Energy Expenditure Responses to Fasting and Overfeeding Identify Phenotypes Associated With Weight Change

Abstract

Because it is unknown whether 24-h energy expenditure (EE) responses to dietary extremes will identify phenotypes associated with weight regulation, the aim of this study was to determine whether such responses to fasting or overfeeding are associated with future weight change. The 24-h EE during energy balance, fasting, and four different overfeeding diets with 200% energy requirements was measured in a metabolic chamber in 37 subjects with normal glucose regulation while they resided on our clinical research unit. Diets were given for 24 h each and included the following: (1) low protein (3%), (2) standard (50% carbohydrate, 20% protein), (3) high fat (60%), and (4) high carbohydrate (75%). Participants returned for follow-up 6 months after the initial measures. The decrease in 24-h EE during fasting and the increase with overfeeding were correlated. A larger reduction in EE during fasting, a smaller EE response to low-protein overfeeding, and a larger response to high-carbohydrate overfeeding all correlated with weight gain. The association of the fasting EE response with weight change was not independent from that of low protein in a multivariate model. We identified the following two independent propensities associated with weight gain: a predilection for conserving energy during caloric and protein deprivation and a profligate response to large amounts of carbohydrates.

Figure 1

Flow diagram of participant progress through the study.

Figure 2

Study diagram of the clinical study.

Figure 3

Macronutrient composition of the dietary interventions (A) and related 24-h EE response (B). Protein, carbohydrate, and fat content of the diets are expressed in grams based on a representative diet for an individual requiring 2,000 kcal for EB and 4,000 kcal for overfeeding (A). The 24-h EE response to each dietary intervention is expressed as the percentage change compared with the 24-h EE measured during EB (B). Error bars represent the mean with SD.

Figure 4

Inverse relationships between the 24-h EE response to FST and the average change in 24-h EE during overfeeding (A) and during low-protein overfeeding (B). The 24-h EE response to FST and to overfeeding is expressed as the percentage change compared with the 24-h EE measured during EB. The average change in 24-h EE during overfeeding was calculated as the mean value across the four overfeeding diets. The best-fit line is displayed in both panels. Vertical and horizontal lines indicate points with no change in 24-h EE compared with EB.

Figure 5

Associations between body weight change after 6 months and the 24-h EE responses to overfeeding and FST. Inverse associations between weight change at 6 months after discharge from the CRU and the change in 24-h EE with FST (A) and during low-protein overfeeding (B). C: Positive relationship between the increase of 24-h EE with high-carbohydrate overfeeding and weight change (two high-carbohydrate diets were excluded as <95% of the food was consumed). D: Inverse relationship between RQ during 24 h of FST and weight change. The mean follow-up time was 6.5 ± 0.9 months with a weight change of 1.2 ± 4.2 kg (range −6.1 to 11.2 kg). No point met the statistical criteria to be an outlier. All associations were still significant (P < 0.05) when excluding the subjects with the greatest weight change. The results for weight change, expressed as a percentage of the baseline weight, are as follows: %EE response to FST (r = −0.36, P = 0.03), low-protein overfeeding (r = −0.51, P = 0.007), high-carbohydrate overfeeding (r = 0.34, P = 0.05), and RQ during FST (r = −0.44, P = 0.006).

Figure 6

Phenotypes of 6-month weight change based on the 24-h EE responses to low-protein and high-carbohydrate overfeeding. Subjects were categorized into four subgroups according to the median %EE during low-protein and high-carbohydrate, normal-protein overfeeding (two high-carbohydrate diets were excluded as <95% of the food was consumed). Subjects with a lower-than-median EE response during low-protein overfeeding and a higher-than-median response during high-carbohydrate, normal-protein overfeeding gained more weight compared with those with the opposing EE responses (mean difference 7% increase in their baseline weight, P = 0.007). The mean follow-up time was 6.5 ± 0.9 months, with a weight change of 1.2 ± 4.2 kg (range −6.1 to 11.2 kg).