Energy expenditure in humans: effects of dietary fat and carbohydrate WILLIAM GIACOMO
G. H. ABBOTT, BARBARA V. HOWARD, RUOTOLO, AND ERIC RAVUSSIN
Clinical Diabetes and Nutrition Section, National Institute of Diabetes, Digestive and Kidney Diseases, National Institutes of Health, Phoenix, Arizona 85016
ABBOTT, WILLIAM G. H., BARBARA V. HOWARD, GIACOMO RUOTOLO, AND ERIC RAVUSSIN. Energy expenditure in hurnarw: effects of dietary fat and carbohydrate. Am. J. Physiol.
METHODS
258 (Endocrinol. Metab. 21): E347-E351, 1990.-A high-dietary fat intake may be an important environmental factor leading to obesity in some people. The mechanism could be either a decrease in energy expenditure and/or an increase in caloric intake. To determine the relative importance of these mechanisms we measured 24-h energy expenditure in a whole body calorimeter in 14 nondiabetic subjects and in six subjects with non-insulin-dependent diabetes mellitus, eating isocaloric, weight-maintenance, high-fat, and high-carbohydrate diets. All subjects were Pima Indians. In nondiabetics, the mean total 24-h energy expenditure was similar (2,436 k 103 vs. 2,359 k 82 kcal/day) on high-fat and high-carbohydrate diets, respectively. The means for sleeping and resting metabolic rates, thermic effect of food, and spontaneous physical activity were unchanged. Similar results were obtained in the diabetic subjects. In summary, using a whole body calorimeter, we found no evidence of a decrease in 24-h energy expenditure on a highfat diet compared with a high-carbohydrate diet.
The subjects for this study were 20 Pima Indians. All subjects had a full medical history and physical examination, with routine hematology and biochemistry screens and urinalysis. Six subjects had noninsulin-dependent diabetes mellitus (NIDDM) by World Health Organization criteria (18). At the time of the study none of the subjects had evidence of significant disease apart from NIDDM, and none had significant complications of NIDDM. The physical characteristics of the subjects are shown in Table 1 and plasma glucose concentrations in Table 2.
Subjects
Protocol
The subjects resided on the metabolic ward for the duration of the studies. Activity was restricted to the ward, and alcohol consumption was strictly forbidden. The ‘subjects were admitted to the research ward on obesity; non-insulin-dependent diabetes mellitus two separate occasions at least 4 wk apart, eating a different diet on each occasion. Eleven subjects ate the high-fat diet on the first admission, and nine subjects ate the high-carbohydrate diet on the first admission. POPULATION STUDIES SHOW that a high-dietary fatintake is associated with obesity in humans (5, 14, 17). The measurement of energy expenditure was made after There are supportive data in animal studies (11, 15). 18 & 2 (mean & SE) days (range 6-32 days) on the highfat diet and after 27 & 2 days (range 5-43 days) on the Thus a high-fat intake may be an important environhigh-carbohydrate diet. mental factor promoting obesity in some individuals. During the 24-h period in the chamber, the caloric Because obesity results from a positive energy balance, i.e., energy intake exceeds energy expenditure, there are intake of the subjects was reduced to 80% of that given on the research ward. This was to allow for the reduction two possible mechanisms for this association. The first in calories expended in spontaneous physical activity is that high-fat diets may increase caloric intake (11,17). because of confinement in the ,small chamber (8.2 m2). The second possibility is that high-fat diets may lower energy expenditure (6, 8, 16), which has been shown to Because it is possible that the effect of diet on energy expenditure could be most marked during physical activpredict weight gain (13). ity, five subjects spent an additional 24-h period in the To determine the relative importance of these mechchamber on each diet. Energy intake was increased to anisms, 20 Pima Indian subjects were admitted to a the same number of calories normally consumed on the metabolic ward and were fed isocaloric high-fat and highresearch ward, and energy expenditure was increased to carbohydrate diets. The prevalence of obesity in the Pima an equivalent level by use of an exercycle (see below). Indian population is high (7). The 24-h energy expenditure was measured on each diet in a whole body indirect calorimeter. This is a sealed room with a floor area of Diets 8.2 m2, in which restricted activity is possible. Energy Weight-maintenance caloric intake was estimated expenditure under these conditions was similar on the from body weight on admission to the research ward and high-fat and the high-carbohydrate diets. was adjusted on the basis of fluctuations in the fasting, E347 Downloaded from www.physiology.org/journal/ajpendo at Macquarie Univ (137.111.162.020) on February 12, 2019.
E348 TABLE
DIETARY
1. Physical characteristics
NIDDM
of subjects in the nondiabetic
and diabetic groups W k
Sex
n
Nondiabetic
FAT AND THERMOGENESIS
WV
14
1212
6
4/z
35A3 (21-57) 41*4 (24-49)
Body Fat, %
HIFAT
HICHO
HIFAT
HICHO
100*7 (63-157) 88*10 (65-136)
lOOk7 (64-153) 87klO (65-133)
29*2 (16-42) 34*4 (22-48)
3Ok2 (14-45) 34*3 (21-46)
Values are means * SE, ranges are in parentheses; n = 20 subjects. HIFAT, high-fat diet; HICHO, high-carbohydrate insulin-dependent diabetes mellitus.
2. Fasting and 2-h plasma glucose concentrations in diabetic and nondiabetic groups
TABLE
Nondiabetic HIFAT
Fasting glucose, w/d1 2-h glucose,
93k3 (77-109) 108k7
HICHO
93Ik3 (77-106) 115*9
HICHO
215&39 (99-344)
216&40 (94-336)
368k44 (234-500)
345k51
(194-480) Values are means & SE, ranges in parentheses; n = 20 subjects. NIDDM, non-insulin-dependent diabetes mellitus; HIFAT, high-fat diet; HICHO, high-carbohydrate diet. w/d1
(70-157)
(50-178)
posturination body weight measured each morning. The composition of the high-fat and high-carbohydrate diets is summarized in Table 3. The fat in the high-fat diet was derived predominantly from dairy products. No subjects developed clinical evidence of lactose intolerance on the high-fat diet. On the high-carbohydrate diet, fat was replaced by predominantly complex carbohydrate of vegetable, legume, and cereal origin, resulting in an increase in fiber intake. Meals were made from readily available foods purchased from a local supermarket and were freshly and individually prepared using standardized recipes with weighed and/or measured ingredients (Mettler PG balance, &l g, Mettler, Greifensee, Switzerland). Diet composition was calculated with a computerized diet analysis program (the CBORD group: CBORD’s Diet Analyzer System, ESHA Database, 1986; Ithaca, NY) as well as manufacturer specifications. Procedures Body composition. Body composition was estimated by underwater weighing with simultaneous measurement of lung volume by helium dilution (13). OraZ gZucose toZerance test. After at least 4 days on the research ward and after a 12-h overnight fast, an oral glucose tolerance test was performed with 75 g of glucose (Koladex, Custom Laboratories, MD). Blood was drawn at -5 min and at 120 min for measurement of the plasma glucose concentration (Beckman glucose analyzer, Fullerton, CA). Energy expenditure. The 24-h energy expenditure was measured in a closed-circuit indirect calorimetry chamber that has previously been described (12). The chamber is a sealed room with a volume of 19,000 liters through which atmospheric air is continuously drawn. The subject’s gas exchange is calculated by measuring the OZ and CO2 concentrations at the inlet and outlet of the chamber and by measuring the flow of air through the chamber. Spontaneous physical activity was measured in 14 non-
non-
TABLE 3. composition (% of total calories) of the high-fat and high-carbohydrate diets
NIDDM HIFAT
diet; NIDDM,
High Fat
Carbohydrate, % Simple, % Complex, % Fat, % Polyunsaturated, % Monounsaturated, % Saturated, % Protein, % Fiber, g /l,OOO kcal Cholesterol, mg/day Nonprotein respiratory quotient of diet
43
High Carbohydrate
22
65 20 45 20 6 8 6
15 10
15 19
14 29 42 7
13
560 0.83
524
0.91
diabetic and five of the six diabetic subjects by two microwave motion sensors (MICD 930, Honeywell, Minneapolis, MN). These units emit a signal, the frequnecy of which is ahered by movement (Doppler effect), and the change is measured by a transreceiver. The sensitivity of this system was set to detect any movement greater than breathing. One activity unit, expressed as percent, represents the percentage of time that the subject was moving, and these units are independent of work intensity. After an overnight fast and after urinating, each subject entered the chamber at 0750 h and was confined until 0700 h the next morning. Vigorous exercise and cigarette smoking were prohibited. Meals were passed through a small air lock at 0800, 1115, and 1615 h, with a snack at 1900 h. Subjects were carefully instructed to eat all provided food and to return any uneaten food to the kitchen so the caloric intake could be computed. Subjects were instructed to collect all urine, and this was used to measure 24-h urinary glucose loss and 24-h urea nitrogen production. At 0700 h the next morning, the chamber door was opened, and while the subject was lying on the bed, a clear plastic hood was placed over his/her head. After 25 min of becoming acclimatized to the hood, the resting metabolic rate was measured for between 9 and 12 min, with the hood connected to the analyzing equipment of the chamber. Five subjects spent an additional day in the chamber on each diet, during which time caloric intake wasmaintained at the weight maintenance level on the research ward, and energy expenditure in the chamber was increased to match energy intake by use of an Exercycle (Collins Pedelmate, Warren E. Collins, Braintree, MA). The exercise was divided into four periods of equal length and was carried out at a rate of 75 W. The’ duration of
Downloaded from www.physiology.org/journal/ajpendo at Macquarie Univ (137.111.162.020) on February 12, 2019.
DIETARY
FAT
AND
the exercise differed for each subject and was calculated on the basis of previously determined energy expenditure at 50,100, and 150 W. Energy Expenditure
Calculations
Oz uptake (vo2) and CO2 (VCO~) production rates were calculated from the difference between the atmospheric air and the outflow Oz and CO2 concentrations, corrected for the changes in Oz and CO2 concentrations in the chamber. Values were corrected for temperature, pressure, and humidity and expressed as STPD. Energy expenditure (EE in kcal/min) during each 15min period was calculated as follows (12) EE = Voz* [4.686 + (RQ-0.707).
(0.361/0.293)]
in which RQ is the respiratory quotient, i.e., the ratio between CO2 production and 02 consumption during the 15min period. Twenty-four-hour energy expenditure was calculated as the average energy expenditure of 92, 15-min periods extrapolated to 24 h. Sleeping metabolic rate was calculated as the average energy expenditure of all 15-min Periods between 2330 and 0500 h, during which the spontaneous physical activity measured by radar was cl%. The average cost of physical activity per unit of measured activity (%) was calculated as the slope of the linear regression line between activity (%) during a 15-min period and the energy expenditure (kcal/min) during that period. The amount of energy expended per 24 h due to spontaneous physical activity was calculated as the product of the energy cost per unit activity (the slope of the above regression line) and the mean 24-h activity, and this product was extrapolated to 24 h. The thermic effect of food was estimated by integrating the energy expenditure corrected for activity above sleeping metabolic rate, and was expressed as a percentage of the 24-h caloric intake. We used the sleeping metabolic rate rather than the resting metabolic rate to calculate the thermic effect of food because we find it is a more reproducible measurement. The measurement of thermic effect of food thus includes the energy cost of arousal. I
24 hr EE
I I I
I I
Resting
metabolic
Sleeping
metabolic
I -25
I
I
I
I
I
-20
-15
-10
-5
oh Decrease on 43% Fat Diet
Energy
I I
I
+5
+lO
I +15
effect
cost per unit activity
Calories
I 0
rate
Thermic of food
I I
rate
expended
I
I
+20
+25
in activity
oh Increase on 43% Fat Diet
FIG. 1. Confidence limits (95%) for mean changes in energy expenditure (EE) on high-fat compared with high-carbohydrate diet in 14 nondiabetic subjects.
E349
THERMOGENESIS
Statistics
All data are expressed as the means & SE. The Kolomogorov D statistic was used to test the normality of distributions, and those variables with nonnormal distributions (sleeping metabolic rate and 24-h respiratory quotient) were logarithmically transformed. The paired t test was used to assess the significance of differences between means on the two diets. Correlation analysis was done by means of the Pearson product-moment correlation coefficient. These statistical analyses were calculated with the Statistical Analysis System (SAS), SAS Institute, Cary, NC. The 95% confidence intervals for the changes in mean energy expenditure on the high-fat diet were calculated to show the magnitude of type 2 error (4). The 95% confidence intervals are shown in Fig. 1. RESULTS
Nondiabetic
Subjects
In the nondiabetic group there was no decrease in either 24-h energy expenditure or in any component of 24-h energy expenditure on the high-fat compared with the high-carbohydrate diet (Table 4). The 95% confidence intervals for the changes in the means are shown in Fig. 1. These are estimates of type 2 error. The purpose of this figure is to show the largest mean decrease in energy expenditure on the high-fat diet that would be consistent with our data. Thus, given the limits of the experimental conditions, we could exclude the possibility that the high-fat diet induced a decrease in mean 24-h energy expenditure of >0.8%, a decrease in mean resting metabolic rate of >6.3%, and a decrease in mean sleeping metabo1ic rate Of >2e5%* There were no correlations between the weight or body composition of the subjects and any of the individual changes in energy expenditure. There was a positive correlation between the duration of the high-fat diet and the measurement of physical activity (r = 0.70, P c 0.01) on the high-fat diet. There were no correlations between the duration of the high-carbohydrate diet and any measurements of energy expenditure made on this diet. Table 5 shows that mean energy balance (food intake - 24-h energy expenditure) tended to be negative on both diets but was not significantly different from zero on either diet. Both the resting respiratory quotient and the 24-h respiratory quotient were significantly higher on the high-carbohydrate diet than on the high-fat diet (Table 4). On the high-fat diet the 24-h respiratory quotient was similar to the respiratory quotient of the food. On the high-carbohydrate diet the 24-h respiratory quotient was significantly less than the respiratory quotient of the food (p c 0.001). Five subjects had a 2nd day in the chamber on each diet, with a 25% increase in caloric intake and an equivalent increase in physical activity. The purpose was to determine whether the failure to find a decrease in energy expenditure on the high-fat diet might be due to the decrease in activity that occurs with confinement in the chamber. The addition of physical exercise caused in-
Downloaded from www.physiology.org/journal/ajpendo at Macquarie Univ (137.111.162.020) on February 12, 2019.
E350
DIETARY
4. Measurements
TABLE
FAT
of energy expenditure
AND
THERMOGENESIS
on the high-fat
and high-carbohydrate
diets
Nondiabetic HIFAT 24-h energy expenditure, kcal/day Resting metabolic rate, kcal/min Sleeping metabolic rate, kcal/min Energy cost/unit activity, kcal min. % activity-’ Physical activity, kcal/day Thermic effect of food, % caloric intake*
NIDDM HICHO
HIFAT
HICHO
2,435&103 1.38kO.06 1.2220.06 0.032&0.002
2,359*82 1.36kO.06 1.17kO.04 0.031*0.002
2,169*207 1.29kO.14 1.17kO.11 0.029~0.004
2,128&185 1.15*0.10 1.11&0.07 0.026&0.002
416&38 19.6k2.3
384&36 19.2kl.7
329k46 16.4kl.6
309k30 14.6kl.7
l
Values are means & SE; nondiabetic n = 14, NIDDM, high-carbohydrate diet. * Energy expenditure corrected
n = 6. NIDDM, non-insulin-dependent for activity above sleeping metabolic
rate
diabetes mellitus; over 24 h.
HIFAT,
high-fat
diet; HICHO,
5. Caloric intake, energy expenditure, and respiratory quotients on high-fat and high-carbohydrate diets in nondiabetic and diabetic subjects
TABLE
Nondiabetic
NIDDM
HIFAT 24-h energy expenditure, kcal/day Food intake, kcal/day Energy balance, kcal/day 24-h urinary glucose loss, g/day R*Q of diet 24-h RQ R*esting RQ Values are means & SE; nondiabetic high-carbohydrate diet; RQ, respiratory
HICHO
2,435&103 2,361&113 -75*63 0.2kO.04 0.83 0.84&0.01 0.82&0.01
n = 14, NIDDM quotient. * P
HIFAT
2,359&82 2,329&102 -30&42 0.3*0*04 0.91 0.88*0.01* 0.84&0.01?
n = 6. NIDDM, non-insulin-dependent T P c 0.01.
HICHO
2,169&207 2,334&212 +164&116 55223 0.83 0.82kO.01 0.77&0.01 diabetes
2,128&185 2,316&210 +189&101 104*34 0.91 0.85*0.02? 0.78&0.01
mellitus;
HIFAT,
high-fat
diet;
HICHO,
< 0.001,
6. Effect of a 20% increase 24-h energy expenditure and isocaloric increase in energy intake on thermogenesis and respiratory quotient in nondiabetic subjects
TABLE
Without
24-h energy intake, kcal/day 24-h energy expenditure, kcal/day Energy balance, kcal/day Mean 24-h RQ Values compared
are means & SE; n = 5 subjects. with food quotient.
HIFAT,
Exercise
With
Exercise
HIFAT
HICHO
HIFAT
HICHO
2,280*75 2,363&110 -83&47 0.83kO.02
2,240&74 2,382&88 -142272 0.87&0.01*
2,820&78 2,960&107 -140&55 0.84kO.01
2,834&74 3,025*91 -192&44 0.89kO.01
high-fat
diet;
HICHO,
creases in energy expenditure of 597 & 81 and 643 & 118 kcal/day for the high-fat and high-carbohydrate diets, respectively. The 24-h energy expenditure was not decreased with exercise on the high-fat compared with the high-carbohydrate diet (Table 6). The 24-h respiratory quotient was similar to the respiratory quotient of the food during the exercise protocol on the high-carbohydrate diet. Diabetic Subjects
In the diabetic group the high-fat diet did not result in a decrease in either 24-h energy expenditure or in any of the components of 24-h energy expenditure (Table 4). DISCUSSION
There is evidence that a high-dietary fat intake is an important environmental factor that leads to obesity in some people. Whether dietary fat-induced obesity is due to a decrease in metabolic rate, an increase in caloric intake, or a combination of both factors is unclear.
high-carbohydrate
diet;
RQ, respiratory
quotient.
*
P < 0.05 24-h RQ
Animal studies have not consistently favored either mechanism (11). The purpose of this study was to determine whether a high-fat diet could reduce 24-h energy expenditure in a group of Pima Indian subjects. There was no decrease in the mean of either the total 24-h energy expenditure or of any components of energy expenditure on the high-fat diet in either the nondiabetic or the diabetic subjects. Given the limits of the experimental conditions, the 95% confidence intervals for the changes in the means suggest that the maximum decrease in 24-h energy expenditure induced by the high-fat diet was 0.8% or -18 kcal/day. A decrease in energy expenditure of this magnitude would not result in significant obesity (9). The data do not totally exclude the possibility that high-fat diets contribute to obesity by reducing energy expenditure. The duration of the acclimatization periods may have been too short for the effect to become apparent. The metabolic chamber restricts the activity of the subjects, which would make a dietary effect on activityrelated energy expenditure more difficult to detect. The
Downloaded from www.physiology.org/journal/ajpendo at Macquarie Univ (137.111.162.020) on February 12, 2019.
DIETARY
FAT AND THERMOGENESIS
majority of the subjects were obese, and it is possible that the effect of dietary fat on energy expenditure is only important in preobese subjects. It is also possible that a high-fat diet-induced decrease in energy expenditure is only apparent with overfeeding (3) compared with weight-maintenance feeding. Lastly, the subjects were all Pima Indians and the results may not apply to other racial groups. The effect of high- and low-fat diets on energy expenditure has been investigated by three other groups. Lean and James (8) found no change in 24-h energy expenditure in groups of lean and obese women fed weight-maintenance diets for 24 h. Hurni et al. (6) compared high-fat and high-carbohydrate diets fed to lean, young subjects for 7 days. The subjects were in negative caloric balance at the time of the energy-expenditure measurement. The sleeping metabolic rate was decreased by the high-fat diet, although 24-h energy expenditure and resting metabolic rates were unchanged. McNeil1 et al. (10) found that the resting metabolic rate in women was the same on high-fat and high-carbohydrate diets fed for 5 days. A decrease in the thermic effect of food was expected on the high-fat diet (8,16). This would be because of the greater obligatory cost of glycogen storage (2) and possibly to a facultative increase in thermogenesis due to glucose stimulation of the sympathetic nervous system (1). We have found that the measurement of thermic effect of food in the metabolic chamber is not very reproducible; and a substantially larger number of subjects would have been needed to estimate this parameter accurately. A decrease in thermic effect of food may have been counterbalanced by small increases in other components of 24-h metabolic rate, resulting in no change in total daily energy expenditure. The 24-h respiratory quotient was less than the respiratory quotient of the food on the high-carbohydrate diet. This could have been caused by three factors. The first is the decrease in activity that occurs during the day in the chamber. Second, there was a tendency for the subjects to be in negative energy balance while in the chamber. Third, the subjects may not have fully acclimatized to the high-carbohydrate diet in 2-6 wk. In summary, using a whole-body calorimeter, we could find no evidence that a weight-maintenance, high fatdiet reduces 24-h energy expenditure in humans. The data suggest that if high-fat diets do predispose to obesity, then the mechanism is either an increase in caloric intake or a decrease in physical activity. We thank Carol dietary staff of the National Institutes Adams for statistical
Lamkin and Vicky Boyce and the nursing and Clinical Diabetes and Nutrition Section of the of Health, Phoenix, AZ. We also thank Beverly assistance, as well as all the volunteers.
E351
Address for reprint requests: W. G. H. Abbott, Dept. of Human Nutrition, Southwestern Medical School, 53‘23 Harry Hines Blvd., Dallas, TX 75235. Received 21 March 1989; accepted in final form 27 September 1989. REFERENCES 1. ACHESON, K. J., E. RAVUSSIN, J. WAHREN, AND E. JEQUIER. Thermic effect of glucose in man: obligatory and facultative thermogenesis. J. Clin. Inuest. 74: 1572-1580, 1984. 2. ACHESON, K. J., Y. SCHUTZ, T. BESSARD, E. RAVUSSIN, AND E. JEQUIER. Nutritional influences on lipogenesis and thermogenesis after a carbohydrate meal. Am. J. Physiol. 246 (Endocrinol. Metub. 9): E62-E70, 1984. E. Diet and obesity. Am. J. CZin. Nutr. 41: 1132-1145, 3. DANFORTH, 1985. 4. DETSKY, A. S., AND D. L. SACKETT. When was a negative clinical trial big enough. Arch. Intern. Med. 145: 709-712, 1985. 5. DREON,D. M.,B. FREY-HEWITT, N.ELLSWORTH, P. T. WILLIAMS, R. B. TERRY, AND P. D. WOOD. Dietary fat: carbohydrate ratio and obesity in middle-aged men. Am. J. CZin. Nutr. 47: 9954000, 1988. 6. HURNI, M., B. BURNAND, PH. PITTET, AND E. JEQUIER. Metabolic effects of a mixed and a high-carbohydrate diet in man, measured over 24 hours in a respiration chamber. Br. J. Nutr. 47: 33-43, 1982. 7. KNOWLER, W. C., D. J. PETTITT, P. H. BENNETT, AND R. C. WILLIAMS. Diabetes mellitus in the Pima Indians: genetic and evolutionary considerations. Am. J. Whys. Anthropology 62: 107114,1983. 8. LEAN, M. E. J., AND W. P. T. JAMES. Metabolic effects of isoenergetic nutrient exchange over 24 hours in relation to obesity in women. Int. J. Obesity 12: 15-27, 1988. 9. LEW, E. A., AND L. GARFINKEL. Variations in mortality by weight among 750,000 men and women. J. Chron. Dis. 32: 563-576,1978. 10. MCNEILL, G., A. C. BRUCE, A. RALPH, AND W. P. T. JAMES. Interindividual differences in fasting nutrient oxidation and the influence of diet composition. Int. J. Obesity 12: 455-463, 1988. 11. OSCAI, L. B., M. B. BROWN, AND W. C. MILLER. Effect of dietary fat on food intake, growth and body composition in rats. Growth 48: 415-424,1984. 12. RAVUSSIN, E., S. LILLIOJA, T. E. ANDERSON, L. CHRISTIN, AND C. BOGARDUS. Determinants of 24-hour energy expenditure in man: methods and results using a respiratory chamber. J. Clin. Inuest. 78: 1568-1578,1986. 13. RAVUSSIN, E., S. LILLIOJA, W. C. KNOWLER, L. CHRISTIN, D. FREYMOND, W. G. H. ABBOTT, V. BOYCE, B. V. HOWARD, AND C. BOGARDUS. Reduced rate of energy expenditure as a risk factor for body-weight gain. N. Engl. J. Med. 318: 467-472, 1988. 14. ROMIEU, I., W. C. WILLETT, M. J. STAMPFER, G. A. COLDITZ, L. SAMPSON, B. ROSNER, C. H. HENNEKENS, AND F. E. SPEIZER. Energy intake and other determinants of relative weight. Am. J. Clin. Nutr. 47: 406-412, 1988. 15. SALMON, D. M. W., AND J. P. FLATT. Effect of dietary fat content on the incidence of obesity among ad libitum fed mice. Int. J. Obesity 9: 443-449, 1985. 16. SCHWARTZ, R. S., E. RAVUSSIN, M. MASSARI, M. O’CONNELL, AND D. C. ROBBINS. The thermic effect of carbohydrate versus fat feeding in man. Metabolism 34: 285-293, 1985. 17. TREMBLAY, A., G. PLOURDE, J-P. DESPRES, AND C. BOUCHARD. Impact of dietary fat content and fat oxidation on energy intake in humans. Am. J. Clin. Nutr. 49: 799-805, 1989. 18. WHO EXPERT COMMITTEE ON DIABETES MELLITUS SECOND REPORT. World Health Organization Tech. Rep. Ser. 646. Geneva: WHO, 1980.
Downloaded from www.physiology.org/journal/ajpendo at Macquarie Univ (137.111.162.020) on February 12, 2019.
G. H. ABBOTT, BARBARA V. HOWARD, RUOTOLO, AND ERIC RAVUSSIN
Clinical Diabetes and Nutrition Section, National Institute of Diabetes, Digestive and Kidney Diseases, National Institutes of Health, Phoenix, Arizona 85016
ABBOTT, WILLIAM G. H., BARBARA V. HOWARD, GIACOMO RUOTOLO, AND ERIC RAVUSSIN. Energy expenditure in hurnarw: effects of dietary fat and carbohydrate. Am. J. Physiol.
METHODS
258 (Endocrinol. Metab. 21): E347-E351, 1990.-A high-dietary fat intake may be an important environmental factor leading to obesity in some people. The mechanism could be either a decrease in energy expenditure and/or an increase in caloric intake. To determine the relative importance of these mechanisms we measured 24-h energy expenditure in a whole body calorimeter in 14 nondiabetic subjects and in six subjects with non-insulin-dependent diabetes mellitus, eating isocaloric, weight-maintenance, high-fat, and high-carbohydrate diets. All subjects were Pima Indians. In nondiabetics, the mean total 24-h energy expenditure was similar (2,436 k 103 vs. 2,359 k 82 kcal/day) on high-fat and high-carbohydrate diets, respectively. The means for sleeping and resting metabolic rates, thermic effect of food, and spontaneous physical activity were unchanged. Similar results were obtained in the diabetic subjects. In summary, using a whole body calorimeter, we found no evidence of a decrease in 24-h energy expenditure on a highfat diet compared with a high-carbohydrate diet.
The subjects for this study were 20 Pima Indians. All subjects had a full medical history and physical examination, with routine hematology and biochemistry screens and urinalysis. Six subjects had noninsulin-dependent diabetes mellitus (NIDDM) by World Health Organization criteria (18). At the time of the study none of the subjects had evidence of significant disease apart from NIDDM, and none had significant complications of NIDDM. The physical characteristics of the subjects are shown in Table 1 and plasma glucose concentrations in Table 2.
Subjects
Protocol
The subjects resided on the metabolic ward for the duration of the studies. Activity was restricted to the ward, and alcohol consumption was strictly forbidden. The ‘subjects were admitted to the research ward on obesity; non-insulin-dependent diabetes mellitus two separate occasions at least 4 wk apart, eating a different diet on each occasion. Eleven subjects ate the high-fat diet on the first admission, and nine subjects ate the high-carbohydrate diet on the first admission. POPULATION STUDIES SHOW that a high-dietary fatintake is associated with obesity in humans (5, 14, 17). The measurement of energy expenditure was made after There are supportive data in animal studies (11, 15). 18 & 2 (mean & SE) days (range 6-32 days) on the highfat diet and after 27 & 2 days (range 5-43 days) on the Thus a high-fat intake may be an important environhigh-carbohydrate diet. mental factor promoting obesity in some individuals. During the 24-h period in the chamber, the caloric Because obesity results from a positive energy balance, i.e., energy intake exceeds energy expenditure, there are intake of the subjects was reduced to 80% of that given on the research ward. This was to allow for the reduction two possible mechanisms for this association. The first in calories expended in spontaneous physical activity is that high-fat diets may increase caloric intake (11,17). because of confinement in the ,small chamber (8.2 m2). The second possibility is that high-fat diets may lower energy expenditure (6, 8, 16), which has been shown to Because it is possible that the effect of diet on energy expenditure could be most marked during physical activpredict weight gain (13). ity, five subjects spent an additional 24-h period in the To determine the relative importance of these mechchamber on each diet. Energy intake was increased to anisms, 20 Pima Indian subjects were admitted to a the same number of calories normally consumed on the metabolic ward and were fed isocaloric high-fat and highresearch ward, and energy expenditure was increased to carbohydrate diets. The prevalence of obesity in the Pima an equivalent level by use of an exercycle (see below). Indian population is high (7). The 24-h energy expenditure was measured on each diet in a whole body indirect calorimeter. This is a sealed room with a floor area of Diets 8.2 m2, in which restricted activity is possible. Energy Weight-maintenance caloric intake was estimated expenditure under these conditions was similar on the from body weight on admission to the research ward and high-fat and the high-carbohydrate diets. was adjusted on the basis of fluctuations in the fasting, E347 Downloaded from www.physiology.org/journal/ajpendo at Macquarie Univ (137.111.162.020) on February 12, 2019.
E348 TABLE
DIETARY
1. Physical characteristics
NIDDM
of subjects in the nondiabetic
and diabetic groups W k
Sex
n
Nondiabetic
FAT AND THERMOGENESIS
WV
14
1212
6
4/z
35A3 (21-57) 41*4 (24-49)
Body Fat, %
HIFAT
HICHO
HIFAT
HICHO
100*7 (63-157) 88*10 (65-136)
lOOk7 (64-153) 87klO (65-133)
29*2 (16-42) 34*4 (22-48)
3Ok2 (14-45) 34*3 (21-46)
Values are means * SE, ranges are in parentheses; n = 20 subjects. HIFAT, high-fat diet; HICHO, high-carbohydrate insulin-dependent diabetes mellitus.
2. Fasting and 2-h plasma glucose concentrations in diabetic and nondiabetic groups
TABLE
Nondiabetic HIFAT
Fasting glucose, w/d1 2-h glucose,
93k3 (77-109) 108k7
HICHO
93Ik3 (77-106) 115*9
HICHO
215&39 (99-344)
216&40 (94-336)
368k44 (234-500)
345k51
(194-480) Values are means & SE, ranges in parentheses; n = 20 subjects. NIDDM, non-insulin-dependent diabetes mellitus; HIFAT, high-fat diet; HICHO, high-carbohydrate diet. w/d1
(70-157)
(50-178)
posturination body weight measured each morning. The composition of the high-fat and high-carbohydrate diets is summarized in Table 3. The fat in the high-fat diet was derived predominantly from dairy products. No subjects developed clinical evidence of lactose intolerance on the high-fat diet. On the high-carbohydrate diet, fat was replaced by predominantly complex carbohydrate of vegetable, legume, and cereal origin, resulting in an increase in fiber intake. Meals were made from readily available foods purchased from a local supermarket and were freshly and individually prepared using standardized recipes with weighed and/or measured ingredients (Mettler PG balance, &l g, Mettler, Greifensee, Switzerland). Diet composition was calculated with a computerized diet analysis program (the CBORD group: CBORD’s Diet Analyzer System, ESHA Database, 1986; Ithaca, NY) as well as manufacturer specifications. Procedures Body composition. Body composition was estimated by underwater weighing with simultaneous measurement of lung volume by helium dilution (13). OraZ gZucose toZerance test. After at least 4 days on the research ward and after a 12-h overnight fast, an oral glucose tolerance test was performed with 75 g of glucose (Koladex, Custom Laboratories, MD). Blood was drawn at -5 min and at 120 min for measurement of the plasma glucose concentration (Beckman glucose analyzer, Fullerton, CA). Energy expenditure. The 24-h energy expenditure was measured in a closed-circuit indirect calorimetry chamber that has previously been described (12). The chamber is a sealed room with a volume of 19,000 liters through which atmospheric air is continuously drawn. The subject’s gas exchange is calculated by measuring the OZ and CO2 concentrations at the inlet and outlet of the chamber and by measuring the flow of air through the chamber. Spontaneous physical activity was measured in 14 non-
non-
TABLE 3. composition (% of total calories) of the high-fat and high-carbohydrate diets
NIDDM HIFAT
diet; NIDDM,
High Fat
Carbohydrate, % Simple, % Complex, % Fat, % Polyunsaturated, % Monounsaturated, % Saturated, % Protein, % Fiber, g /l,OOO kcal Cholesterol, mg/day Nonprotein respiratory quotient of diet
43
High Carbohydrate
22
65 20 45 20 6 8 6
15 10
15 19
14 29 42 7
13
560 0.83
524
0.91
diabetic and five of the six diabetic subjects by two microwave motion sensors (MICD 930, Honeywell, Minneapolis, MN). These units emit a signal, the frequnecy of which is ahered by movement (Doppler effect), and the change is measured by a transreceiver. The sensitivity of this system was set to detect any movement greater than breathing. One activity unit, expressed as percent, represents the percentage of time that the subject was moving, and these units are independent of work intensity. After an overnight fast and after urinating, each subject entered the chamber at 0750 h and was confined until 0700 h the next morning. Vigorous exercise and cigarette smoking were prohibited. Meals were passed through a small air lock at 0800, 1115, and 1615 h, with a snack at 1900 h. Subjects were carefully instructed to eat all provided food and to return any uneaten food to the kitchen so the caloric intake could be computed. Subjects were instructed to collect all urine, and this was used to measure 24-h urinary glucose loss and 24-h urea nitrogen production. At 0700 h the next morning, the chamber door was opened, and while the subject was lying on the bed, a clear plastic hood was placed over his/her head. After 25 min of becoming acclimatized to the hood, the resting metabolic rate was measured for between 9 and 12 min, with the hood connected to the analyzing equipment of the chamber. Five subjects spent an additional day in the chamber on each diet, during which time caloric intake wasmaintained at the weight maintenance level on the research ward, and energy expenditure in the chamber was increased to match energy intake by use of an Exercycle (Collins Pedelmate, Warren E. Collins, Braintree, MA). The exercise was divided into four periods of equal length and was carried out at a rate of 75 W. The’ duration of
Downloaded from www.physiology.org/journal/ajpendo at Macquarie Univ (137.111.162.020) on February 12, 2019.
DIETARY
FAT
AND
the exercise differed for each subject and was calculated on the basis of previously determined energy expenditure at 50,100, and 150 W. Energy Expenditure
Calculations
Oz uptake (vo2) and CO2 (VCO~) production rates were calculated from the difference between the atmospheric air and the outflow Oz and CO2 concentrations, corrected for the changes in Oz and CO2 concentrations in the chamber. Values were corrected for temperature, pressure, and humidity and expressed as STPD. Energy expenditure (EE in kcal/min) during each 15min period was calculated as follows (12) EE = Voz* [4.686 + (RQ-0.707).
(0.361/0.293)]
in which RQ is the respiratory quotient, i.e., the ratio between CO2 production and 02 consumption during the 15min period. Twenty-four-hour energy expenditure was calculated as the average energy expenditure of 92, 15-min periods extrapolated to 24 h. Sleeping metabolic rate was calculated as the average energy expenditure of all 15-min Periods between 2330 and 0500 h, during which the spontaneous physical activity measured by radar was cl%. The average cost of physical activity per unit of measured activity (%) was calculated as the slope of the linear regression line between activity (%) during a 15-min period and the energy expenditure (kcal/min) during that period. The amount of energy expended per 24 h due to spontaneous physical activity was calculated as the product of the energy cost per unit activity (the slope of the above regression line) and the mean 24-h activity, and this product was extrapolated to 24 h. The thermic effect of food was estimated by integrating the energy expenditure corrected for activity above sleeping metabolic rate, and was expressed as a percentage of the 24-h caloric intake. We used the sleeping metabolic rate rather than the resting metabolic rate to calculate the thermic effect of food because we find it is a more reproducible measurement. The measurement of thermic effect of food thus includes the energy cost of arousal. I
24 hr EE
I I I
I I
Resting
metabolic
Sleeping
metabolic
I -25
I
I
I
I
I
-20
-15
-10
-5
oh Decrease on 43% Fat Diet
Energy
I I
I
+5
+lO
I +15
effect
cost per unit activity
Calories
I 0
rate
Thermic of food
I I
rate
expended
I
I
+20
+25
in activity
oh Increase on 43% Fat Diet
FIG. 1. Confidence limits (95%) for mean changes in energy expenditure (EE) on high-fat compared with high-carbohydrate diet in 14 nondiabetic subjects.
E349
THERMOGENESIS
Statistics
All data are expressed as the means & SE. The Kolomogorov D statistic was used to test the normality of distributions, and those variables with nonnormal distributions (sleeping metabolic rate and 24-h respiratory quotient) were logarithmically transformed. The paired t test was used to assess the significance of differences between means on the two diets. Correlation analysis was done by means of the Pearson product-moment correlation coefficient. These statistical analyses were calculated with the Statistical Analysis System (SAS), SAS Institute, Cary, NC. The 95% confidence intervals for the changes in mean energy expenditure on the high-fat diet were calculated to show the magnitude of type 2 error (4). The 95% confidence intervals are shown in Fig. 1. RESULTS
Nondiabetic
Subjects
In the nondiabetic group there was no decrease in either 24-h energy expenditure or in any component of 24-h energy expenditure on the high-fat compared with the high-carbohydrate diet (Table 4). The 95% confidence intervals for the changes in the means are shown in Fig. 1. These are estimates of type 2 error. The purpose of this figure is to show the largest mean decrease in energy expenditure on the high-fat diet that would be consistent with our data. Thus, given the limits of the experimental conditions, we could exclude the possibility that the high-fat diet induced a decrease in mean 24-h energy expenditure of >0.8%, a decrease in mean resting metabolic rate of >6.3%, and a decrease in mean sleeping metabo1ic rate Of >2e5%* There were no correlations between the weight or body composition of the subjects and any of the individual changes in energy expenditure. There was a positive correlation between the duration of the high-fat diet and the measurement of physical activity (r = 0.70, P c 0.01) on the high-fat diet. There were no correlations between the duration of the high-carbohydrate diet and any measurements of energy expenditure made on this diet. Table 5 shows that mean energy balance (food intake - 24-h energy expenditure) tended to be negative on both diets but was not significantly different from zero on either diet. Both the resting respiratory quotient and the 24-h respiratory quotient were significantly higher on the high-carbohydrate diet than on the high-fat diet (Table 4). On the high-fat diet the 24-h respiratory quotient was similar to the respiratory quotient of the food. On the high-carbohydrate diet the 24-h respiratory quotient was significantly less than the respiratory quotient of the food (p c 0.001). Five subjects had a 2nd day in the chamber on each diet, with a 25% increase in caloric intake and an equivalent increase in physical activity. The purpose was to determine whether the failure to find a decrease in energy expenditure on the high-fat diet might be due to the decrease in activity that occurs with confinement in the chamber. The addition of physical exercise caused in-
Downloaded from www.physiology.org/journal/ajpendo at Macquarie Univ (137.111.162.020) on February 12, 2019.
E350
DIETARY
4. Measurements
TABLE
FAT
of energy expenditure
AND
THERMOGENESIS
on the high-fat
and high-carbohydrate
diets
Nondiabetic HIFAT 24-h energy expenditure, kcal/day Resting metabolic rate, kcal/min Sleeping metabolic rate, kcal/min Energy cost/unit activity, kcal min. % activity-’ Physical activity, kcal/day Thermic effect of food, % caloric intake*
NIDDM HICHO
HIFAT
HICHO
2,435&103 1.38kO.06 1.2220.06 0.032&0.002
2,359*82 1.36kO.06 1.17kO.04 0.031*0.002
2,169*207 1.29kO.14 1.17kO.11 0.029~0.004
2,128&185 1.15*0.10 1.11&0.07 0.026&0.002
416&38 19.6k2.3
384&36 19.2kl.7
329k46 16.4kl.6
309k30 14.6kl.7
l
Values are means & SE; nondiabetic n = 14, NIDDM, high-carbohydrate diet. * Energy expenditure corrected
n = 6. NIDDM, non-insulin-dependent for activity above sleeping metabolic
rate
diabetes mellitus; over 24 h.
HIFAT,
high-fat
diet; HICHO,
5. Caloric intake, energy expenditure, and respiratory quotients on high-fat and high-carbohydrate diets in nondiabetic and diabetic subjects
TABLE
Nondiabetic
NIDDM
HIFAT 24-h energy expenditure, kcal/day Food intake, kcal/day Energy balance, kcal/day 24-h urinary glucose loss, g/day R*Q of diet 24-h RQ R*esting RQ Values are means & SE; nondiabetic high-carbohydrate diet; RQ, respiratory
HICHO
2,435&103 2,361&113 -75*63 0.2kO.04 0.83 0.84&0.01 0.82&0.01
n = 14, NIDDM quotient. * P
HIFAT
2,359&82 2,329&102 -30&42 0.3*0*04 0.91 0.88*0.01* 0.84&0.01?
n = 6. NIDDM, non-insulin-dependent T P c 0.01.
HICHO
2,169&207 2,334&212 +164&116 55223 0.83 0.82kO.01 0.77&0.01 diabetes
2,128&185 2,316&210 +189&101 104*34 0.91 0.85*0.02? 0.78&0.01
mellitus;
HIFAT,
high-fat
diet;
HICHO,
< 0.001,
6. Effect of a 20% increase 24-h energy expenditure and isocaloric increase in energy intake on thermogenesis and respiratory quotient in nondiabetic subjects
TABLE
Without
24-h energy intake, kcal/day 24-h energy expenditure, kcal/day Energy balance, kcal/day Mean 24-h RQ Values compared
are means & SE; n = 5 subjects. with food quotient.
HIFAT,
Exercise
With
Exercise
HIFAT
HICHO
HIFAT
HICHO
2,280*75 2,363&110 -83&47 0.83kO.02
2,240&74 2,382&88 -142272 0.87&0.01*
2,820&78 2,960&107 -140&55 0.84kO.01
2,834&74 3,025*91 -192&44 0.89kO.01
high-fat
diet;
HICHO,
creases in energy expenditure of 597 & 81 and 643 & 118 kcal/day for the high-fat and high-carbohydrate diets, respectively. The 24-h energy expenditure was not decreased with exercise on the high-fat compared with the high-carbohydrate diet (Table 6). The 24-h respiratory quotient was similar to the respiratory quotient of the food during the exercise protocol on the high-carbohydrate diet. Diabetic Subjects
In the diabetic group the high-fat diet did not result in a decrease in either 24-h energy expenditure or in any of the components of 24-h energy expenditure (Table 4). DISCUSSION
There is evidence that a high-dietary fat intake is an important environmental factor that leads to obesity in some people. Whether dietary fat-induced obesity is due to a decrease in metabolic rate, an increase in caloric intake, or a combination of both factors is unclear.
high-carbohydrate
diet;
RQ, respiratory
quotient.
*
P < 0.05 24-h RQ
Animal studies have not consistently favored either mechanism (11). The purpose of this study was to determine whether a high-fat diet could reduce 24-h energy expenditure in a group of Pima Indian subjects. There was no decrease in the mean of either the total 24-h energy expenditure or of any components of energy expenditure on the high-fat diet in either the nondiabetic or the diabetic subjects. Given the limits of the experimental conditions, the 95% confidence intervals for the changes in the means suggest that the maximum decrease in 24-h energy expenditure induced by the high-fat diet was 0.8% or -18 kcal/day. A decrease in energy expenditure of this magnitude would not result in significant obesity (9). The data do not totally exclude the possibility that high-fat diets contribute to obesity by reducing energy expenditure. The duration of the acclimatization periods may have been too short for the effect to become apparent. The metabolic chamber restricts the activity of the subjects, which would make a dietary effect on activityrelated energy expenditure more difficult to detect. The
Downloaded from www.physiology.org/journal/ajpendo at Macquarie Univ (137.111.162.020) on February 12, 2019.
DIETARY
FAT AND THERMOGENESIS
majority of the subjects were obese, and it is possible that the effect of dietary fat on energy expenditure is only important in preobese subjects. It is also possible that a high-fat diet-induced decrease in energy expenditure is only apparent with overfeeding (3) compared with weight-maintenance feeding. Lastly, the subjects were all Pima Indians and the results may not apply to other racial groups. The effect of high- and low-fat diets on energy expenditure has been investigated by three other groups. Lean and James (8) found no change in 24-h energy expenditure in groups of lean and obese women fed weight-maintenance diets for 24 h. Hurni et al. (6) compared high-fat and high-carbohydrate diets fed to lean, young subjects for 7 days. The subjects were in negative caloric balance at the time of the energy-expenditure measurement. The sleeping metabolic rate was decreased by the high-fat diet, although 24-h energy expenditure and resting metabolic rates were unchanged. McNeil1 et al. (10) found that the resting metabolic rate in women was the same on high-fat and high-carbohydrate diets fed for 5 days. A decrease in the thermic effect of food was expected on the high-fat diet (8,16). This would be because of the greater obligatory cost of glycogen storage (2) and possibly to a facultative increase in thermogenesis due to glucose stimulation of the sympathetic nervous system (1). We have found that the measurement of thermic effect of food in the metabolic chamber is not very reproducible; and a substantially larger number of subjects would have been needed to estimate this parameter accurately. A decrease in thermic effect of food may have been counterbalanced by small increases in other components of 24-h metabolic rate, resulting in no change in total daily energy expenditure. The 24-h respiratory quotient was less than the respiratory quotient of the food on the high-carbohydrate diet. This could have been caused by three factors. The first is the decrease in activity that occurs during the day in the chamber. Second, there was a tendency for the subjects to be in negative energy balance while in the chamber. Third, the subjects may not have fully acclimatized to the high-carbohydrate diet in 2-6 wk. In summary, using a whole-body calorimeter, we could find no evidence that a weight-maintenance, high fatdiet reduces 24-h energy expenditure in humans. The data suggest that if high-fat diets do predispose to obesity, then the mechanism is either an increase in caloric intake or a decrease in physical activity. We thank Carol dietary staff of the National Institutes Adams for statistical
Lamkin and Vicky Boyce and the nursing and Clinical Diabetes and Nutrition Section of the of Health, Phoenix, AZ. We also thank Beverly assistance, as well as all the volunteers.
E351
Address for reprint requests: W. G. H. Abbott, Dept. of Human Nutrition, Southwestern Medical School, 53‘23 Harry Hines Blvd., Dallas, TX 75235. Received 21 March 1989; accepted in final form 27 September 1989. REFERENCES 1. ACHESON, K. J., E. RAVUSSIN, J. WAHREN, AND E. JEQUIER. Thermic effect of glucose in man: obligatory and facultative thermogenesis. J. Clin. Inuest. 74: 1572-1580, 1984. 2. ACHESON, K. J., Y. SCHUTZ, T. BESSARD, E. RAVUSSIN, AND E. JEQUIER. Nutritional influences on lipogenesis and thermogenesis after a carbohydrate meal. Am. J. Physiol. 246 (Endocrinol. Metub. 9): E62-E70, 1984. E. Diet and obesity. Am. J. CZin. Nutr. 41: 1132-1145, 3. DANFORTH, 1985. 4. DETSKY, A. S., AND D. L. SACKETT. When was a negative clinical trial big enough. Arch. Intern. Med. 145: 709-712, 1985. 5. DREON,D. M.,B. FREY-HEWITT, N.ELLSWORTH, P. T. WILLIAMS, R. B. TERRY, AND P. D. WOOD. Dietary fat: carbohydrate ratio and obesity in middle-aged men. Am. J. CZin. Nutr. 47: 9954000, 1988. 6. HURNI, M., B. BURNAND, PH. PITTET, AND E. JEQUIER. Metabolic effects of a mixed and a high-carbohydrate diet in man, measured over 24 hours in a respiration chamber. Br. J. Nutr. 47: 33-43, 1982. 7. KNOWLER, W. C., D. J. PETTITT, P. H. BENNETT, AND R. C. WILLIAMS. Diabetes mellitus in the Pima Indians: genetic and evolutionary considerations. Am. J. Whys. Anthropology 62: 107114,1983. 8. LEAN, M. E. J., AND W. P. T. JAMES. Metabolic effects of isoenergetic nutrient exchange over 24 hours in relation to obesity in women. Int. J. Obesity 12: 15-27, 1988. 9. LEW, E. A., AND L. GARFINKEL. Variations in mortality by weight among 750,000 men and women. J. Chron. Dis. 32: 563-576,1978. 10. MCNEILL, G., A. C. BRUCE, A. RALPH, AND W. P. T. JAMES. Interindividual differences in fasting nutrient oxidation and the influence of diet composition. Int. J. Obesity 12: 455-463, 1988. 11. OSCAI, L. B., M. B. BROWN, AND W. C. MILLER. Effect of dietary fat on food intake, growth and body composition in rats. Growth 48: 415-424,1984. 12. RAVUSSIN, E., S. LILLIOJA, T. E. ANDERSON, L. CHRISTIN, AND C. BOGARDUS. Determinants of 24-hour energy expenditure in man: methods and results using a respiratory chamber. J. Clin. Inuest. 78: 1568-1578,1986. 13. RAVUSSIN, E., S. LILLIOJA, W. C. KNOWLER, L. CHRISTIN, D. FREYMOND, W. G. H. ABBOTT, V. BOYCE, B. V. HOWARD, AND C. BOGARDUS. Reduced rate of energy expenditure as a risk factor for body-weight gain. N. Engl. J. Med. 318: 467-472, 1988. 14. ROMIEU, I., W. C. WILLETT, M. J. STAMPFER, G. A. COLDITZ, L. SAMPSON, B. ROSNER, C. H. HENNEKENS, AND F. E. SPEIZER. Energy intake and other determinants of relative weight. Am. J. Clin. Nutr. 47: 406-412, 1988. 15. SALMON, D. M. W., AND J. P. FLATT. Effect of dietary fat content on the incidence of obesity among ad libitum fed mice. Int. J. Obesity 9: 443-449, 1985. 16. SCHWARTZ, R. S., E. RAVUSSIN, M. MASSARI, M. O’CONNELL, AND D. C. ROBBINS. The thermic effect of carbohydrate versus fat feeding in man. Metabolism 34: 285-293, 1985. 17. TREMBLAY, A., G. PLOURDE, J-P. DESPRES, AND C. BOUCHARD. Impact of dietary fat content and fat oxidation on energy intake in humans. Am. J. Clin. Nutr. 49: 799-805, 1989. 18. WHO EXPERT COMMITTEE ON DIABETES MELLITUS SECOND REPORT. World Health Organization Tech. Rep. Ser. 646. Geneva: WHO, 1980.
Downloaded from www.physiology.org/journal/ajpendo at Macquarie Univ (137.111.162.020) on February 12, 2019.
