The Journal of Nutrition Vol. 128 No. 12 December 1998, pp. 2374-2382

Energy Available from Corn Oil Is Not Different than that from Beef Tallow in High- or Low-Fiber Diets Fed to Humans^1,2

William V. Rumpler³, David J. Baer, and Donna G. Rhodes

Diet and Human Performance Laboratory, Beltsville Human Nutrition Research Center, Agricultural Research Service, United States Department of Agriculture, Beltsville, MD 20705 USA

	ABSTRACT

Abstract Introduction Methods Results Discussion References

The impact of the source of dietary fat and the level of dietary fiber on digestibility and energy metabolism were studied in human (six male, six female) volunteers. Subjects were divided into two diet treatment groups, high fiber [29.0 g total dietary fiber (TDF)/d] and low fiber (18.6 g TDF/d), for the duration of the study. Each participated in three, 2-wk controlled feeding periods. Either beef tallow (BT), corn oil (CO) or carbohydrate (CHO) was added (25% of diet energy) to a base diet in a three-way crossover study. Energy expenditure, substrate oxidation and digestibility determinations were conducted at the end of each period. The protein, fat and CHO digestibility of the base diet was significantly different between the fiber levels. The digestibility (high-fiber/low-fiber) averaged 82%, 90% for protein, 96% and 98% for fat. After adjusting for TDF, the CHO digestibility averaged 96% and was not different between fiber levels. The digestibility of the added CO and BT was 99.6 and 99.8% respectively, and was not significantly different between the fiber levels. No significant differences in 24-h energy expenditure existed nor the thermic effect of food due either to fiber level or between the CHO, BT or CO. Fat oxidation in subjects consuming the low-fiber diet was 14% higher (P < 0.03) with the BT treatment than with the CO treatment but not different in those that consumed the high-fiber diet. The energy value of the two fat sources was not different but their utilization by individuals near energy balance may lead to differences in long-term weight maintenance.

	INTRODUCTION

Abstract Introduction Methods Results Discussion References

For more than a decade, recommendations for healthy eating have included decreasing the intake of fat from animal products and increasing dietary fiber intake. Isocaloric replacement of fat from animal products with fat from vegetable sources markedly increases the intake of mono- and polyunsaturated fatty acids. Outside of the purported benefit on cardiovascular disease risk factors, several reports suggest that isocaloric substitution of vegetable oils for animal fat may have an impact on the maintenance of a healthy weight. Mercer and Trayhurn (1987) reported lower weight gain in rats fed corn oil (CO)⁴-supplemented diets than when fed beef tallow (BT)-supplemented diets. Shimormura et al. (1990) found a similar response when comparing safflower oil and BT in the diets of rats. These observed differences in energy value might be explained by differences in digestibility of the dietary fat or effects of the different fat sources on energy expenditure (EE) or substrate oxidation.

The studies by Mercer and Trayhurn (1987) and Shimomura et al. (1990) used growth as the only end point. From these data, it is impossible to determine if the changes were due to difference in digestibility or energy metabolism. However, subsequent reports from Takeuchi et al. (1995) and Matsuo et al. (1995) observed an increase in EE following a meal with the substitution of safflower or linseed oil for either BT or lard. In addition, Jones et al. (1985) and Layton et al. (1987) showed that poly- and mono-unsaturated fats are oxidized more quickly than saturated fats. Because vegetable oils have a much higher level of unsaturated fatty acids, they may be oxidized more readily, thereby sparing other energy sources.

Several authors reported differences in digestibility between fat from animal or vegetable sources and an interaction with other dietary components. Digestibility in rats is reported to be lower for fats high in stearic acid, such as BT, than for fish or vegetable oils (Dougherty et al. 1995, Kaplan and Greenwood 1998, Monsma et al. 1996) In addition, some reports (Denke et al. 1993, Fang and Kies 1993, Hanson et al. 1993, Schrijver et al. 1991) suggest that dietary fiber and dietary fortification with calcium depress digestibility more in BT-supplemented diets than with vegetable- or fish oil-supplemented diets.

The lower digestibility of BT reported in these data seems to contradict the growth data since the growth response was better on BT than on the vegetable oils, which suggests that fat from corn or vegetable oils may have some effect on energy metabolism beyond the differences in digestibility compared to BT. In this study, we determine the digestibility of diets with additions of BT, CO, or sugar and/or sweetened beverage and measure the substrate oxidation and EE of subjects consuming these diets. In addition, total dietary fiber (TDF) content is manipulated to determine if an interaction exists with the two types of fat. Based on these data, we discuss the impact of dietary fiber on fat digestion and the possible implications of different sources of fat on substrate use and EE.

	MATERIALS AND METHODS

Abstract Introduction Methods Results Discussion References

Experimental design and statistical analysis.  Six adult male and six adult female human volunteers were randomly, within gender, assigned to either a high- or low-fiber diet for the duration of the study. Each subject participated in three controlled feeding periods separated by at least 2 wk of unrestricted eating. During each of the periods, subjects consumed the same food each day for 12 to 14 d. Carbohydrate (CHO) or two different sources of diet fat were incorporated into the diet as described below. Two principal measurements were made during each of the periods on each subject: a 5-d total collection of food, feces and urine, and a 24-h calorimeter measurement.

Statistical analysis was performed with the SAS statistical package (SAS Institute, Cary, NC). A mixed model analysis of variance was used to examine treatment and diet differences. Interaction terms for gender, type of supplement and diet were included in the model. Since the design was balanced with no missing, cells variance estimates used in the analysis were essentially separate variance estimates for comparisons within diet and a mean variance for comparisons across diet. The mixed model analysis of variance was also used to develop empirical relationships between the intake and composition of the various dietary components and their appearance in the feces along with digestibility values. Slope estimates for each diet and gender vs. type of supplement were tested for significant differences. No significant interactions of gender vs. type of supplement or diet vs. type of supplement were significant for the slope of the various relationships examined. Thus, the slopes reported represent values across gender and diet.

Diets and metabolizable energy determinations.  Two 1-d menus were formulated to provide a basis for the substitution of different fat sources or CHO. The diets were formulated based on USDA Handbook 8 values [United States Department of Agriculture (USDA, 1976-1989] by the Beltsville Human Nutrition Research Center Human Study Facility (HSF) dietitians to meet the RDA. The two base menus were formulated to have different TDF levels and an energy content distribution of 15% from protein, 20% from fat and the balance from CHO. Substituting high-fiber items for low-fiber items in the menus (Appendix 1) attained different fiber levels for the two base diets, similar with respect to food items with substitutions principally made in the fruit, vegetables, bread and cereal items. At least 72% of the fat and 60% of the protein in the two base diets were composed of food items common to both diets. Primarily the bread and crackers contributed the remaining 28% of the fat and 40% of the protein in the substituted items. The majority of the fiber in both diets came from the substituted items. Items common to both diets represented only 10.6% of the TDF in the high-fiber diet and 29% in the low-fiber diet.

The three diets consumed, within each fiber group, varied only in the addition of either CHO or two sources of dietary fat, which were calculated to increase energy intake by 25% beyond the base diet. Example menus are presented in Appendix 1 for the 5.6 MJ/d energy level. Added fat (+CO or +BT) was incorporated into foods, dressings and baked goods, as indicated by the asterisk in Appendix 1. The base diets with added fat from the two sources represented two of the three treatment groups. The third treatment (+CHO) was accomplished with the addition of an equivalent amount of energy from a combination of sugar or noncaffeine-containing soft drinks. Each individual was allowed to choose the makeup of the CHO supplement with some subjects choosing all soft drinks and some choosing all cane sugar packets. Thus, the CHO supplement ranged in composition from mostly high-fructose corn syrup to mostly disaccharide of glucose.

The subjects consumed breakfast and dinner in the dining room of the HSF under the supervision of a registered dietitian or dietary technician. Lunches and snacks were packed for take-out from the HSF. Weekend meals were packed into ice-filled coolers for consumption away from the HSF. Daily beverage (noncaloric: i.e., coffee, tea) and medication records were maintained throughout the diet study. Foods and calorie-containing beverages were limited to those provided by the study. Daily menus were formulated in one MJ increments from 4.6 to 12.5 MJ/d. Energy intake level, as formulated, was set for each individual to be equal to the EE determined during a 24-h calorimeter measurement made prior to the start of the study. Energy intake, as formulated, was the same for all three periods. We reported (Seale et al. 1990) that calorimeter measurements, on average, underestimate the free-living EE of individuals due to the sedentary nature of a calorimeter measurement. However, the slightly negative energy balance of the subjects during the 2-wk period was necessary since we wished the subjects to be near energy balance in the calorimeter. In addition, we wanted the balance (digestibility experiment) determinations conducted on the same intake level as in the calorimeter. Subjects were weighed weekly during the three periods as an index of their energy balance status. Due to the short period of study, these weights were only intended to detect gross differences between expected energy requirements and actual expenditure.

On d 7 of each diet period, subjects consumed a marker capsule containing brilliant blue dye with their evening meal and were instructed to collect all feces and urine voided over the next 5 d. At the end of the collection period, they consumed another dose of brilliant blue. During the collection period, total urine was collected, subsampled (10% of daily production) and pooled. Total fecal output was collected, frozen, pooled and homogenized with water and ice in a Waring blender prior to freeze-drying. Duplicate meals, during each period, were prepared for chemical analyses by homogenizing the food in a Waring blender with ice and water before freeze-drying. Diets, feces and urine were analyzed for combustible energy by adiabatic bomb calorimetry (Parr Instrument Company, Moline, IL) and nitrogen by combustion (Leco Corp., St. Joseph, MI). Diets and feces also were analyzed for fat (methylene chloride extraction; CEM Corp., Matthew, NC) and ash (muffle furnace). Diets were analyzed for TDF following the Prosky et al. (1985) method. All composition data are expressed on a dry matter basis.

Calorimetry and substrate oxidation.  Twenty-four-hour EE was determined by indirect calorimetry using the Beltsville room calorimeter (Seale et al. 1991) and was conducted during the 8th d of each period for three measurements on each subject. EE and substrate (CHO, protein and fat) oxidation rates are calculated using the equations described by Livesey and Elia (1988), which use the daily production of CO₂, consumption of O₂, the excretion of total urinary nitrogen. We assumed a proportional distribution of 90:5:5 for urine nitrogen as urea, ammonia and creatinine, respectively (Livesey and Elia 1988). The average values for heat equivalent of oxygen (kJ/L O₂) and the energy density (kJ/g) and respiratory quotient (L CO₂/L O₂), respectively, for fat (19.61, 39.5, 0.71), protein (19.48, 19.68, 0.83) and CHO (21.12, 17.5, 1.0) used for the computations reported in this paper are the same as reported by Livesey and Elia (1988). However, we made an adjustment in the factors for the high simple-sugar content of the added CHO (21.03, 1.0 and 16.85).

Each subject became familiarized with the calorimeter during at least one 24-h stay in the chamber prior to the measurements made during this experiment. Height, weight and body composition by bioelectric impedance (Kushner and Schoeller, 1986) were determined in the morning following a 12-h fast and immediately before the subjects entered the calorimeter at 0800 h and remained for 23.5 h, where initial and final gas composition of the air in the chamber was recorded. Data on gas composition and airflow were collected for the entire time the subject was in the chamber. The 23.5-h measured EE was corrected for differences in initial and final gas composition in the chamber and excretion of urinary nitrogen, and this value was adjusted to 24-h. Multiple alcohol combustion measurements were used to estimate the recovery of CO₂ and O₂ that averaged 101 and 99%, respectively. CO₂ production was also corrected for the CO₂ content of the carbonated beverages (2 L CO₂/405-mL can) consumed in the calorimeter. While in the calorimeter, each subject followed the same activity protocol (Rumpler et al. 1990), which included 4 h of desk work, 1.5 h for meal consumption, 1 h of exercise, 7.5 h for sleeping and the balance of the time spent as the subject desired (i.e., watching television, listening to radio, reading, desk work). The 1 h of exercise was divided into two 30-min sessions on an exercise bicycle. Sleep energy expenditure (EE_sleep) was calculated by multiplication of the hourly rate during sleep by 24. Doppler or infrared motion detectors mounted in each of the chambers monitored total movement in the chamber. Sedentary energy expenditure (EE_sed) and the thermic effect of food (TEF) were calculated as described by Ravussin et al. (1986). EE_sed was calculated as the intercept of the regression of EE and number of counts from the activity monitors in the chamber and was used to estimate the TEF in Equation 1:

Subject characteristics.  Twelve healthy male and female human subjects (6 female, 6 male) were selected from the general population. Individuals between the ages of 24 and 57 y were recruited from whom measurements of height, weight, and lean body mass were taken during the initial screening. Physical characteristics of the subjects are presented in Table 1. A physician conducted physical evaluations on all subjects prior to selection and reviewed and monitored all procedures for possible medical implications and health hazards. All subjects, normotensivenon-diabetic and with normal blood chemistry values, were informed of the general purpose of the study, all procedures were explained and written informed consent declarations were obtained. The Institutional Review Board of Johns Hopkins University approved all procedures.

	RESULTS

Abstract Introduction Methods Results Discussion References

Body weight.  Body weight decreased in all subjects, reflecting the negative energy balance that was a natural consequence of the assignment of energy intake as discussed above. Average weight loss was not different between the high- and low-fiber groups and averaged 1.19 and 1.37 kg, respectively. Average weight loss by treatment was 0.9, 1.34 and 1.6 kg (treatment standard error 0.21) when subjects consumed the CHO, BT and CO, respectively. Weight loss was significantly less when subjects consumed the CHO treatment than either of the fat treatments. The differences in weight loss between the two fat treatments were not significant.

Diet composition and digestibility.  Metabolizable energy value of each diet was calculated as the difference between intake energy and energy excreted in feces and urine. All calculations were based on 100% dry samples. Total energy, protein and fat digestibility were calculated as the difference between nutrient intake and fecal loss, expressed as a percentage of nutrient intake. Total CHO was calculated in Equation 2:

Subjects consuming the +CHO diet did not have significantly higher levels of fecal CHO than those fed the additional fat. Therefore, the digestibility of the supplements was determined by assuming that the entire CHO supplement, added to each of the basal diets, was digested, which seemed appropriate since the composition of the CHO supplement consisted essentially of highly soluble simple sugars. We then computed the difference between fecal output of fat, protein and CHO for each individual while consuming the CHO-supplemented diet (high- and low-fiber) and the fat-substituted diets. An example of the computation for the digestibility of the fat supplement is in Equation 3:

A similar computation was used to draw conclusions regarding the effect of fat supplements on the other components. Differences between the CHO- and fat-supplemented diets were attributed to the effect of fat supplementation. In this paper, all references to digestibility actually refer to apparent digestibility since there was no adjustment for fecal materials from endogenous sources.

Table 2 contains the average diet composition, as determined by the chemical analysis of the basal diets with the added CHO, BT or CO. Protein intake was slightly higher (P < 0.01) in subjects that consumed the low-fiber diet than in those that consumed the high-fiber diet, but as designed, was not different among the three treatment groups within diet.

Fat and CHO intake reflected the experimental design with fat intake higher on the fat treatments and CHO intake higher on the CHO treatment. The additional fat intake ranged from 42 to 85 g/d and represented ~26% of the total energy consumed. The high-fiber diet contained 56% more TDF than the low-fiber diet and intake ranged from 16.4 to 32.3 g/d. Measured total energy content of the diet was lower than values estimated from Handbook 8 (USDA, 1976-1989) by ~10% in subjects that consumed the low-fiber diet, and by 15% in those that consumed the high-fiber diet. This difference was a consequence of three factors: i) A number of papers reported an overestimation of the energy value of diet from Handbook 8 (Livesey 1990, Miles et al. 1992) by as much as 8%, particularly in high-fiber diet. This difference probably accounts for much of the discrepancy between our calculated and measured values. ii) The lower energy value of our substitution of some of the low-fiber items with high-fiber items was not sufficiently taken into account and resulted in a greater discrepancy in the high-fiber diet. iii) The base diet was lower in fat content than we anticipated, based on our calculations, and thereby had a lower energy density.

Table 3 contains the digestibility and the contribution of protein, CHO and fat to total digestible energy. The digestible energy content of the high-fiber diets was lower than the low-fiber diets. This difference reflects the lower digestibility of both protein and fat in the base diet. The lower digestibility of all measured components of high-fiber diet compared to the low-fiber diet was significant (P < 0.01). Digestibility was decreased an average of 4.5% for energy, 6% for fat, 9.8% for protein and 2.1% for CHO.

The digestibility of the added fat was nearly 100%, was significantly greater than the digestibility of the fat in either the high- or low-fiber base diet and was not significantly different between the two types of fat and the two diets.

The digestibility of the CHO in the diets with the additional fat was lower than in the diet with the added CHO. This difference reflected the highly digestible nature of the added CHO. The differences in CHO digestibility between the +CHO diet and the added fat diets disappear when the digestibility of the CHO in the +CHO diets is adjusted for the added CHO. Fecal CHO averaged 7.8 g/d for the low-fiber diet and 17.8 g/d for the high-fiber diet. There were no differences between the diets supplemented with CHO and the fats or between the two fat sources.

Calorimeter measurements.  Data collected for all subjects during the 23.5 h, adjusted to 24 h, in the calorimeter during each period are presented in Table 4. Metabolizable energy intake was higher on the low-fiber diet than on the high-fiber diet. The differences between treatments, within diets, were small and insignificant. There were no differences in any of the EE parameters measured. EE_sleep accounted for ~60% of the 24-h EE. EE_sed averaged ~12% above sleep rate EE. This difference represents the TEF that ranged from 4 to 14% of the ME intake. As a percentage of ME intake, TEF was greater in subjects who consumed the high-fiber diet than in those who consumed the low-fiber diet. However, on an absolute basis no differences existed between the two diet groups. In addition, no differences existed in TEF between the treatments within diet. All measurements in this study were made near or below maintenance as reflected in the negative energy balance values (Table 4). The intake, oxidation and balance of substrates are presented in Table 5. The substrate oxidation reflected changes in diet composition. The net balance of fat was consistently negative, and CHO and protein were positive on the low-fat diet and negative on the high-fat diet. However, the values were generally small relative to total intake or oxidation, reflecting the near energy balance on the low-fiber diet and ~90% of energy balance on the high-fiber diet. Therefore, these estimates of TEF represent predominantly the energy cost associated with complete oxidation of the absorbed substrates with a small contribution, 0-10% of the total, coming from body stores. Significant differences existed in the substrates oxidized as a result of the supplement consumed. As expected, CHO oxidation was higher when subjects consumed the CHO supplement and fat oxidation when the fat supplements were consumed. No differences existed for protein oxidized, which reflected no differences in protein intake between the three supplements. Significant differences existed for fat and CHO oxidized when comparing the two fat supplements when the low-fiber diet was consumed. Fat oxidation was higher (P < 0.01) and CHO oxidation tended (P < 0.06) to be lower when BT was consumed compared to CO.

	DISCUSSION

Abstract Introduction Methods Results Discussion References

Digestibility.  Neither the addition of the two different fats nor the addition of the CHO to the base diets had any measurable impact on the digestibility of the base diet fat, protein or CHO. A profound effect resulted from the substitution of the high-fiber items for the low-fiber items in the base diet on the digestibility of the fat, protein and CHO: it decreased the digestibility of the original items, particularly the base diet fat. The majority of the fat in both base diets was from the same, unsubstituted fat sources. In addition, the substituted items containing the additional fat were principally baked goods in which the fat would have been added during the baking process and would likely be similar to the original items. Thus, it is likely that the dietary fiber from one food item is making the dietary fat from another food item unavailable for absorption.

Protein digestibility was also affected by the substitution of high-fiber items for low-fiber items. However, nearly 40% of the protein in the diet was contained in the substituted items. In addition, the substituted protein would more likely be from somewhat different sources, unlike the above comparison of the fat sources. For example, the protein in the bran flakes would likely have a different digestibility than corn flakes. Thus, it is more difficult to draw conclusions regarding the fiber effect on protein digestibility. With this qualifier in mind, these data support the observations by other authors (Baer et al. 1997, Calloway and Kretsch 1978, Farrell et al. 1978, Livesey 1990, Miles et al. 1988, Southgate and Durnin 1970) that the protein digestibility is lower in high-fiber diets than low-fiber diets.

An interesting aspect of the data from this study is the strong correlation of dietary fiber to the digestibility of all components of the diet. The observations in this study are consistent with our previously published study (Baer et al. 1997), which demonstrated lower protein and fat digestibility with high-fiber diets. Fiber intake was found by least squares regression to be a strong predictor of fecal levels of fat, protein and CHO, which suggests that the amount of digestible fat, protein and CHO in a diet can be predicted by knowing the amount of each component consumed and correcting for TDF: for example, the digestibility of the fat contained in the base diets was substantially lower than the added fat. However, we can adjust the fat digestibility based on the least squares relationship of digestibility to intake and fiber level (i.e., fat digested (g) = 0.998 · fat intake- 0.137 · TDF intake) to 0 g of TDF intake. The digestibility of the base diet fat is 99.8% after adjustment, which is essentially the same as the added fat. Another approach is to adjust the data of Miles et al. (1992) to a zero TDF intake and the fat digestibility is 98% for the high-fat low-fiber diet and 100% for the low-fat high-fiber diet. These calculated values are consistent with those observed in the current study. Livesey (1991) proposed a similar approach for the total DE content of the diet and determined that digestible energy content could be estimated by a simple empirical approach in which energy intake is adjusted by a factor times the unavailable CHO intake. Livesey's equation (DEI = 0.96-9 · UC) assumes an overall digestibility of 96% in human diets, which is then adjusted down by 9 kJ/g of unavailable CHO. We reported in an earlier study (Baer et al. 1997) that TDF could be used as an estimate of unavailable CHO in the Livesey equation. Application of the Livesey equation to the current study, using TDF as an estimate of unavailable CHO, gives DE values not different from the measured values.

The TDF itself would be considered mainly part of the CHO fraction and represents the majority of the indigestible CHO. Total unavailable CHO, as determined by the amount of fecal CHO, averaged 7.8 and 18.1 g/d in subjects consuming the high- and low-fiber diets, respectively. We did not analyze for TDF in the feces; however it is logical to assume that a large portion of the CHO in the feces is undigested dietary fiber. We reported earlier (Baer et al. 1997) that TDF constituted about 61% of the CHO in feces. Thus, digestibility of the TDF in the diet would have been ~62 and 75% in the high- and low-fiber diets, respectively. The extent of digestion is higher than observations of other workers (Farrell et al. 1978, Kelsay et al. 1981, Marlet and Johnson 1985) that report neutral detergent fiber digestibility, similar to our previous report of TDF digestibility (Baer et al, 1997) that ranged between 66 and 82% on a variety of diets.

Thermic effect of food.  The term "TEF" is used to describe the rise in EE associated with the consumption and processing of food. The TEF is a consequence of changes in a number of physiologic and metabolic processes including mastication, digestion, absorption and the utilization and/or storage by various tissues within the body (Blaxter 1989, Ravussin et al. 1986). Comparison of one or more 24-h measurements in which the conditions are similar except for alterations in the energy intake or composition of the diet was made by a number of investigators. The differences in EE and energy balance between the measurements can be attributed to the energy cost and balance of substrates associated with the diet change. This method was used to examine the TEF and maintenance requirements in weight-stable individuals (Van Es et al. 1984) and before and after weight loss (DeBoer et al. 1986, Rumpler et al. 1991). The sensitivity of the method as well as the design of the experiment limits the usefulness of the data to examining differences in TEF between the +CHO treatment and the two fat treatments. Different individuals received different fiber levels throughout the study. Thus, the statistical power of comparisons between the two fiber levels is weak. Differences between the two fiber levels would have had to be greater than 1300 kJ to be significant. However, each of the participants in the study received each of the three supplements. Differences between the supplements of 350 kJ would have been significant. This difference is well within the theoretical differences between the fat treatments and the CHO treatment discussed below. Differences between the TEF of the two fat sources were also considered a possibility. Takeuchi et al. (1995) and Matsuo et al. (1995) observed an increase in the EE of rats following a meal with the substitution of safflower or linseed oil for either BT or lard. Oxygen consumption was 20 or 81 mL/min higher 1 h following a meal containing BT or linseed oil, respectively. This observation does not directly translate into higher 24-h EE on the linseed oil diet, due to possible compensatory responses later in the day. However, their data also demonstrated slower growth and a lower accumulation of total body fat in rats consuming equal caloric amounts of the vegetable oil diet vs. the BT diet. The combined observations of higher TEF and lower growth suggest that the observed higher TEF translated into higher EE by rats consuming the vegetable oil.

We found the TEF to be nearly identical between the base diets with the +CHO or the two sources of fat and accounted for ~13% of the energy supplied in the diet. Based on the stoichiometry of the oxidation of fat and CHO, some observable differences should exist in the TEF associated with the metabolism of CHO and fat substrates. Flatt (1991) presented a theoretical estimate of the adenosine triphosphate (ATP) yield of the oxidation of a mole of starch/glycogen or fatty acid to CO₂ and water. Included in these calculations were some assumptions regarding the amount of sympathetic nervous system stimulation and the proportion of substrate movement through various metabolic pathways. Flatt estimated the production of ATP equivalent to be about 80% for CHO and 92% for fat of the energy contained in the original molecule. This difference should produce a higher TEF of ~12 percentage points (i.e. 20 vs. 8%) when an equivalent amount of ATP is produced from CHO substrate vs. fat. In this study, no observable differences in the TEF existed between the high- and low-fiber diets or between the CHO and fat additions. Given the sensitivity of the method used to estimate TEF, a difference between the two different fiber levels would have been unlikely to be statistically significant. However, the difference expected between the fat and +CHO treatment, based on the theoretical estimates above, would have been well within the sensitivity of the methodology. The lack of observed difference between the fat and +CHO treatments suggests two possibilities: i) The majority of the TEF was associated with some aspect of metabolism not specific to the substrate. ii) Counter-balancing responses resulted in a muting of the difference and a coincidental similarity between fat and CHO in the energy cost associated with the change in ME intake. A probable explanation is that since subjects were in slightly negative energy balance the substitution of fat, regardless of source, for CHO simply reflected a shift between body stores and the diet source of fat and CHO. This difference in energy cost associated with the use of body CHO and body fat vs. diet CHO and diet fat would be the cost of absorption. The highly digestible nature of the diet CHO and the two sources of diet fat must have not incurred a large energy cost of absorption.

Substrate oxidation.  The objective of this study was to determine if the differences in body composition and energy metabolism between diets containing BT or vegetable oil observed in rats (Shimomura et al. 1990) and mice (Pariza et al. 1996) might also apply in humans. We found no significant differences in overall EE between the two fat sources. In addition, the net absorption from the diet of the fat was only slightly higher from the BT diet than from the CO diet.

We expected some differences in fat oxidation between the two fat sources. Leyton et al. 1987 and Jones et al. 1985 showed that poly- and monounsaturated fats are oxidized more quickly than saturated fats. Since vegetable oils have a much higher level of unsaturated fatty acids, they may be oxidized more readily, thereby sparing other energy sources. A significant difference in fat oxidation existed between the BT vs. the CO treatment in the subjects who consumed the low-fiber diet but not in the high-fiber diet group. The inclusion of BT in the low-fiber diet increased fat oxidation 243% over the +CHO treatment while the +CO diet increased fat oxidation 210% in subjects who consumed the low-fiber diet. This translated into a 10 g/d increase in net fat loss when subjects consumed the BT diet compared to the CO diet.

The highest fat oxidation occurring when subjects consumed the BT diet seems to contradict the rat data that showed greater fat deposition in rats fed BT diets (Shimomura et al. 1990) and higher oxidation rates of unsaturated fatty acids (Leyton et al. 1987). This contradiction raises the question of whether this is just a type I error due to inadequate statistical power. The observation of the highest fat oxidation occurring while consuming the BT was consistent across subjects. Five of six individuals consuming the low-fiber diet and three of six fed the high-fiber diet had the greatest fat oxidation while consuming the BT. Of the twelve individuals, only two had the greatest fat oxidation when consuming the CO and both were consuming the high-fiber diet. The consistency of the response suggests that the data collected reflected some physiological response, not just a statistical anomaly.

If we accept higher fat oxidation during BT diet consumption, as observed, two questions arise: Why is the response much smaller in the high-fiber group? Why would the consumption of BT result in higher fat oxidation than CO? The smaller response in the high-fiber group may be related to a greater negative energy balance in the high-fiber group relative to the low-fiber group. Individuals in the low-fiber group were within 1.5% of energy balance while individuals in the high-fiber group were below energy balance by as much as 10%. The greater negative energy balance in the high-fiber diet group was a consequence of an overestimation of the energy value of the high-fiber diet when formulated. The formulated value of the high-fiber diet was much higher than that determined by direct measurement. However, the measured energy value of the low-fiber diet was close to its formulated value. Thus, the actual energy intake of subjects consuming the high-fiber diet was lower than the target intake, which resulted in the energy intake of subjects consuming the high-fiber diet being negative in the chamber while the low-fiber group was nearly in energy balance or slightly positive. This may have moderated the differences between the two fat sources when consumed by the high-fiber group. Diet substrates contributed ~90% of the total substrates required to meet energy demand of subjects consuming the high-fiber diet. Thus, in the high-fiber group, all of the diet substrates would be oxidized to meet energy demands with the remainder coming from body stores. In subjects consuming the low-fiber diet, substrate intake and EE were nearly equal. Thus, small changes in use or storage could be reflected in a net change over the 24-h calorimeter measurement.

All of the subjects in the study lost some body weight during each of the diet periods, indicating that they were in negative energy balance for most of the study. However, because of the sedentary nature of a calorimeter measurement, subjects in the low-fiber group were at or near energy balance while in the calorimeter. When low-fiber diet subjects consumed either the BT or the CHO supplement, they were in positive CHO balance in the calorimeter. The positive CHO balance suggests a sparing of CHO at the expense of greater fat oxidation. A short-term increase in CHO stores in response to an increase in intake or decrease in EE is common in calorimetry studies particularly following a period of negative energy balance (Burstein et al. 1996). CHO utilization is the principal mechanism by which the body adjusts energy intake to EE over the short term. However, after low-fiber diet subjects consumed the CO diet, they did not exhibit a positive CHO balance and had lower fat oxidation, suggesting that this treatment did not result in a sparing of CHO. This difference in CHO balance between CO and BT could have been a consequence of the small difference (not significant) in energy balance (Table 4) between the two treatments. However, the magnitude of the differences in substrate oxidation was five times as great, on an energy basis, as the differences in energy balance. These results suggest that the rate of fat, oxidation in subjects consuming either the CHO or BT supplement, persisted at a rate more like that present outside of the calorimeter. However, when subjects consumed the CO diet, fat oxidation was lower in response to the lower EE occurring in the calorimeter.

Several compositional differences between CO and BT exist that might influence fat oxidation. Aside from the differences in fatty acid composition alluded to earlier, the presence of higher levels of conjugated linoleic acid (CLA) in the BT-supplemented diets may have impacted net fat oxidation. BT may contain 0.8% of total fatty acids as CLA, whereas CO would be expected to contain only trace amounts. Thus, the BT-supplemented diets would have a concentration of about 0.2% CLA. Several recent studies reported both decreases in fat mass in mice (Park et al. 1997) and rats (Chin et al. 1994) consuming diets supplemented with 0.5% CLA. In addition, the work by Park et al. (1997) demonstrated increased levels of carnitine palmitoyltransferase activity in several fat and muscle sites in rats that would indicate higher fat oxidation. They also reported increases in vitro hormone-sensitive lipase activity and enhanced norepinephrine-induced lipolysis in rat adipocytes. To extrapolate the results of these rat studies to the results seen in the current work is difficult. However, if CLA or some other characteristic differences between BT and CO lead to a greater persistence of fat oxidation following either a change in intake or activity, the accumulated effect could result in a substantial difference in fat utilization over time.

These results leave open the possibility that BT may have some effect on net fat balance. However, it does not appear to be a consequence of differences in digestibility or effects on metabolic rate. The observed higher rate of fat oxidation when subjects consumed the BT may have consequences in the regulation of body weight when compared to fat from vegetable sources.

View this table: [in this window] [in a new window]   APPENDIX 1 High- and low-fiber menus with food items with added fat (*) and food substitutions () identified

	FOOTNOTES

Manuscript received 15 April 1998. Initial reviews completed 28 May 1998. Revision accepted 11 August 1998.

	LITERATURE CITED

Abstract Introduction Methods Results Discussion References

• Baer D. J., Rumpler W. V., Miles C. W., Fahey G. C. Dietary fiber decreases the metabolizable energy content and nutrient digestibility of mixed diets fed to humans. J. Nutr. 1997; 127:579-586[Abstract/Free Full Text]
• Blaxter, K. (1989) Energy Metabolism in Animals and Man. Cambridge University Press, Cambridge, UK.
• Burstein R., Prentice A. M., Goldberg G. R., Murgatroyd P. R., Harding M., Coward W. A. Metabolic fuel utilization in obese women before and after weight loss. Int J. Obesity 1996; 20:253-259
• Calloway D. H., Kretsch M. J. Protein and energy utilization in men given a rural Guatemalan diet and egg formulas with and without added oat bran. Am. J. Clin. Nutr. 1978; 31:1118-1126[Abstract/Free Full Text]
• Chin S. F., Storkson J. M., Albright K. J, Cook M. E., Pariza M. W. Conjugated linoleic acid is a growth factor for rats as shown by enhanced weight gain and improved feed efficiency. J. Nutr. 1994; 124:2344-2349
• DeBoer J. O, van Es A.J.H., Roovers L.C.A., van Raaij J.M.A., Hautvast J.G.A.J. Adaptation of energy metabolism of overweight women to low-energy intake, studied with whole body calorimeters. Am. J. Clin. Nutr. 1986; 44:585-595[Abstract/Free Full Text]
• Denke M. A., Fox M. M., Schulte M. C. Short-term dietary calcium fortification increases fecal saturated fat content and reduces serum lipid in men. J. Nutr. 1993; 123:1047-1053
• Dougherty R. M., Allman M. A., Iacono J. M. Effects of diets containing high or low amounts of stearic acid on plasma lipoprotein fraction and fecal fatty acid excretion of men. Am. J. Clin. Nutr. 1995; 61:1120-1128[Abstract/Free Full Text]
• Fang C. S., Kies C. Serum lipid levels and fecal fat losses of humans: Effects of level of dietary fiber and kind of fat. The FASEB J. 1993; 7:4209
• Farrell D. J., Girle L., Arthur J. Effects of dietary fibre on the apparent digestibility of major food components and on blood lipids in men. Aust. J. Exp. Biol. Med. Sci. 1978; 56:469-479[Medline]
• Flatt, J. P. (1991) The biochemistry of energy expenditure. Recent Advances in Obesity Research: Proceedings of the 2nd International Congress on Obesity. (G. Bray, ed.) pp. 211-228, Newman Publishers, Los Angeles.
• Hanson C., Khalil D., Owens F. Effect of calcium level or fiber on energy use and body composition of mice. FASEB J. 1993; 7:4208
• Jones P.H.H., Pencharz P. B., Clandinin M. T. Whole body oxidation of dietary fatty acids: implication for energy utilization. Am. J. Clin. Nutr. 1985; 42:769-777[Abstract/Free Full Text]
• Kaplan R. J., Greenwood C. E. Poor digestibility of fully hydrogenated soybean oil in rats: a potential benefit of hydrogenated fats and oils. J. Nutr. 1998; 128:875-880[Abstract/Free Full Text]
• Kelsay J. L., Goering H. K., Behall K. M., Prather E. S. Effect of fiber from fruits and vegetables on metabolic responses of human subjects: fiber intakes, fecal excretions, and apparent digestibilities. Am. J. Clin. Nutr. 1981; 34:1849-1852[Abstract/Free Full Text]
• Kushner R. F., Schoeller D. A. Estimation of total body water by bioelectric impedance analysis. Am. J. Clin. Nutr. 1986; 44:417-424[Abstract/Free Full Text]
• Leyton J., Drury P. J., Crawjoed M. A. Differential oxidation of saturated and unsaturated fatty acids in vivo in the rat. Br. J. Nutr. 1987; 57:383-393[Medline]
• Livesey G. Energy values of unavailable carbohydrate and diets: an inquiry and analysis. Am. J. Clin. Nutr. 1990; 51:637
• Livesey G. Calculating the energy values of foods: Towards new empirical formulae based on diets with varied intakes of unavailable complex carbohydrates. Eur. J. Clin. Nutr. 1991; 49:1-12
• Livesey G., Elia M. Estimation of energy expenditure, net carbohydrate use and net fat oxidation and synthesis by indirect calorimetry: evaluation of errors with special reference to the detailed composition of fuels. Am. J. Clin. Nutr. 1988; 47:608-626[Abstract/Free Full Text]
• Marlett J. A., Johnson E. J. Composition of fecal fiber from human subjects. J. Nutr. 1985; 115:650-660
• Matsuo T., Shimomura Y., Saitoh S., Tokuyama K., Takeuchi H., Suzuki M. Sympathetic activity I lower in rats fed a beef tallow diet than in rats fed a safflower oil diet. Metab. 1995; 44:934-939[Medline]
• Mercer S. W., Trayhurn P. Effect of high fat diets on energy balance and thermogenesis in brown adipose tissue of lean and genetically obese ob/ob mice. J. Nutr. 1987; 117:2147-2153
• Miles C. W., Kelsay J. L., Wong N. P. Effect of dietary fiber on the metabolizable energy of human diets. Nutr. 1988; 118:1075-1081
• Miles C. W., Webb P., Bodwell C. E. The metabolizable energy of diets differing in dietary fat and fiber measured in humans. J. Nutr. 1992; 122:306-311
• Monsma C. C., Gallaher D. D., Ney M. N. Reduced digestibility of beef tallow and cocoa butter affects bile acid excretion and reduces hepatic esterified cholesterol in rats. J. Nutr. 1996; 126:2028-2035
• Pariza M., Park Y., Cook M., Albright K., Liu W. Conjugated linoleic acid (CLA) reduced body fat. FASEB J. 1996; 10:A560
• Park Y., Albright K. J., Liu W., Storkson J. M., Cook M. E., Pariza M. W. Effect of conjugated linoleic acid on body composition in mice. Lipids 1997; 32:853-858[Medline]
• Prosky L., Asp N. G., Schweizer T. F., DeVries J. H., Furda I. Determination of insoluble, soluble and total dietary fiber in foods and food products: interlaboratory study. J. Assoc. Off. Anal. Chem. 1985; 71:1017-1023
• Ravussin E., Lillioja S., Anderson T. E., Christin L., Bogardus C. Determinants of 24 h energy expenditure in man. Methods and results using a respiratory chamber. J. Clin. Invest. 1986; 43:257-269
• Rumpler W. V., Seale J. L., Conway J. M., Moe P. W. Repeatability of 24 hour energy expenditure measurements in humans by indirect calorimetry. Am. J. Clin. Nutr. 1990; 51:147-142[Abstract/Free Full Text]
• Rumpler W. V., Seale J. L., Miles C. W., Bodwell C. E. Energy intake restriction and diet-composition effects on energy expenditure in men. Am. J. Clin. Nutr. 1991; 53:430-436[Abstract/Free Full Text]
• Schrijver R. D., Vermeulen D., Viaene E. Lipid metabolism responses in rats fed beef tallow, native or randomized fish oil and native or randomized peanut oil. J. Nutr. 1991; 121:948-955
• Seale J. L., Rumpler W. V., Conway J. M. Comparison of energy expenditure estimations using doubly labeled water, intake balance and direct indirect calorimetry methods in adult men. Am. J. Clin. Nutr. 1990; 52:66-71[Abstract/Free Full Text]
• Seale J. L., Rumpler W. V., Moe P. W. Description of a direct-indirect room sized calorimeter. Am. J. Physiol. 1991; 260:E306-320[Abstract/Free Full Text]
• Shimormura Y., Tamura T., Suzuki M. Less body fat accumulation in rats fed a safflower oil diet than in rats fed a beef tallow diet. J. Nutr. 1990; 120:1291-1296
• Southgate D.A.T., Durnin J.V.G.A. Calorie conversion factors. An experimental reassessment of the factors used in the calculation of the energy value of human diets. Br. J. Nutr. 1970; 24:517-535[Medline]
• Takeuchi H., Matsuo T., Tokuyama K., Shimomura Y., Suzuki M. Diet-induced thermogenesis is lower in rats fed a lard diet than in those fed a high oleic acid safflower oil diet, a safflower oil diet or a linseed oil diet. J. Nutr. 1995; 125:920-925
• Van Es A.J.H., Vogt J. E., Niessen C. H., Veth N. J., Rodenburg L., Teeuwse V., Dhuyvetter J., Deurenberg P., Hautvast J.G.A.J., Van Der Beek E. Human energy metabolism below, near and above energy equilibrium. Brit. J. Nutr. 1984; 52:429-442[Medline]
• Watt, B. K. & Merrill, A. L. (1963) Composition of foods: raw processed, prepared. Agricultural Handbook no. 8, Washington, D.C.: U.S. Government Printing Office.

0022-3166/98 $3.00 ©1998 American Society for Nutritional Sciences