. Finally, since the respiratory quotient (RQ) equals the food quotient (FQ) in a state of energy and macronutrient balance (1), then RQ is independent of energy intake when body weight is stable. This fundamental constraint contradicts the paper’s title and main conclusion, suggesting that RQ measurements appropriately adjusted for weight changes can predict energy intake. The observed correlations in the residuals of the adjusted RQ and adjusted energy intake (see Figure 3D in Ref. 5) is likely due to the fact that the statistical adjustment of the data is incomplete, as suggested above. The more important observation is that the state of energy imbalance influences the relationship between unadjusted RQ and FQ, such that RQ increases with increasing energy intake (Figure 3A in Ref. 5), a fact that has been known for more than 20 years (1). As an alternative to adjusting indirect calorimetry data to correct for altered mouse behavior during the procedure, we recently described a novel mathematical method for estimating long-term dynamics of energy expenditure and RQ in mice (2). Our method is based on the law of energy conservation and uses the measured time course of body weight, body composition, and food intake to calculate the underlying energy balance and fuel selection dynamics over many weeks, while mice are housed in their home cages. Mathematical models that predict mouse energy expenditure, RQ, and weight changes have also been developed (3), and we believe that such methods will become increasingly important as the challenges of indirect calorimetry in mice are more widely recognized.