Energy expenditure during overfeeding

When digested foods enter the bloodstream there is an oxidative hierarchy. The macronutrient that is most easily stored (fat) is oxidized last, while macronutrients that can not be stored at all (alcohol), or that can only be stored under certain circumstances (protein) or in limited amounts (carbohydrate) are oxidized first [10]. Alcohol ingestion directly increases alcohol oxidation, which is maintained until all alcohol is cleared. Protein and carbohydrate oxidation closely follow intake. In contrast, fat intake does not stimulate fat oxidation. Moreover, fat oxidation is inhibited by high intakes of the other macronutrients [10‐12]. The thermic effect of the separate macronutrients is 20 to 30% for protein, 5 to 15% for carbohydrate, and 0–3% for fat [12]. The figure for the thermic effect of alcohol is not clear, values range between 6 and 30% in different studies [13].

The intake of any macronutrient in excess of energy needs will lead to fat storage, but a reduced capacity for fat oxidation could particularly predispose to obesity. Diaz et al. [4] overfed subjects 50% above baseline energy requirements for 42 d, which suppressed fat oxidation by 37% in lean subjects, but by 64% in overweight subjects. These results were confirmed by Horton et al. [14] who overfed 50% above baseline energy intake with isoenergetic amounts of fat and carbohydrate for 14 d. During both overfeeding periods, obese subjects had a higher average RQ and oxidized proportionally more carbohydrate than lean subjects. However, EE increased proportionally with the increased body size and tissue gain leaving no evidence for adaptive thermogenesis. The capacity for fat oxidation, therefore, does not seem to relate to the capacity for adaptive thermogenesis.

DIT is increased on a high-protein, high-carbohydrate diet compared to a high-fat diet [11]. In contrast, low-protein diets result in increased DIT as well. This apparent contradiction is attributed to a mechanism for enriching nutrient-deficient diets while dissipating the excess energy on low-protein diets, whereas high-protein diets result in increased thermogenesis due to the high cost of metabolizing protein [15, 16]. In this context it is important to note that the term DIT is not only used for the increase in EE above BMR during the first hours after a meal, but also includes adaptive changes in BMR in response to the diet. Dulloo and Jacquet [16] reviewed the results of the normal- (15 energy%) and low- (3 energy%) protein overfeeding of Miller [17, 18] who overfed five young adults of normal body weight with 4.2 MJ/d or more for 3–6 weeks. The energy costs of weight gain on the low-protein diet (80 to >300 MJ/kg body weight) were much higher than on the high-protein diet (25–45 MJ/kg body weight) suggesting that low-protein overfeeding induces adaptive changes in EE. They concluded that the capacity for adaptive thermogenesis is individually determined, as the energy costs of weight gain on normal- and low-protein overfeeding were positively related. Therefore, Stock [15] and Dulloo and Jacquet [16] suggested low-protein overfeeding as a tool to discriminate between metabolically efficient and metabolically inefficient persons by maximizing differences in thermogenesis. However, we overfed healthy females 50% above baseline energy requirements for 14 d with a low-protein (7 energy%) diet (Table 1) and did not find adaptive changes in energy expenditure [19].

The limited storage capacity for carbohydrates forces an increase in carbohydrate oxidation with carbohydrate overfeeding, which together with a decrease in fat oxidation results in a positive fat balance [14]. However, the influence of the carbohydrate content of the (overfeeding) diet on metabolic efficiency is less clear. Though not always intentionally, overfeeding diets are generally high in carbohydrates. The effects of carbohydrates are thus only comparable between diets supplying the energy excess entirely as fat (or protein) or as carbohydrates, or respectively relatively low- and high-carbohydrate diets (Table 1; refs: [14, 20, 21]). Lammert et al [20] overfed subjects a high-fat (energy percentages from protein:fat:carbohydrate were 11:58:31) or a high-carbohydrate diet (en% P:F:CHO 11:11:78). Calculated from the mean overfeeding of 118 (high-CHO) and 101 MJ (high-F) and the mean weight gains of 1.35 (high-CHO) and 1.58 kg (high-F), the costs of weight gain were 87 and 63 MJ/kg respectively. In contrast, the cost of weight gain on a high-protein/high-fat diet (en% P:F:CHO 20:50:30) were 72 MJ/kg compared to ~47 MJ/kg on both an average (en% P:F:CHO 14:41:45) and a high-carbohydrate (en% P:F:CHO 10:30:60) in the study of Webb and Annis [21]. Results from the study of Horton et al [14] appear to point towards the same direction with costs of weight gain 100 MJ/kg on high-fat and 90 MJ/kg on high-carbohydrate overfeeding. While the first study suggests that costs of weight gain are increased with high-carbohydrate overfeeding which might be caused by de novo lipogenesis, the last two studies suggest that costs of weight gain are rather increased when the carbohydrate content is relatively low which could be explained by increased gluconeogenesis. However, it should be noted that comparison between studies is difficult as macronutrient composition and measurement techniques differed substantial. This is also shown in the large range in costs of weight gain of 23 to 54 MJ/kg with 'average/mixed diet' overfeeding.

The component of daily energy expenditure most affected by changes in body weight is the BMR [1], any adaptive changes in total energy expenditure are therefore likely to appear in this component. Several studies reported an increased BMR after overfeeding [4, 19, 22‐29]. This increase is due to the energy cost of fat and fat-free mass gains as well as the costs of maintaining a larger body weight [1].

Another component, DIT, will increase due to the increased amount of food that has to be digested and absorbed. Yet, several studies did not find a significant increase in DIT, independent of dietary composition and duration of the experiment [22, 28‐30]. Others could explain significant increases in DIT solely by the increased amount of EI, as reflected by the percentage of the EI found in the DIT component being similar before and after overfeeding [27] or the response to a fixed meal being unaltered [25]. Pasquet et al. [26] reported a similar increase in DIT with long-term high-carbohydrate overfeeding compared to overfeeding with a typical western, mixed diet [22, 28‐30], but concluded that this increase included an adaptive component as the increase was even larger after adjusting for a reduction in physical activity.

The last component, AEE, is the most variable component of TEE between persons [31], and thus is most likely the main contributor to variation in weight gain during overfeeding. Indeed, several overfeeding experiments show that those subjects with the largest increase or decrease in AEE have respectively the lowest and highest weight gains [4, 25]. But relatively large changes in AEE (as percentage of TEE) above increased costs of performing physical activity due to an increased body weight, might reflect behavioral changes rather than adaptive thermogenesis.

It should be noted that the division of energy expenditure into its components may induce over- or underestimations of the separate components. AEE is particularly hard to determine, as measurement errors in TEE, BMR and DIT are accumulated in AEE [2]. SMR might be confounded by DIT; the influence of a large evening meal has been shown to continue well into the night [32], which might confound measurement of BMR in the morning as well [4, 19]. In addition, there is an interaction between DIT and physical activity both at high and low levels of activity [33, 34], which will not only affect DIT but will also influence determination of the energy costs of physical activity [4].