I From the Department of Human Biology. University of Limburg.Maastricht. The Netherlands.

2 Address reprint requests to KR Westerterp. Department of HumanBiology. University of Limburg. P0 Box 616. 6200 MD, Maastricht.

The Netherlands.

.Ini J (l/ii \zuir 1993:57(suppl):759S-65S. Printed in USA. � 1993 American Society for Clinical Nutrition 759S

Food quotient, respiratory quotient, and energy balance1’2

Klaas R H ‘e.sterierp

ABSTRACT This paper reviews evidence that the macro-

nutrient composition ofthe diet and the maintenance of energy

balance are correlated. Intervention studies show that subjects

lose weight on low-fat diets and gain weight on high-fat diets.

Descriptive studies show that overweight subjects eat relatively

more fat but have the same total energy intake as nonoverweight

subjects. The body has a limited ability to oxidize fat compared

with its ability to oxidize carbohydrate and protein. The con-

elusion is that becoming overweight can be prevented by reducing

the fat content of the diet. Studies on nutrient utilization show

a ready increase in carbohydrate oxidation whereas fat oxidation

does not change after meals enriched with, respectively, carbo-

hydrate or fat. However, in the long term, the respiratory quotient

(RQ) is closer to the food quotient (FQ) for subjects eating high-

fat diets than it is for subjects eating high-carbohydrate diets.

For high-carbohydrate diets, the RQ is lower than is the FQ.

indicating that subjects must mobilize body fat. This is supported

by data on body weight loss in subjects changing from a standard

maintenance diet to a low-fat diet, even while energy intake was

increased with nearly 20% . Direct evidence for a higher energy

expenditure for low-fat diets is not yet available. ..1,;i J (liii

Nutr I 993:57(suppl):759S-65S.

KEY WORDS Food quotient, respiratory quotient, energy

balance, diets, carbohydrate, protein. fat

Introduction

Humans consume food to maintain energy balance. Food en-

ergy is consumed in the form ofcarbohydrate, protein, and fat,

the macronutrients of our diet. The distribution of food energy

between the macronutrients differs between cultures and coun-

tries. There is a fairly wide range of carbohydrate-protein-fat

ratios (c:p:f) at which energy balance can be maintained. as

shown by nationwide studies of nutrient intake (Table I). The

carbohydrate intake ranges from 3 to 82 en% and the fat intake

ranges from 6 to 54 en#{176}%. whereas the protein intake is at least

1 1 en%. Within one culture or country. dietary changes take

place with changes in food supply. There is an increasing avail-

ability of fats in Western countries. In the United States the

contribution of fats to the diet has steadily increased from 32

en% in 1910 to 43 en% in 1985 (Fig I). The shift in the com-

position of the diet to a higher contribution of fats has often

been quoted as the reason for the increasing incidence of over-

weight. ie, of a positive energy balance. This paper will discuss

some of reasons and evidence for the correlation between the

macronutnient composition ofthe diet and the maintenance of

energy balance.

Dietary fat-carbohydrate ratio

Eneigi balance

Dietary fat is the main determinant ofthe energy density of

our diet. The metabolizable energy for dietary carbohydrate.

protein. and fat is. respectively. 16. 16. and 37 kJ/g (3). Fat has

an important function as an energy depot in the body as well.

Fat can be stored with minimal additional weight. The energy

density ofthe fat stores is approximately eight times higher than

is the energy density of the carbohydrate (glycogen) stores (4).

Thus. in most circumstances energy balance can be reached with

the minimum bulk consuming high-fat diets. and an energy sur-

plus is mainly stored as body fat.

Some studies show that people change energy intake when

they change to a diet with a lower or a higher energy density.

Duncan et al (5) allowed subjects to eat to satiety from a diet

low in energy density (3 kJ/g) and from a diet high in energy

density (6.5 kJ/g). Each diet was provided for 5 d in a randomized

cross-over design with a weekend in between. Subjects ate nearly

twice as much on the high-energy diet (12.5 MJ/d), or. con-

versely. twice as little on the low-energy diet (6.5 MJ/d). Sur-

prisingly. there was no trend toward a higher energy intake on

subsequent days on the low-energy diet or to a lower energy

intake on subsequent days on the high-energy diet. Lissner et al

(6) provided subjects with three different diets. one with a low.

one with a medium. and one with a high energy density by

exchanging carbohydrate and fat. All subjects got each diet for

14 d in a balanced sequence. Energy intake increased from a

mean value of 8.7 MJ/d on the low-fat diet ( 1 5-20 en% fat) to

9.8 MI/d on the medium-fat diet (30-35 en% fat) and to I 1.4

MJ/d on the high-fat diet (45-50 en% fat). Again. there was no

systematic trend ofintake over time in any ofthe subjects during

any 14-d dietary treatment. indicating that there was no adap-

tation to the diet during the observation period. On average,

subjects were in energy balance on the medium-fat diet. lost

weight on the low-fat diet. and gained weight on the high-fat

diet.

Animals like the laboratory rat compensate in time ifthe en-

ergy density of the food is changed to achieve the same energy

intake and reach the same hod�’ weight as controls (7). The failure

by guest on March 9, 2014

ajcn.nutrition.orgD

ownloaded from

* Adapted from reference I.

50’

40’

30

C

a)

a)

a)

0

C0

C00

.- -4�-- � S.

20

10

01920 I 940 1 960 1 980

year

760S WESTERTERP

TABLE 1

Ratio of macronutrients in the diet of people of different countries*

Country Carbohydrate Protein Fat

en%

Nigeria 82 12 6

Japan 77 12 11

India 77 11 12CSSR 59 11 30

United States 46 12 42Greenland 3 43 54

of humans to compensate for a change in the energy density of

the food may be a way for them to avoid becoming overweight.

Energy dilution with carbohydrate or fat analogues like aspar-

tame (8, 9) and sucrose polyesthers has its application in the

new market for “light” products. On the other hand. there are

indications that humans compensate for energy dilution when

not all high-energy-density food items are reduced-energy ver-

sions by increasing the consumption of nonmanipulated items

(10). Subjects failed to compensate for increases in energy intake

by covertly substituting part ofthe food items with high-energy

versions.

Obesiti

There are quite a few studies on the macronutrient ratio of

the diet and obesity or being overweight. Here we restrict our-

selves to studies where subjects were allowed to freely choose

their own foods under ordinary living conditions. Under those

circumstances, overweight subjects take diets with a higher energy

density. Jiang and Hunt ( 1 1) asked 1 1 adult men to collect a

double portion of all they ate over 7 d for analysis. The energy

density of the diet, including drinks, ranged from 1 .8 kJ/g in

normal-weight subjects to 3.9 kJ/g in overweight subjects. Sun-

prisingly, the researchers mention that the energy density of the

diet in overweight subjects was not higher because of a higher

fat content, and no alternative explanation is presented. Dreon

et al ( 12) related diet composition, as measured with 7-d food

records, to body composition. as measured with hydrostatic

weighing, in 1 55 middle-aged men. Subjects with a higher per-

centage body fat consumed a diet with relatively more fat and

less carbohydrate. Tremblay et al ( 1 3) performed a comparable

study in 244 male adults, measuring diet composition with a 3-

d food record. They also found that the en% fat of the diet was

positively correlated and the en% carbohydrate was negatively

correlated with the fatness of the subjects. Finally, Miller et al

(14) made similar observations in 2 16 adult subjects, 50% men

and 50% women, and found that adiposity was positively con-

related with dietary fat content and negatively correlated with

dietary carbohydrate consumption. In all four studies mentioned

above ( 1 1 - I 4), there was no correlation between energy intake

and indexes for overweight or obesity.

There have been indications that self-reported intake tends to

be an underestimate of true habitual intake. Reported intakes

tend to be lower than energy expenditure as measured simul-

taneously with doubly labeled water, especially in obese subjects

and subjects with a high energy intake. such as those engaged

in endurance exercise (15, 16). We do not know whether the

results of those studies, mainly using self report to measure di-

etary intake. were influenced by this phenomenon. On the other

hand. it is nearly impossible to measure habitual intake while

subjects freely choose their own foods under ordinary living

conditions with a more accurate alternative technique. The

double-portion technique as used by Jiang and Hunt ( 1 1) is one

ofthe best but is not really feasible in large groups and interferes

with daily routines when subjects are not at home every time

they eat something. Thus, Romieu et al ( 17) and Miller (18)

came to the conclusion that the role offat intake in obesity may

be independent of energy intake.

Substrate utilizatioui

Knowing that obese subjects generally do not eat more than

do normal-weight subjects but do differ from normal-weight

subjects with respect to the fat-carbohydrate ratio of their diet,

the next question is. why is an isoenergetic diet fattening when

it contains relatively more fat and less carbohydrate? In terms

ofenergy, we can make carbohydrate from protein and fat from

carbohydrate and thus fat and carbohydrate are not essential

nutrients. However the conversion processes are very energy

consuming and produce a lot ofwaste products such as ammonia

and urea. Ideally. the body covers its energy needs with a mixture

of fat and carbohydrate. Some tissues preferentially use carbo-

hydrate whereas fat is less bulky to consume, as mentioned ear-

her. Additionally, humans are periodic eaters and continuous

metabolizers (ie. part ofthe energy intake is stored before usage

as well).

The storage capacities for fat and carbohydrate are very dif-

ferent and this may have consequences for the regulation of

body weight. Flatt ( 19) proposed a model with a regulated car-

bohydrate (glycogen) store and with the fat store as a function

of nutrient intake and nutrient utilization. An increase in the

fat content of the diet needs an increase in fat oxidation and it

is hypothesized that the latter needs an expansion of the body-

FIG I . Distribution of dietary energy intake between carbohydrate(H). protein (#{149}).and fat (5) in the United States from 1910 to 1985(from ref 2).

by guest on March 9, 2014

ajcn.nutrition.orgD

ownloaded from

FQ, RQ, AND ENERGY BALANCE 76 1S

fat mass, explaining the increased adiposity in subjects consum-

ing high-fat diets.

Several groups studied the effect ofa change in nutrient intake

on nutrient utilization. Nutrient utilization can be measured

with indirect calorimetry from oxygen consumption, carbon

dioxide production, and urinary nitrogen excretion (20). Some-

times protein oxidation is measured by primed infusion of a

‘3C-labeled amino acid: measuring the elimination ofthe label

in respiratory gas provides a more rapid, more responsive index

(2 1). Studies of nutrient utilization can be split in short-term

studies. usually measuring the effects ofa single meal, and long-

term studies. covering at least one 24-h cycle on a fixed dietwith or without an adaptation period beforehand. The indirect-

calorimetry system in short-term studies is a ventilated hood,

with subjects lying or sitting. Long-term studies are performed

in a respiration chamber that allows subjects to move around

on 10-20 m2 offloor space.

Short-term studies of nutrient utilization

Short-term studies of nutrient utilization are usually started

in the postabsorptive condition after an overnight fast. Food is

consumed as a breakfast, after which baseline measurements

and observations are continued for several hours. Acheson et al

(22) measured the effect of a large carbohydrate meal (9 Mi.

Table 2) on nutrient utilization. The high-carbohydrate load

lowered the fat oxidation rate: however. there were no indications

of the conversion of carbohydrate to fat. Assuming complete

processing of the meal in the subsequent 10-h observation in-

terval, the storage capacity for carbohydrate as glycogen was

probably sufficient to accommodate the intake surplus because

only 1 33 g carbohydrate was oxidized. In their next experiment.

Acheson et al (23) controlled the diet over the 3-6 d preceding

a test comparable to the one mentioned above. Subjects con-

sumed a low-carbohydrate diet, a mixed diet, or a high-carbo-

hydrate diet. The higher the carbohydrate content ofthe initial

diet. the higher the carbohydrate oxidation rate after the car-

bohydrate meal: 177 ± 5. 241 ± 11, and 258 ± 9 g over the 14

h after consumption of 500 g in the low, medium, and high

carbohydrate group, respectively. Carbohydrate oxidation was

higher than it was in the first experiment (22), possibly because

of the use of a different dietary carbohydrate. Again, most of

the carbohydrate surplus was stored as glycogen, although there

was a significant lipid synthesis from carbohydrate of 3.4 ± 0.6

and 9.0 ± 1 .0 g in the medium- and high-carbohydrate groups,

respectively. However. the limited lipogenesis combined with

the extremely high carbohydrate intake was judged to be un-

important in daily life. Finally, Flatt et al (24) studied the effects

of the addition of fat to a standard, mixed meal. Fat and car-

bohydrate oxidation was not influenced by the increased fat

consumption (Table 2). In the postabsorptive state the main fuel

of the body is fat (2 1 ), with a rapid shift to carbohydrate when

feeding begins. independent of the fat content of the food con-

sumed.

Long-term studies of nutrient balance

Observations of nutrient balance should ideally cover at least

a 24-h interval or a multiple of 24 h because of the diurnal

pattern of nutrient utilization (25). Nowadays. several labora-

tories have facilities such as respiration chambers to perform

these studies. There have been at least four studies on the con-

sequences of nutrient exchange for nutrient utilization as mea-

sured over � 24 h (Table 3).Hunni Ct al (26) measured nutrient utilization in subjects con-

suming a mixed diet and. subsequently. a high-carbohydrate,

TABLE 2Food intake and nutrient utilization

Studs’ (ref) and c:p:f’� IntakeObservation

time

Oxidation

Carbohydrate

Rate Percent Rate

Protein

Percent Rate

Fat

Percent

fIji Ii g//i en% g/h en% g//z en%

Acheson et al (22)93:5:2 en% 9.0 10 13.3 66 2.9 15 1.7 19

Acheson et al (23)100:0:0 en�t� 8.4 14 12.6 60 2.6 12 2.6 28l00:0:0enc� 8.4 14 17.2 75 2.4 II 1.4 14100:0:0 en� � 8.4 14 18.4 84 2.4 1 1 0.5 5

Flatt et al. (24)62:27:11 en� 2.0 9 9.4 42 3.1 14 4.3 4435:l5:50enc� 3.6 9 9.4 41 2.8 12 4.6 47

35:15:50 en� 3.6 9 9.4 40 3.2 14 4.7 46

* Carhohydrate:protein:fat ratio.

t Dextrin-maltose solution flavored with fruit juice.:t c:p:fdiet 3-6 d preceding test. 14:1 1:75 en%.

§ c:p:f diet 3-6 d preceding test. 60: 12:28 en�.

II c:p:fdiet 3-6 d preceding test. 80:1 1:9 en%.#{182}Three-quarters of the fat in the form of medium-chain triglycerides.

by guest on March 9, 2014

ajcn.nutrition.orgD

ownloaded from

762S WESTERTERP

TABLE 3Mean 24-h food quotient (FQ) and respiratory quotient (RQ) for dietswith different carbohydrate:protein:fat ratios (c:p:f) from four studies

Study (ref)and c:p:f

Durationofdiet

Energybalance FQ RQ

d .41J/d

26

44: 16:40 en% 7 - 1.4 ± 0.4 0.85 0.80 ± 0.0 1�78:16:6 en% 7 -1.9 ± 0.5 0.95 0.88 ± 0.01*

28

45:15:40 en% 0 -0.1 0.85 0.82 ± 0.01

82:15:3 en% 0 -0.1 0.96 0.87 ± 0.01

29

43:15:42 en% 5-43 -0.3 ± 0.3 0.83 0.84 ± 0.0165:15:20 en% 6-32 -0.1 ± 0.2 0.91 0.88 ± 0.011’

3020:20:60 en% 3 - 0.77 0.75 ± 0.0120:20:60 en% 7 - 0.77 0.75 ± 0.01

35:20:45 en% 3 - 0.83 0.79 ± 0.0135:20:45 eri% 7 - 0.83 0.78 ± 0.0060:20:20 en% 3 - 0.92 0.86 ± 0.0260:20:20 en% 7 - 0.92 0.86 ± 0.01

*t Significantly different from FQ: *P < 0.05, tP < 0.001.

low-fat diet. Both diets were consumed for 7 d, with a 2-wk

washout period in between, and measurements of nutrient uti-

lization took place on the last day of each 7-d period. Unfor-

tunately. subjects were in negative energy balance: intake was

on average 14% lower than expenditure during the 24-h respi-

ration-chamber measurements. Despite the negative energy bal-

ance, carbohydrate and protein oxidations were lower than their

intake, even on the mixed diet. However, these results are still

within the error range of the applied methods (27). The carbo-

hydrate oxidation more than doubled on the high-carbohydrate

diet compared with the low-carbohydrate diet. illustrating the

flexibility ofthe body to use carbohydrate for energy metabolism.

Lean and James (28) measured the metabolic effect ofan iso-

energetic exchange of nutrients over one 24-h interval after the

consumption of a standard evening meal. Subjects were mea-

sured fasting and, subsequently, after a low-fat diet and a high-

fat diet. Mean energy intake, based on 24-h fasting energy ex-

penditure plus 5%, was 1 .7% higher than was mean energy ex-

penditure, with individual differences all < 10% (ie, subjects were

in energy balance). Nutrient utilization was closer to nutrient

intake for the high-fat diet than for the low-fat diet. comparing

the calculated FQ with the presented RQ value (Table 2).

Abbott et al (29) measured energy metabolism in subjects

after 5-43 d on high-fat and low-fat diets in a metabolic ward.

Part of the subjects ate the high-fat diet on the first admission

and the others started with the low-fat diet. The interval between

first and second admission was more than 4 wk. Subjects were

in energy balance: the mean difference between energy intake

and energy expenditure (< 5%) was within the error range of

both measurements. FQ and RQ were not different on the high-

fat diet, indicating that nutrient intake and nutrient utilization

were the same. On the high-carbohydrate diet, RQ was system-

atically lower than FQ (P < 0.00 1 ), indicating that fat oxidation

was higher than fat intake, assuming there was no net protein

synthesis or oxidation.

Hill et al (30) measured nutrient utilization in subjects after

3 and 7 d on three different diets: a mixed diet, a low-fat diet,

and a high-fat diet. Unfortunately. they did not report whether

the subjects were in energy balance: they presented only data

on diet composition and nutrient utilization. For all diets there

was a tendency for RQ to be lower than FQ. Hill et al’s expla-

nation for the discrepancy is the difficulty in using indirect cal-

orimetry combined with food-table nutrient analysis (27).

Nutrient addition and nutrient balance

Dallosso and lames (3 1 ) increased energy intake for 1 wk by

50% with fat, mainly as double cream. after a l-wk observation

on a maintenance diet with a c:p:f of 57: 1 3:30 en%. Subjects

were observed on days when they had a low activity level and

on days when they had a high activity level. The carbohydrate

and protein balances were not affected by the addition of fat to

the diet. The 5-MI fat supplement was mainly stored as fat,

judging from the change in fat balance from + 14 ± 25 to +163

± 27 g/d and from -14 ± 19 to +151 ± 33 g/d on days with

low activity and high activity. respectively. The mean net fat

storage in terms of energy calculated from these figures is 5.5

Mi on low-active days and 5. 1 Ml on high-active days. Schutz

et al (32) did a comparable experiment in which they measured

energy expenditure and substrate utilization over 2-d intervals

in subjects getting a maintenance diet on the first day (c:p:f 50:

1 5:35 en%) and the diet with the same amounts of carbohydrate

and protein but twice the amount of fat on the second day. On

the first day, the mean difference between intake and expenditure

was 0.1 ± 0.2 MI. indicating subjects were in energy balance.

The 24-h FQ and mean RQ values were 0.87 and 0.85 ± 0.01,

respectively. indicating that subjects were in nutrient balance as

well. On the second day the energy expenditure and substrate

utilization were not changed by the fat supplement. The energy

balance became +4. 1 ± 0.3 Mi, compared with the extra intake

of4. I ± 0. 1 Ml, that is, the fat supplement was stored as observed

before (3 1 ). Acheson et al (33) overfed subjects for I wk with

carbohydrate after a 3-d interval with a restricted diet to deplete

the glycogen stores. The energy-restricted diet was 6.7 Ml/d.

with c:p:fof 10:15:75 en%. whereas energy expenditure on the

third day was 9.6 Ml. Subsequently. subjects got a diet with a

c:p:fof86: 1 1:3 en%, increasing energy intake over 7 d from I 5.5

to 2 1 .0 MI/d. On the first day ofoverfeeding. the energy surplus

was fully stored as glycogen: then, de novo lipogenesis started

making up all of the energy surplus from the fifth day of over-

feeding onwards. At the end of the overfeeding period. there

was a glycogen gain of 0.7 kg and a fat gain of 1 . 1 kg, together

representing 75% of the energy consumed in excess of mainte-

nance requirements.

Discussion and conclusion

Based on the literature referred to. there is evidence that a

change to a fattier diet leads to an increase in body weight. Com-

bining this with the fact that obese people tend to eat more fat

leads to the conclusion that becoming overweight can be pre-

vented by reducing the fat content of the diet. Secondly. there

by guest on March 9, 2014

ajcn.nutrition.orgD

ownloaded from

FQ, RQ. AND ENERGY BALANCE 763S

is evidence that the body has a limited ability to oxidize fat

compared with the ability to oxidize carbohydrate and protein.

It is often suggested that this limitation is more pronounced in

obesity-susceptible individuals (34). Thus some individuals are

more likely to become obese consuming a high-fat diet than are

others. On the other hand. intervention studies do not support

unequivocally the body’s limitations to burn fat.

Short-term studies, measuring substrate utilization over up to

14 h after a meal, show how the carbohydrate oxidation is in-

creased after a high-carbohydrate meal whereas addition of fat

to a meal does not influence fat oxidation (Table 2). In the long

term. measuring substrate utilization over � 24 h, there are in-

dications of the opposite. At diets higher in carbohydrate there

is a bigger discrepancy between FQ and RQ such that RQ is

lower than FQ (Table 3). This should mean that substrate uti-

lization is closer to substrate intake for diets higher in fat than

for diets higher in carbohydrate. a phenomenon that cannot be

readily explained.

Under conditions ofperfect energy and nutrient balance. FQ

must equal RQ. Under conditions ofenergy or nutrient imbal-

ance, a normal-weight adult stores or mobilizes, in the long term.

near1� all energy in the form of body fat. Then. the body does

not use protein and carbohydrate reserves for energy storage or

energy mobilization. The carbohydrate store in the form of liver

and muscle glycogen fluctuates between 250 and 500 g, or 4 and

8 Ml. Reference man has a muscle mass of 30 kg with 7.5 kg

protein, or an energy equivalent of 120 Mi. but changes in mus-

cle mass are insignificant in terms of energy compared with

changes in fat mass, unless body fat is nearly depleted. Thus, in

the long term, an RQ higher than the FQ implicates conversion

of carbohydrate or protein to body fat. and an RQ lower than

the FQ, a mobilization ofenergy from body fat.

Combining the observation of the lower-than-expected RQ

for a high-carbohydrate diet with the fact that an RQ measured

over a long-term interval can only be lower than the FQ by

mobilizing body fat leads to the conclusion that high-carbohy-

drate diets induce body fat loss. Apart from studies referred earlier

(5. 6), a recent study suggests that the macronutrient composition

of the diet plays a role in the energy requirement for weight

maintenance. Prewitt et al (35) reported data on 18 women con-

suming a standard diet for 4 wk (c:p:f44:19:37 en%), followed

by 20 wk of low-fat diet (c:p:f 60: 19:2 1 en%). Energy intake of

the subjects was adjusted to maintain body weight throughout

the study period. that is. intake was increased or decreased when

body weight decreased or increased by > 1 kg, respectively.

Comparing the initial 4-wk interval for the standard diet with

the last 4 wk for the low-fat diet, the mean energy intake increased

with 19% and the mean body weight decreased with 2 kg. Thus.

a high-carbohydrate diet resulted in a significant reduction in

body weight despite a substantial increase in energy intake aimed

at weight maintenance.

This review leads to the suggestion that energy expenditure is

higher for low-fat, high-carbohydrate diets than for high-fat, low-

carbohydrate diets. Convincing evidence for this at this point is

not yet available. Current studies include measurement of energy

expenditure over short time intervals, up to 24 h in a respiration

chamber, and calculation of energy expenditure from energy

intake and changes in body composition. Future studies should

include direct measurement of energy expenditure for � 1 wk

under normal living conditions. #{163}3

References

I . Fabry P. Feeding pattern and nutritional adaptations. Prague: Aca-

demia. and London: Butterworths. 1969.2. National Research Council (US). Committee on Diet and Health.

Diet and health: implications for reducing chronic disease risk.

Washington. DC: National Academy Press. 1989.

3. Atwater WO. Bryant AP. The availability and food values of food

materials. In: The I 2th annual report ofthe Storrs. CT. Agricultural

Experimental Station. Storrs, CT: Storrs. CT. Agricultural Experi-

ment Station, 1900:73-I 10.

4. Bray GA. Bethune JE. eds. Treatment and management of obesity.Hagerstown. MD: Harper & Row, 1974.

5. Duncan KH. Bacon IA. Weinsier RL. The effects of high and low

energy density diets on satiety. energy intake. and eating time of

obese and nonobese subjects. Am I Clin Nutr 1983:37:763-7.

6. Lissner L, Levitsky DA. Strupp BJ. Kalkwarf HI. Roe DA. Dietary

fat and the regulation of energy intake in human subjects. Am JClin Nutr 1987:46:886-92.

7. Adolph EF. Urges to eat and drink in rats. Am I Physiol 1947:151:

110-25.

8. Porikos KP. Booth G. Van Itallie TB. Effect ofcovert nutritive di-lution on the spontaneous food intake of obese individuals: a pilot

study. Am I Clin Nutr 1977:30:1638-44.

9. Porikos KP. Control offood intake in man: response to covert caloric

dilution of a conventional and palatable diet. In: Cioffi LA, JamesWPT, Van Itallie TB, eds. The body weight regulatory system: normal

and disturbed mechanisms. New York: Raven Press, 198 1:83-7.

10. Foltin RW. Fischman MW, Emurian CS, Rachlinski. Compensation

for caloric dilution in humans given unrestricted access to food ina residential laboratory. Appetite I 988:10:13-24.

I 1. Jiang C-L. Hunt IN. The relation between freely chosen meals andbody habitus. Am I Clin Nutr 1983:38:32-40.

12. Dreon DM, Frey-Hewitt B, Ellsworth N, Williams PT, Terry RB,Wood PD. Dietary fat:carbohydrate ratio and obesity in middle-

aged men. Am I Clin Nutr 1988:47:995-1000.

I 3. Trembla� A. Plourde G. Despres I-P. Bouchard C. Impact of dietary

fat content and fat oxidation on energy intake in humans. Am IClin Nutr 1989:49:799-805.

14. Miller WC. Lindeman AK. Wallace I. Niederpruem M. Diet com-

position, energy intake. and exercise in relation to body fat in menand women. Am I Clin Nutr 1990:52:426-30.

I 5. Schoeller DA. How accurate is self-reported dietary energy intake?Nutr Rev 1990:10:373-9.

16. Westerterp KR. Verboeket-van de Venne WPHG. Meijer GAL. HoorF ten. Self-reported intake as a measure for energy intake. a validation

against doubly labelled water. In: Ailhaud G. Guy-Grand B. LafontanM. Riequier D. eds. Obesity in Europe 91 . London: John Libbey.

1992:17-22.

I 7. Romieu I. Willett WC. Stampfer MI, et al. Energy intake and other

determinants of relative weight. Am I Clin Nutr I 988:47:406-12.

18. Miller WC. Diet composition. energy intake, and nutritional status

in relation to obesity in men and women. Med Sci Sports ExercI991:23:280-4.

19. Flatt IP. The difference in the storage capacities for carbohydrate

and for fat. and its implications in the regulation of body weight.Ann NY Acad Sci 1987:499:104-23.

20. Frayn KN. Calculation of substrate oxidation rates in vivo fromgaseous exchange. I AppI Physiol 1983:55:628-34.

2 1 . Garlick P1, McNurlan MA. McHardy KC. Factors controlling the

disposition of primary nutrients. Proc Nutr Soc 1988:47:169-76.

22. Acheson KJ. Flatt IP, l#{233}quierE. Glycogen synthesis versus lipogenesis

after a 500 gram carbohydrate meal in man. Metabolism 1982:31:

I 234-40.

by guest on March 9, 2014

ajcn.nutrition.orgD

ownloaded from

7645 WESTERTERP

23. Acheson KI. Schutz Y. Bessard T. Ravussin E. l#{233}quierE. Flatt IP.Nutritional influences on lipogenesis and thermogenesis after a car-bohydrate meal. Am I Physiol 1984:246:E62-70.

24. Flatt JP. Ravussin E. Acheson 10. l#{233}quierE. Effects of dietary faton postprandial substrate oxidation and on carbohydrate and fat

balance. I Clin Invest 1985:76:1019-24.25. Verboeket-van de Venne PHG. Westerterp KR. Influence of the

feeding frequency on nutrient utilization in man: consequences for

energy metabolism. Eurl Clin Nutr 1991:45:161-9.26. Hunri M. Burnand B, Pittet Ph. lequier E. Metabolic effects

ofa mixed and a high-carbohydrate low-fat diet in man. mea-

sured over 24 h in a respiration chamber. Br I Nutr 1982:47:

33-43.

27. Livesey G. Elia M. Estimation of energy expenditure, net carbo-hydrate utilization. and net fat oxidation and synthesis by indirectcalorimetry: evaluation of errors with special reference to the detailedcomposition offuels. Am J Clin Nutr 1988:47:608-28.

28. Lean MEl, James WPT. Metabolic effects of isoenergetic nutrient

exchange over 24 hours in relation to obesity in women. Int I Obes1988: 12:15-27.

29. Abbott WGH, Howard BV, Ruotolo G. Ravussin E. Energy expen-

diture in humans: effects of dietary fat and carbohydrate. Am I

Physiol 1990:258:E347-5 I.

30. Hill 10. Peters IC. Reed GW. Schlundt DG. Sharp T. Greene HL.Nutrient balance in humans: effects ofdiet composition. Am I Clin

Nutr 1991:54:10-7.

31. Dallosso HM. James WPT. Whole-body calorimetry studies in adult

men: 1 . The effect of fat over-feeding on 24 h energy expenditure.Br I Nutr 1984:52:49-64.

32. Schutz Y. Flatt IP. J#{233}quierE. Failure ofdietary fat intake to promotefat oxidation: a factor favoring the development of obesity. Am I

Clin Nutr 1989:50:307-14.33. Acheson KI. Schutz Y. Bessard T, Anantharaman K. Flatt IP, and

J#{233}quierE. Glycogen storage capacity and de novo lipogenesis duringmassive carbohydrate overfeeding in man. Am I Clin Nutr I 988:48:

240-7.

34. Zurlo F. Lillioja S. Esposito-Del Puente A. et al. Low ratio of fat tocarbohydrate oxidation as predictor of weight gain: study of 24-h

RQ. Am I Physiol l990:259:E650-7.35. Prewitt TE. Schmeisser D. Bowen PE. et al. Changes in body weight.

both composition. and energy intake in women fed high- and low-fat diets. Am J Clin Nutr 199 1:54:304-10.

Discussion

Michael A Crawford: This may be a naive question. but I wasinterested in the analogy that you gave us-your models of the

test tube and the barrel. the test tube of carbohydrate and the

barrel of fat. What seems to be slightly out of focus to me was

the size of the input you showed. You showed two little pots.

One ofcarbohydrate and one offat ofdifferent proportions and

different diets. It seems to me that if you had represented those

inputs in proportion to the daily amount ofstuffthat was going

in, one would have seen that the amount of carbohydrate that

went in over a 24-hour period would be really very much larger

than the amount of carbohydrate that was stored in the test

tube. A 24-hour input of carbohydrate is quite a lot in a high-

carbohydrate diet. At the same time. the barrel is full of fat, so

that carbohydrate. although coming through the carbohydrate

route, is actually filling up the fat barrel because it doesn’t store

that amount of carbohydrate. So. it has to be converted to fat

and fill up the fat barrel. The energy process ofconverting that

carbohydrate to fat needs to be a sort of negative consideration

in terms ofthe overall energy balance ofthe individual. I wonder

to what extent that contributes to disturbing these relationships

and explaining why fat is going straight into the barrel and doesn’t

really have any effect of energy. The carbohydrate going in de-

mands energy to convert it into fat and destorts it as such. Does

that confound the equation with regard to a rich-carbohydrate

vs a rich-fat diet?

Klaas R Westerterp: I think you are stressing an interesting point.

I don’t know whether this model was really on the right scale

and, as you say. we eat a lot of carbohydrate in proportion to

our carbohydrate stores. On the other hand, there is not much

information on whether we really have to convert carbohydrate

to fat, apart from the one study I showed you, and maybe one

or two others in the literature in which you can see that at least

there is a day-night rhythm in substrate utilization. That rhythm

suggests that we might store part ofthe carbohydrate we consume

during the day for overnight use. That has been seen in small-

animal studies. but you can’t really compare those with human

studies because we as humans have a big intestine for food storage

and we have got a relatively far lower energy metabolism per

unit body weight. Maybe the reason for having a higher energy

expenditure at a high-carbohydrate diet is the conversion of car-

bohydrate to fat.

Eric Ravussin: I would like to expand a little on this comment.

I think that in men there is less and less evidence that de novo

lipogenesis plays an important role. and in the study that you

showed from Kevin Acheson, in which he gave an enormous

amount ofcarbohvdrate (500 g). We tried that first as one meal.

but after the subject had problems, we offered it at 250 grams

and at two times 125 grams. It was very, very difficult to induce

a significant amount ofde novo lipogenesis. I agree that indirect

calorimetry doesn’t differentiate between oxidation and going

into fat. but I think that the point is that there is little room for

de novo lipogenesis in men. Therefore. you are right when you

say that maybe 50- 100% of the glycogen stores are replaced on

a daily basis vs < 1% for fat. I think that most ofthe carbohydrate

ingested is either stored as glycogen or oxidized into CO2. I

think that the enzymatic activities ofthe lipogenic enzymes are

very low in men, and recent studies using stable isotope showed

that de novo lipogenesis is very insignificant in man.

Westerterp: I agree with that, but on the other hand. the studies

ofAcheson were short-term studies. In the long term. when you

consume a high-carbohydrate diet. you can’t really store this

amount of carbohydrate in your glycogen stores because they

are just not big enough. So. you have to do something with it

when you are in a positive energy balance. but apparently you

tend to be in a negative energy balance on a high-carbohydrate

diet. We do not yet have any reason for that. Maybe you have

a suggestion for why you tend to be in a negative energy balance

on a high carbohydrate diet. Why would your energy expenditure

go up?

Ravussin: In all the studies ofoverfeeding. for example. Schutz’sstudy. you need to push a lot ofcarbohydrate through the system.

by guest on March 9, 2014

ajcn.nutrition.orgD

ownloaded from

FQ. RQ. AND ENERGY BALANCE

When we talk about weight gain in obese people. they are ob-

viously in positive energy balance. but by a minute amount on

a daily basis. There is limited storage capacity in men, and there-

fore carbohydrate is oxidized: maybe there is a little lipogenesis.

There was the British school in the l960s that said that lipogenesis

might be higher in lean people because it is an energy-costly

process. I don’t know what is the status of that now.

Nlartijn Katan: Could I sum a little note of caution about the

effectiveness of high-carbohydrate diets in reducing obesity be-

cause the final tests are long-term control trials and there have

been a few ofthose? In 1970 the National Diet Heart Study was

published in (‘irculation, where large groups of middleaged

American men were given diets of 40% or 20% energy as fat.

There was some weight loss in the 20-en%-fat group but that

was very minor, 1 or at most 2 kg in the first few months and

765S

than no change for the rest of the year. There is a recent pub-

lication from the pilot study for the breast-cancer-dietary-fat

trial, where you see a similar phenomenon. that indeed the

women on the very-low-fat diet lost some weight but nothing

like the amounts you would expect from the short-term trials,

again something like 1 or 2 kg over the year. There are also the

data from the Seattle dietary alternative trial. a 2-year study that

I think is still in press, where again you see in middleaged man

only � I kg weight loss on 20 vs 40 en% of fat. I think there is

more going on than just the mechanisms that we have been

talking about. The body seems to resist weight loss even on the

more carbohydrate-rich diets.

Westerterp: Maybe you are right. The only problem is the control

offood intake in these types ofstudies. On the other hand. long-

term experimental feeding studies are costly. Therefore, we have

to accept the results as they are.

by guest on March 9, 2014

ajcn.nutrition.orgD

ownloaded from