original communications

290 The American Journal of Clinical Nutrition 24: MARCH 1971, pp. 290-296. Printed in U.S.A.

EfFect on body composition and other parametersin obese young men of carbohydrate levelof reduction diet12Charlotte M. Young,3 Ph.D., Sonia S. Scan/an,4 M.N.S., Hae Sook Jm,5 M.N.S., andLeo Lutwak,6 Ph.D., M.D.

Very limited studies have been made ofchanges in body composition during weightreduction on various dietary regimens, par-ticularly over any extended period of timeand with subjects engaged in their usual ac-tivities. Practical interests are concernedwith reduction regimens that can be fol-lowed for extended periods as individualspursue their normal routines.

In this laboratory considerable experiencehas been accumulated with the weight re-duction of both young men and young womenwho are moderately obese, using a low carbo-hydrate, high protein, moderate fat type ofdiet originally promulgated by Ohlson (1-5).This type of diet seems to have certain ad-vantages for long-term weight loss: a) it canbe adequate in all nutrients except calories;b) it is simple to follow both at home andaway from home; and c) it may be expandedeasily to a lifetime pattern of eating whenone is ready for weight maintenance. Afurther asset is that patients are notably freefrom hunger, and fatigue, though presentduring the reduction process, seems as limitedas for any regimen used. For moderatelyobese young males in the usual sedentary oc-cupations, an 1 ,800-kcal level has been usedwith 115 g protein, 104 g carbohydrate, and103 g fat; 50% of the calories are from fat.Recently there have been vigorous advocates,especially in the popular press, for stillgreater restriction in carbohydrate in the re-duction diet. Is there any advantage to begained?

The present study was undertaken to corn-

pare results of the low carbohydrate dietpreviously used extensively with the resultsof two isocaloric, isoprotein diets in whichcarbohydrate was reduced by approximatelyone-half to 60 g and then one-half to 30 gdaily. Thus, the study compares the effect ofvarying levels of carbohydrate and fat in amoderate reduction regimen over an ex-tended period on moderately obese youngmen engaged in their usual activities. Com-parisons will be made in terms of weightloss; changes in body composition measuredby density, anthropometric (envelope) meas-urements, skin-fold thicknesses, and balancesof nitrogen, sodium, and potassium; bloodlipid levels; qualitative excretion of ketonebodies in the urine; and subjective response.Of the three levels of carbohydrate used,which appeared to be the most desirable?

Experimental

The subjects were eight young college men aged20.6 to 27.8 years (median, 22.8; mean 23.3)(Table 1). All were science majors; four weregraduate students and four were upper-class under-graduates. Originally there were nine subjects butthe data from one were discarded because of pos-

From the Graduate School of Nutrition, Cor-nell University, Ithaca, New York 14850.

2 Supported in part by grants from the NationalLivestock and Meat Board, Chicago, Frank E.Gannett Newspaper Foundation, Rochester, NewYork, and state appropriations through the StateUniversity of New York.

Professor of Medical Nutrition. ResearchAssistant. 6 Graduate Assistant. Jami-son Professor of Clinical Nutrition.

by guest on August 5, 2013ajcn.nutrition.org

Dow

nloaded from

EFFECT OF CARBOHYDRATE LEVEL ON BODY COMPOSITION 291

TABLE 1Description of subjects

Mean Median Range

Age, years 23.3 22.8 20.6 - 27.8Height, cm 176.2 179.1 163.8 -182.4Weight, kg 98.0 98.6 78.9 -113.6Density, g/cc 1.0317 1.0298 1.0179- 1.0531Specific gravity 1.0379 1.0363 1.0237- 1.0590Fat, % body wt 30.20 30.94 19.49 - 37.56Skin-fold thickness, mm 305.4 306.2 266.8 -353.2

sible irregularities in collections. Median bodyweight was 98.6 kg; median body density, 1.0298;and median percent body weight as fat, using theRathbun and Pace formula, was 30.94%, roughlythree times the usual percent body weight as fat forthis age group.

The general experimental plan was as follows.For approximately 3 weeks subjects were on amaintenance diet for base-line observations. Thenthree as comparable as possible subgroups of threeeach were formed on the basis of maintenancecaloric requirements and percent body weight asfat. Because of the discard of data for one subjectoriginally assigned to subgroup A (leaving only twosubjects), the resulting average initial character-istics of the subgroup are not as comparable aswhen they were formed (Table 2).

TABLE 2Characteristics of diet subgroups

MeanDiet subgroups

A B C

Weight, kgDensity, g/mlFat, % body wtSkin-fold thick-

ness, mm

Maintenance,kcal

98.551.0349

28.40275.4

3,700

95.821.0319

30.20314.9

3,666

102.21.0314

31.39315.8

3,666

For 9 weeks (interrupted after 3 weeks for 1week of spring vacation), one subgroup each wasplaced on isocaloric, isoprotein diets (1,800 kcal,115 g protein) at three different levels of carbo-hydrate: Diet A, 104 g; Diet B, 60 g; and Diet C,30 g/day. Table 3 gives the details of the diet plan.

All food intake was weighed and consumed atour Special Diet Table Unit while students attendedschool and participated in their usual noneating,nondrinking activities. Physical activity was notcontrolled, though a retrospective questionnaire, re-lated to these and other matters and completed atthe end of the study, gave estimates of energy cx-penditures. Only glass-distilled water was used in

cooking, to drink, and to prepare tea and coffee,which with water were allowed ad libitum and re-corded daily. Food was prepared without salt andeach man had his daily supply to be used in totoin 24 hr. Sodium intake during maintenance wascalculated to be an NaC1 equivalent of 11 g/day;during weight reduction, of S g/day.

Each subject was weighed in fasting conditionand without shoes or jacket, prior to any othermorning activity.

Nitrogen, sodium, and potassium balances wereconducted during the last 10 days of the 3-weekweight maintenance period and during the first 3and last 3 weeks of the 9-week weight reductionperiod. Account was not taken of cutaneous lossesexcept in the interpretation of sodium and potas-sium balances. All periods were 7 days in lengthexcept the 10-day maintenance period and the thirdweek of reduction, which was a 6-day period.

Estimates of all other parameters of body com-position were made at four intervals: the last weekof weight maintenance and the 3rd, 6th and 9thweek of weight reduction. Urinary creatinineswere measured daily and ketone bodies in theurine were estimated qualitatively daily duringbalances of weight reduction by means of Ketostixand of the nitroprusside test.

Body density was determined by underwaterweighing with simultaneous measurement of resi-dual air (6). For calculating body fat from density,the Rathbun and Pace formula was used (7). Some14 skeletal and 9 envelope measurements weremade using techniques described elsewhere (8).Skeletal measurements included: height; iliocristalheight; sitting height; biacromial, biiliac, and bi-trochanteric diameters; chest width at xiphoidlevel; knee to knee, knee, ankle, and wrist widths;and knee, ankle, and wrist circumferences.Envelope measurements included shoulder, chest(xiphoid level), biceps, upper arm, forearm, calf,abdominal at hypochondriac and abdominal atumbilicus level circumferences and deltoid di-ameter. In addition, 12 skin-fold measurementswere made as previously described (8) except thatthe skin fold over the biceps was added and thathalfway between the umbilicus and the pubis onthe midabdominal line was deleted.

Data will be viewed as comparisons betweendiet subgroups by carbohydrate level: A (104 g), B

by guest on August 5, 2013ajcn.nutrition.org

Dow

nloaded from

292 YOUNG ET AL.

TABLE 3Diet plan

Xutrient

Maintenance Reduction

Grams % Calories

A

Grams % Calories

B

Grams % Calories

C

Grams % Calories

ProteinCarbohydrateFatCalories

115425160

3,600

12.847.240.0

100

115104103

1,800

25.523.151.4

100

11560122

1,800

25.613.361.1

100

11530

135.51,800

25.56.7

67.8100

(60 g) and C (30 g) on an average basis, thoughin most instances individual data are included.

Results

Weight lossWeight maintenance caloric needs varied

from 3,400 to 4,000 kcal/day. Table 4 givesthe weight losses by isocaloric, isoproteinsubgroups with diminishing carbohydrate in-take during the 9 weeks reduction period.Losses ranged from 9.19 to 18.14 kg; themean was 13.82; the median, 12.87 kg.Group A with 104 g carbohydrate daily lostan average of 11.85 kg; group B with 60 gcarbohydrate daily lost an average of 12.78kg; group C (30 g carbohydrate daily) anaverage of 16.18 kg. Thus, there seems tohave been a slight increase in weight lossas carbohydrate in the diet decreased. How-ever, it must be remembered that physicalactivity was not controlled. The retro-spective reporting of activity would indicatethat the average caloric expenditures ofall groups were very comparable, about3,400 kcal/day for groups A and C; 3,300for group B.

TABLE 4Weight losses in kilograms, duringreduction period

SubjectsGroup and carbohydrate level

A (104 g/day) B (60 g/day) C (30 g/day)

123

MeanMedianRange

9.1914.52

11.8511.85

9.19-14.52

13.0412.5912.7012.7812.70

12.59-13.04

12.2518.1418.1416.1818.14

12.25-18.14

Ketone excretion

An additional factor that may contributesomewhat to the differences in weight losseson the isocaloric diets is the variation in lossof energy through the excretion of ketonebodies in the urine (Table 5). Unfortunatelythe ketone measurements were only qualita-tive tests, but the two tests used agreed verywell in their relative estimations. In general,it can be said that the daily tests of the urineshowed increasing ketone bodies with de-creasing carbohydrate intake and that ketonebodies continued to increase and for a longertime as the carbohydrate level in thediet decreased. For group A, during the first2 weeks, a low to moderate level of ketonebodies appeared in the urine but this wasmuch diminished by the 3rd week; by the 7ththrough 9th week, they had disappeared en-tirely. For group B the picture was much thesame as group A in the first 3 weeks, but inthe last 3 weeks, ketones were absent for onesubject, questionable for another, and at thelowest detectable level for the third. In con-trast, all of group C had the heaviest load ofketones in the urine of any of the subjectsand the level remained high throughout thefirst 3 weeks. By the last 3 weeks, the levelwas somewhat lighter but still high in allthree subjects. Continued heavy loss of ke-tone bodies in the urine on the very re-stricted carbohydrate intake ultimately wouldresult in a loss of base from the body andpossible acidosis, as well as some loss ofenergy.

Changes in body fatnessIn Table 6 are presented changes in body

weight, and in fatness or lean body mass, orboth, based on differences between the first set

by guest on August 5, 2013ajcn.nutrition.org

Dow

nloaded from

TABLE 5Qualitative ketone losses in urine by diet group

Diet group\Veek of reduction

23789

A

++

++

+000

B

++++

+

C

++++++++

+++++

++

Subjectby group

Weightloss, kg

Fat loss

%body kweight g

Weight lostas lat, %

Density oftissue lost,

g/ml

Skin-fold loss Circumferenceloss

mm % cm C.

Al

2

Mean

Bi23

Mean

Cl23

Mean

8.50

13.88

11.19

13.3811 .5611.80

12.25

10.7715.2515.99

14.00

5.71

6.67

6.19

8.427.118.12

7.88

11.3

10.1

10.7

6.56

10.2

8.38

9.899.94

10.9

10.2

14.7

15.0

14.85

77.2

73.5

75.3

73.986.092.4

84.1

96.4

93.8

95.1

0.9424

0.9572

0.9498

0.95440.93530.9240

0.9379

0.9181

0.9206

0.9193

118.5

142.7

130.6

155.6164.7155.6

158.6

137.2199.8173.5

170.2

44.4

50.2

47.3

59.749.844.1

51.2

46.762.751.8

53.7

Changes increatinine

excretion, g/24 hr

+0.371

-0.025

+0. 173

+0.043+0.863-0.012

+0.298

+0. 232+0.360+0.553

+0. 382

45.6

52.8

49.2

57.456.451.9

55.2

56.361.663.3

60.4

7.9

8.1

8.0

9.19.18.3

8.8

9.19.69.4

9.4

EFFECT OF CARBOHYDRATE LEVEL ON BODY COMPOSITION 293

a Experimental error in density measurement at 9th week of weight reduction for subject Cl. Averagesfor C subgroup based on only C2 and C3.

of tests (approximately the last week of themaintenance period) and the fourth set oftests (the last week of the reduction period).Calculations are based on changes in nudeweight, density, total of skin-fold thicknesses,total of envelope or circumference measure-ments, and average creatinine excretion. Thetotal number of days between tests is notexactly equal, being slightly greater on theaverage for group A than groups B and C,which are essentially equal.

Note that body fat losses on the averageincreased progressively with less carbohy-drate. The same is true for the percent ofweight loss that is fat. Over the 9-week re-duction period close to 100% of the weightlost by group C, the lowest carbohydrate

TABLE 6

group, was fat. The decreases in total of skin-fold thicknesses and of body circumferences,whether expressed in absolute amounts orpercentage of maintenance values, againshowed the greatest losses in the lowest car-bohydrate group, C, and least in the highestgroup, A, with the intermediate carbohy-drate level group in the intermediate position.Curiously, the groups fell in the same relativepositions with regard to increases in averagecreatinine excretions between the two testperiods.

Nitrogen balances

Table 7 gives the average nitrogen re-tentions per day and the balance status ofsubjects by period and by diet groups. Duringweight maintenance all the subjects were re-taining nitrogen except one in group B whowas in balance. In the first week of weightreduction all but one subject in group Bwere losing nitrogen regardless of diet group.Thereafter, all subjects retained nitrogen orwere in balance except one in group B wholost nitrogen throughout the reduction period.(In the 3rd week of reduction the loss was notquite great enough to be classified L.) Thehighest carbohydrate group, A, initially re-tained more nitrogen early in reduction, i.e.,2nd and 3rd week. However, by the 7th

Changes in body fatness

by guest on August 5, 2013ajcn.nutrition.org

Dow

nloaded from

294 YOUNG ET AL.

TABLE 7Average nitrogen retentions and balance status by period and diet group

Period

Diet group

A B C

N, g/day Balance N, g/day Balance N, g/day Balance

MaintenanceReduction 1

23

789

0.61-1.58

1.593.22

1.480.960.63

2 R2 Lc2R2R

2 RR BRB

1.93-1.74

0.001.19

0.890.610.25

2R BblB 2LRBL2R1B

2R lL2R 1L2R1L

2.69-1.55

0.001.48

1.772.511.37

3R3L3B1R2B

3R2R lB2R1B

a R = Retention >5% intake. B = Balance, within 5% intake. c L = Loss >5% intake.

through the 9th week the lowest carbohydrategroup, C, had the highest average nitrogenretentions. The individual who had the low-est percent weight loss as fat (Table 6) andwho also reached the lowest percent bodyweight as fat, had the greatest nitrogen lossesthroughout. Group C, with the highest per-cent weight loss as fat, also had the highestaverage nitrogen retentions as weight lossprogressed.

Thus, there was no marked difference indiet groups with regard to nitrogen balance,and one cannot say that any one level of car-bohydrate was better than any other insofaras nitrogen retention is concerned.

Sodium balance

On the weight maintenance diet with a cal-culated sodium intake equivalent to 11 gNaCl/day, urinary sodium excretions rangedfrom 2.7 to 3.27 g/day (mean, 3.10; median,3.17). On the isocaloric reduction diets witha calculated sodium intake equivalency of 5g NaCl/day, average daily sodium urinaryexcretions ranged from 1.2 to 2.0 g in all 3diet subgroups. The averages kept falling inthe first 3 weeks so gradually that a slightlymore positive balance was obtained. How-ever, in some cases in the 7th through 9thweek, the sodium excretion was slightlyhigher and sodium balances lower.

There were no marked group differencesin sodium balances. The subject known toexercise most vigorously had consistentlylower urinary sodium excretions and ap-

parently higher balances, undoubtedly due tomore sodium lost in sweat. Overall sodiumbalances were only slightly positive exceptfor this one individual, but the slight posi-tive balances (range 0 to 47 mEq on a dailybasis; mean, 18.69 mEq; median 17.0 mEq)were easily accountable by loss in insensibleperspiration. Least sodium retention wasfound in the 1st week of reduction (after thechange in sodium intake), with the greatesturinary sodium excretion then for all subjects,particularly in group C, the lowest carbohy-drate group.

Potassium balances

During weight maintenance the urinarylosses of potassium ranged from 3.06 to 4.04g/day; mean, 3.29; median, 3.26. Duringweight reduction urinary losses ranged from1.7 to 3.29 g/day; mean, 2.24; median, 2.11,except for one subject.

All subjects except one were in slightlynegative potassium balance throughoutweight reduction. Potassium loss in the 1stweek of reduction was markedly higher in al-most all instances; after the 1st week it wasreasonably constant except in one subject inthe last 2 weeks when he was nearing normalbody fatness and was probably breakingdown lean tissue. He was the only subject innegative nitrogen balance during these weeks.One subject who exercised most appeared tobe retaining small amounts of potassium inthe last 3 weeks of reduction while he wasalso retaining a fair amount of nitrogen. Po-

by guest on August 5, 2013ajcn.nutrition.org

Dow

nloaded from

EFFECT OF CARBOHYDRATE LEVEL ON BODY COMPOSITION 295

tassium losses appeared to approximatelyparallel weight losses in subjects, period byperiod.

Blood lipidsLimited data are available on blood lipid

trends in weight reduction as compared withweight maintenance. In general, the bloodlipids during weight reduction as comparedwith two sets of values obtained duringweight maintenance showed no clear-cut pattern by diet group. Cholesterol valuestended to decrease a bit with diet A. Fordiets B and C, in each case for two of thethree subjects, cholesterol tended to increaseand for one, to decrease. There was a littlemore tendency for total lipids to rise forsubjects on diet C (lowest carbohydrate andgreatest weight loss) than for those of diet Aand to some slight extent than diet B. Serumphospholipids did not increase in any group.Triglycerides tended to drop in two of threecases on both diets B and C. Esterified fattyacids tended to decrease or be the same ondiets A and B and were slightly up in onecase on diet C.

Subjects responsesExcept under circumstances of unusual

physical activity, after the first week hungerwas not a problem to any of the subjects. Sixof the eight commented on a feeling of tired-ness or weakness. This was true for all threesubjects on diet C, two on diet B and one ondiet A. The subjects on diet C expresseda desire for more roughage, especially saladsand vegetables.

Discussion

The literature with regard to the effects ofthe level of carbohydrate and of fat in thediet on rate of weight loss has been ade-quately reviewed in this journal (9). Thesupposed superiority of an isocaloric, lowcarbohydrate diet in promotion of weightloss has been explained by sodium and fluidexcretion in what often have been very short-term studies. Total body water techniques donot seem to be sufficiently precise to measurethe changes in fluid balance.

In the current practical study using three

levels of low carbohydrate diets of isocaloric,isoprotein content, some differences in aver-age weight loss have been experienced over a9-week period with decreasing levels of car-bohydrate intake. No adequate explanationcan be offered. These were students pursuingtheir normal activities. Groups were matchedoriginally in terms of caloric requirements forweight maintenance over a three-week periodand percent body weight as fat. Differencesin energy expenditure might be an explana-tion, though retrospective comparisons viaquestionnaire at the conclusion of the studydid not appear to indicate that this was true.There were no outstanding differences innitrogen, sodium, or potassium balancesamong the diet groups. The outstanding dif-ference between diet groups was in the dailyexcretion of ketone bodies in the urine,which increased with decreasing carbohy-drate, and on the lowest carbohydrate levelcontinued in abundance throughout thestudy. Heavy loss of ketone bodies on themost restricted carbohydrate intake may wellhave contributed to the greater weight losses;however, it is unlikely to be a total explana-tion.

In terms of practical interests, in retro-spect it would appear that of the low carbo-hydrate diets used in this study, the one atthe 104-g carbohydrate level might be themost suitable for long-term use for a numberof reasons: a) good weight loss may beachieved; b) a more varied and nutritionallyadequate diet might be planned; c) there isless loss of ketone bodies in the urine and lessdanger of loss of base with the potential foracidosis; d) there may be less likelihood ofadverse effects on blood lipids; and e) therewould be an easier transition from reductionto a normal weight maintenance diet.

From the appearance of ketone bodies inthe urine over more than a transient period,it would appear that the carbohydrate levelon the 1,800-kcal, 1 15-g protein diet proba-bly should be slightly above 60 g.

Summary

Moderately obese young college men pur-suing their usual activities were studied firstduring a 3-week prereduction weight main-

by guest on August 5, 2013ajcn.nutrition.org

Dow

nloaded from

296 YOUNG ET AL.

tenance period and subsequently were dis-tributed into three isocaloric, isoprotein dietsubgroups, which varied as to level of carbo-hydrate in the diet. On the 1,800-kcal reduc-tion diet consumed over a 9-week period,diet A contained 104 g carbohydrate/day;diet B, 60 g; diet C, 30 g. The three-mansubgroups were matched as closely as possi-ble on the basis of maintenance caloric re-quirement and percent body weight as fat.

Weight loss, fat loss, and percent weightloss as fat appeared to be inversely related tothe level of carbohydrate in the isocaloric,isoprotein diets. No adequate explanationcan be given for weight loss differences.

There were no outstanding differences innitrogen, sodium, or potassium balancesamong the diet groups. The little effect onblood lipids that was observed was probablymost favorable to the highest carbohydratediet group A.

The outstanding difference among dietgroups was in the daily urinary excretion ofketone bodies. Although initially all sub-jects showed some ketone bodies in the urine,the bodies increased with decreasing carbo-hydrate and the urine continued to showthem longer (throughout the 9-week period)and in greater abundance on the most re-stricted carbohydrate intake.

Any of the low carbohydrate levels in thereduction diet under study were effective incontrolling hunger.

From a practical point of view, it wouldappear that of the low carbohydrate dietsused, the one at the 104-g level would bemost suitable for long-term use for the rea-sons cited.

From the appearance of ketone bodies inthe urine over a more than transient period,it would appear that the carbohydrate levelon the 1,800-kcal, 1 15-g protein diet shouldbe slightly above 60 g.

References

1. YOUNG, C. M. Weight reduction using amoderate fat diet. 1. Clinical responses andenergy metabolism. I. Am. Dietet. Assoc. 28:410, 1952.

2. YouNG, C. M. Weight reduction using amoderate fat diet. 2. Biochemical responses. J.Am. Dietet. Assoc. 28: 529, 1952.

3. YOUNG, C. M., I. RINGLER AND B. J. Gmt.Reducing and post-reducing maintenance on themoderate fat diet. Metabolic studies. I. Am.Dietet. Assoc. 29: 890, 1953.

4. YOUNG, C. M., E. L. EMPEY, V. U. SERRAONAND Z. H. PIERCE. Weight reduction in obeseyoung men. Metabolic studies. I. Nutr. 61: 437,1957.

5. YOUNG, C. M., AND M. M. DIGIAc0M0. Proteinutilization and changes in body composition dur-ing weight reduction. Metab. Clin. Exptl. 14:1084, 1965.

6. YOUNG, C. M., B. A. GEmUN0, S. H. MEluuu.AND M. E. Kriut. Metabolic responses of youngwomen while reducing. Body fatness and nitro-gen metabolism. I. Am. Dietet. Assoc. 36: 447,1960.

7. RATHBUN, E. N., AND N. PACE. Studies on bodycomposition. 1. The determination of total bodyfat by means of the body specific gravity. I.Biol. Chem. 158: 667, 1945.

8. YOUNG, C. M., M. E. K. MARTIN, M. CumAN,M. MCCARTHY, M. J. MANNIELLO, E. H.HARMUTH AND J. H. FRYER. Body compositionof young women. Some preliminary findings. I.Am. Dietet. Assoc. 38: 332, 1961.

9. BORTZ, W. M., A. WROLD5ON, P. Moiuus ANDB. IssEKuTz, JR. Fat, carbohydrate, salt, andweight loss.Am. I. Cliii. Nutr. 20: 1104, 1967.

by guest on August 5, 2013ajcn.nutrition.org

Dow

nloaded from