Sugar as a slimming agent?

Does dietary composition in itself have any importance for weight gain and obesity? Meta-analyses of dietary intervention trials comparing ad libitum normal-fat diets with low-fat diets clearly demonstrate that a reduction in the dietary fat content decreases body fat stores (Bray & Popkin, 1998; Yu-Poth et al. 1999; Astrup et al. 2000). Energy from fat has a weaker satiating power than energy from carbohydrates, and individuals are unconsciously more likely to consume more energy from fat-rich diets than from carbohydrate-rich diets. In addition, there also seems to be important differences in the digestion and metabolism of fat and carbohydrates which may influence energy balance. In order to gain a better understanding of the effects of fat and carbohydrate on energy metabolism, it is necessary to eliminate the effect of appetite regulation and induce overeating in paid experimental subjects. This is a method used by Lammert et al. (2000), who studied the response to overfeeding for 21 d with 5 MJ/d of either a fat-rich or an extremely-carbohydrateand sugar-rich diet. Previous overfeeding studies using extreme carbohydrate-rich lowfat diets have shown that the conversion of the glucose to fat by the de novo lipogenesis does not occur before most of the oxidation of the body is covered by glucose, and the glycogen stores are filled. The conversion of glucose to fat is however energetically a very costly synthesis, and based on the stoichiometry it can be predicted that overfeeding with carbohydrate should result in a 21 % lower fat deposition than overfeeding with an isoenergetic amount of fat (Flatt, 1992). Consequently, it should be less fattening to overeat carbohydrates than fat. This is exactly what Lammert et al. (2000) have tested. However, they do not think that they have revealed any differences between fat and carbohydrate overfeeding, but they overlook the main findings: overfeeding by carbohydrate compared with fat showed a mean net conversion of carbohydrate to fat of 15 ́8 g/d in contrast to 0 g/d on fat overfeeding. This de novo lipogenesis was estimated to account for 40 % of the increase in fat mass about 332 g fat. In addition they found a 30 % higher faecal energy loss equivalent to 8 MJ during the 21 d. These increased energy outputs should be expected to result in a lower fat deposition on the carbohydrate overfeeding and actually they do find a 30 % lower increase in fat mass despite an 18 % higher energy intake during the carbohydrate-overfeeding regimen. The subjects should therefore eat 68 % more energy in order to increase body fat stores by 1 kg on carbohydrate overfeeding than on fat overfeeding (155 MJ/kg v. 42 MJ/kg). This difference was not significant, but the study does not possess the sufficient statistical power to demonstrate a difference of such magnitude. As can also be seen from the study, overfeeding of non-related individuals results in substantial differences in body-fat gain, which can be attributed to genetic variation in the ability (to increase energy expenditure; Bouchard et al. 1990; Levine et al. 1999). When the study was designed, a power calculation would have shown, that a statistically significant difference of 20 % in fat gain would have required at least twice as many subjects. Alternatively a cross-over design or identical twins could have been used. The authors do not comment on this point and one may assume that the lack of significance of the 68 % higher energy cost of fat deposition on the carbohydrate overfeeding is due to a type 2 error. Is it then correct when Lammert et al. (2000) conclude that they do not find any evidence for increased thermogenesis during carbohydrate overfeeding? No, an assessment of the energy balance on the two overfeeding regimens speaks for itself (Fig. 1). The extra energy intake is 5 ́6 MJ/d over 21 d, i.e. 118 MJ, of which 8 ́6 % is lost as faecal energy. Gain of fat mass and fat-free mass can at best explain 34 MJ/kg for 1 ́36 kg, i.e. 46 MJ (Forbes et al. 1986). The additional energetic cost of de novo lipogenesis of 332 g fat can explain an additional 4 MJ, but where is the remaining 118 2 …9‡ 46‡ 4† MJ, i.e. 59 MJ or 50 % of the energy intake during the overfeeding? The authors overlook the possibility of increased thermogenesis (luxury consumption) during day and evening time, when the experimental subjects were awake (Levine et al. 1999). Unfortunately, energy expenditure was only measured at night time. Two alternatives should also be considered: