Energy expenditure, sex, and endogenous fuel availability in humans (original) (raw)
Subject characteristics. The characteristics of the subjects are provided in Table 1. The expected differences between men and women in body composition, body fat distribution, and REE were observed. Plasma palmitate concentrations were not different, and with the exception of the expected greater concentrations of growth hormone in women than in men (P < 0.005), plasma hormone and catecholamine concentrations were not different. The relationship between REE and FFM was assessed using multiple linear regression analysis using a sex and a sex × FFM interaction term. This relationship was not different between men and women (Figure 1; simple _r_2 = 0.69, P < 0.0001).
REE is plotted vs. FFM for the men and women participating in the study.
Clinical characteristics, body composition, and hormone concentrations of the men and women participating in this study
Day-to-day variability of indirect calorimetry and substrate measurements. The coefficient of variation (CV) for REE was 4% ± 2% and for the respiratory quotient (RQ) was 3% ± 1%. Perhaps thanks to the greater dietary control imposed by the study design, the CVs for average fasting plasma palmitate concentrations and flux were 16% ± 8% and 14% ± 8% (which is approximately one-half of what we have observed previously) (26). The CV for plasma glucose concentrations over the 4 days was 3% ± 1%, whereas the difference between glucose flux between the 2 study days for which it was measured averaged 13% ± 13%.
Relationship between FFA release and REE versus FFM. For the entire group, palmitate release rates were positively correlated with REE (r = 0.54, P < 0.0001), but were not significantly correlated with FFM (see inset of Figure 2; r = 0.25, P = 0.09). Because the multivariate regression analysis (see below) indicated a significant sex effect, palmitate release was plotted in relationship to REE separately for men and women (Figure 2) and sex-specific univariate regression analyses were performed. The relationship between palmitate kinetics and REE was stronger for each sex than for the combined group. Palmitate release rates relative to REE were significantly greater in women than in men, such that for a given level of energy expenditure, the average fuel availability from FFA was approximately 40% more in women. Note that the greater FFA release rates were not associated with higher plasma FFA (palmitate) concentrations in women (Table 1). Nor were these sex differences in FFA release associated with differences in resting fuel use; the RQ, which provides a measure of the proportion of fat and carbohydrate being oxidized, was the same in men and women (Table 1).
Resting postabsorptive palmitate release rate is plotted vs. REE for the men and women participating in the study. The inset depicts the relationship between palmitate release and FFM for these same volunteers.
The initial multivariate stepwise regression analysis results disclosed that REE, sex, and plasma epinephrine concentrations were significant predictors of average palmitate flux. The parameter estimates for the model are provided in Table 2. The _r_2 for this model was 0.57. Percentage of body fat and indices of central fat distribution (visceral fat area and WHR) were not included in the final model.
Parameter estimates for stepwise (stepping up) multivariate regression analysis
To ensure variables were not inappropriately left out of the model by the stepwise regression analysis approach we performed regression analysis using a nonstepping model that included REE, sex, FFM, fat mass, and our measures of adrenergic tone (plasma epinephrine and norepinephrine concentrations) as independent variables. The results were the same as for the stepwise model. The _r_2 for this model was 0.61. The parameter estimates and associated P values for this model are provided in Table 3. Consistent with the stepwise approach, REE, sex, and epinephrine significantly contributed to the ability to predict the interindividual differences in palmitate flux. Total fat mass, FFM, and plasma norepinephrine did not contribute significantly to this model. Although the parameter estimate for REE appears small, the REE values are expressed in the table per 100 kcal/day (range of values 1,190–2,238 kcal/day), creating a numerically small parameter estimate that was highly statistically significant. The parameter estimate for male sex was –31, consistent with the visual separation of the regression lines for men and women depicted in Figure 2. The parameter estimate for plasma epinephrine should be interpreted in the context of the need to log transform the plasma epinephrine concentrations because of their skewed distribution.
Parameter estimates for nonstepped multivariate regression analysis
Despite the good relationship between REE and palmitate release, there remained individual variation from the sex-specific group relationships (Figure 2). Because the multivariate regression analysis indicated that epinephrine contributed to the interindividual differences in palmitate release rates relative to REE, we examined this interaction in a different manner to allow a better visualization of the interactions. To do this the residual variance in palmitate release was determined for each participant. This value represents how each participant’s average palmitate release rate differed from their sex-specific group and is calculated by subtracting the predicted palmitate release (based on REE and sex-specific regression formulas) from the observed palmitate release for each subject.
Consistent with the multivariate regression analysis, the mean fasting plasma epinephrine concentrations were positively correlated (r = 0.33, P < 0.05) with residual palmitate release. Although plasma norepinephrine concentrations were not a significant contributor to the model, they were also positively correlated (r = 0.31, P < 0.05) with residual palmitate release. There was no significant correlation between plasma epinephrine and norepinephrine concentrations. Neither the mean fasting plasma insulin or growth hormone concentrations were correlated with the residual variance in palmitate release.
Although indices of body fatness were not significant predictors of palmitate flux in the regression models, we have found previously that upper-body obesity is associated with elevated FFA concentrations and flux (13). We therefore explored whether visceral fat area might be associated with the residual variance in palmitate flux. In this analysis the residual variance in palmitate release was correlated (r = 0.44, P = 0.03) with visceral fat area in men (Figure 3). Percentage of body fat and WHR tended (P = 0.06 and P = 0.07, respectively) to be positively correlated with residual variance in palmitate release in men. To our surprise, abdominal fat cell size did not significantly correlate (r = 0.24, P = 0.24) with residual palmitate flux in men. None of the body fat parameters (percentage of body fat, kilograms of body fat, WHR, visceral or abdominal subcutaneous fat area, or gluteal or abdominal subcutaneous fat cell size) correlated with residual palmitate release rates in women.
The relationship between visceral fat area and residual differences between sex-specific REE/palmitate release (flux) rate relationships for men and women are shown. In men visceral fat area showed a significant, positive relationship (r = 0.45, P < 0.05) with the residual palmitate flux values. In women visceral fat area was not correlated with the residual palmitate flux values; neither were percentage of body fat, WHR, or fat cell size.
If REE is important for determining what constitutes “normal” FFA release rates, we anticipated that plasma palmitate concentrations would vary proportionately with the residual variance in palmitate release relative to REE. Indeed, there was a strong, positive correlation between residual palmitate release and palmitate concentrations (Figure 4). Thus, individuals in whom palmitate release rates were greater than their sex-specific group relationship had higher plasma concentrations and vice versa.
The relationship between plasma palmitate concentration and residual differences between sex-specific REE/palmitate release rate relationships for all volunteers.
Glucose appearance rates. The sex difference in the relationship between fuel mobilization and REE was specific to FFA. Glucose release rates correlated directly with both FFM (r = 0.71, P < 0.00001) and REE (r = 0.62, P < 0.0001). Multivariate regression analysis failed to disclose a sex effect on the relationship between glucose release rates and FFM or REE (Figure 5). The relationship between glucose appearance and REE in women (r = 0.50) was glucose flux (mg/min) = 70 + [REE (kcal/day) × 0.034]; in men (r = 0.34) was glucose flux = 82 + [REE (kcal/day) × 0.039]. The 95% confidence intervals for both the intercept and the slope overlapped for men and women.
Resting postabsorptive glucose release rate is plotted vs. REE for the men and women participating in the study. The inset depicts the relationship between glucose release and FFM for these same volunteers.