Glycogen storage capacity and de novo lipogenesis massive carbohydrate overfeeding in man13

Kevin JAcheson, Jean-Pierre Flail, PhD; PhD; Yves Schutz, PhD; Thierry and Eric J#{233}quier, MD The metabolic balance excesses ofcarbohydrate.
MJ [1994 65 kcal]

during
PhD;

Bessard,

MD;

Krishna

Anantharaman,

ABSTRACF

the fate oflarge

d, 8.35 0.27

method was performed on three men to investigate Glycogen stores, which first depleted were by diet (3
decreasing to 5.70 1.03 MJ [1361 247 kcal], 15%

protein,

75%

fat,

10% carbohydrate)

and

exercise,

were

repleted

during
.

7 d carbohydrate

overfeeding (1 1% protein, 3% fat, and 86% carbohydrate) providing 15.2510 MJ 1 (3642 263 kcal) on the first day, increasing progressively to 20.64 1 .30 MJ (4930 3 1 1 kcal) on the last day of overfeeding. Glycogen depletion again was accomplished with 2 d of carbohydrate restriction (2.52 MJ/d [602 kcal/d], 85% protein, and 1 5% fat). Glycogen storage capacity in man is 15 g/kg bodyweight and can accommodate gain of -5#{174} before a g net lipid synthesis contributes to increasing body fat mass. When the glycogen stores are saturated, massive intakes ofcarbohydrate are disposed ofby high carbohydrate-oxidation rates and sub-

Downloaded from www.ajcn.org by guest on April 14, 2012

stantial

de novo

lipid
Am

synthesis
JClin

(1 50 g lipid/d
Nutr 1988;48:240-7.

using

-475

g CHO/d)

without

postabsorptive

hyperglycemia.
KEY rect WORDS calorimetry,

body

Carbohydrate composition

overfeeding,

glycogen stores,

de novo

lipid

synthesis,

mdi-

a liver weighing 1 .8 kg, one can estimate that .3 mol gjycosyl residues or almost 500 g of glycogen are stored It has long been known from studies involving carcass in the body. If the highest reported literature values (1, and organ analysis in animals that the composition of are extrapolated 3) to the whole body, then up to 4.3 mol the diet and its availability effect the organisms glycogen glycosyl residues or some 700 g of glycogen could be reserves. The work of Bergstrom et al (1) and Huitman stored in the body. and Nilsson (2) using muscle and liver biopsies provided Although it is known that this value can be increased direct measurements ofsuch changes in man. The high- markedly during glycogen loading, when muscle glycoest muscle glycogen levels were observed when glycogen gen levels can reach 2.4 g/lOO g wet muscle or more (1), stores are first depleted by sustained exercise followed by it is often believed that the glycogen stores are normally ingestion of a high-fat diet and then repleted by con-maintained within a relatively narrow range. However, sumption of a carbohydrate-rich diet, ie, the glycogen the capacity for storing large amounts dietary of carboloading technique (1). hydrate by conversion to glycogen is in fact considerable Skeletal muscles and liver are the principal sites for the (5-7). To assess the upper limit for glycogen storage in storage of glycogen in the body. Liver glycogen concen- man, we performed continuous metabolic balance studtrations vary with the diet with values in the range of 50500 mmol glycosyl residues/kg tissue in the postabsorpI From the Institute ofPhysiology, Faculty ofMedicine, University tive state (mean 270 mmol [44 g] glycosyl residues/kg Lausanne, Switzerland; the Nestl#{233} Research Center, Nesliver) (3). Liver glycogen varies appreciably during theofLausanne, tee Lausanne, Switzerland; and the Departday in relation to the patterns of eating and fasting (2). Ltd, Vers-chez-les-Blanc, ment of Biochemistry, University of Massachusetts, Medical Center, Glycogen concentrations in biopsy samples from the Worcester, MA. quadriceps femoris muscle were found to be in the range 2 Supported by the Nestl#{233} Switzerland. Co. of6O-l20 mmol glycosyl residues/kg with a mean of 85 3 Address reprint requests to KJ Acheson, Institute of Physiology, mmol (14 g) glycosyl residues/kg tissue (4). However, the University ofLausanne, Rue du Bugnon 7, CH-l005 Lausanne, Switglycogen concentration in skeletal muscle also depends zerland. upon the muscle group being investigated (4). For a 70- ReceivedJune 15, 1987. October 6, 1987. kg man with -40% ofhis weight as skeletal muscle and Accepted for publication Introduction 240
Am

JC/in

Nuir

l988;48:240-7.

Printed

in USA.

1988

American

Society forClinical

Nutrition

CARBOHYDRATE ies for excesses 10 d, which included 7 d during which

DISPOSAL massive

IN OVERFEEDING

241

ofcarbohydrates
and methods

were

ingested.

Subjects

Subjects
Three

healthy

swimmer

young at university

men, level,

and
sity

1 1-14% body fat)with and who were not taking

one of whom was a competition (21-22 y, 62-72 kg, 174-180 no family history ofdiabetes or obe-

The frozen sample was dried to constant weight to establish its water content. The residue was ground to a powder thatwas analyzed for nitrogen (10), extractable fat by the Soxlet method (1 1), and ash contents. Dietary protein was taken 6.25 be to x total N and the carbohydrate content was calculated by difference (ie, dry weight [weight protein + fat + ash]). Much of the excess carbohydrate was provided by sugared fruitjuices ofknown uniform composition and energy content. in the 20% duplicate sample but were cm, These were not included
-

analyzed

separately;

their

energy

contents

were

found

to agree

any

medication,

participated in

this study. The subjects were each given a detailed account the protocol, which had previously been reviewed and accepted by the institutes ethical committee, before they gave their consent to participate. Protocol

closely content of ofthe Energy

energy

with those indicated by the manufacturers. The energy ofthe fruit juices was added to the gross energy values high-carbohydrate diets.

expenditure
low-carbohydrate diet, physical activity was a pedometer (Pedoboy #10, Barigo Barometer Schweninger, FRG). Heart rate was monitored using a portable heart-rate monitoring instru2 d on

During the first expenditure

the

high-fat,

was not measured

but

The experiment
3 d the subjects

lasted
consumed

14 consecutive
a restricted

days.
diet,

During

first the

high

in fat and

in carbohydrate, and followed an exercise program. through this period the subjects were admitted into a respiration chamber in which respiratory exchange measurements

recorded by Gmbh, lowFabrik Halfway continuously

ment

(HRM,

Difa,

Breda,

Holland).

The

subjects

performed
At

Downloaded from www.ajcn.org by guest on April 14, 2012

various types of physical activity, ie, running and swimming, during this period to deplete their muscle glycogen stores. were to be continued for 10 d. After 36 h in the chamber the 2, each subject was admitted into an open-circuit diet was changed to a high-carbohydrate, low-fat diet that was the end ofday calorimeter chamber (12) where energy expenditure ingested for the following 7 d. During the last 2 d while still in indirect

the chamber,

the subjects

received

limited

amounts

of a high- was

measured

continuously

for the

next

10 d. Resting

meta-

with a ventilatedprotein diet (protein-sparing modified-fast [PSMF], 2.5 MJ bolic rate was measured for 1 h each morning hood system within the chamber (12). Each day the subjects or 600 kcal) essentially devoid of carbohydrate. The subjects for two 30-mm periods (beginning 1 at 30 and 1 1730) then left the respiration chamber but continued to consume walked (Quinton Inst Co. Seattle, WA) at 2 mph (3.22 the high-fat, low-carbohydrate diet restricted in amounts for a on a treadmill km/h), 5% slope in the morning and 2 mph (3.22 km/h), 10% further 2 d. slope in the afternoon. During the remainder ofthe day, spontaneous physical activity was allowed within the confines imEnergy intake posed by the chamber but strenuous physical exertion not was The diets were prepared by trained dieticians at the Institute permitted. of Physiology. The restricted high-fat, low-carbohydrate diet The subjects were allowed to leave the chambertwice a day consumed on days 1-3 and 1 3-14 provided --6.70 MJ (1600 for 30 mm after the resting metabolic rate measurements in kcal) composed of 1 5% protein, 75% fat, and 10% carbohythe morningand again in the afternoon at 1630 during which drate. During the overfeeding period (days 4-10 inclusive) the time the calibrations of the analyzers were verified and other high-carbohydrate, low-fat diet provided 1 5 MJ (3600 kcal, measurements were performed, eg, body weight, urine colleccomposed of 1 1% protein, 3% fat, and 86% carbohydrate) on tion, etc. day 4. Energy intake was then increased progressively each day while the composition was kept constant to provide 6.28 MJ Nutrient balances (1500 kcal) in excess ofthe previous days energy expenditure, were collected from day until 1 several days after the which was measured in the respiration chamber (8). By day 10 Stools test. At each change in the diet the subject consumed 1 g carthe energy intake had thus increased to -21 Mi(5000 kcal). Stools were weighed, frozen, freeze For the next (and last) 2 d in the respiration chamber, the sub- mine red as afecal marker. for N, fat, and ash, and carbohydrate was jects consumed a high-protein, low-calorie diet (PSMF, --2.5 dried, and analyzed by difference as described above. MJ or 600 kcal). The diet was then changed to the same re- calculated nutrient and metabolizable energy instricted high-fat, low-carbohydrate diet eaten on days 1-3 for Twenty-four-hour from the food tables (9). The 24-h gross the last 2 d ofthe experiment spent outside the chamber. Each takes were calculated energy intake was also determined on the basis of the direct food item consumed wasweighed to the nearest gram with a analysis of the 20% dried duplicate sample and corrected for Mettler PlO balance (Mettler, Greifensee, Switzerland) and its losses in the stool collections between the appearance intake was corrected for any residue left on the plate. Energy nutrients intake and dietary composition were calculated from food taofthe fecal markers. Urine was collected during the day (14 h) and the night (10 bles (9) with a desk-top computer (HP 9830, Hewlett Packard was tested for glucose (Gluketur-Test, Boehringer (Schweiz) AG, Schlieren, Switzerland). The factors 16.74, h) and 37.67, and 16.74 kJ/g (4, and 4 kcal) were used to calculate 9, Mannheim Gmbh, Mannheim, FRG) and total N was ana--

oxidation was calculated by summing energy contents ofprotein, fat, and carbohy- lyzed (10). Daily protein Twenty percent, by weight, of each food the urinary N excretions during the day and the night and mulitem (except for the sugared fruit juices) ere set aside and w tiplying this value by 6.25. Twenty-four-hour carbohydrate were calculated according to classical formupooled. At the end of each 24-h period, the pooled samples and fat oxidations were homogenized. An aliquot was immediately freeze dried las (13) from the nonprotein respiratory quotient. (Virtis automatic freeze-drier, Gardiner, NY) and the remainFrom a knowledge ofthe composition ofenergy entering the body, that oxidized, and that leaving the body in urine and feder was frozen and stored at -20 #{176}C.

the metabolizable drate, respectively.

242 TABLE
Bodyweight 1 changes

ACHESON

ET AL

Subject

Age
y

Height day 1
cm 180 175 173.5 176.2

Start

high day 4 kg 71.9 70.8 61.8 68.2

fat

Start

high CHO day 1 1 kg 71.3 68.6 62.3 67.4

Start

PSMF
13

End PSMF
day kg 71.5 69.6 61.7 67.6 15 End kg 71.9 68.8 61.4 67.4 test
,

day kg

2
3

22 21 21 21.3

76.9 73.5 65.4 71.9

I
SD
S

0.6 High fat, high-fat,

3.4 low-carbohydrate

5.5 diet; high CHO, high-carbohydrate,

4.6 low-fat

5.9 diet; PSMF, protein-sparing

5.2 modified fast.

5.4

ces (hair

and

cutaneous

losses
daily the

were

assumed

to be negligible),

drate

diet,

the

increase

over

the

preceding

days

energy

it was possible to calculate in body composition) during

Blood

variables

Blood samples were at the beginning ofand periment. They were free fatty acids (16) on and blood urea N (19). phoresis (Readysystem AG, Bad-Zurzack, thyroid hormone concentrations were radioimmunoassay ARIA II (Becton burg, NY).

intake rose to 0.95 MJ (227 kcal). This pattern was slightly different when intake was based on the data obtamed by direct analysis, where on the fifth day of overfeeding (day 8) the additional energy was 1.94 MJ taken in the postabsorptive fasting state (464 kcal) and fell to 0 on the last day overfeeding of (day at regular intervals throughout the cx10). The food tables overestimated protein and carbohyanalyzed for glucose (14), insulin (15), drate intakes during each ofthe different diets except for the Dole extract (17), triglycerides (18), the 2 d on the low-energy, high-protein (PSMF) diet Lipoproteins were separated by electroSwitzerland)
analyzed

nutrient balances (ie, changes 10 d ofthe experiment.

Downloaded from www.ajcn.org by guest on April 14, 2012

and Orange-

(days

1 1 and

12).

The

tables

overestimated

fat

intake

by automated

Dickinson,

during the low-energy high-fat, low-carbohydrate but considerably underestimated it during the bohydrate, low-fat diet by -40-80 g/d (670-1339
160-320 kcal/d).

diets high-carkJ/d;

The substrate balances calculated from substrates entering (direct analysis) and leaving the body are presented in Table 3. With the onset of carbohydrate Mean body weight decreased by 0.8 1.4 kg during there was a dramatic increase in carbohythe 3 d on the restricted, high-fat, low-carbohydrate diet overfeeding, drate oxidation (Fig 2) from 40 g/d (day 74 3) to 398 (Table 1). After the 7 d ofoverfeeding the high-carbohy 87 g/d (day 4). Thereafter carbohydrate utilization (ie, drate, low-fat diet (day 10), body weight had increased and that used for de novo lipid synthesis) inby 4.6 1 .3 kg (ie,5.6, 4.9, and 3.2 kg). During the 2 d oxidation creased progressively in response to the increase in caron the restricted high-protein, low-energy diet (600 kcal/ bohydrate ingestion, attaining 1010 37 g/d on the last d) 4.4 0.9 kg were lost. Two days later, ody b weights day of overfeeding. Concomitant with increase the in were the same as at the start of the overfeeding phase of utilization therewas a rapid suppression of the experiment (ie, 7 1 .3 and 7 1 .9, 68.6 and 68.8, 62.3carbohydrate lipid oxidation. and 61 .4 kg, respectively). After an initial decrease in protein oxidation at the beEnergy and nutrient intake ginning of carbohydrate overfeeding (from 104 12 to 82 7 g/d from day 4 to day 5), protein oxidation reThe composition ofthe diets over the l4-d experiment maimed relatively constant until the last 2 d of overfeedare presented in Table 2. From the ratio metabolizable of it increased in proportion with the protein in energy obtained from direct analysis and food tables, iting when the diet so that a positive N balance was maintained at can be seen that the food tables tended to overestimate g N during the last 5 d of overfeeding. energy intake with the high-fat, low-carbohydrate diet 3.8 As the diet passed from a hypocaloric high-fat, lowand underestimate during the high-carbohydrate, low-fat carbohydrate to a hypercaloric high-carbohydrate, lowoverfeeding and the high-protein low-energy diets. there was not only a large increase in When the high-carbohydrate, low-fat diet was initi- fat composition, oxidation but also in glycogen storage (339 ated, it was necessary to increase the energy intake by 8.8 carbohydrate each successive day the amount of carboMi (2100 kcal, food table data) to obtain a positive en- 82 g/d). With hydrate that was stored decreased even though the erg)1 balance of6.28 MJ (1 500 kcal) during first 24-h the amount that was ingested increased (Fig 2). After 4 d of period of overfeeding. On the second day an increase of overfeeding, the glycogen stores had become saturated 1 .54 MJ (370 kcal, food tables) was necessary to mainthat they had increased by -770 thin the same positive energy balance, the increment de- and it was calculated g. carbohydrate balance was maintained near creasing gradually to 0.42 MJ/d (100 kcal/d) over the Thereafter equilibrium. When the diet as devoid w of carbohydrate next 3 d (Fig 1). During the last 2 d ofthe high-carbohyResults

CARBOHYDRATE TABLE
Daily

DISPOSAL

IN OVERFEEDING

243

2
composition ofthe diet calculated by direct analysis and from food
(n = tables

3, 1

SD)

Nutrients Ingested
Day

excreted
direct

in eces from f
Available CHO g 139 129 2417 2013 2017 188 225 146 73 54 101 101 101 1610 Protein g 725 5016 4924 8218 934 1049 10214 11723 1226 12411 12610 12610 6420 6518 nu trients Fat g 1667 12848 1176 5619 9713 8124 7814 9423 10223 6923 102 112 11328 13521 from direct analysis5 CHO g 2918 4320 2816 73712 81378 84335 86858 93384 93245 98143 0 0 3314 497

nu trients
Fat g

from di analysis rect

CHO g 4213 5628 522 75711 83492 86141 89063 94778 93947 98741 0 0 4313 647 Available Protein g 1311 1311 2716 2113 2215 203 233 156 95 96 112 112 122 148 nutrients Fat g 17117 16312 1429 121 171 130 155 161 204 173 51 51 15629 15512 based on directanalysis
= -

analysis Fat g 63 63 118 72 74 74 72 41 32 43 61 51 61 73

Protein g

1 2 3 4 5 6 7 8 9 10 11 12 13 14

8612 636 759 10310 11616 12412 12513 13119 1312 13311 13610 13610 7722 7925

1728 13451 1283 6320 10411 8827 8516 9822 10522 7420 162 162 11829 14222

Downloaded from www.ajcn.org by guest on April 14, 2012

from

foodtablest CHO g 458 5912 472 83665 90089 94861 99468 98990 103874 107344 0 0 4314 6517

Metabolizable
Direct analysis/
food

energy

Day

Protein g

tables
%

1 2 3 4 5

6
7 8 9 10 11 12 13 14
S

798 708 694 1148 1245 1244 1273 12914 13810 14213 1124 1113 7519 7511

96 75 77 94

104 100
98 106 104 97 126 116 72 83

Nutrients

absorbed by the organism

offood

and feces.
losses
+

t Available nutrients in foodusing Atwaters coefficients which allow for partial t From direct analysis, Metabolizable Energy Gross Energy (Fecal Energy

ineces (20). f

Urinary

Energy).

(days cipal

1 1 and 12), carbohydrate was still used as the prim- During the 6 d during which lipid synthesis exceeded energy substrate to the detriment of the glycogen fat oxidation, net de novo lipogenesis amounted to a tostores, which decreased by -7#{174}g during these 2 d. tal of -580 g. Because in addition to de novo lipogenesis The initial increase in the glycogen stores by 500 g some fat was provided in the diet (--85 g/d), the overall was accompanied by an increase in the mean 24-h non- fatgainwas-.l.l kg. protein respiratory quotient(Fig 3). The mean 24-h nonThe principal blood variables are presented in Table 4. The hypocaloric high-fat, low-carbohydrate diet caused protein respiratory quotient exceeded 1.00 (indicative of net de novo lipid synthesis) on day 2 of carbohydrate both plasma glucose and insulin concentrations to decrease and free fatty acids to increase. During carbohyoverfeeding; it continued to increase and reached a value of 1. 1 5 during the last 3 d of overfeeding when daily drate overfeeding plasma glucose rose initially but was at the control value obtained at the beginfat synthesis from glucose averaged 142 g/d. Even on day maintained 1 1 when no carbohydrate was present in the diet, the ning ofthe experiment by the rising plasma insulin conmean nonprotein respiratory quotient was still just centrations. Plasma triglycerides increased 10-fold dur> 1.00. ing carbohydrate overfeeding. This increase was realso
-

244

ACHESON

ET AL
1000
80o

Respiratory

Exchange

_I

600
0
>. I

g
C
C.)

400

200

FIG 2. Daily carbohydrate intake(and its disposal (oxidation, glycogen storage, and conversion tolipid)during 7 d ofprogressive bohydrate overfeeding(n 3).
-) =

car-

10

11

12

Days

Downloaded from www.ajcn.org by guest on April 14, 2012

protocol (PSMF = protein-sparing modified fast; HFLE = high-fat, low-energy diet [--7 MJ]) and the changes daily metabolizable energy intake -, energy expenditure D, positive energy balance 0, and negative energy balance U (n = 3, 1 SD).

FIG 1. Experimental

bodys During in

glycogen

stores

can

be

assumed

to

be

very

low.

flected by changes can be seen that

in

the

lipoprotein

fractions

where

the very-low-density lipoproteins (VLDLS) increased from 20 to 70%. Blood urea was N decreased on the high-carbohydrate, low-fat diet, (end of day 2 of overfeeding), carbohydrate oxidation creased markedly on the hypocaloric high-protein diet, and storage became insufficient to dispose of all of the and returned to control concentrations at the end of the ingested carbohydrate. The excess was disposed of by experiment. Only very slight changes were observed in conversion to fat, ie, de novo lipogenesis. During the last the thyroid hormone concentrations and these can be cx. -

carbohydrate overfeeding the rate of glycogen storage was initially large, decreasing as the stores became saturated. The maximum increases in stored glycogem observed were 1 146, 629, and 654 g (in subjects 1,2, and 3, respectively) with a mean of 810 g. Saturation of it the glycogen stores occurred on day 4 of carbohydrate overfeeding for subjects 2 and 3 and on day 5 for subject 1 When the glycogen stores had increased by 500 g in-

plained

by the

short

duration

ofthe

experiment.

3 d ofoverfeeding,

total

carbohydrate

utilization

(ie,

oxi-

Discussion After 3 d of diet combined

TABLE 3 Daily substrate

hypocaloric high-fat, low-carbohydrate with a rigorous exercise program,

dation and glucose conversion to fat) was very that which was ingested. Within the limitations methods used, these results demonstrate that the drate balance was again achieved.

similar to of the carbohy-

balance

(g) ofthe three

subjects

(1 SD)

Substrate
Substra Day 1 2 3 4 5 6 7 8 9 10 11 12 13 14
S

te entering Fat 1667 12848 1176 5619 9713 8124 7814 9423 10223 6923 102 112 11328 13521

the system CHO 2918 4320 2816 73712 81378 84335 86858 93384 93245 98143 0 0 3314 497 Protein

oxidation and disappearance from the system

Fat CHO Protein

Balance
Fat CHO

Protein 725 5016 4924 8218 934 1049 10214 11723 1226 12411 12610 12610 6420 6518
Negative

10116 10412 827 8217 8123 8823 1017 987 11511 1454 13721 1216

16446 4962 -3038 -4537 -8117 -14052 -13726 -14914 -4384 8420

7440 39887 62296 67798 79283 950158 962102 101037 49161 22359

-5231 -2329 1127 238 2216 2910 223 265 111 -207

-4741 743 12725 12628 16012 23566 23917 21820 5486 -7322

-4648 33982 19253 16674 7631 -18123 -3064 -296 -49161 -22359

values

for fat oxidation

represent

de novo fat synthesis.

CARBOHYDRATE

DISPOSAL

IN

OVERFEEDING

245

exhaustion
445
C

and

he calculated
However,

values
by

ranging
using known

from

3 1 5 to
for

g carbohydrate.

values

the glycogen content of muscle and liver, he calculated that his subjects (mean body weight72 kg) should have had maximal values of 700 glycogen. g Bj#{246}rntorp and 0 Sj#{246}str#{246}m (22) came to the same conclusion using similar Ca a calculations but suggested that a further 100 g could be Cl, a, stored with 2 wk ofcarbohydrate overfeeding or by using C a, the glycogen-loading technique. Bergstrom et al (1) re0 ported values in the range 500-800 g in some of their a 5 subjects who followed the glycogen-loading technique. If z this glycogen was derived from muscle, as suggested by Olsson and Saltin (23), the addition of liver glycogen would increase their values to those observed in the presFIG 3. Average 24-h nonprotein respiratory quotients day 3 of a on ent study. high-fat, low-carbohydrate diet, during carbohydrate overfeeding, and Although the values reported in this study may seem for 2 d while on a PSMF devoid ofcarbohydrate. (n 3, i SD). surprising the metabolic balance data does agree with the observed body weight changes. At the end of the period after the glycogen stores had been reBy adding the negative carbohydrate balances at theoverfeeding slightly by the high rate oflipogenesis and carboend of the experiment, it was possible to obtain a value duced oxidation, body weight had increased by 4.6 kg for the amount of glycogen utilized. During this period hydrate and 700 g glycogen remained. Assuming that glycogen is 897, 89 1, and 752 g were utilized by subjects 1, 2, and 3, with two to four times its weight ofwater (23, 24), respectively. These results suggest that subjects 2 and stored 3 kg of the change in body weight can be achad not depleted their glycogen stores completely before --2. 1-3.5 for. Cumulative gains of body fat by de novo carbohydrate overfeeding began and that they still con- counted lipogenesis and from that which was provided in the diet tamed 100-200 at that time. g to 1. 1 kg fat. Thus with the 665 g increase in From these data it would seem that the glycogen stores amounted mass indicated by a gain of 1 33 g protein, we can maximally accommodate 800-900 g of carbohy- lean body can account for the 4.6 kg increase body in weight. In a drate and perhaps as much as 1-1 kg in trained 1 athmanner it is possible to account for 70% of the letes. These results are among the highest glycogen- similar lost during the2-d hypocaloric diet at the end of storage values reported in the literature. Hedman (21)weight the experiment. used respiratory-exchange measurements to calculate carbohydrate oxidation and hence glycogen depletion in Initially large glycogen storage (340 g/d) on the day 1 ofcarbohydrate overfeeding was observed during carbofour well-fed, trained cross-country skiers who skied to
0 = .

a,

Downloaded from www.ajcn.org by guest on April 14, 2012

TABLE 4 Blood variables

measured during

the experiment Day

(i 4 end

SD)

Day
Glucose (mmol/L) Insulin(pmol/L)
Free

high fat
4.0 0.4 507 984537

Day 6
5. 1 0.2 11536 321335 1.40.4 4.2 0.8
-

Day 9
4.9 0.2 13629 399434 5.30.8 4.7 0.8 5616 2811 164 4.2 2.53 87 3.5

Day 1 1 end high CHO

4.8 0.3 14450 488504 8.6 1.0

Day 13 end PSMF

4.9 0.2 57 14 700500 3.5 1.2

Day

1 5 end

test
4.8 0.3 5714 687283 0.90.3 6.6 1.2

4.9 0.3 8643 587399 1.3 1.2

fatty acids
(imol/L)

Triglycerides

(mmol/L) Cholesterol (mmol/L) Lipoproteins VLDLprefi(%) LDL(%) HDLa(%) BUN (mmol/L)

0.80.4 5.3 1.2 188 5910 234 6.9 1.4 1.86 0.38 11720 2.0 1.3 diet; high lipoprotein; CHO, HDL,

4.8 0.5 185 538 296 6.3 2.3

5.8 0.9 6824 2016 129 4.2 0.9 2.39 0.58 10036 2.7 1.4

6.4 0.9
-

T3 total (nmol/L)
T4total(nmol/L) TSH(mU/L)

2.22

0.27

9620 2.60.9

4.4 1.0 2.26 0.2 1 10421 3.4 1.5

0.8 0.28 11 1.6

10.0

1.4
2.0.3 1 10

87 18 2.20.8

163 6410 208 7.0 1.5 2.29 0.46 101 13 2.2 1.0 fast; VLDL, T4, thyroxine; very-low-

S High fat, high-fat, low-carbohydrate density lipoprotein; LDL, low-density

high-carbohydrate, low high-density lipoprotein;

fat diet; PSMF, protein-sparing modified BUN, blood nitrogen; urea T3, triodothyronine;

TSH,

thyroid

stimulatmng

hormone.

246

ACHESON

ET AL

hydrate refeeding after starvation (2) and after exhaus- is limited even during carbohydrate overfeeding, we tive exercise (25). The decreasing ability of the body to demonstrated that this pathway can readily dispose of store glycogen may be mediated by inhibition of glyco- nearly 500 g of glucose per day. Furthermore, the large gem synthetase activity with increasing glycogem concem- excess of carbohydrate entering the organism did not even cause excessive increases in circulating glucose contratioms (25, 26). We (6, 7) and others (5) demonstrated that humans centration. cam ingest relatively large amounts ofcarbohydrate withAs shown in Figure 1, the subjects energy expendiout initiating de novo lipid synthesis at rates exceeding tures increased markedly during the carbohydrate concomitant fat oxidation. These results are consistent overfeeding period. Because the experimental protocol with in vitro data demonstrating very low fatty acid syn- aimed at initially depleting their glycogen stores, the subthase activity in human liver and adipose tissue (27) even jects energy expenditure on day 3 is somewhat less than after the ingestion of a carbohydrate-rich diet for 3 d. their maintenance energy expenditure. The latter cam be However, these authors did observe elevated fatty acid estimated at 10 MJ/d (2400 kcal/d), ie, the observed ensynthase activities in certain situations where long-term ergy expenditure on day 3 (9.66 MJ/d) plus 10% of the deficit on that day (ie, 3.96 MJ X 10%0.4 MJ), fat-free diets were being received, eg, parenteral nutri- energy for the thermic effect ofthat amount of food. tion. It is precisely under such conditions that high rates to account of de novo lipid synthesis were observed with indirect By day 7 on the high-carbohydrate, low-fat diet, their 24h energy expenditures hadincreased by 3.5 MJ/d (840 calorimetry (28). By extrapolating from in vitro data, Bj#{246}rntorp and kcal/d). This 35% increase is one ofthe most substantial Sj#{246}str#{246}malso concluded (22) that de movo fatty acid syn- diet-induced increases in energy expenditure demonstrated in man. It is of interest to assess how much of thesis from carbohydrate is a quantitatively insignificant pathway in the whole human organism. They presumed this food-induced thermogenesis is due to the obligatory that it remained so even during carbohydrate overfeed- costs incurred for nutrient storage. Considering that fasting where excess carbohydrate would cause hyperglyceing blood glucose levels remained in the normal range occurred during the entire day, we mia and hyperinsulinemia and eventually glucose intol- and that lipogenesis erance. In the present study, postpramdial plasma glucose assumed that 80% of the carbohydrate consumed was and insulin concentrations were not measured during converted to glycogen before being used for energy proor lipogemesis. Carbohydrate stored as glycogem the carbohydrate overfeeding days but no glucosuria was duction the expenditure of 2 mol ATP per glucose moiever observed. However, fasting concentrations were requires measured every other morning. Fasting glycemia as w ety converted into glycogen plus 0.5 mol for the cost of normal (up to 5. 1 mmol/L) but insulin concentrations active transport in the gut and other phenomena, such rose from 50 7 to 144 50 pmol/L. as digestive enzyme synthesis andut motility g (3 1). BeIn addition to providing an assessment of the bodys cause glucose oxidation yields 36 mol ATP, the cost of maximal glycogem storage capacity, this study also demglycogen synthesis consumes 2.5/36 or 7% ofthe glucose stored as glycogem. The transformation ofglycogem into onstrates that de novo lipogemesis can become a major metabolic pathway for the disposal ofexcess glucose car- fatty acids, the subsequent esterification before export bons. This is not only evident from the respiratory cx-from the liver, and them triglyceride storage in adipose tissue consume additional ATP, estimated at 18%. Thus change data but also from the increases observed in of the glucose channelled into de plasma triglyceride concentrations and the proportion of-25% of the energy plasma VLDL. Because very little triglyceride was pro- novo lipogemesis can be expected to be needed for this process. vided in the diet at this time, it must have originated as newly formed triglyceride in the liver, which is the princiOfthe energy consumed in excess ofmaintenance en75% was retained and 25% dissipated. Such a high pal site ofde novo lipid synthesis in man (29). Our labo- ergy, ofenergy consumed in excess can only ratory and others (5, 27, 30) showed that de novo lipo- rate ofdissipatiom be about by conditions leading to the induction genesis does not contribute to increasing the body fat brought ofhigh rates ofcarbohydrate conversion into fat. Indeed, stores even when very large amounts of carbohydrate in daily energy expenditure was only 8% (500 g) are occasionally consumed. Glycogen storage fol- the increase the day 1 before lipogenesis became necessary for lowed by high subsequent rates ofglucose oxidation camduring of some of the excess carbohydrate calories easily accommodate the daily ingestion ofrelatively large the storage consumed. Subsequently, when the nonprotein respiraamounts of carbohydrate without there being a need to tory quotient became markedly > 1 .0, energy expendiconvert carbohydrate to fat. causing the dissipation of nearly 30% Our data suggest that glycogen stores must increase byture rose further, 500 g before appreciable de novo lipogenesis begins. ofthe calories consumed in excess. our findings indicate that the bodys glycogen Provided that massive amounts of carbohydrate con- Finally, filled under normal ad libitinue to be ingested, the glycogen stores become satu- stores are far from completely tum conditions. Ifthe glycogen stores are not limited by rated so that the only way of disposing of additional cxsaturation ofthe glycogen storage capacity, one cess carbohydrate is by fat synthesis in addition to maxi- physical mal use ofglucose for energy generation. Although it has can more readily envision that individual differences and responsiveness to food palatability and accessibility may been suggested that the capacity for de movo lipogenesis
=

Downloaded from www.ajcn.org by guest on April 14, 2012

CARBOHYDRATE

DISPOSAL

IN OVERFEEDING
6-phosphate
enzymic

247
of

influence to a considerable extent the range within which glycogen stores are spontaneously maintained. In turn this will affect the relative contributions that glucose and 15. free fatty acids tend to make to the metabolic fuel mix used for energy generation and the conditions for which 16. the steady state of body weight maintenance tends to be achieved (32). 0
17.

Herbert

dehydrogenase. In: Bergmeyer HU, ed.: Methods New York: Academic Press, 1965: 17-23. 1 V, Lau KS, Gottlieb CW, Bleicher Si. Coated charcoal
analysis.

immunoassay ofinsulin. J Clin Endocrinol Metab l965;25: 137584. Heindel JJ, Cushman SW, Jeanrenaud B. Cell associated fatty acid levels and energy requiring processes in mouse adipocytes. Am J

Physiol

l974;226:

16-24.

Dole VP, Meinertz H. Microdetermination of long chain fatty We thank the dieticians Fiona Hunter, Carolyne Summerbell, and acids in plasma and tissues. J Biol Chem l960;235:2595-9. Nicole Baudat; A Beccarelli, J Braissant, D Kock, and K Rocafi for 18. Bucolo 0, David H. Quantitative determination ofserum triglyctheir technical assistance; and Dr J Frei, Dr T Lemarchand, E Temler, erides by the use ofenzymes. Gin Chem 1973; 19:476-82. and D Penseyres for the blood analyses. 19. Technicon. Technicon autoanalyser methodology. Simultaneous

glucose/BUN
20. Southgate

Method
DAT, Durnin

N-l6b.
JVOA.

Tarrytown,
Calorie

NY: Technicon,
conversion factors.

1967.
An cx-

References
1. Bergstrom

cogen and 140-50.

2. 3. 4. Hultman

J, Hermansen L, Hultman E, Saltin physical performance. Acts Physiol E, Nilsson LH.

B. Diet,

muscle

gly21.

perimental the energy

Hedman

reassessment value ofhuman

R. The available

of the factors used in the calculation diets. BrJ Nutr l970;24:517-35.

glycogen in man and the

of

Scand

l967;7 1:

between muscularexercise.

rate ofoxygen

intake

and carbohydrate

usage. Acta

connection Phys-

dietsand

5.

6.

7.

Liver glycogen in man. Effect of different 22. Adv Exp Med Biol 1971; 11:143-51. Nilsson LH. Liver glycogen content in man in the postabsorptive state. Scand J Clin Lab Invest l973;32:3 17-23. Hultman E. Muscle glycogen in man determined in needle biopsy 23. specimens method and normal values. Scand J Gin Lab Invest 24. 1967; 19:209-17. Passmore R, Swindells YE. Observations on the respiratory quotients and weight gain ofman after eatinglarge quantities of carbo25. hydrate. BrJ Nutr 1963; 17:33 1-9. 26. Acheson KJ, Flatt JP, J#{233}quierOlycogen synthesis versus lipoE. genesis after a 500 gram carbohydrate meal man. Metabolism in 1982;3l: 1234-40. Acheson IU, Schutz Y, Bessard T,Ravussin E, J#{233}quier Hart E,

iol Scand l957;40:305-2l. Bj#{246}rntorp P, Sj#{246}str#{246}m L. Carbohydrate storage in man: speculations and some quantitative considerations. Metabolism l978;27(suppl 2): 1853-65. Olsson KE, Saltin B. Variations in total body water with muscle glycogen changes in man. Acts Physiol Scand l970;80:1 1-8. GarrowJS. Energy balance and obesityn man. Amsterdam: i North Holland Publishing Co. 1974. Hultman E, Bergstrom J, Roch-Norland AE. Olycogen storage in human skeletal muscle. Adv Exp Med Biol 197 ; 1 1:273-88. 1 Huijing F, Nuttal FQ, Villar-Palasi C, Lamer J. UDP glucose: l,4-glucan a-4-glucosyltransferase in heart regulation ofthe activity of the transferase in vivo and in vitro in rat. A dissociation the action of insulin on transport and on transferase conversion.

Downloaded from www.ajcn.org by guest on April 14, 2012

ain

influences on lipogenesis and thermogenesis after a Biochim Biophys Acts 1969; 177:204-12. meal. Am J Physiol l984;246:E62-70. 27. Weiss L, Hoffmann OE, Schreiber R, et al. Fatty-acid biosynthesis 8. Ravussin E, Burnand B, Schutz Y, J#{233}quier Twenty-four-hour E. in man, a pathway ofminor importance. Purification, optimal assay conditions and organ distribution of fatty-acid synthase. Biol energyexpenditure and resting metabolic rate in obese, moderately obese and control subjects. Am J Gin Nutr 1982; 35:566-73. Chem Hoppe Seyler 1986; 367:905-12. 9. Kaltenbach M. Mangercorrectement; maiscomment?Zurich: Fed- 28. Elwyn DH, Oump FE, Munro HN, Iles M, Kinney JM. Changes in nitrogen balance ofdepleted patients with increasing infusions Cration des cooperatives Migros, 1984. 10. Hawk PB. Practical physiological chemistry. 12th ed. Toronto: ofglucose. Am J Gin Nutr 1979;32: 1597-6 1 1. Blackiston, 1947. 29. Angel A, Bray GA. Synthesis offatty acids cholesterol by liver, and I 1. Carpenter Ki, Anantharaman K. The nutritional value of poor adipose tissue and intestinal mucosa from obese and control paproteins fed at high levels. I The growth ofrats. Br J Nutr 1968;22: tients. Euri Gin Invest l979;9:355-62.
30. Sj#{246}str#{246}m L Adult human adipose tissue cellularity and metaboM, Burnand B, PittetPh, J#{233}quier Metabolic E. effects of a lism. Acts Med Scand [Suppl] l972;544: 1-52. mixed and ahigh-carbohydrate low-fat diet in man measured over 31. Hart JP. The biochemistry of energy expenditure. In: Bray GA, 24 h in a respiration chamber. BrJ Nutr 1982;47:33-43. ed. Recent advances in obesity research. Vol 2. London: Newman, 1978:21 1-28. 13. Lusk 0. Animal calorimetry analysis ofthe oxidative mixtures of 32. Hart JP. Dietary fat, carbohydrate balance, and weight maintecarbohydrate and fat. J Biol Chem l924;59:4l-2. nance: effectsofexercise. AmJ Gin Nutr l987;45(suppl):296-306. 14. 51cm MW. D-Glucose determination withhexokinase and glucose12. 183-97. Hurni

JP. Nutritional carbohydrate