Human Nutrition and Metabolism - Qivana Products

Human Nutrition and Metabolism

Increased Dietary Protein Modifies Glucose and Insulin Homeostasisin Adult Women during Weight Loss1,2

Donald K. Layman,*†3 Harn Shiue,† Carl Sather,† Donna J. Erickson* and Jamie Baum†

*Department of Food Science and Human Nutrition, †Division of Nutritional Sciences, University of Illinoisat Urbana-Champaign, Urbana, IL 61801

ABSTRACT Amino acids interact with glucose metabolism both as carbon substrates and by recycling glucosecarbon via alanine and glutamine; however, the effect of protein intake on glucose homeostasis during weight lossremains unknown. This study tests the hypothesis that a moderate increase in dietary protein with a correspondingreduction of carbohydrates (CHO) stabilizes fasting and postprandial blood glucose and insulin during weight loss.Adult women (n � 24; �15% above ideal body weight) were assigned to either a Protein Group [protein: 1.6g/(kg � d); CHO �40% of energy] or CHO Group [protein: 0.8 g/(kg � d); CHO �55%]. Diets were equal in energy(7100 kJ/d) and fat (50 g/d). After 10 wk, the Protein Group lost 7.53 � 1.44 kg and the CHO Group lost 6.96 � 1.36kg. Plasma amino acids, glucose and insulin were determined after a 12-h fast and 2 h after a 1.67 MJ test mealcontaining either 39 g CHO, 33 g protein and 13 g fat (Protein Group) or 57 g CHO, 12 g protein and 14 g fat (CHOGroup). After 10 wk, subjects in the CHO Group had lower fasting (4.34 � 0.10 vs 4.89 � 0.11 mmol/L) andpostprandial blood glucose (3.77 � 0.14 vs. 4.33 � 0.15 mmol/L) and an elevated insulin response to meals (207� 21 vs. 75 � 18 pmol/L). This study demonstrates that consumption of a diet with increased protein and a reducedCHO/protein ratio stabilizes blood glucose during nonabsorptive periods and reduces the postprandial insulinresponse. J. Nutr. 133: 405–410, 2003.

KEY WORDS: ● insulin ● amino acids ● leucine ● obesity ● syndrome X

Dietary requirements for amino acids remain controversial.Most studies are focused on criteria to define a minimumrequirement to maintain short-term nitrogen balance. Thisconcept is particularly useful for a limiting amino acid such aslysine, which serves as an essential amino acid for peptidestructures and has limited use as a metabolic substrate (1,2).At the other end of the spectrum, the branched-chain aminoacids (BCAA) are essential amino acids for protein synthesisand also participate in critical metabolic processes (3,4). Thesedifferences in roles among amino acids suggest that a singledefinition of requirements may not be adequate to encompassthe full range of human needs for each of the nine indispens-able amino acids.

The three BCAA, leucine, valine and isoleucine, supportnumerous metabolic processes ranging from the fundamentalrole as substrates for protein synthesis to metabolic roles asprecursors for synthesis of alanine and glutamine (5,6) and asa modulator of the insulin-signaling pathway (7–9). The po-tential for the BCAA to participate in each of these metabolicprocesses appears to be in proportion to their availability.

Experimental evidence comparing the priority of use of theBCAA for each of these individual processes is limited, butsuggests that the first priority is for aminoacylation of tRNAfor protein synthesis (10), whereas their contribution to theproduction of alanine and glutamine or their effect on thesignaling pathway is dependent on increasing intracellularconcentrations (5,11,12). The potential effect of these aminoacids on metabolic processes under physiologic conditionsremains to be explored.

The interrelationship between BCAA and glucose metab-olism was first reported to be associated with the glucose-alanine cycle (5,6). These investigators found that there was acontinuous flux of BCAA from visceral tissues through theblood to skeletal muscle where transamination of the BCAAprovides the amino nitrogen to produce alanine from pyruvatewith a corresponding movement of alanine from muscle toliver to support hepatic gluconeogenesis. Although the impor-tance of the glucose-alanine cycle has been debated, Ahlborget al. (5) reported that it accounted for �40% of endogenousglucose production during prolonged exercise.

More recently, the overall contribution of dietary aminoacids to glucose homeostasis received further support on thebasis of quantitative evaluations of hepatic glucose production.Jungas et al. (13) provided an elegant argument that aminoacids serve as a primary fuel for the liver and the primarycarbon source for hepatic gluconeogenesis. Other investigators(14,15) extended this thinking with the findings that endog-enous glucose production in the liver is a critical factor in the

1 Presented in part at Experimental Biology 2001, April 2001, Orlando, FL[Shiue, H., Sather, C. & Layman, D. K. (2001) Reduced carbohydrate/proteinratio enhances metabolic changes associated with weight loss diet. FASEB J. 15:A301 (abs.)].

2 Supported by the Cattlemen’s Beef Board, National Cattlemen’s Beef As-sociation, Kraft Foods, USDA/Hatch, and the Illinois Council on Food and Agri-culture Research.

3 To whom correspondence should be addressed.E-mail: [email protected].

0022-3166/03 $3.00 © 2003 American Society for Nutritional Sciences.Manuscript received 18 July 2002. Initial review completed 13 August 2002. Revision accepted 20 November 2002.

405

at AC

ES

Library-E on A

ugust 13, 2008 jn.nutrition.org

Dow

nloaded from

http://jn.nutrition.org

maintenance of blood glucose. After an overnight fast, glu-coneogenesis provides � 70% of hepatic glucose release, withamino acids serving as the principal carbon source (16). Thesestudies provide further evidence for a linkage between dietaryprotein and glucose homeostasis.

Recent reports have highlighted the critical need to en-hance regulation of blood glucose in overweight adults (17)and during weight loss (18). This study tests the hypothesisthat a moderate increase in dietary protein with a correspond-ing reduction of carbohydrates (CHO) improves glucose andinsulin homeostasis during weight loss. We propose thatleucine is a critical substrate in the relationship of dietaryprotein to glucose homeostasis. Currently, the minimumleucine requirement for nitrogen balance is reported to be� 38 mg/(kg � d) (19,20), whereas positive oxidative balancerequires intakes of 89 mg/(kg � d) (20). We evaluated the effectof changing the dietary CHO/protein ratio from 3.5 to 1.4 andwith doubling of dietary leucine on plasma BCAA, plasmalevels of gluconeogenic precursors alanine and glutamine, in-sulin response to meals and maintenance of fasting and post-prandial blood glucose.

SUBJECTS AND METHODS

Women (n � 24; 45–56 y old) with body weights �15% aboveideal body weight (21) were recruited from the University of Illinoiscommunity. Subjects were screened using a medical history and a24-h diet recall; subjects with known medical conditions, routine useof medications or smokers were excluded from the study. Subjectsselected for the study had minimal daily physical activity, had main-tained a stable body weight during the past 6 mo and consumed a dietthat contained 12–17% of energy as protein. These conditions wereselected as representative of average U.S. food intake (22) and tostandardize prestudy conditions. All protocols and consent forms werereviewed and approved by the Institutional Review Board of theUniversity of Illinois Urbana-Champaign.

After the initial screening period, subjects had an additionalbaseline evaluation period that included a 3-d weighed dietary recordand measurement of plasma glucose, insulin and amino acids. Thisperiod served as an initial control period for each subject. After the

baseline evaluation, subjects were divided into two groups (n � 12)based on age (50.1 � 1.1 y) and body weight (85.2 � 3.6 kg).

One group of 12 women was assigned to a moderate protein diet(Protein Group) with protein intake of 1.5 g/(kg � d) and a CHO/protein ratio �1.4. This diet provided 30% of dietary energy asprotein, 40% as carbohydrates and 30% as fats (Table 1). The secondgroup was designated a control group and consumed a high carbohy-drate diet (CHO Group) similar to their baseline diet with proteinintake at 0.8 g/(kg � d) and a CHO/protein ratio �3.5. This dietprovided 15% of the energy as protein, 55% as carbohydrates and30% as fats. Daily intakes of individual amino acids reflected totalprotein intake such that the relative amino acid composition for thetwo diets remained constant. For example, leucine accounted for7.85% of the protein consumed by the Protein Group and 7.86% forthe CHO Group. The Protein Group consumed 9.9 g/d of leucine and22.3 g/d of BCAA and the CHO Group consumed 5.4 g/d of leucineand 12.3 g/d of BCAA. Both diets were designed to produce a dailyenergy deficit of �2.09 MJ and generate weight loss of �0.6 kg/wk(23). Subjects were instructed to maintain a constant level of physicalactivity throughout the study.

The overall experiment lasted for 11 wk with wk 1 used as anInitial Control period providing baseline data for all subjects. The testperiod consisted of a 10-wk diet study. During the first 4 wk of thestudy, all food was prepared in the food research laboratory and allmeals were weighed by the research staff and also by the subjects toevaluate reliability and reproducibility of the subject weighed foodrecords. Also during the laboratory-based diet period, subjects re-ceived daily instruction by our dietitian about the menus, foodsubstitutions, portion sizes and procedures for maintaining weigheddiet records. During the final 6 wk of the study, subjects continued touse the 2-wk menu rotation while preparing meals at home. Eachweek, subjects were required to report to the research laboratory formeasurement of body weight and to review their 3-d food recordswith the research dietitian.

At wk 0, 2, 4 and 10 during the weight loss, subjects reported tothe research facility at 0700 h after a 12-h overnight fast. Bodyweights were determined by electronic scale and blood samples weredrawn. Subjects were then given a test meal providing 1.67 MJ withthe macronutrient composition similar to the respective diet treat-ments. The test meals were designed to represent common breakfastmeals and to provide energy levels similar to standard oral glucosetolerance tests (24,25). The test meals were designed to be diet

TABLE 1

Energy and macronutrient compositions of weight loss diets1

Daily intakes2

Energy

Macronutrients

Leu

Selected amino acids

ThrCHO Protein Fat Ile Val Phe

MJ/d g/d

Protein group 6.98 � 0.19 171 � 7 125 � 3 54 � 2 9.89 � 0.19 5.99 � 0.11 6.42 � 0.12 5.42 � 0.11 5.03 � 0.09CHO group 6.94 � 0.17 239 � 5 68 � 2 48 � 2 5.39 � 0.10 3.20 � 0.06 3.67 � 0.07 3.12 � 0.06 2.67 � 0.05

Test meal3

Energy CHO Protein Fat Leu Ile Val Phe Thr

MJ g

Protein group 1.69 39 33 13 2.70 1.71 1.92 1.69 1.36CHO group 1.57 57 10 12 0.66 0.36 0.41 0.44 0.28

1 Weight loss diets were designed to be equal in energy and fat. The protein group had a daily intake of protein of 1.5 g/(kg � d) and a ratio ofCHO/protein � 1.4 and the CHO group had a protein intake of 0.8 g/(kg � d) and CHO/protein � 3.5.

2 Values represent intakes from 3-d weighed records; means � SEM, n � 12.3 Values represent defined intakes fed at test meal.

LAYMAN ET AL.406

at AC

ES

Library-E on A


Dow

nloaded from


specific to evaluate the normal metabolic response to the chronicdiet. Two hours after completion of the meal, a postprandial bloodsample was drawn. Plasma samples were analyzed for amino acids,glucose and insulin. Plasma amino acids were determined by HPLC asdescribed previously (26). Plasma glucose was analyzed by a glucoseoxidase-peroxidase automated method (YSI model 2300 analyzer,Yellow Springs Instruments, Yellow Springs, OH) and insulin wasdetermined by commercial RIA kit (07–26102 ICN Pharmaceuticals,Costa Mesa, CA).

Data were evaluated using a one-way ANOVA with repeatedmeasures with diet treatment and time as independent variables(SAS Institute, Cary, NC). When significant treatment or treatment� time effects were observed (P � 0.05), differences were evaluatedusing Fisher’s Least Significant Difference test to determine differ-ences between diet treatments or differences within each diet treat-ment over time. Values are means � SEM.

RESULTS

Daily intakes of macronutrients were determined fromweekly 3-d weighed food records (Table 1). The ProteinGroup consumed 6.98 � 0.19 MJ/d with 125 g of protein and171 g of carbohydrates. The CHO Group consumed 6.94� 0.16 MJ/d with 68 g of protein and 240 g of carbohydrates.After consuming the respective diets for 10 wk, subjects in theProtein Group lost 7.53 � 1.44 kg of body weight and subjectsin the CHO Group lost 6.96 � 1.36 kg (27).

At the beginning of the study (wk 0), fasting blood glucosedid not differ between groups (Fig. 1). After 10 wk, the CHOGroup exhibited fasting blood glucose of 4.34 � 0.10 mmol/L,which was 11% lower than that of the Protein Group (P� 0.05) and represented a significant decline over time for theCHO Group (Fig. 1).

For both groups, the 2-h postprandial blood glucose valueswere �15% lower than the values after the 12-h fast, with theCHO Group reduced to 3.77 � 0.14 mmol/L (Fig. 2A). Thetiming of these measurements at 2 h after the meal was basedon established oral glucose tolerance curves, which indicatethat a 2-h time point reflects the end of the absorptive periodand the return of glucose to baseline nonabsorptive levels(24,25).

Fasting plasma insulin concentrations did not differ be-tween the groups (Fig. 2B). The test meal increased insulinlevels, which was still evident 2 h after the meal. In theProtein Group, insulin was 42% higher than fasting levels,

whereas in the CHO Group, insulin was 115% above fastinglevels. The insulin response to the test meal at 2, 4 and 10 wkincreased over time in the CHO Group (Fig. 3) and the time� diet interaction was significant.

Plasma concentrations for the indispensable amino acidsdid not differ between treatment groups after an overnight fast(Table 2). Subjects in the CHO Group had higher fastingblood levels for glutamine and for the sum of alanine plusglutamine. After the test meal, changes in plasma amino acidsreflected the indispensable amino acid content of the respec-tive diets. The Protein Group received a meal containing 33 gof protein (Table 1) and plasma concentrations of amino acidsremained significantly above fasting levels 2 h after the com-pletion of the meal (Table 2). The magnitude of the increasesvaried among the amino acids with the BCAA increasing most(68–108%) and threonine increasing least (22%). In theCHO Group, the test meal provided 10 g of protein; 2 h afterthe meal, most amino acids exhibited concentrations thatwere not different from values after a 12-h fast. Alanine andphenylalanine concentrations tended to be greater (P � 0.61and 0.78, respectively), whereas glutamine concentration was� 19% lower than the fasting concentration (P � 0.05).

Meal responses for the BCAA and the two nonessentialamino acids, alanine and glutamine, reflected fundamentaldifferences in metabolism between the groups (Table 2). Afterthe higher protein meal, the sum of the BCAA increased by76% (248 � 10 �mol/L, P � 0.05) for the Protein Group witha corresponding increase in the nitrogen-transporting mole-

FIGURE 2 Plasma glucose (A) and insulin (B) concentrations aftera 12-h fast and 2 h after a 1.67 MJ test meal in women assigned to amoderate protein diet (Protein Group) or a high carbohydrate diet (CHOGroup) during 10 wk of weight loss. Values are means � SEM, n � 12.Means without a common letter differ, P � 0.05.

FIGURE 1 Plasma glucose concentrations in fasting women as-signed to a moderate protein diet (Protein Group) or a high carbohy-drate diet (CHO Group) during 10 wk of weight loss. Values are means� SEM, n � 12. Means without a common letter differ, P � 0.05. Thelinear decline in the CHO Group was significant (R2 � 0.982; P� 0.0086).

DIETARY PROTEIN MODIFIES GLUCOSE HOMEOSTASIS 407

at AC

ES

Library-E on A


Dow

nloaded from


cules, alanine plus glutamine (42%, 296 � 22 �mol/L). Be-cause virtually no dietary alanine or glutamine escapes gut andliver metabolism during meal absorption (28), these valuesmust reflect changes in plasma amino acid flux resulting fromeither increased peripheral production or reduced visceral uti-lization of these nonessential amino acids. In the CHO Group,2 h after the test meal, the sum of the BCAA and the sum ofalanine and glutamine did not differ from the fasting concen-trations. Although plasma amino acid contractions do notreflect quantitative flux measurements, plasma concentrationsdo reflect changes in intracellular concentrations and rates ofBCAA catabolism (29,30).

DISCUSSION

The potential for amino acids to interact with glucosemetabolism is well established; however, the effect of pro-longed modification of protein intake on glucose homeostasisis unknown. In the 1970s, researchers reported that amino acidavailability supported glucose metabolism during prolongedaerobic exercise (5) or during intravenous infusions (31).Subsequently, quantitative measures of amino acid flux estab-lished the importance of liver gluconeogenesis in the mainte-nance of blood glucose during nonabsorptive periods

(13,15,16). The present study evaluated the potential of mod-erate changes in the dietary CHO/protein ratio to affect themetabolic balance between glucose and insulin homeostasisand the availability of gluconeogenic amino acids. After 10 wkof diet modification, increasing dietary protein and reducingthe CHO/protein ratio minimized postprandial and fastingchanges in blood glucose during weight loss.

Maintenance of blood glucose within the normal range of4.4–6.0 mmol/L requires a precise balance between hepaticglucose release (16) and peripheral tissue glucose use (32). Theliver regulates glucose release by balancing the disposal ofexogenous dietary glucose with endogenous production fromgluconeogenesis and glycogenolysis. The balance achieved bythe liver among absorption, de novo synthesis and storedglycogen is dependent on diet composition and stage of ab-sorption (14,16,33,34). Similarly, the use of blood glucose byperipheral tissues is a balance among both insulin dependentand insulin independent tissues and varies widely, dependingon glucose availability, hormone status and tissue energyneeds. This balance between hepatic glucose release and pe-ripheral clearance must be able to extend from minimumneeds of �80 to 120 g/d for obligate glycolytic tissues such asthe brain, nerve tissue and blood cells to levels � 400 g/dduring conditions of high dietary carbohydrate intakes.

To test the influence of the dietary CHO/protein ratio onglucose homeostasis, one approach would be to examine glu-cose absorption curves and peak insulin levels. This approachwould be essential to evaluate the meal response to a carbo-hydrate load including the potential to reduce hepatic endog-enous production (16) or to increase insulin-driven peripheralclearance (35). In nondiabetic subjects, the balance of theseresponses maintains blood glucose within normal ranges. Onthe other hand, another key regulatory challenge occurs dur-ing periods between meals when exogenous glucose is notavailable. During these periods, the body must rebalance he-patic glucose production and peripheral clearance to protectblood glucose from hypoglycemic responses. The two mostlikely periods for hypoglycemia would be in the morning afteran overnight fast or at the end of a postprandial period wheninsulin is elevated and blood glucose returns to nonabsorptivelevels (24,36). These two “nonabsorptive” periods were thefocus of the current study.

After 10 wk of controlled dietary intakes, subjects consum-ing a diet with adequate protein [0.8 g/(kg � d)] and a high

FIGURE 3 The 2-h postprandial insulin response to a 1.67 MJ testmeal in women assigned to a moderate protein diet (Protein Group) ora high carbohydrate diet (CHO Group). The insulin response was de-termined as the 2-h postprandial value minus the fasting value for eachsubject. Values represent means � SEM, n � 12. Means without acommon letter differ, P � 0.05.

TABLE 2

Blood amino acid values in women consuming either a moderate protein or high carbohydrate (CHO) diets during weight loss1

Protein group CHO Group

Fasting Test meal Change Fasting Test meal Change

�mol/L % �mol/L %

Leucine 102 � 5b 181 � 9a 77 99 � 4b 93 � 4b �6Isoleucine 50 � 3b 104 � 6a 108 50 � 2b 49 � 2b �2Valine 171 � 8b 287 � 12a 68 174 � 8b 158 � 7b �9Phenylalanine 41 � 1.5c 66 � 2.6a 61 43 � 1.8c 51 � 2.3b 18Threonine 104 � 7b 127 � 11a 22 106 � 12b 113 � 11b 7Alanine 324 � 16b 485 � 28a 50 388 � 21b 452 � 24a,b 16Glutamine 378 � 15b 513 � 24a 36 449 � 12a 363 � 19b �19�BCAA 323 � 5b 571 � 9a 77 322 � 5b 300 � 4b �7�ala gln 702 � 16c 998 � 26a 42 837 � 16b 815 � 22b �3

1 Values represent means � SEM of three trials for each of the 12 women (n � 12) at wk 2, 4 and 10. Means in a row without a common letter differ,P � 0.05.

LAYMAN ET AL.408

at AC

ES

Library-E on A


Dow

nloaded from


CHO/protein ratio (� 3.5) had lower fasting and postprandialblood glucose than subjects consuming the diet with moreprotein [1.5 g/(kg � d)] and a lower CHO/protein ratio (� 1.4).As outlined above, regulation of blood glucose must beachieved through changes in either hepatic output or periph-eral clearance, or both. In the CHO Group at 2 h after the testmeal, lower blood glucose was associated with blood insulinlevels higher than either the fasting concentration or those ofthe Protein Group. This response to the test meal seemsreasonable because the CHO Group consumed more carbohy-drates. However, although the direction of the insulin re-sponse to the test meal was expected, the magnitude of theinsulin response appeared to be disproportionate to the carbo-hydrates consumed. Carbohydrate intake from the test mealwas 46% higher for the CHO Group (57 g vs. 39 g), whereasthe insulin response to the meal was 115% higher in the CHOGroup (Fig. 2B) and the magnitude of the insulin responseincreased over the 10-wk feeding period (Fig. 3). These dataare consistent with reduced insulin sensitivity in the CHOGroup during the study. On the other hand, subjects in theCHO Group had lower fasting blood glucose than subjects inthe Protein Group with similar fasting insulin levels. If insulinsensitivity is lower in the CHO Group, reduced peripheralsensitivity should produce higher fasting blood glucose. Thesedata suggest that dietary changes in the CHO/protein ratioproduce changes in endogenous glucose regulation that likelyinclude changes in both peripheral glucose clearance andhepatic glucose production.

Estimates of the contribution of amino acid carbon to denovo glucose synthesis range from 0.5 to 0.7 g of glucose from1 g of dietary protein (28,37). In addition to the direct con-version of amino acid carbon to gluconeogenic precursors,there is also the contribution of the BCAA to the glucose-alanine cycle (5,29). Because BCAA are catabolized in skel-etal muscle, the amino nitrogen is transferred from the BCAAto �-keto glutarate, forming glutamate. The nitrogen is thentransferred from glutamate to alanine via aminotransferase, orglutamate is further aminated to glutamine. The net effect is adirect stoichiometric relationship between catabolism of theBCAA in skeletal muscle and production of alanine or glu-tamine. The alanine and glutamine produced in muscle arereleased to the blood circulation and extracted by splanchnictissues, predominately the liver and gut. Production of thesenonessential amino acids affects glucose homeostasis by therecycling of glucose carbon. Production of alanine capturespyruvate in skeletal muscle (29,31); during catabolism of glu-tamine in the gut, at least 50% of the �-amino group ofglutamine is converted to alanine via transamination (28,29).These metabolic pathways suggest that skeletal muscle uptakeof BCAA is directly linked to the quantity of alanine andglutamine produced and provision of 3-carbon substrates forhepatic gluconeogenesis.

Changes in plasma concentrations of amino acids after themeal reflect diet composition and known metabolic responsesin the gut, liver and peripheral tissues (1,28). After a mixedmeal containing protein, there is an increased rate of disposalof amino acids for protein synthesis and amino acid degrada-tion. Catabolism of individual amino acids is tissue specific.For example, after a meal, the gut is highly active in aminoacid catabolism and disposes of most glutamine and threoninebefore they reach the portal circulation (28). At the otherextreme, the gut and liver have minimal capacity to initiateamino acid degradation of the BCAA, resulting in increasedmovement of the BCAA through the blood to peripheraltissues (11,28,30). Plasma appearance of phenylalanine wouldbe expected to be intermediate between Thr and BCAA, with

minimal disposal in the gut and a relatively slow rate ofcatabolism in liver (28). Changes in the BCAA, Phe and Thrreflected these metabolic differences in tissue specificity, withpostprandial plasma Thr increased by 21%, Phe by 60% andthe sum of the BCAA by 93% in the Protein Group.

Postprandial changes in plasma levels of BCAA, alanineand glutamine after the test meal were consistent with ourproposed relationship of dietary amino acids to glucose ho-meostasis. Specifically, we proposed that postprandial increasesin BCAA are associated with increased production of alanineand glutamine and enhanced hepatic gluconeogenesis tomaintain fasting blood glucose. We found that the higherprotein meal produced anticipated increases in plasma BCAAwith corresponding increases in alanine and glutamine. Forthe CHO Group, with dietary protein at levels designed tomeet minimum needs for nitrogen balance, the 2-h postpran-dial values for BCAA were not different from fasting levelswith no change in the sum of alanine plus glutamine.

Changes in plasma amino acid concentrations are notequivalent to quantitative measurements of amino acid flux.However, concentration differences for BCAA, alanine andglutamine reflect metabolic differences. Increases in BCAAlevels in the blood relate directly to changes in intracellularconcentrations (11,29,38). Similarly, increases in tissue con-centrations of BCAA increase catabolism via the aminotrans-ferase and dehydrogenase (11,39), and increased BCAA fluxto muscle relates directly to increased production and releaseof alanine and glutamine (5,29,30,39). Summing each of theelements of the pathway, this study demonstrated that pro-longed dietary modification resulting in increased postprandiallevels of BCAA, and increased levels of alanine and glutaminecan affect glucose homeostasis in free-living subjects.

The relationship of the dietary CHO/protein ratio to he-patic glucose production was also evident in subjects afterfasting overnight. After 10 wk of consuming the energy-restricted diets, subjects receiving the higher CHO diet (239g/d) had 12% lower fasting blood glucose and the level de-clined throughout the study (Fig. 1). Associated with thereduced blood glucose, subjects in the CHO Group had com-bined alanine plus glutamine levels that were 20% higher thanthose of the Protein Group. This increase in plasma alanineand glutamine associated with a diet containing more CHOand less protein was unexpected. Increases in peripheral pro-duction of alanine and glutamine are unlikely with lowerdietary intake of the BCAA. A possible explanation could beincreased rates of protein breakdown associated with the highCHO diet. A more likely explanation for the increased alanineand glutamine levels would be a reduction in the rate ofhepatic clearance. This hypothesis is supported by reports thatthe rate of clearance of plasma alanine is proportional toplasma alanine concentration and the rate of gluconeogenesis(40). Investigators have shown that high CHO feeding orelevated insulin result in down-regulation of hepatic glucone-ogenesis on the basis of measurements of flux (16) as well asdown-regulation of gene expression of key regulatory enzymes(41).

This study evaluated the effect of sustained dietary changesin the ratio of CHO to protein intake on plasma amino acidprofiles and maintenance of blood glucose during energy re-striction. We found that a diet with increased protein andreduced levels of CHO stabilizes blood glucose during nonab-sorptive periods and reduces postprandial insulin response.Additional research utilizing substrate flux measurements isrequired to define the quantitative relationship between di-etary intake of glucose and amino acids and hepatic vs. pe-ripheral management of blood glucose levels. However, this

DIETARY PROTEIN MODIFIES GLUCOSE HOMEOSTASIS 409

at AC

ES

Library-E on A


Dow

nloaded from


study supports the hypothesis that the ratio of dietary proteinand carbohydrates can have a significant effect on metabolicbalance and specifically on glucose homeostasis during weightloss.

LITERATURE CITED

1. Waterlow, J. C., Garlick, P. J. & Millward, D. J. (1978) Protein Turnoverin Mammalian Tissues and in the Whole Body. Elsevier-North Holland, Amster-dam, The Netherlands.

2. Wolfe, R. R., Wolfe, M. H., Nadel, E. R. & Shaw, J.H.F. (1984) Isotopicdetermination of amino acid-urea interactions in exercise in humans. J. Appl.Physiol. 56: 221–229.

3. Hutson, S. M. & Harris, R. A. (2001) Leucine as a nutritional signal. J.Nutr. 131: 839S–840S.

4. Layman, D. K. (2002) Role of leucine in protein metabolism duringexercise and recovery. Can. J. Appl. Physiol. 27:592–608.

5. Ahlborg, G., Felig, P., Hagenfeldt, L., Hendler, R. & Wahren, J. (1974)Substrate turnover during prolonged exercise in man. J. Clin. Investig. 53: 1080–1090.

6. Ruderman, N. B. (1975) Muscle amino acid metabolism and glucone-ogenesis. Annu. Rev. Med. 26: 245–258.

7. Patti, M.-E., Brambilla, E., Luzi, L., Landaker, E. J., & Kahn, C. R. (1998)Bidirectional modulation of insulin action by amino acids. J. Clin. Investig. 101:1519–1529.

8. Xu, G., Kwon, G., Marshall, C. A., Lin, T-A., & Lawrence, J. C. (1998)Branched-chain amino acids are essential in the regulation of PHAS-I and p70 S6kinase by pancreatic �-cells. J. Biol. Chem. 273: 28178–28184.

9. Anthony, J. C., Anthony, T. G., Kimball, S. R., & Jefferson, L. S. (2001)Signaling pathways involved in translational control of protein synthesis in skeletalmuscle by leucine. J. Nutr. 131: 856S–860S.

10. Tischler, M. E., Desautels, M., & Goldberg, A. L. (1982) Does leucine,leucyl-tRNA, or some metabolite of leucine regulate protein synthesis in degra-dation in skeletal and cardiac muscle? J. Biol. Chem. 257: 1613–1621.

11. Harper, A. E., Miller, R. H., & Block, K. P. (1984) Branched-chainamino acid metabolism. Annu. Rev. Nutr. 4: 409–454.

12. Anthony, J. C., Yoshizawa, F., Gautsch-Anthony, T., Vary, T. C., Jeffer-son, L. S., & Kimball, S. R. (2000) Leucine stimulates translation initiation inskeletal muscle of postabsorptive rats via a rapamycin-sensitive pathway. J. Nutr.130: 2413–2419.

13. Jungas, R. L., Halperin, M. L. & Brosnan, J. T. (1992) Quantitativeanalysis of amino acid oxidation and related gluconeogenesis in humans. Physiol.Rev. 72: 419–448.

14. Pascual, M., Jahoor, F. & Reeds, P. J. (1997) Dietary glucose isextensively recycled in the splanchnic bed of fed adult mice. J. Nutr. 127:1480–1488.

15. Katz, J. & Tayek, J. A. (1998) Gluconeogenesis and the Cori cycle in12-, 20- and 40-h-fasted humans. Am. J. Physiol. 38: E537–E542.

16. Balasubramanyam, A., McKay, S., Nadkarni, P., Rajan, A. S., Farza, A.,Pavlik, V., Herd, J. A., Jahoor, F. & Reeds, P. J. (1999) Ethnicity affects thepostprandial regulation of glycogenolysis. Am. J. Physiol. 40: E905–E914.

17. Reaven, G. M. (1993) Role of insulin resistance in human disease(Syndrome X): an expanded definition. Annu. Rev. Med. 44: 121–131.

18. Ludwig, D. S. (2000) Dietary glycemic index and obesity. J. Nutr. 130:280S–283S.

19. World Health Organization (1985) FAO/WHO/UNU Expert Consulta-tion. Energy and protein requirements. WHO Technical Report no. 724. WHO,Geneva, Switzerland.

20. El-Khoury, A. E., Fukagawa, N. K., Sanchez, M., Tsay, R. H., Gleason,R. E., Chapman, T. E. & Young, V. R. (1994) The 24h pattern and rate of leucineoxidation, with particular reference to tracer estimates of leucine requirements inhealthy adults. Am. J. Clin. Nutr. 59: 1000–1011.

21. Metropolitan Life Insurance Company (1983) Stat. Bull. Metrop. LifeInsur. Co. 64: 4.

22. Third National Health and Nutrition Examination Survey (1994) Energyand macronutrient intakes for persons age 2 and over in the United States.NHANES, Phase 1, 1988–91, CDC publication no. 255, U.S. Government PrintingOffice, Washington, DC.

23. National Heart, Lung and Blood Institute (1998) Clinical guidelines onthe identification, evaluation and treatment of overweight and obesity in adults.NIH publication no. 98–4083. Washington, DC.

24. Bantle, J. P., Laine, D. C., Castle, G. W., Thomas, J. W., Hoogwerf, B. J.& Goetz, F. C. (1983) Postprandial glucose and insulin responses to mealscontaining different carbohydrates in normal and diabetic subjects. N. Engl.J. Med. 309: 7–12.

25. Wolever, T.M.S. & Bolognesi, C. (1996) Prediction of glucose andinsulin responses of normal subjects after consuming mixed meals varying inenergy, protein, fat, carbohydrate and glycemic index. J. Nutr. 126: 2807–2812.

26. Paul, G. L., Rokusek, J. T., Dykstra, G. L., Boileau, R. A. & Layman, D. K.(1996) Oat, wheat or corn cereal ingestion before exercise alters metabolism inhumans. J. Nutr. 126: 1372–1381.

27. Layman, D. K., Boileau, R. A., Erickson, D. J., Painter, J. E., Shiue, H.,Sather, C. & Christou, D. D. (2003) A reduced ratio of dietary carbohydrate toprotein improves body composition and blood lipid profiles during weight loss inadult women. J. Nutr. 133: 411–417.

28. Reeds, P. J., Burrin, D. G., Davis, T. A. & Stoll, B. (1998) Amino acidmetabolism and the energetics of growth. Arch. Anim. Nutr. 51: 187–197.

29. Wagenmakers, A.J.M. (1998) Muscle amino acid metabolism at restand during exercise: role in human physiology and metabolism. Exerc. Sport Sci.Rev. 26: 287–314.

30. Rennie, M. J. & Tipton, K. D. (2000) Protein and amino acid metabo-lism during and after exercise and the effects of nutrition. Annu. Rev. Nutr. 20:457–483.

31. Ferrannini, E., Bevilacqua, S., Lanzone, L., Bonadonna, R., Brandi, L.,Oleggini, M., Boni, C., Buzzigoli, G., Ciociaro, D., Luzi, L. & DeFronzo, R. A.(1988) Metabolic interactions of amino acids and glucose in healthy humans.Diabetes Nutr. Metab. 3: 175–186.

32. Defronzo, R. A., Bonnadonna, R. C. & Ferrannini, E. (1992) Pathogen-esis of NIDDM. Diabetes Care 15: 318–368.

33. Nuttall, F. Q., Mooradian, A. D., Gannon, M. C., Billington, C. J. &Krezowski, P. A. (1984) Effect of protein ingestion on the glucose and insulinresponse to a standardized oral glucose load. Diabetes Care 7: 465–470.

34. Krezowski, P. A., Nuttall, F. Q., Gannon, M. C. & Bartosh, N. H. (1986)The effect of protein ingestion on the metabolic response to oral glucose innormal individuals. Am. J. Clin. Nutr. 44: 847–857.

35. Hansen, B. C. (1999) The metabolic syndrome. Ann. N.Y. Acad. Sci.892: 1–24.

36. Shapiro, E. T., Tillil, H., Miller, M. A., Frank, B. H., Galloway, J. A.,Rubenstein, A. H. & Polonsky, K. S. (1987) Insulin secretion and clearance:comparison after oral and intravenous glucose. Diabetes 36: 1365–1371.

37. Gannon, M. C., Nuttall, J. A., Gamberg, G., Gupta, V. & Nuttall, F. Q.(2001) Effect of protein ingestion on the glucose appearance rate in people withtype 2 diabetes. J. Clin. Endocrinol. Metab. 86: 1040–1047.

38. Biolo, G., Tipton, K. D., Klein, S. & Wolfe, R. R. (1997) An abundantsupply of amino acids enhances the metabolic effect of exercise on muscleprotein. Am. J. Physiol. 273: E122–E129.

39. Suryawan, A., Hawes, J. W., Harris, R. A., Shimomura, Y., Jenkins, A. E.& Hutson, S. M. (1998) A molecular model of human branched-chain aminoacid metabolism. Am. J. Clin. Nutr. 68: 72–81.

40. Fafournoux, P., Remesy, C. & Demigne, C. (1983) Control of alaninemetabolism in rat liver by transport processes or cellular metabolism. Biochem. J.210: 645–652.

41. Yoon, J. C., Pulgserver, P., Chen, G., Donovan, J., Wu, Z., Rhee, J.,Adelmant, G., Stafford, J., Kahn, C. R., Granner, D. K., Newgard, C. B. &Spiegelman, B. M. (2001) Control of hepatic gluconeogenesis through thetranscriptional coactivator PGC-1. Nature (Lond.) 413: 131–138.

LAYMAN ET AL.410

at AC

ES

Library-E on A


Dow

nloaded from



A Reduced Ratio of Dietary Carbohydrate to Protein ImprovesBody Composition and Blood Lipid Profiles during WeightLoss in Adult Women1,2

Donald K. Layman,*†3 Richard A. Boileau,†** Donna J. Erickson,* James E. Painter,*†

Harn Shiue,† Carl Sather† and Demtra D. Christou**

*Department of Food Science and Human Nutrition, †Division of Nutritional Sciences and **Departmentof Kinesiology, University of Illinois at Urbana-Champaign, Urbana, IL 61801

ABSTRACT Claims about the merits or risks of carbohydrate (CHO) vs. protein for weight loss diets are extensive,yet the ideal ratio of dietary carbohydrate to protein for adult health and weight management remains unknown.This study examined the efficacy of two weight loss diets with modified CHO/protein ratios to change bodycomposition and blood lipids in adult women. Women (n � 24; 45 to 56 y old) with body mass indices �26 kg/m2

were assigned to either a CHO Group consuming a diet with a CHO/protein ratio of 3.5 (68 g protein/d) or a ProteinGroup with a ratio of 1.4 (125 g protein/d). Diets were isoenergetic, providing 7100 kJ/d, and similar amounts of fat(�50 g/d). After consuming the diets for 10 wk, the CHO Group lost 6.96 � 1.36 kg body weight and the ProteinGroup lost 7.53 � 1.44 kg. Weight loss in the Protein Group was partitioned to a significantly higher loss of fat/lean(6.3 � 1.2 g/g) compared with the CHO Group (3.8 � 0.9). Both groups had significant reductions in serumcholesterol (�10%), whereas the Protein Group also had significant reductions in triacylglycerols (TAG) (21%) andthe ratio of TAG/HDL cholesterol (23%). Women in the CHO Group had higher insulin responses to meals andpostprandial hypoglycemia, whereas women in the Protein Group reported greater satiety. This study demon-strates that increasing the proportion of protein to carbohydrate in the diet of adult women has positive effects onbody composition, blood lipids, glucose homeostasis and satiety during weight loss. J. Nutr. 133: 411–417,2003.

KEY WORDS: ● obesity ● body fat ● blood lipids ● insulin

Obesity is an important public health concern in theUnited States (1). Accumulating body fat is associated withthe onset of diverse health risks including type 2 diabetes,cardiovascular disease, cancer and osteoarthritis (2). Althoughobesity is recognized as a disorder of energy balance, causes andsolutions remain elusive. At the center of the debate aboutcontrolling obesity is the optimal balance of macronutrientsfor adult health.

To maintain or reduce body weight, diets must controlenergy intake. Diets high in fat are assumed to contribute toobesity because they are usually highly palatable and energydense. Current dietary guidelines advocate a daily intake ofmacronutrients with carbohydrates accounting for �55% ofdietary energy, fats limited to �30% of dietary energy, andprotein at �15% of energy (3–5). However, the putativebalance of dietary fat, carbohydrates and protein leading to

obesity has been challenged by evidence from epidemiologic(6,7), clinical (8–10) and experimental studies (11–14).These researchers report that high carbohydrate diets reduceoxidation of body fat (11,12), increase blood triglycerides(8,10,13) and reduce satiety (14). These reports raise newquestions about the ideal ratios of macronutrients to balanceenergy needs for adults.

Generally, the debate about an optimal ratio of macronu-trients for adults focuses on carbohydrates vs. fat; however,there are increasing questions about the role of protein in theadult diet (7,8,15). Three independent groups reported bene-ficial effects on body composition and blood lipids derivedfrom direct substitution of protein for carbohydrates in adultdiets (8–10). Although these researchers found positive effectsof increasing dietary protein and reducing carbohydrates, thereports lack a fundamental hypothesis to explain the metabolicneed for protein above current recommended dietary allow-ances (RDA)4 levels. In a companion paper (16), we proposedthat an optimal ratio of carbohydrate to protein could beevaluated on the basis of a theory of glucose homeostasis.

1 Presented in part at the Experimental Biology 2000, April 2000, San Diego,CA [Layman, D. K., Boileau, R., Painter, J., Erickson, D., Shiue, H. & Sather, C.(2000) Carbohydrates versus protein in diets for mid-life women. FASEB J. 14:A564 (abs.)].

2 Supported by the Cattlemen’s Beef Board, National Cattlemen’s Beef As-sociation, Kraft Foods, USDA/Hatch, and the Illinois Council on Food and Agri-culture Research.


4 BMI, body mass index; BHB, �-hydroxybutyrate; CHO, carbohydrate; DXA,dual X-ray absorptiometry; GNG, gluconeogenesis; HDL-C, HDL cholesterol;LDL-C, LDL cholesterol; TAG, triacylglycerol; T4, thyroxine; T3, triiodothyronine.

0022-3166/03 $3.00 © 2003 American Society for Nutritional Sciences.Manuscript received 18 July 2002. Initial review completed 31 August 2002. Revision accepted 20 November 2002.

411

at AC

ES

Library-E on A


Dow

nloaded from


Under conditions of high CHO intake (55% of energy), thebody must adapt to the disposal of large quantities of dietaryglucose and relies on insulin to manage changes in bloodglucose. Under conditions of lower carbohydrate (CHO) in-take (�200 g/d), the body relies on hepatic production tomaintain blood glucose. Glucose released from the liver can bederived from glycogenolysis and gluconeogenesis (GNG);however the larger component is GNG, with dietary aminoacids providing the primary source of carbon substrates(17,18). Applying this hypothesis to a weight loss study, wefound that obese women consuming a diet with a CHO/protein ratio �1.5 for 10 wk minimized fasting and postpran-dial changes in blood glucose and enhanced insulin sensitivity(16). These findings provide a conceptual basis on which totest the efficacy of diets with reduced ratios of CHO/protein.

The objective of this study was to compare the effects oftwo reduced energy diets with modified ratios of protein andcarbohydrates on changes in body weight, body composition,blood lipids and satiety. This study used a 10-wk, highlycontrolled nutrition protocol with adult women to evaluatethe effects of weight loss diets with moderate differences inprotein and carbohydrates.


Women (n � 24; 45 to 56 y old) with body weights �15% aboveideal body weight were recruited to participate in a weight loss studyas previously described (16). All subjects participated in a baselineevaluation period that included a 3-d weighed dietary record andmeasurement of body composition, blood lipids, glucose and insulin.This period served as an initial control period for each subject. Afterthe baseline evaluation, subjects were divided into two groups (n� 12) based on age (50.1 � 1.1 y), body weight (85.2 � 3.6 kg) andbody mass index (BMI; 30.3 � 1.0 kg/m2).

One group of 12 women was assigned to a moderate protein diet(Protein Group) designed to provide dietary protein at 1.6 g/(kg � d)with a CHO/protein ratio of �1.4 and dietary lipids at �30% energyintake. The second group was assigned to a high carbohydrate diet(CHO Group) designed to provide dietary protein at 0.8 g/(kg � d)with a CHO/protein ratio �3.5 and dietary lipids at �30% energyintake. The two diets were designed to be equal in energy (�7100kJ/d; 1700 kcal/d), total fat intake (�50 g/d) and fiber (� 20 g/d).With these general criteria, we developed a 2-wk menu plan for eachgroup with meals for each day meeting established nutritional re-quirements and lipid guidelines for the STEP I diet. The dietarydifferences between groups were designed to reflect a direct substitu-tion of foods in the protein groups (meats, milk, cheese, eggs andnuts) for foods in the refined grain/starch groups (breads, rice, cereals,pasta and potatoes). For the CHO Group, the diet followed theguidelines of the USDA Food Guide Pyramid diet (4), which em-phasizes the use of breads, rice, cereals and pasta. For the ProteinGroup, the diets substituted foods from the protein groups thatemphasized animal proteins including combinations of red meats,milk, cheese and eggs with a requirement for a minimum of sevenbeef-containing meals each week. Both diets included extensive useof vegetables (5 to 6 servings/d). Physical activity was monitored withwritten questionnaires and was kept constant for each subjectthroughout the study.

The overall experimental design included a 1-wk Initial Controlperiod providing baseline data for all subjects about their usual dietarypatterns. The baseline period was followed by a 10-wk diet study thatconsisted of an initial 4 wk of a highly controlled diet with subjectsreceiving all food in our food research laboratory followed by 6 wkwith subjects continuing to follow the 2-wk diet rotation at home.During the first 4 wk of the study, all food was prepared in the foodresearch laboratory and all meals were weighed by the research staffand also by the subjects to evaluate reliability and reproducibility ofthe subject weighed food records. During the laboratory-based dietperiod, subjects also received daily instruction by a research dietitianabout the menus, food substitutions, portion sizes and procedures for

maintaining weighed diet records. During the final 6 wk of the study,subjects continued to use the 2-wk menu rotation at home. Eachweek, subjects were required to report to the research laboratory formeasurement of body weight and to review their 3-d food recordswith the dietitian.

At times 0, 2, 4 and 10 wk, body composition, blood and urinarychemistries, and food records were evaluated. Measurements includedbody weights by electronic scale, body composition by dual X-rayabsorptiometry (DXA), plasma glucose, insulin, thyroid hormones(T4, thyroxine and T3, triiodothyronine), urea, �-hydroxybutyrate(BHB) and lipid profiles and urinary acetoacetate and urea. Plasmaglucose was analyzed by a glucose oxidase-peroxidase automatedmethod (YSI model 2300 analyzer, Yellow Springs Instruments, Yel-low Springs, OH). Insulin and thyroid hormones were determined bycommercial RIA kits (07–26102 and 06B254221, respectively, ICNPharmaceuticals, Costa Mesa, CA). Serum total cholesterol, HDLcholesterol (HDL-C) and triacylglycerols (TAG) were determined bystandardized methods (19) by the Washington University School ofMedicine Core Laboratory for Clinical Studies (St. Louis, MO). LDLcholesterol (LDL-C) was calculated using the Friedewald equation(20). Plasma urea nitrogen and urinary urea were determined with ablood urea nitrogen (BUN) kit (Sigma, St. Louis, MO), BHB wasdetermined by enzymatic conversion to acetoacetate with measure-ment of NADH (Ketone Kit, Sigma) and urinary acetoacetate wastested with a clinical urine stick (Sigma). Nutrient intakes wereevaluated as mean daily intakes from the 3-d weighed records usingNutritionist V software (First DataBank, San Bruno, CA). Individualfeelings and satisfaction were evaluated with a questionnaire com-pleted on d 3 of the 3-d food records throughout the study. Thequestionnaire contained 11 quantitative questions with a 7-pointresponse range addressing feelings about satiety, energy level andhunger.

On measurement days, subjects reported to the research facility at0700 h after a 12-h overnight fast. Body weights were determined andblood samples were drawn. Then subjects were given a standardbreakfast meal providing 1670 kJ (400 kcal) with the macronutrientcomposition designed to reflect the respective diet treatments (e.g.,Protein Group received 33 g protein, 39 g carbohydrate and 13 g fat;CHO Group received 10 g protein, 57 g of carbohydrate and 12 g fat).Subjects had a postprandial blood sample drawn 2 h after completionof the meal.

Data were evaluated using a one-way ANOVA with repeatedmeasures with diet treatment and time as independent variables(SAS Institute, Cary, N.C.). When significant treatment or treat-ment � time effects were observed (P � 0.05), differences werefurther evaluated using Fisher’s Least Significant Difference test todetermine differences between diet treatments or differences withineach diet treatment over time. A paired t test was used for fasting vs.postprandial comparisons. Values are means � SEM.

RESULTS

The experimental diets provided similar intakes of energy,fat and fiber (Table 1) and produced weight loss in both groups(Table 2). The protein intake for the high protein diet aver-aged 125 g/d [�1.5 g/(kg � d)] with carbohydrate intake of 171g/d. The CHO Group received an average of 68 g of proteinper day [0.8 g/(kg � d)] and 239 g of carbohydrate. The relativeproportions of energy in the Protein Group were 30% protein,41% carbohydrate and 29% fat with a ratio of carbohydrate toprotein (CHO/protein) of 1.4, whereas the proportion in theCHO Group was 16% protein, 58% carbohydrate and 26% fatwith a ratio of 3.5. The Protein Group consumed 239 � 8mg/d of cholesterol, whereas the CHO Group consumed 115� 5 mg/d. Saturated fatty acid intake was higher in the ProteinGroup, but both groups reduced their intakes of saturated fattyacids compared with the initial baseline period.

Body weight changes did not differ between the groups(Table 2). After consuming the diets for 10 wk, the ProteinGroup had a total weight loss of 7.53 � 1.44 kg and the CHOGroup lost 6.96 � 1.36 kg.

LAYMAN ET AL.412

at AC

ES

Library-E on A


Dow

nloaded from


Changes in body composition indicated that the weight losswas predominately body fat (Table 2). After 10 wk, women inthe Protein Group lost 5.60 � 0.52 kg of body fat or 14.4% ofinitial body fat, whereas women in the CHO Group lost 4.74� 0.65 kg of body fat or 12.2% of initial body fat. Loss of leanbody mass tended to be greater for the women in the CHOGroup (1.21 � 0.58 kg) compared with those in the ProteinGroup (0.88 � 0.33 kg) (P � 0.07). When changes in bodycomposition were expressed as a ratio of fat/lean loss (Fig. 1),weight loss was targeted progressively to body fat in bothgroups. The ratio of fat/lean loss demonstrated that the higherprotein diet partitioned a significantly greater percentage ofthe weight loss to body fat while sparing lean tissue. Specifi-cally, the Protein Group achieved a fat/lean loss of 6.36 � 0.85at 10 wk compared with the CHO Group with a fat/lean lossof 3.92 � 0.79 (P � 0.05).

Thyroid hormones in the two groups responded differently.Throughout the study, plasma T4 concentrations (Fig. 2A)increased in both groups from initial baseline values, whereasthe energy restriction decreased plasma T3 levels in bothgroups. T3 levels differed between the two groups at wk 2 and4 with the Protein Group having a greater concentrationduring weight loss. Although T3 levels declined, the precursorT4 levels increased with the increase at wk 10 greater in theProtein Group.

Both weight loss diets produced changes in serum lipids(Table 3). The initial values indicated that women assigned tothe Protein Group had baseline values for total cholesterol andLDL-C that were significantly greater than those of the CHOGroup. HDL-C and TAG did not differ between groups.Changes in blood cholesterol associated with the dietary treat-ments reflected the energy restriction and weight loss withsignificant decreases in both total cholesterol and LDL-C.After the first 4 wk of the treatments with food intake mon-itored in the research facility, total cholesterol was decreasedin the Protein Group by 10.0% (0.58 mmol/L) and LDL-C by10.5% (0.40 mmol/L). Similarly, total cholesterol was de-creased by 11.2% (0.55 mmol/L) in the CHO Group andLDL-C by 14.3% (0.45 mmol/L). HDL-C values decreased inboth groups at wk 2 and 4, but were not different from baselineat wk 10. Fasting TAG were reduced significantly in theProtein Group, with the values ranging from 16 to 23% belowinitial baseline values.

Blood and urinary ketone levels were not altered by thedietary treatments. The initial fasting concentration of plasmaBHB had a mean of 98.9 � 11.6 �mol/L for both groups andconcentrations of 112.8 � 15.2 and 166.2 � 10.1 �mol/L after10 wk of consuming the Protein and CHO diets, respectively.Acetoacetate was not detectable in the urine at any timepoint.

Blood urea concentrations were affected by diet treatment,by fed vs. fasting state and by the duration of the study (Table4). The baseline values for the two groups differed, but themagnitude of this difference declined over the course of thestudy. At wk 4 and 10, fasting blood urea did not differbetween groups. Postprandial blood urea concentrations re-flected the nitrogen intakes. For the protein Group, the post-prandial responses at wk 2, 4 and 10 had mean increases of43% (P � 0.05) and the CHO Group had a mean increase of27% (P � 0.05). Urinary urea also reflected differences innitrogen intake between the two diet groups with a mean forwk 2, 4 and 10 for the Protein Group of 511 � 39 mmol/d andthe CHO Group of 279 � 37 mmol/d (P � 0.05).

Plasma glucose and insulin were measured under both fast-ing and postprandial conditions. After 10 wk, fasting insulinlevels did not differ between groups (Table 5). However,fasting glucose was lower in the CHO Group. Insulin waselevated 2 h after ingestion of the test meal, whereas plasmaglucose was lower than the corresponding fasting concentra-tion in both groups. In the CHO Group, the 2-h postprandialinsulin concentration remained �2 times the fasting value andwas associated with a lower level of blood glucose.

Both Groups expressed general satisfaction with the dietplans, but the Protein Group consistently reported a higherlevel of satisfaction (Table 6). Subjects in the Protein Groupexpressed a feeling of more energy, and they reported a higherlevel of satiety. There were no differences in hunger betweenthe two groups.

DISCUSSION

The ideal ratios of dietary protein, carbohydrate and fat foradult health and weight management remain unknown.Claims about the merits or risks of diets with high protein areextensive (21), but there are few experimental studies. Thisstudy evaluated the effect of substitution of protein for carbo-hydrate as components of weight loss diets.

Changes in body weight require changes in energy balance(energy intake � energy expenditure). In this study, physicalactivity was constant; therefore, changes in body weightshould be in proportion to food intake. Subjects in this studyconsumed diets with equal energy, fat and fiber but withdifferent CHO/protein ratios of either 3.5 or 1.4. After 10 wk,subjects in both groups lost significant body weight.

Although changes in body weight did not differ betweendiet groups, the higher protein diet was more effective inimproving body composition. Changes in the ratio of fatloss/lean loss (Fig. 1) indicated that the higher protein dietimproved utilization of body fat while maintaining lean bodymass. The mechanism for these differential effects on bodycomposition is unknown. However, we found that substitutingdietary protein for carbohydrate in an energy-restricted dietmaintained levels of thyroid hormones T3 and T4 (Fig. 2) andreduced the insulin response to a test meal (Table 5). Theseendocrine differences are consistent with higher rates of lipol-ysis (22). Further, an increased amount of dietary protein hasbeen shown to reduce nitrogen losses associated with very lowenergy diets (23), and we reported that increased use of pro-teins or specifically branched-chain amino acid serves to main-

TABLE 1

Dietary intakes for adult women at baseline and during weightloss while consuming reduced energy diets with a ratio

of carbohydrate (CHO)/protein of 3.5 (CHO Group)or 1.4 (Protein group)1

Baseline2Proteingroup CHO group

Energy, kJ/d 8196 � 267 6987 � 197 6941 � 167(kcal/d) (1959) (1670) (1659)

Protein, g/d 75 � 3 125 � 3 68 � 2Carbohydrate, g/d 246 � 10 171 � 7 239 � 5Total fat, g/d 75 � 5 54 � 2 48 � 2Cholesterol, mg/d 207 � 17 239 � 8 115 � 5SFA, g/d 25 � 2 18 � 1 14 � 1Dietary fiber, g/d 14 � 1 19 � 1 25 � 1

1 Values are means � SEM, n � 12.2 Initial values were not different between groups, and data for all 24

subjects were combined for the baseline period (wk 0). SFA � satu-rated fatty acids.

DIETARY PROTEIN EFFECTS ON WEIGHT LOSS 413

at AC

ES

Library-E on A


Dow

nloaded from


tain muscle protein synthesis during catabolic conditions (24–26). Hence, the changes in body composition associated withthe higher protein diet may be associated with either targetingof body fat or sparing of muscle protein, or both.

Similar findings for changes in body composition have beenreported in other studies in which dietary fat was constant(9,10). Parker et al. (10) examined weight loss in 66 subjectswith type 2 diabetes and BMI of 34 kg/m2. These investigatorsutilized diets with equal energy (6690 kJ/d) and equal fat (�27% of energy) with CHO/protein ratios of 3.4 or 1.5. After 8wk, subjects had similar weight losses (4.5 kg) but thoseconsuming the high protein diet lost more body fat (5.3 kg vs2.8 kg; P � 0.05).

In the study from Denmark (9), investigators selected 65individuals with a mean age of 39 y and BMI of 30 kg/m2.Individuals were assigned to either a high carbohydrate diet(CHO/protein � 4.9, fat � 29% of energy) or high proteindiet (CHO/protein � 1.9, fat � 29%) with all food providedby the researchers but with self-selection by the subjects andhome preparation. Subjects were allowed to consume theirfood ad libitum. After 6 mo, the subjects receiving the higherprotein diet consumed 17% less energy per day, lost more body

weight and lost more body fat than the high carbohydrategroup. In the Denmark study, subjects self-selected energyintake based on appetite, whereas subjects in the present studywere restricted to equal energy intakes. However, both studies

FIGURE 1 Time course changes for the ratio of loss of body fatcompared with loss of lean body mass (fat/lean) during weight loss foradult women consuming diets with a carbohydrate (CHO)/protein ratioof 3.5 (CHO Group) or 1.4 (Protein Group). Values are means � SEM, n� 12. Means without a common letter differ, P � 0.05.

FIGURE 2 Response of thyroid hormones thyroxine (T4) andtriiodothyronine (T3) during weight loss in adult women consumingdiets with a carbohydrate (CHO)/protein ratio of 3.5 (CHO Group) or 1.4(Protein Group). Values are means � SEM, n � 12. *Different from theCHO Group, P � 0.05.

TABLE 2

Body weight and composition of adult women consuming either moderate protein or high carbohydrate (CHO) weight loss diets1

Time, wk

Total loss0 2 4 10

kg

Body weightProtein group 84.83a � 3.64 82.13a,b � 2.70 80.97b � 3.85 77.30b � 3.50 7.53 � 1.44CHO group 85.68a � 2.77 83.47a,b � 2.21 82.46b � 3.22 78.72b � 2.46 6.96 � 1.36

Body fatProtein group 38.97a � 2.86 37.15a,b � 2.80 35.92b � 2.77 33.37b � 2.81 5.60 � 0.52CHO group 38.92a � 2.18 37.41a,b � 2.18 36.32b � 2.07 34.18b � 1.99 4.74 � 0.65

Lean body massProtein group 42.69 � 1.04 42.14 � 1.05 42.14 � 0.99 41.81 � 0.99 0.88 � 0.33CHO group 43.53 � 1.24 42.66 � 1.17 42.51 � 1.23 42.32 � 1.16 1.21 � 0.58

1 Values are means � SEM, n � 12. Means for a variable without a common letter differ, P � 0.05.

LAYMAN ET AL.414

at AC

ES

Library-E on A


Dow

nloaded from


found that diets with CHO/protein ratios �2.0 partitionedweight loss toward body fat. In a subsequent study (27), theseresearchers reported that individuals consuming a high proteindiet had higher 24-h energy expenditure than when consum-ing a high carbohydrate diet, and that energy expenditure washigher when the protein was derived from animal proteinscompared with plant proteins.

A major concern about using higher protein diets, partic-ularly those rich in animal products, has been the associationof cholesterol and saturated fatty acids with cardiovasculardisease. There is a clear relationship of abnormal blood li-poprotein patterns with onset of heart disease. Further, there isincreasing evidence concerning the causal relationship of totaldietary lipids and saturated fatty acids with the onset of ath-erosclerosis (28). However, the independent roles of proteinand carbohydrates have received less attention. It is oftenassumed that animal proteins are atherogenic, whereas dietshigh in complex carbohydrates reduce the risk of heart disease.However, both of these associations are confounded by othercomponents of the diet including total fat, total energy anddietary fiber. A recent report from the Nurses’ Health Study by

Hu and colleagues (7) provides epidemiologic evidence thatreplacing carbohydrates in the diet with protein enhanced theoverall quality of the diet and lowered the risk of ischemicheart disease in adult women.

The current study directly evaluated the relationship be-tween protein and carbohydrate by substituting foods in theprotein groups (meats, dairy, eggs and nuts) for foods in thehigh carbohydrate group (breads, rice, pasta, and cereals)while maintaining total energy, total fat and fiber constant.Using this protocol, both diet groups exhibited �10% reduc-tions in total cholesterol and LDL-C even though the ProteinGroup had more than twice the cholesterol intake. These datasuggest that other factors such as energy balance (weight loss),total energy intake, total fat or saturated fat were more impor-tant than dietary cholesterol. These findings are consistentwith previous clinical reports (8,29) and with the currentunderstanding of regulation of endogenous cholesterol synthe-sis (30).

Although changes in total cholesterol and LDL-C weresimilar in the two groups, fasting serum TAG concentrationsdiffered. After 2 wk, the Protein Group exhibited a 20%

TABLE 3

Serum lipid concentrations in adult women at baseline (wk 0) and during weight loss while consuming reduced energy dietswith a ratio of carbohydrate (CHO)/protein of 3.5 (CHO group) or 1.4 (protein group)1

Time, wk

0 2 4 10

mmol/L

Total cholesterolProtein group 5.80a � 0.16 5.24b � 0.26 5.22b � 0.19 5.38b � 0.21CHO group 4.92c � 0.29 4.46d � 0.27 4.37d � 0.27 4.40d � 0.29

LDL-CProtein group 3.81a � 0.17 3.46b � 0.23 3.41b � 0.19 3.45b � 0.19CHO group 3.15c � 0.24 2.77d � 0.21 2.70d � 0.22 2.64d � 0.22

HDL-CProtein group 1.43a � 0.07 1.35b � 0.08 1.33b � 0.09 1.46a � 0.10CHO group 1.33a � 0.09 1.23b � 0.07 1.20b � 0.07 1.28a � 0.07

TriacylglycerolProtein group 1.21a � 0.15 0.92b � 0.13 1.02a,b � 0.17 0.95b � 0.13CHO group 1.05a � 0.09 1.01a,b � 0.11 1.01a,b � 0.09 1.05a � 0.11

1 Values are means � SEM, n � 12. Means for a variable without a common letter differ, P � 0.05.

TABLE 4

Fasting and post-prandial plasma and urinary urea in adult women at baseline (wk 0) and during weight loss while consumingreduced energy diets with a ratio of carbohydrate (CHO)/protein of 3.5 (CHO group) or 1.4 (Protein group)1

Time, wk

0 2 4 10

Fasting plasma urea, mmol/LProtein group 4.32a � 0.21 3.75b � 0.21 3.39c � 0.14 3.14c � 0.18CHO group 3.46b,c � 0.28 2.39d � 0.21 3.32b,c � 0.25 3.07c � 0.28

Post-prandial plasma urea, mmol/LProtein group 4.49b � 0.21 5.53a* � 0.18 4.57b* � 0.14 4.60b* � 0.18CHO group 3.50c � 0.28 3.25c* � 0.21 4.60b* � 0.25 3.32c � 0.28

Urinary urea, mmol/dProtein group 339b � 28 468a � 32 482a � 25 585a � 64CHO group 307b � 25 218c � 28 253b,c � 53 368b � 32

1 Values are means � SEM, n � 12. Means for a variable without a common letter differ, P � 0.05. * Indicates the value for post-prandial is differentfrom fasting, P � 0.05.


at AC

ES

Library-E on A


Dow

nloaded from


reduction in TAG, whereas the CHO Group did not change.These changes in TAG are in agreement with other studies(7–9,13,31) that found that high carbohydrate diets are linkedto increased hepatic production of TAG (11,31) and increasedlevels of circulating TAG (7,8,13). Similarly, dietary substi-tution of protein for carbohydrates appears to have positiveeffects on the ratio of TAG/HDL-C. The Bezafibrate CoronaryAtherosclerosis Intervention study (32) reported that de-creases in the ratio of TAG/HDL-C correlate directly withreductions in coronary heart disease without changes in LDL-C levels. In the present study, the ratio of TAG/HDL-Cdecreased in the Protein Group (0.846 � 0.021 to 0.651� 0.016 mmol/L; P � 0.031), whereas the ratio did not changein the CHO Group (0.789 � 0.004 to 0.820 � 0.004 mmol/L).

The effect of dietary protein intake on renal function hasbeen debated for 50 y (33). In cases with compromised renalfunction, reduced levels of dietary protein can retard the

progression to renal failure. Conditions such as diabetes, hy-pertension, infection or renal surgery often lead to changes inrenal physiology, including increases in glomerular capillarypressure and blood flow rates (34). Restriction of dietary pro-tein appears to reduce the “renal workload” and minimizeglomerular perfusion. By extrapolation, it is often suggestedthat adults avoid high protein intakes to minimize glomerularfiltration rates. However, there is no known association ofprotein intake with progressive renal insufficiency during aging(34,35). Further, a recent study by Poortmans and Dellalieux(36) reported no negative effects on renal function of long-term daily protein intakes ranging from 1.2 to 2.0 g protein/kgbody weight. Evaluation of urea data in the current studysuggests that subjects adapted rapidly to the different nitrogenintakes. There were no differences between the groups infasting blood urea, with the values becoming more consistentthroughout the course of the study. Both groups exhibited apostprandial increase in blood urea that was in proportion tothe protein intake. Similarly, urinary excretion of urea was inproportion to the dietary intake. These data suggest that withdietary intakes of protein ranging from 0.8 to 1.5 g/kg, renalclearance of nitrogen is rapid and efficient.

The roles of specific macronutrients in satiety and controlof energy intake remain poorly understood. There is increasingevidence that diets high in some carbohydrates, particularlyrefined grains, produce high postprandial levels of blood glu-cose. This meal response of blood sugar is termed the glycemicindex (37). More important, the rapid rise in blood glucoseevokes an equally intensive response of insulin that can lead toa period of hypoglycemia and increased hunger (37–39). Thishypoglycemic response typically occurs �2 h after a meal,depending on meal size and rates of gastric emptying (38).Lipids appear to increase satiety both as dietary lipids that slowgastric emptying (40) and as fatty acids free in the blood thatreduce hunger (41). Proteins made up of 20 individual aminoacids are more complex, but clearly have the potential to effectneurotransmitters and satiety (42). Still other factors affectingsatiety include dietary fiber and fluid intakes (43). This studywas designed to maintain levels of fiber, fat, fluids and totalenergy constant and compare effects of protein vs. carbohy-drates. Overall, subjects in the Protein Group reported highersatiety and greater energy when consuming the diet with alower CHO/protein ratio. Subjects in the CHO Group exhib-ited increased postprandial insulin and reduced glucose con-centrations. These findings are consistent with the hypothesisabout the role of carbohydrates in postprandial hypoglycemiaand hunger (39) and the effect of the glycemic index on bloodglucose and insulin (38).

TABLE 5

Fasting and post-prandial plasma glucose and insulin concentrations in adult women after 10 wk consuming reduced energyweight loss diets with a ratio of carbohydrate (CHO)/protein of 3.5 (CHO group) or 1.4 (protein group)1

Glucose Insulin

Fasting Post-prandial2 Fasting Post-prandial2

mmol/L pmol/L

Protein group 4.89a � 0.11 4.34b � 0.15 176c � 18 251b � 21CHO group 4.33b � 0.10 3.77c � 0.14 178c � 18 384a � 27

1 Values represent means � SEM, n � 12. Means for a variable without a common letter differ, P � 0.05.2 Post-prandial values were determined 2 h after consumption of a 1670 kJ breakfast. The macronutrient content of the breakfast reflects the

respective dietary treatments. Data adapted from Layman et al. (16).

TABLE 6

Satisfaction surveys for adult women consuming reducedenergy weight loss diets with a ratio of carbohydrate

(CHO)/protein of 3.5 (CHO group: CHO) or 1.4(protein group: protein)1

My satisfaction level with this dietary plan is:(Very unsatisfied 1–7 Very satisfied)

Protein 6.08 � 0.13*CHO 5.48 � 0.13

My energy level is:(Very low 1–7 Very high)

Protein 5.72 � 0.08*CHO 4.90 � 0.11

Between meals on this dietary plan my hunger is:(Very low 1–7 Very high)

Morning:Protein 2.84 � 0.16CHO 2.96 � 0.15

Afternoon:Protein 3.42 � 0.14CHO 3.30 � 0.14

Evening:Protein 3.07 � 0.14CHO 2.87 � 0.15

The satiety value of this dietary plan is:(Very low 1–7 Very high)

Protein 6.05 � 0.12*CHO 5.32 � 0.14

1 Values are means � SEM for averages of weekly surveys, n � 12.* Different from CHO group, P � 0.05.

LAYMAN ET AL.416

at AC

ES

Library-E on A


Dow

nloaded from


In summary, both diets were effective weight loss plans andimproved the blood lipid profile. However, the protein dietproduced greater improvements in body composition with anincreased ratio of fat/muscle loss. Further, the protein diet,with a reduced CHO/protein ratio, produced positive changesin blood lipids with reduction of TAG levels and the ratio ofTAG/HDL-C. Subjects consuming the higher protein dietreported greater satiety. Although it is unlikely that any onediet will be ideal for all individuals, these results indicate thatchanges in the ratio of protein to carbohydrate toward a higherprotein diet can be effective in the control of body weight withparallel improvements in blood lipids.

LITERATURE CITED1. Kuczmarski, R. J., Flegal, K. M., Campbell, S. M. & Johnson, C. L. (1994)

Increasing prevalence of overweight among US adults. J. Am. Med. Assoc. 272:205–211.

2. Must, A., Spadano, J., Coakley, E. H., Field, A. E., Colditz, G. & Dietz,W. H. (1999) The disease burden associated with overweight and obesity.J. Am. Med. Assoc. 282: 1523–1529.

3. Surgeon General’s Report on Nutrition and Health (1988) USDHHSPublication no 88–50210. Public Health Service, Washington, DC.

4. U. S. Department of Agriculture/Department of Health and Human Ser-vices (1995) Dietary Guidelines for Americans, 4th ed., Home and Gardenbulletin 232. DHHS, Washington, DC.

5. American Heart Association Scientific Statement (2000) AHA DietaryGuidelines Revision 2000. Circulation 102: 2284–2299.

6. Seidell, J. C. (1998) Dietary fat and obesity: an epidemiologic perspec-tive. Am. J. Clin. Nutr. 67 (suppl.): 546S–550S.

7. Hu, F. B., Stampfer, M. J., Manson, J. E., Rimm, E., Colditz, G. A.,Speizer, F. E., Hennekens, C. H. & Willett, W. C. (1999) Dietary protein and riskof ischemic heart disease in women. Am. J. Clin. Nutr. 70: 221–227.

8. Wolfe, B. M. & Giovannetti, P. M. (1991) Short-term effects of substi-tuting protein for carbohydrate in diets of moderately hypercholesterolemic hu-man subjects. Metabolism 40: 338–343.

9. Skov, A. R., Toubro, S., Ronn, B., Holm, L. & Astrup, A. (1999) Ran-domized trial on protein vs carbohydrate in ad libitum fat reduced diet for thetreatment of obesity. Int. J. Obes. 23: 528–536.

10. Parker, B., Noakes, M., Luscombe, N. & Clifton, P. (2002) Effect of ahigh-protein, monounsaturated fat weight loss diet on glycemic control and lipidlevels in type 2 diabetes. Diabetes Care 25: 425–430.

11. McGarry, J. D. (1998) Glucose-fatty acid interactions in health anddisease. Am. J. Clin. Nutr. 67 (suppl.): 500S–504S.

12. Wolfe, R. R. (1998) Metabolic interactions between glucose and fattyacids in humans. Am. J. Clin. Nutr. 67 (suppl.): 519S–526S.

13. Sidossis, L. S., Mittendorfer, B., Walser, E., Chinkes, D. & Wolfe, R. R.(1998) Hyperglycemia-induced inhibition of splanchnic fatty acid oxidation in-creases hepatic triacylglycerol secretion. Am. J. Physiol. 275: E798–E805.

14. Ludwig, D. S., Majzoub, J. A., Al-Zahrani, A., Dallal, G. E., Blanco, I. &Roberts, S. B. (1999) High glycemic index foods, overeating, and obesity.Pediatrics 103: E261–E266.

15. El-Khoury, A. E., Fukagawa, N. K., Sanchez, M., Tsay, R. H., Gleason,R. E., Chapman, T. E. & Young, V. R. (1994) The 24-h pattern and rate ofleucine oxidation, with particular reference to tracer estimates of leucine require-ments in healthy adults. Am. J. Clin. Nutr. 59: 1012–1020.

16. Layman, D. K., Shiue, H., Sather, C., Erickson, D. J. & Baum, J. (2003)Increased dietary protein modifies glucose and insulin homeostasis in adultwomen during weight loss. J. Nutr. 133: 405–410.

17. Jungas, R. L., Halperin, M. L. & Brosnan, J. T. (1992) Quantitativeanalysis of amino acid oxidation and related gluconeogenesis in humans. Physiol.Rev. 72: 419–448.

18. Balasubramanyam, A., McKay, S., Nadkarni, P., Rajan, A. S., Farza, A.,Pavlik, V., Herd, J. A., Jahorr, F. & Reeds, P. J. (1999) Ethnicity affects thepostprandial regulation of glycogenolysis. Am. J. Physiol. 40: E905–E914.

19. Myers, G., Cooper, G., Winn, C. & Smith, S. (1989) The Centers forDisease Control-National Heart, Lung and Blood Institute Lipid StandardizationProgram. An approach to accurate and precise lipid measurements. Clin. Lab.Med. 9: 105–135.

20. Friedewald, W., Levy, R. & Frederickson, D. (1972) Estimation of theconcentration of low-density lipoprotein cholesterol in plasma, without use of thepreparative ultracentrifuge. Clin. Chem. 18: 499–502.

21. American Heart Association Science Advisory (2001) Dietary proteinand weight reduction. Circulation 104: 1869–1873.

22. Tepperman, J. (1980) The thyroid. In: Metabolic and Endrocrine Phys-iology, 4th ed. Year Book Medical Publishers, Chicago, IL.

23. Bistrian, B. R., Winterer, J., Blackburn, G. L., Young, V. & Sherman, M.(1977) Effect of a protein-sparing diet and brief fast on nitrogen metabolism inmildly obese subjects. J. Lab. Clin. Med. 89: 1030–1035.

24. Hong, S. C. & Layman, D. K. (1984) Effects of leucine on in vitroprotein synthesis and degradation in rat skeletal muscle. J. Nutr. 114: 1204–1212.

25. Gautsch, T. A., Anthony, J. C., Kimball, S. R., Paul, G. L., Layman, D. K.& Jefferson, L. S. (1998) Availability of eIF4E regulates skeletal muscle proteinsynthesis during recovery from exercise. Am. J. Physiol. 274: C406–C414.

26. Anthony, J. C., Gautsch-Anthony, T. & Layman, D. K. (1999) Leucinesupplementation enhances skeletal muscle recovery in rats following exercise. J.Nutr. 129: 1102–1106.

27. Mikkelsen, P. B., Toubro, S. & Astrup. A. (2000) Effect of fat-reduceddiets on 24-h energy expenditure: comparisons between animal protein, vegeta-ble protein, and carbohydrate. Am. J. Clin. Nutr. 72: 1135–1141.

28. National Cholesterol Education Program (2001) Third Report of theExpert Panel on Detection, Evaluation, and Treatment of High Blood Cholesterolin Adults (Adult Treatment Panel III) NIH Publication no. 01–3670, Washington,DC.

29. Wolf, R. N. & Grundy, S. M. (1983) Influence of exchanging carbohy-drate for saturated fatty acids on plasma lipids and lipoproteins in men. J. Nutr.113: 1521–1528.

30. Spady, D. K., Woollett, L. A. & Dietschy, J. M. (1993) Regulation ofplasma LDL-cholesterol levels by dietary cholesterol and fatty acids. Annu. Rev.Nutr. 13: 355–381.

31. Katan, M. B. (1998) Effect of low-fat diets on plasma high-densitylipoprotein concentrations. Am. J. Clin. Nutr. 67: 573S–576S.

32. Ericsson, C. G., Mamsten, A., Nilsson, J., Grip, L., Svane, B. & De Faire,U. (1996) Angiographic assessment of effects of bezafibrate on progression ofcoronary artery disease in young male postinfarction patients. Lancet 347: 849–853.

33. Addis, T. (1948) Glomerular Nephritis: Diagnosis and Treatment. Mac-millan, New York, NY.

34. Brenner, B. M., Meyer, T. W. & Hostetter, T. H. (1982) Dietary proteinintake and the progressive nature of kidney disease. N. Engl. J. Med. 307:652–659.

35. Brandle, E., Sieberth, H. G. & Hautmann, R. E. (1996) Effect of chronicdietary protein intake on the renal function in healthy subjects. Eur. J. Clin. Nutr.50: 734–740.

36. Poortmans, J. R. & Dellalieux, O. (2000) Do regular high protein dietshave potential health risks on kidney function in athletes? Int. J. Sport Nutr. Exerc.Metab. 10: 28–38.

37. Ludwig, D. S. (2000) Dietary glycemic index and obesity. J. Nutr. 130:280S–283S.

38. Roberts, S. B. (2000) High-glycemic index foods, hunger, and obesity:is there a connection? Nutr. Rev. 58: 163–169.

39. Cryer, P. E. (1999) Symptoms of hypoglycemia thresholds for theiroccurrence, and hypoglycemia unawareness. Endocrinol. Metab. Clin. N. Am. 28:495–500.

40. Sidery, M. B., Macdonald, I. A. & Blackshaw, P. E. (1994) Superiormesenteric artery blood flow and gastric emptying in humans and the differentialeffects of high fat and high carbohydrate meals. Gut 35: 186–190.

41. Ludwig, D. S., Majzoub, J. A., Al-Zahrani, A., Dallal, G. E., Blanco, I. &Roberts, S. B. (1999) High glycemic index foods, overeating, and obesity.Pediatrics 103: E261–E266.

42. Fernstrom, J. D. (1994) Dietary amino acids and brain function. J. Am.Diet. Assoc. 94: 71–77.

43. Burton-Freeman, B. (2000) Dietary fiber and energy regulation. J. Nutr.130: 272S–275S.


at AC

ES

Library-E on A


Dow

nloaded from



Dietary Protein and Exercise Have Additive Effects on Body Compositionduring Weight Loss in Adult Women1,2

Donald K. Layman,*†3 Ellen Evans,†** Jamie I. Baum,† Jennifer Seyler,†

Donna J. Erickson,* and Richard A. Boileau**

*Department of Food Science and Human Nutrition; †Division of Nutritional Sciences;and **Department of Kinesiology, University of Illinois, Urbana, IL 61801

ABSTRACT This study examined the interaction of 2 diets (high protein, reduced carbohydrates vs. low protein,high carbohydrates) with exercise on body composition and blood lipids in women (n � 48, �46 y old, BMI � 33kg/m2) during weight loss. The study was a 4-mo weight loss trial using a 2 � 2 block design (Diet � Exercise).Diets were equal in total energy (7.1 MJ/d) and lipids (�30% energy intake) but differed in protein content and theratio of carbohydrate:protein at 1.6 g/(kg � d) and �1.5 (PRO group) vs. 0.8 g/(kg � d) and �3.5 (CHO group),respectively. Exercise comparisons were lifestyle activity (control) vs. a supervised exercise program (EX: 5 d/wkwalking and 2 d/wk resistance training). Subjects in the PRO and PRO � EX groups lost more total weight and fatmass and tended to lose less lean mass (P � 0.10) than the CHO and CHO � EX groups. Exercise increased lossof body fat and preserved lean mass. The combined effects of diet and exercise were additive for improving bodycomposition. Serum lipid profiles improved in all groups, but changes varied among diet treatments. Subjects inthe CHO groups had larger reductions in total cholesterol and LDL cholesterol, whereas subjects in the PRO groupshad greater reductions in triacylglycerol and maintained higher concentrations of HDL cholesterol. This studydemonstrated that a diet with higher protein and reduced carbohydrates combined with exercise additivelyimproved body composition during weight loss, whereas the effects on blood lipids differed between diettreatments. J. Nutr. 135: 1903–1910, 2005.

KEY WORDS: ● obesity ● low-carbohydrate diets ● blood lipids ● insulin ● adiponectin

Obesity is an important public health problem associatedwith multiple chronic health conditions including heart dis-ease, hypertension, hyperlipidemia, diabetes, hyperinsulin-emia, and cancer. Recommendations for treatment of adultswho are overweight or obese focus on energy balance withlifestyle modifications designed to reduce daily energy intakeand increase physical activity (1). Although these recommen-dations represent public health policy, there are few studiesthat provide direct evidence about the combined merits andinteractions of specific diet and exercise choices. Furthermore,the ideal balance of macronutrients (carbohydrates, lipids, andprotein) necessary to optimize the combined effects of exerciseand energy restriction is unknown.

There is general consensus that the most critical factor indetermining weight loss is total energy intake (1–3). The idealbalance of macronutrients for adult weight loss remains widelydisputed (1–5). However, evidence is accumulating that en-ergy-restricted diets with reduced levels of carbohydrates andhigher levels of protein are beneficial during weight loss (6–16). These studies report that diets with carbohydrate intake� 150 g/d and protein intake � 1.4 g/(kg � d) result in in-creased weight loss (6–8,12), increased loss of body fat(7,8,10), attenuated loss of fat-free mass (FFM)4 (6,8,10,13,14),improved glycemic control (6,10–12), improved blood lipid pro-files (10,12,16,17), and enhanced satiety (15).

The importance of exercise in weight loss and prevention ofweight regain is well accepted (1–3). In general, exerciseduring weight loss appears to target loss of fat mass whilepreserving lean mass. The effects of exercise on total bodyweight are variable, but generally proportional to the accumu-lated total energy expenditure (2,3). Variables including du-ration of activity per day, frequency of activity per week, andintensity ultimately determine energy expenditure and the

1 Presented in part at the 2004 North American Association for the Study ofObesity meeting, November 15, 2004, Las Vegas, NV [Evans, E. M., Layman,D. K., Baum, J. I., Seyler, J. E., Erickson, D. J. & Heinrichs, K. L. (2004) Additiveeffects of dietary protein and exercise on body composition and endocrinechanges during weight loss in adult women. Obes. Res. 12 (suppl.): A48 (abs.)];and at Experimental Biology 05, April 2–6, 2005, San Diego, CA [Layman, D.,Evans, E., Baum, J., Seyler, J. & Erickson, D. (2005) Interaction of dietaryprotein and exercise during weight loss in adult women. FASEB J. 19: A419(abs.)].

2 Support for this research was provided by the Illinois Council on Food andAgricultural Research, National Cattlemen’s Beef Association, The Beef Boardand Kraft Foods.


4 Abbreviations used: ANCOVA, analysis of covariance; CHD, coronary heartdisease; CHO, higher carbohydrate diet; DXA, dual energy X-ray absorptiometry;FFM, fat-free mass; HDL-C, HDL cholesterol; LDL-C, LDL cholesterol; PRO,higher protein diet; RDA, Recommended Dietary Allowance; TAG, triacylglycerol;TC, total cholesterol.

0022-3166/05 $8.00 © 2005 American Society for Nutritional Sciences.Manuscript received 16 February 2005. Initial review completed 7 March 2005. Revision accepted 7 May 2005.

1903

potential for weight loss. Studies suggest that low-intensityexercise (i.e., walking) at �150 min/wk has minimal potentialto produce weight loss (18,19). Similarly, resistance exerciseappears to have beneficial effects on body composition, main-taining FFM while increasing loss of body fat, but has minimaleffect on body weight (20,21). In addition to changes in bodycomposition, exercise also has positive effects on blood lipids.Regular endurance exercise is associated with decreases inblood concentrations of triacylglycerol (TAG) and increasesin HDL cholesterol (HDL-C) (22). In total, exercise-inducedchanges in body composition and blood lipids are similar tochanges observed with reduced-carbohydrate, high-protein,weight loss diets (7,10,12,17). However, the combined effectsof a low-carbohydrate, higher-protein diet with exercise duringweight loss have not been examined.

This study was designed to evaluate interactions of dietcomposition and exercise on body composition changes andblood lipids during weight loss. Based on our review of thescientific literature, we hypothesized that a diet with increasedprotein and reduced carbohydrates combined with an exerciseprogram of aerobic activity and resistance training would pro-duce an additive effect on loss of body fat and maintenance ofFFM. Further, we measured the interactions of diet treatmentsand exercise on changes in blood lipids and blood concentra-tions of leptin, adiponectin, ghrelin, and insulin as indicatorsof metabolic change during weight loss.


Design. This study was a randomized 4-mo weight loss trial usinga 2 � 2 block design (Diet � Exercise). Diet treatments consisted ofeither a low carbohydrate:protein ratio (PRO group) or a high car-bohydrate:protein ratio (CHO group). Exercise treatments consistedof a control group participating in light walking activity or anexercise group (designated EX) that required walking a minimum of5 d/wk plus 2 sessions/wk of resistance exercise.

Subjects. Women (n � 48) aged 40–56 y were recruited toparticipate in a weight loss study. Exclusion criteria were BMI � 26kg/m2, body weight � 140 kg (due to DXA scanning bed constraints),smoking, any existing medical conditions requiring medications thatwould affect primary or secondary outcomes of the study, use of oralsteroids or use of antidepression medication. Due to resource andpersonnel constraints, the first 24 women recruited were randomlyassigned to either PRO � EX or CHO � EX groups (n � 12 each).The next 24 women recruited were randomly assigned to either thePRO or CHO groups (n � 12 each). Note that the randomization wasblocked on age and BMI for each wave of recruitment. Baselinecharacteristics did not differ among subjects in each of the treatmentgroups (Table 1). The protocol and aim of the study were fullyexplained to the subjects, who gave their written informed consent.This study was approved by the Institutional Review Board at theUniversity of Illinois at Urbana-Champaign.

Diet treatments. The CHO diet provided dietary protein at 0.8g/(kg � d) (�15% of energy intake) with a carbohydrates:proteinratio � 3.5 and dietary lipids at �30% energy intake. The PRO diet

provided dietary protein at 1.6 g/(kg � d) (�30% of energy intake)with a carbohydrate:protein ratio � 1.5 and dietary lipids at �30%energy intake. These diets were designed to fall within the Accept-able Macronutrient Distribution Range established by the Institute ofMedicine (23) with minimum Recommended Dietary Allowance(RDA) intakes for carbohydrates � 130 g/d and protein � 0.8 g/kgand with upper limits for carbohydrates � 65% and protein � 35%of total energy intake. The 2 diets were formulated to be equal inenergy (7100 kJ/d; 1700 kcal/d), total fat intake (�57 g/d) and fiber(�17 g/d). With these general criteria, we developed a 2-wk menuplan for each group with meals for each day meeting establishednutritional requirements (24) and lipid guidelines (25). The dietarydifferences between groups were designed to reflect a direct substitu-tion of foods in the protein groups (meats, dairy, eggs, and nuts) forfoods in the refined grain/starch groups (breads, rice, cereals, pasta,and potatoes). The education guidelines for the CHO group followedthe USDA Food Guide Pyramid (26) and emphasized restrictingdietary fat and cholesterol with use of breads, rice, cereals, and pasta.For the PRO group, the education guidelines emphasized use ofhigh-quality proteins including meats, dairy, and eggs. Both dietsincluded 5 servings/d of vegetables and 2–3 servings of fruits.

Exercise treatments. Exercise treatments consisted of a controlgroup that emphasized lifestyle recommendations for physical activitybased on the NIH Guidelines for Weight Management (1). Theseguidelines recommend a minimum of 30 min of walking 5 d/wk.Participation in physical activity for the control groups was voluntaryand the education program for this group focused on diet compliancewith limited attention to exercise. Exercise was monitored using dailyactivity logs; on 3 d/mo subjects wore armband accelerometers (Body-Media). Based on these measurements, we estimated that subjects inthe control groups averaged �100 min/wk of added exercise. The EXtreatment group required a minimum of walking 5 d/wk for 30 min/dplus a resistance training program 2 d/wk consisting of 30 min ofstretching and resistance exercise using 7 Nautilus® weight machines.EX subjects were required to complete a minimum of one 12-repeti-tion set on each machine with the resistance weight selected to elicitfatigue by the final repetition using full range of motion with bothconcentric and eccentric pressure and within the comfort level of theparticipant. Exercise compliance in the EX group was supervised andaveraged �200 min/wk.

Protocol. All subjects participated in a baseline evaluation pe-riod that included a 24-h food recall, instructions for weighing andrecording of foods, two 3-d weighed food records and measurementsof body weight and height, and blood lipids. This evaluation periodfrom first contact with the subjects was 10–20 d and served as aninitial control period for each subject. During this baseline period,subjects were instructed to maintain stable body weight and toconsume a diet similar to that of the past 6 mo. After the baselineperiod, subjects reported to the nutrition research laboratory at0630 h after a 12–14 h overnight fast for weighing and blood sam-pling.

Body composition and blood measures. Body weight was mea-sured using an electronic scale (Tanita, Model BWB-627A). Whole-body composition and trunk fat (regionally defined with analysis linesdrawn through the axilla, proximal femur neck and at the chin lineto segregate the arms, legs, and head, respectively) was determined bydual energy X-ray absorptiometry (DXA; Hologic) and scans for agiven individual were analyzed by the same technician using standardmanufacturer guidelines. The CVs for DXA outcomes of interest were1.5%. Plasma leptin, adiponectin, and ghrelin (Linco Research; Cat-alogue # HL-81HK, HADP-61HK and GHRA-88HK, respectively)and insulin (MP Biomedicals; catalogue # 07260105) were deter-mined by commercial RIA kits with all samples for a given partici-pant analyzed in batch. The intra-assay CVs for leptin, adiponectin,ghrelin and insulin were 6.1 1.5, 3.0, and 5.5%, respectively. Serumtotal cholesterol (TC), HDL-C and TAG were determined by stan-dardized methods (27) by the Washington University School ofMedicine Core Laboratory for Clinical Studies. LDL cholesterol(LDL-C) was calculated using the Friedewald equation (28).

Diet education and measures. Subjects were provided electronicfood scales and instructed to weigh all food servings at all meals; theywere required to report a 3-d weighed food record for each week

TABLE 1

Baseline characteristics of subjects1

Group PRO PRO � EX CHO CHO � EX

Age, y 47.0 � 1.7 46.5 � 1.2 45.2 � 1.4 47.9 � 1.4Height, cm 161.5 � 2.5 165.8 � 1.3 162.2 � 1.4 163.2 � 1.6Weight, kg 91.1 � 5.1a,b 86.1 � 4.6a,b 93.7 � 3.5a 79.8 � 2.7b

BMI, kg/m2 34.8 � 1.8a 31.4 � 1.7a,b 35.4 � 1.1a 30.2 � 1.3b

1 Values are means � SEM, n � 12. Means in a row without acommon letter differ, P � 0.05.

LAYMAN ET AL.1904

throughout the study. Nutrient intakes were evaluated as mean dailyintakes from the 3-d weighed records using Nutritionist Pro software(First DataBank). After baseline data collection, subjects receivedinstruction from a research dietitian about their specific diet andwalking program including the menus, food substitutions, portionsizes, and procedures for maintaining weighed diet records. Duringthe 16-wk weight loss program, subjects were required to attend a 1-hmeeting each week at the weight management research facility.Meetings were specific for each treatment group and directed byresearch dietitians who provided diet and exercise information,answered questions, and reviewed diet records for treatment com-pliance. Each week, subjects were weighed in light clothing with-out shoes, and turned in the 3-d weighed food records and dailyactivity logs.

Statistics. All data analyses were conducted using SPSS version12.0 (SPSS). Differences among groups in baseline measurements ofrandomization block variables, age and BMI, were evaluated usingone-way ANOVA with differences further evaluated using Fisher’sLeast Significant Test. As a preliminary analysis, changes that oc-curred in response to treatments (pretest compared to post-test value)within a group were evaluated by a Student’s t test. The primaryanalysis, conducted to evaluate the relative effect of the diet andexercise treatments, utilized a two-way ANOVA (Diet � Exercise)and the change over time as the dependent variable. Note that withthis analysis, significant main effects in the absence of a significantinteraction effect indicate additive effects of the treatment. To eval-uate treatment effects on hormones controlling for changes in bodycomposition, an analysis of covariance (ANCOVA) was used whereindicated. The percentage change was calculated as follows: {[(post-test) � (pretest)/(pretest)] � 100}. Differences were considered sig-nificant at P �0.05. Values are presented as means � SEM.

RESULTS

Subjects. Baseline characteristics did not differ amongsubjects in each of the treatment groups (Table 1). Subjects

had a mean age of 46.6 y, weight of 87.7 kg, BMI of 32.9kg/m2, and relative body fat of 44.1%. Due to the waverecruitment and block randomization strategy, there was adifference in BMI (P � 0.047) among groups, largely associ-ated with the difference between the CHO and CHO � EXgroups.

Dietary compliance. Daily energy intakes determinedfrom 3-d weighed food records during the prestudy baselinewere 8.94 MJ/d (2138 kcal/d) for all subjects (Table 2). Dailymenus for the weight loss diets were designed to provideenergy at 7.1 MJ/d (1700 kcal); however, subjects were free-living and ultimately determined final choices of daily energyintakes. Reductions in energy intake determined from weekly3-d weighed food records did not differ among the groups asindicted by nonsignificant Diet and Exercise main effects andnonsignificant interactions.

A summary of the dietary intakes for the 16-wk periodillustrates how the subjects applied the 2 diets during theintervention (Table 2). Subjects in the PRO groups main-tained protein intakes of 107 g/d and �30% of energy intake.The ratio of carbohydrate:protein was �1.24. Total dietarylipid intake decreased to 49 g/d (32% of dietary energy) andSFA decreased to 19.1 g/d (12.4% of energy). Subjects in theCHO groups maintained carbohydrate intake at 198 g/d (61%of energy) and reduced total lipid intake to 37 g/d (25.5% ofenergy) and SFA to 11.0 g/d (7.5% of energy). Protein intakein the CHO group was 57 g/d (18% of energy intake) and theratio of carbohydrate:protein was �3.5.

Body weight. All groups lost significant body weight dur-ing the 16-wk treatment period. Body weight changes werelarger (P � 0.05) in the groups consuming the higher-protein,reduced-carbohydrate diet (Table 3). The PRO and PRO

TABLE 2

Dietary intakes for adult women at baseline and during weight loss (16 wk) while consuming reduced-energy dietswith a carbohydrate:protein ratio �3.5 (CHO group) or �1.5 (PRO group)1


P-value

D2 E3,4

Energy intake, kJ/dBaseline 8888 � 384 8362 � 322 8479 � 452 7977 � 339Treatment5 6062 � 117* 5540 � 134* 5377 � 179* 5644 � 180* 0.79 0.34

Protein, g/dBaseline 79.3 � 4.0 72.6 � 3.3 82.1 � 5.5 73.9 � 4.7Treatment5 110.0 � 3.0* 102.1 � 3.3* 57.6 � 1.3* 56.5 � 1.5* �0.001 0.44

Carbohydrate, g/dBaseline 262.3 � 22.3 266.7 � 9.2 246.0 � 16.2 266.2 � 17.8Treatment5 141.3 � 4.7* 126.8 � 4.2* 197.0 � 6.2* 201.7 � 5.8* �0.001 0.34

Lipids, g/dBaseline 84.7 � 4.1 73.9 � 6.7 79.3 � 6.0 65.0 � 3.7Treatment5 52.2 � 1.3* 46.3 � 1.4* 34.0 � 2.3* 40.8 � 2.1* 0.40 �0.05

Cholesterol, mg/dBaseline 259.0 � 44.8 168.4 � 21.4 236.2 � 39.1 224.4 � 28.3Treatment5 313.5 � 23.0 283.4 � 18.9* 103.7 � 8.5* 95.0 � 6.0* �0.001 0.35

SFA, g/dBaseline 29.5 � 2.3 23.6 � 1.9 28.0 � 3.0 21.8 � 2.1Treatment5 20.6 � 0.6* 17.5 � 0.7* 10.3 � 0.6* 11.6 � 0.6* �0.01 �0.05

Fiber, gBaseline 16.7 � 3.1 21.2 � 2.0 16.6 � 2.1 19.2 � 1.6Treatment5 18.6 � 1.2 16.0 � 1.2* 22.5 � 1.5* 23.3 � 1.3* �0.01 �0.05

1 Values are means � SEM, n � 12. * Different from baseline, P � 0.05.2 D � Test for significant main effect of diet (PRO; n � 24: CHO; n � 24).3 E � Test for significant main effect of exercise (EX; n � 24: control; n � 24).4 D � E was not significant for any variable.5 Values represent means for the sum of 3-d weighed records for wk 4, 8, 12, and 16.

DIETARY PROTEIN AND EXERCISE DURING WEIGHT LOSS 1905

� EX groups had a weight loss of 9.3 � 0.8 kg after 16 wk,whereas the CHO and CHO � EX groups reduced bodyweight by 7.3 � 0.5 kg (P � 0.05). The PRO � EX group hadthe largest relative weight loss at 11.2% of initial weight andthe CHO group had the smallest relative weight loss at 8.4%.The exercise treatment did not affect body weight change.The interaction between Diet and Exercise treatments was notsignificant for any measurements including body weight, bodycomposition, hormones, or blood lipids.

Body composition. Changes in body composition indi-cated that weight loss was predominately fat mass, and thatthere were increased fat losses associated with the higher-protein diet and the exercise program (Table 3). Subjects inthe PRO and the PRO � EX groups reduced fat mass by7.3 � 0.8 kg and subjects in the CHO and CHO � EX groupsreduced fat mass by 5.3 � 0.3 kg (P � 0.05). Subjects in thePRO � EX and CHO � EX groups reduced body fat by7.2 � 0.7 kg, whereas the subjects not participating in thesupervised exercise program reduced body fat by 5.5 � 0.5 kg(P � 0.05). The significant main effects of Diet and Exercisein the absence of a significant interaction indicate that thePRO and EX effects were independent and additive. Specifi-cally, the combined effects of the PRO diet and the EXprogram produced a 21.4% decrease in absolute body fat.Subjects in the CHO group without exercise had a 12.8%reduction in absolute body fat. Changes in absolute trunk fatwere similar to changes in total fat mass, with reductions intrunk fat greater in the PRO groups compared with the CHOgroups. When the change in trunk fat was expressed relative tothe change in whole-body fat, all treatment groups experi-enced a significant reduction with PRO � EX having thegreatest ratio reduction (�0.028 � 0.007; P � 0.001) andCHO experiencing the least change (�0.015 � 0.005; P� 0.009); however, there was no significant main effects fordiet (P � 0.32) or exercise (P � 0.12).

Changes in lean mass reflected a significant positive effectof the exercise program (P � 0.001) and a trend for a bene-

ficial effect of the PRO diet (P � 0.10) during weight loss(Table 3). Notably the PRO � EX group had no significantchange in lean mass (�0.9%; P � 0.39), whereas theCHO group had the largest decrease in lean mass (�5.4%; P� � 0.001).

Net changes in body composition were reflected in relativebody fatness (Fig. 1). The additive effects of the PRO and EXtreatments on body composition were apparent with signifi-cant main effects of Diet (P � 0.01) and Exercise (P � 0.001)in the absence of a significant interaction. Subjects in thePRO and PRO � EX groups reduced relative body fatness by4.3% (i.e., change in the percentage of body fat), whereassubjects in the CHO and CHO � EX groups reduced relative

TABLE 3

Body weight and composition for adult women at baseline and after 16 wk of consuming reduced-energy diets with acarbohydrate:protein ratio �3.5 (CHO) or �1.5 (PRO) with or without a supervised exercise program

(EX: 5 d/wk walking and 2 d/wk resistance training)1


P-value

D2 E3,4

kg

Body weightBaseline 91.1 � 5.1 86.1 � 4.6 93.7 � 3.5 79.8 � 2.7Post-test 82.4 � 4.4* 76.3 � 3.9* 85.9 � 3.5* 73.1 � 2.8* �0.05 0.98

Fat massBaseline 39.0 � 3.0 40.9 � 3.6 40.6 � 2.0 36.3 � 2.2Post-test 33.1 � 2.4* 32.1 � 2.9* 35.6 � 2.1* 30.8 � 2.3* �0.05 �0.05

Trunk fatBaseline 19.6 � 2.0 20.1 � 2.2 20.1 � 1.3 17.0 � 1.2Post-test 16.0 � 1.6* 15.1 � 1.9* 17.1 � 1.3* 13.8 � 1.2* �0.05 0.11

Lean massBaseline 50.6 � 2.5 42.6 � 1.4 51.7 � 1.7 40.6 � 0.8Post-test 48.6 � 2.4* 42.2 � 1.4 49.0 � 1.8* 39.6 � 0.8* 0.10 �0.001

1 Values are means � SEM; n � 12. * Different from baseline, P � 0.05.2 D � Test for significant main effect of diet (PRO; n � 24: CHO; n � 24).3 E � Test for significant main effect of exercise (EX; n � 24: control; n � 24).4 D � E was not significant for any variable.

FIGURE 1 Changes in relative body fatness (%Fat) for adultwomen after 16 wk of consuming reduced-energy diets with a ratio ofcarbohydrates:protein � 3.5 (CHO) or � 1.5 (PRO) with or without asupervised exercise program (EX: 5 d/wk walking and 2 d/wk resistancetraining). Values are means � SEM, n � 12. *Significant main effect ofdiet, P � 0.05; #significant main effect of exercise, P � 0.05.

LAYMAN ET AL.1906

body fatness by 2.9%. For the main effect of exercise, subjectsin the PRO � EX and CHO � EX groups reduced rela-tive body fat by 4.7%, whereas subjects not participating inthe supervised exercise program reduced relative body fatnessby 2.5%.

Endocrine markers of obesity. Each of the 4 hormonesresponded to weight loss (Table 4). Plasma concentrations ofleptin and insulin decreased and adiponectin and ghrelinincreased, but only adiponectin exhibited treatment-specificchanges. Blood concentrations of adiponectin more than dou-

TABLE 4

Endocrine changes for adult women at baseline and after 16 wk of consuming reduced-energy dietswith a carbohydrate:protein ratio �3.5 (CHO) or �1.5 (PRO) with or without a supervised exercise program

(EX: 5 d/wk walking and 2 d/wk resistance training)1


P-value

D2 E3,4

Leptin, �g/LBaseline 38.6 � 4.9 26.2 � 3.0 47.6 � 5.8 37.5 � 5.6Post-test 17.9 � 4.2* 11.2 � 2.3* 24.6 � 3.2* 17.4 � 4.5* 0.52 0.45

Adiponectin, mg/LBaseline 2.8 � 0.4 2.8 � 0.4 4.4 � 0.7 3.1 � 0.4Post-test 3.8 � 0.6 5.5 � 1.0* 5.0 � 1.9 7.5 � 1.7* 0.64 �0.05

Ghrelin, ng/LBaseline 30.9 � 2.0 24.2 � 0.9 32.4 � 3.0 24.6 � 1.4Post-test 45.2 � 5.4* 40.5 � 2.9* 51.2 � 5.1* 44.6 � 4.3* 0.34 0.71

Insulin,5 pmol/LBaseline 150.6 � 26.5 189.4 � 17.2 153.5 � 27.9 184.4 � 15.8Post-test 125.5 � 10.8 138.5 � 7.9* 141.3 � 9.3 133.4 � 8.6* 0.78 0.16

1 Values are means � SEM; n � 12. * Different from baseline, P � 0.05.2 D � Test for significant main effect of diet (PRO; n � 24: CHO; n � 24).3 E � Test for significant main effect of exercise (EX; n � 24: No-EX; n � 24).4 D � E was not significant for any variable.5 Group size: PRO, n � 9, PRO � EX, n � 9, CARB, n � 9 and CHO � EX, n � 10.

TABLE 5

Serum lipid concentrations for adult women at baseline and after 16 wk of consuming reduced-energy dietswith a carbohydrate:protein ratio �3.5 (CHO) or �1.5 (PRO) with or without a supervised

exercise program (EX: 5 d/wk walking and 2 d/wk resistance training)1


P-value

D2 E3,4

mmol/L

TCBaseline 5.59 � 0.26 5.00 � 0.23 5.46 � 0.24 5.09 � 0.18Post-test 5.35 � 0.28 4.80 � 0.22 4.91 � 0.22* 4.63 � 0.16* 0.06 0.70

LDL-CBaseline 3.61 � 0.19 3.20 � 0.22 3.52 � 0.19 3.24 � 0.15Post-test 3.54 � 0.22 3.11 � 0.20 3.07 � 0.15* 2.93 � 0.12* �0.05 0.70

HDL-CBaseline 1.33 � 0.09 1.20 � 0.06 1.30 � 0.06 1.36 � 0.07Post-test 1.30 � 0.10 1.25 � 0.09 1.20 � 0.04* 1.28 � 0.07 �0.05 0.19

TAGBaseline 1.42 � 0.15 1.31 � 0.21 1.40 � 0.14 1.08 � 0.13Post-test 1.12 � 0.11* 0.98 � 0.16* 1.38 � 0.18 0.91 � 0.13 �0.05 0.41

LDL-C:HDL-CBaseline 2.84 � 0.23 2.78 � 0.26 2.79 � 0.20 2.48 � 0.19Post-test 2.88 � 0.27 2.64 � 0.26 2.59 � 0.13 2.35 � 0.14 0.35 0.64

TAG:HDL-CBaseline 1.11 � 0.13 1.18 � 0.25 1.12 � 0.13 0.84 � 0.11Post-test 0.90 � 0.10 0.86 � 0.17 1.15 � 0.14 0.75 � 0.12 �0.05 0.25

1 Values are means � SEM; n � 12. * Different from baseline, P � 0.05.2 D � Test for significant main effect of diet (PRO; n � 24: CHO; n � 24).3 E � Test for significant main effect of exercise (EX; n � 24: No-EX; n � 24).4 D � E was not significant for any variable.


bled in the EX groups (P � 0.05). Furthermore, after control-ling for change in fat mass, the main effect of EX remainedsignificant (P � 0.02).

Blood lipids. Initial serum lipid concentrations did notdiffer among all groups (Table 5). After 16 wk of weight loss,blood lipid profiles changed in all groups; however, the pat-terns of changes in TC, LDL-C, HDL-C, and TAG wereaffected by diets. Changes in TC and LDL-C were greater inthe CHO and CHO � EX groups than in the groups thatconsumed the PRO diets. Serum TC in the CHO and CHO� EX groups decreased by 0.51 � 0.09 mmol/L (9.2%) andLDL-C decreased by 0.38 � 0.08 mmol/L (10.4%) from thebaseline values. In the PRO and PRO � EX groups, theseconcentrations decreased by 0.21 � 0.12 mmol/L (3.7%) and0.08 � 0.11 mmol/L (1.7%), respectively. Changes in TAGwere largest in the groups consuming the PRO diet. SerumTAG in the PRO and PRO � EX groups decreased by0.32 � 0.06 mmol/L (20.2%). In the CHO and CHO � EXgroups, TAG concentration decreased by 0.10 � 0.07 mmol/L(5.2%). HDL-C concentrations changed in opposite directionsin the 2 diet treatments producing a significant main effect(P � 0.05). The PRO groups had a net increase of 0.01 � 0.03mmol/L in HDL-C and the CHO groups decreased by0.08 � 0.03 mmol/L. No significant main effects of diet wereevident for the ratio of LDL/HDL. Ratios of TAG:HDL weremore responsive to the PRO diet than the CHO diet(�0.27 � 0.07 vs. �0.03 � 0.07 mmol/L; P � 0.05). Nosignificant effects of EX were apparent for any lipid outcomes.

DISCUSSION

This study demonstrates interactions between the macro-nutrient content of the diet and exercise during weight loss.Subjects consuming diets with more protein and less carbohy-drate (PRO and PRO � EX) lost more total weight and fatmass and tended to lose less lean mass (P � 0.10) than thegroups consuming diets with more carbohydrates and lessprotein (CHO and CHO � EX). Exercise increased the loss ofbody fat and preserved lean mass. The combined effects of dietand exercise appeared to be additive for correcting body com-position expressed as change in the percentage of body fat (Fig.1). The PRO � EX group exhibited the greatest loss of bodyfat (�8.8 kg) with minimal change in lean mass (�0.4 kg),whereas the CHO group had the least change in body fat(�5.0 kg) and the greatest loss of lean mass (�2.7 kg).Changes in body weight and body fat appear to reflect meta-bolic differences between the diets because total energy intakeand daily energy deficits were comparable across all treatmentgroups.

Although this is the first study to examine the combinedeffects of exercise with a reduced-carbohydrate, higher-proteindiet, the findings are consistent with studies examining theindividual effects of diet or exercise. In general, short-termweight loss studies have shown that diets with reduced levelsof carbohydrates and increased protein result in increasedweight loss (6,7,9,12), increased loss of body fat (7,9), andreduced loss of lean body mass (6,9,13). Similarly, supplemen-tal exercise tends to increase weight loss, but has greater effectson body composition through preserving lean body mass whileincreasing fat loss (3,18–21,29). In the present study, therewas a main effect of diet on total weight loss with no addedeffect of exercise. The absence of a main effect of exercise onweight loss is consistent with other reports that found thatweight loss produced by exercise requires daily activity � 30min/d and that increases in lean tissue may also reducechanges in total body weight (18–21). For loss of body fat,

there were main effects of both diet and exercise, indicatingthat effects of the higher-protein, reduced-carbohydrate dietand exercise on fat mass loss are independent and additive.Exercise had a main effect of attenuating loss of lean tissue.

Potential explanations for the differential effects of thedietary macronutrient content on body composition andweight loss include: 1) protein-sparing effects of increasedprotein on lean body mass (10,14), 2) increased loss of body fatassociated with lower insulin response to the reduced carbo-hydrate diet (30), and 3) reduced metabolic efficiency pro-duced by the increased dietary ratio of protein to carbohy-drates (31). Although the present study did not directlyevaluate fundamental mechanisms, the outcomes are consis-tent with the combined effects of each of these mechanisms.

Hormone changes in this study largely reflect decreases inbody fat mass. Consistent with previous reports, plasma levelsof leptin decreased with weight loss (32–34), whereas fastinglevels of ghrelin increased (35,36). Only adiponectin exhib-ited exercise-specific changes.

Adiponectin is an adipocytokine thought to be involved inobesity and insulin sensitivity (37,38). Low plasma concentra-tions of adiponectin are positively associated with the meta-bolic Syndrome X characteristics of increased BMI, percentagebody fat, and blood TAG and reduced levels of HDL-C (39).Weight reduction achieved by energy-restricted diets increasesplasma adiponectin in obese and diabetic patients (38–40).Exercise of short duration that does not alter body weight orbody fat does not change adiponectin levels (40,41). Thepresent study found that an exercise regimen that reducedbody fat increased adiponectin levels. In addition, when con-trolling for changes in body fat (ANCOVA), exercise pro-duced positive changes in adiponection concentrations.

As expected, blood lipids were influenced more by diettreatments than exercise. After 4 mo of energy restriction andweight loss, all subjects exhibited improvements in blood lipidprofiles, but changes in specific lipoprotein fractions variedamong diet treatments. Subjects in the CHO groups had largerreductions in total cholesterol and LDL-C, whereas subjects inthe PRO groups had greater reductions in TAG and main-tained higher circulating levels of HDL-C. These patterns ofchanges in blood lipids are consistent with previous studiescomparing low- vs. high-carbohydrate diets (9,10,31,42,43).

The ATP III guidelines (22) emphasize that coronary heartdisease (CHD) risk is positively related to obesity, TC, LDL-C,and TAG concentrations and inversely related to HDL-Cconcentrations. However, the relative effect of each of thesediet-induced lipid changes on reducing CHD risk is not clear.A meta-analysis of 70 studies evaluating blood lipid changesduring weight loss found that weight reduction produced sig-nificant correlations with changes in blood lipids (44). Theyfound that for each kilogram decrease in body weight there wasa 0.05 mmol/L decrease in TC, a 0.02 mmol/L decrease inLDL-C, a 0.015 mmol/L decrease in TAG, and a 0.007mmol/L decrease in HDL-C. These relations are consistentwith the outcomes of the CHO groups in the present studywhen values are expressed per kilogram decrease in bodyweight (0.069, 0.052, 0.014, and 0.011 mmol/L, respectively).Outcomes from the PRO groups differed in magnitude anddirection of change. The PRO groups reduced TC, LDL-C,and TAG concentrations by 0.023, 0.009, and 0.034 mmol/L,respectively, when expressed per kilogram of weight loss whileincreasing the HDL-C concentration by 0.001 mmol/L. Thesedata indicate that blood lipids respond to changes in both bodyweight and dietary macronutrient content.

The importance of these changes in lipoprotein patterns onlong-term health is unknown. In this study, subjects at base-

LAYMAN ET AL.1908

line were obese, but blood lipids and insulin values werewithin normal ranges. Both dietary approaches were successfulfor weight loss, improving body composition, and improvingblood lipids patterns. However, differential effects of the dietand exercise treatments on blood lipids suggest the possibilitythat specific treatments may be more beneficial for specificindividuals. For example, an individual expressing familialhypercholesterolemia may obtain greater health benefit fromuse of a low-fat, CHO diet which is likely to produce largerchanges in TC and LDL-C (10,30,45), whereas an individualexhibiting hypertriacylglycerolemia and low HDL-C, charac-teristic of the metabolic syndrome or type 2 diabetes (30,45),may derive greater benefit from use of the PRO diet. Furtherresearch is required to determine whether individualization ofdietary macronutrient ratios based on phenotype characteris-tics can improve long-term treatment of obesity and associatedhealth risks.

Dietary outcomes observed in this study raise questionsabout the relative merits of expressing dietary intakes of ma-cronutrients as a percentage of energy vs. absolute amounts.Diets used in this study were designed to provide protein at 0.8and 1.6 g/(kg � d) based on specific targets for dietary leucine(10,11). At baseline, subjects in the CHO groups had a meanprotein intake of 78.0 g/d or 0.89 g/(kg � d). With an energyintake of 8.23 MJ/d, protein accounted for 15.8% of energyintake. During weight loss, using the teaching model of theFood Guide Pyramid (26), subjects reduced total energy intakeemphasizing reduced total lipids and SFA. This energy restric-tion produced a decrease in protein intake to 57 g/d or 0.71g/(kg � d). Although this level of protein represents 17% of thereduced energy intake (5.51 MJ/d) it is below the minimumRDA value of 0.8 g/(kg � d) (24). This may explain in part thegreater loss of lean mass associated with the CHO groups.

A similar concern about the percentage of energy vs. ab-solute amounts exists for dietary lipids. Subjects in the PROgroups reduced dietary intake of total lipids by 30.1 g/d andSFA by 7.5 g/d. However, expressed as a percentage of energy,total lipids account for 31.9% of energy intake and SFAaccount for 12.4% of total energy. Hence, expressing dietaryintakes as a percentage of energy intakes would result in theconclusion that subjects in the PRO groups have dietary fatintakes above current guidelines of the AHA (25). Contraryto this conclusion, the PRO groups significantly reduced con-sumption of total lipids and SFA below baseline values andbelow national averages (46).

In summary, subjects in both diet treatments were success-ful in reducing daily energy intakes, achieving the macronu-trient goals of the respective diets, and reducing body weightand body fat mass. Subjects consuming the PRO diet with aCHO:PRO ratio � 1.5 lost more total weight and body fatmass and tended to lose less lean mass compared with thegroups consuming the CHO diet with a CHO:PRO ra-tio � 3.5. Exercise increased loss of body fat and preservedlean mass. The combined effects of diet and exercise appear tobe independent and additive for enhancing body compositionexpressed as the absolute or change in the percentage of bodyfat. This study adds to the increasing body of evidencesupporting protein-sparing effects derived from maintain-ing higher protein intakes during energy restriction(6,8,10,13,14).

LITERATURE CITED

1. National Heart, Lung and Blood Institute (1998) Clinical guidelines onthe identification, evaluation and treatment of overweight and obesity in adults:the evidence report. Obes. Res. 6: 51S–209S.

2. Wing, R. R. & Hill, J. O. (2001) Successful weight loss maintenance.Annu. Rev. Nutr. 21: 323–341.

3. Jakicic, J. M., Clark, K., Coleman, E., Donnelly, J. E., Foreyt, J., Melanson,E. L., Volek, J., Volpe, S. L. & American College of Sports Medicine (2001)American College of Sports Medicine position stand. Appropriate interventionstrategies for weight loss and prevention of weight regain for adults. Med. Sci.Sports Exerc. 33: 2145–2156.

4. Hirsch, J., Hudgins, L. C., Leibel, R. L. & Rosenbaum, M. (1998) Dietcomposition and energy balance in humans. Am. J. Clin. Nutr. 67: 551S–555S.

5. Golay, A., Allaz, A. F., Morel, Y., de Tonnac, N., Tankova, S. & Reaven, G.(1996) Similar weight loss with low- or high-carbohydrate diets. Am. J. Clin.Nutr. 63: 174–178.

6. Piatti, P. M., Monti, F., Fermo, I., Baruffaldi, L., Nasser, R., Santambrogio,G., Librenti, M. C., Galli-Kienle, M., Pontiroli, A. E. & Pozza, G. (1994) Hypoca-loric high-protein diet improves glucose oxidation and spares lean body mass:comparison to hypocaloric high-carbohydrate diet. Metabolism 43: 1481–1487.

7. Skov, A. R., Toubro, S., Ronn, B., Holm, L. & Astrup, A. (1999) Ran-domized trial on protein vs carbohydrate in ad libitum fat reduced diet for thetreatment of obesity. Int. J. Obes. Relat. Metab. Disord. 23: 528–536.

8. Parker, D., Noakes, M., Luscombe, N. & Clifton, P. (2002) Effect of ahigh-protein, high-monosaturated fat weight loss diet on glycemic control andlipid levels in type 2 diabetes. Diabetes Care 25: 425–430.

9. Westman, E. C., Yancy, W. S., Edman, J. S., Tomlin, K. F. & Perkins, C. E.(2002) Effect of a 6-month adherence to a very low carbohydrate diet program.Am. J. Med. 113: 30–36.



12. Foster, G. D., Wyatt, H. R., Hill, J. O., McGuckin, B. G., Brill, C., Moham-med, B. S., Szapary, P. O., Rader, D. J., Edman, J. S. & Klein, S. (2003) Arandomized trial of a low-carbohydrate diet for obesity. N. Engl. J. Med. 348:2082–2090.

13. Farnsworth, E., Luscombe, N. D., Noakes, M., Wittert, G., Argyiou, E. &Clifton, P. M. (2003) Effect of a high-protein, energy-restricted diet on bodycomposition, glycemic control, and lipid concentrations in overweight and obesehyperinsulinemic men and women. Am. J. Clin. Nutr. 78: 31–39.

14. Bistrian, D. R., Winterer, J., Blackburn, G. L., Young, V. & Sherman, M.(1977) Effect of a protein-sparing diet and brief fast on nitrogen metabolism inmildly obese subjects. J. Lab. Clin. Med. 89: 1030–1035.

15. Latner, J. D. & Schwartz, M. (1999) The effects of a high-carbohydrate,high-protein or balanced lunch upon later food intake and hunger ratings. Appe-tite 33: 119–128.

16. Samaha, F. F., Iqbal, N., Seshadri, P., Chicano, K. L., Daily, D. A.,McGrory, J., Williams, T., Williams, M., Gracely, E. J. & Stern, L. (2003) Alow-carbohydrate as compared with a low-fat diet in severe obesity. N. Engl.J. Med. 348: 2074–2081.

17. Volek, J. S. & Sharman, M. J. (2004) Cardiovascular and hormonalaspects of very-low-carbohydrate ketogenic diets. Obes. Res. 12: 115S–123S.

18. Jakicic, J. M., Winters, C., Lang, W. & Wing, R. R. (1999) Effects ofintermittent exercise and use of home exercise equipment on adherence, weightloss, and fitness in overweight women: a randomized trial. J. Am. Med. Assoc.282: 1554–1560.

19. Ross, R., Dagnone, D., Jones, P. J., Smith, H., Paddags, A., Hudson, R.& Janssen, I. (2000) Reduction in obesity and related comorbid conditionsafter diet-induced weight loss or exercise-induced weight loss in men. A random-ized, controlled trial. Ann. Intern. Med. 133: 92–103.

20. Ballor, D. L., Katch, V. L., Becque, M. D. & Marks, C. R. (1988) Re-sistance weight training during caloric resistance enhances lean body weightmaintenance. Am. J. Clin. Nutr. 47: 19–25.

21. Geliebter, A., Mahar, M. M., Gerace, L., Gutin, B., Heymsfield, S. B. &Hashim, S. A. (1997) Effects of strength or aerobic training on body compo-sition, resting metabolic rate, and peak oxygen consumption in obese dietingsubjects. Am. J. Clin. Nutr. 66: 557–563.

22. National Cholesterol Education Program (2001) Third Report of theExpert Panel on Detection, Evaluation, and Treatment of High Blood Cholesterolin Adults (Adult Treatment Panel III). NIH Publication no. 01–3670, Washington,DC. www.nhlbi.nih.gov/cvd_frameset.htm [accessed January 2, 2005].

23. Institute of Medicine, Food and Nutrition Board (2002) Dietary Refer-ence Intakes for Energy, Carbohydrate, Fiber, Fat, Fatty Acids, Cholesterol,Protein and Amino Acids. National Academy Press, Washington, DC.

24. Food and Nutrition Board, National Research Council (1989) Recom-mended Dietary Allowances. National Academy Press, Washington, DC.

25. American Heart Association Scientific Statement (2000) AHA DietaryGuidelines Revision 2000. Circulation 102: 2284–2299.

26. U.S. Department of Agriculture, Department of Health and Human Ser-vices (1995) Dietary Guidelines for Americans. Home and Garden Bulletin 232,4th ed. DHHS, Washington, DC.

27. Myers, G. L., Cooper, G. R., Winn, C. L. & Smith, S. J. (1989) TheCenters for Disease Control-National Heart, Lung and Blood Institute Lipid Stan-dardization Program. An approach to accurate and precise lipid measurements.Clin. Lab. Med. 9: 105–135.

28. Friedewald, W., Levy, R. I. & Fredrickson, D. S. (1972) Estimation of


the concentration of low-density lipoprotein cholesterol in plasma, without theuse of the preparative ultracentrifuge. Clin. Chem. 18: 499–502.

29. Ross, R., Janssen, I., Dawson, J., Kungl, A. M., Kuk, J. L., Wong, S. L.,Nguyen-Duy, T. B., Lee, S., Kilpatrick, K. & Hudson, R. (2004) Exercise-induced reduction in obesity and insulin resistance in women: a randomizedcontrolled trial. Obes. Res. 12: 789–798.

30. Reaven, G. M. (1993) Role of insulin resistance in human disease(syndrome X): an expanded definition. Annu. Rev. Med. 44: 121–131.

31. Fine, E. & Feinman, R. (2004) Thermodynamics of weight loss diets.Nutr. Metab. 8: 1–15.

32. Baile, C. A., Della-Fera, M. A. & Martin, R. J. (2000) Regulation ofmetabolism and body fat mass by leptin. Annu. Rev. Nutr. 20: 105–127.

33. Thong, F. S., Hudson, R., Ross, R., Janssen, I. & Graham, T. E. (2000)Plasma leptin in moderately obese men: independent effects of weight loss andaerobic exercise. Am. J. Physiol. 279: E307–E313.

34. Kozlowska, L. & Rosolowska-Huszcz, D. (2004) Leptin, thyrotropin,and thyroid hormones in obese/overweight women before and after two levels ofenergy deficit. Endocrine 24: 147–153.

35. Cummings, D. E., Weigle, D. S., Frayo, R. S., Breen, P. A., Ma, M. K.,Dellinger, E. P. & Purnell, J. Q. (2002) Plasma ghrelin levels after diet-inducedweight loss or gastric bypass surgery. N. Engl. J. Med. 346: 1623–1630.

36. Leidy, H. J., Gardner, J. K., Frye, B. R., Snook, M. L., Schuchert, M. K.,Richard, E. L. & Williams, N. I. (2004) Circulating ghrelin is sensitive to changesin body weight during diet and exercise program in normal-weight young women.J. Clin. Endocrinol. Metab. 89: 2659–2664.

37. Beltowski, J. (2003) Adiponectin and resistin—new hormones of whiteadipose tissue. Med. Sci. Monit. 9: RA55–RA61.

38. Yang, W. S., Lee, W. J., Funahashi, T., Tanaka, S., Matsuzawa, Y., Chao,C. L., Chen, C. L., Tai, T. Y. & Chuang, L. M. (2001) Weight reduction increases

plasma levels of an adipose-derived anti-inflammatory protein, adiponectin.J. Clin. Endocrinol. Metab. 86: 3815–3819.

39. Kazumi, T., Kawaguchi, A., Hirano, T. & Yoshino, G. (2004) Serumadiponectin is associated with high-density lipoprotein cholesterol, triglycerides,and low-density lipoprotein particle size in young healthy men. Metabolism 53:589–593.

40. Hulver, M. W., Zheng, D., Tanner, C. J., Houmard, J. A., Kraus, W. E.,Slentz, C. A., Sinha, M. K., Pories, W. J., MacDonald, K. G. & Dohm, G. L. (2002)Adiponectin is not altered with exercise training despite enhanced insulin action.Am. J. Physiol. 283: E861–E865.

41. Ryan, A. S., Nicklas, B. J., Berman, D. M. & Elahi, D. (2003) Adiponec-tin levels do not change with moderate dietary induced weight loss and exercisein obese postmenopausal women. Int. J. Obes. Relat. Metab. Disord. 27: 1066–1071.

42. Sharman, M. J., Gomez, A. L., Kraemer, W. J. & Volek, J. S. (2004)Very low-carbohydrate and low-fat diets affect fasting lipids and postprandiallipemia differently in overweight men. J. Nutr. 134: 880–885.

43. Volek, J. S., Sharman, M. J., Gomez, A. L., DiPasquale, C., Roti, M.,Pumerantz, A. & Kraemer, W. J. (2004) Comparison of a very low-carbohydrateand low-fat diet on fasting lipids, LDL subclasses, insulin resistance, and post-prandial lipemic responses in overweight women. J. Am. Coll. Nutr. 23: 177–184.

44. Datillo, A. M. & Kris-Etherton, P. M. (1992) Effects of weight reductionon blood lipids and lipoproteins: a meta-analysis. Am. J. Clin. Nutr. 56: 320–328.

45. Hu, F. B., Stampfer, M. J., Manson, J. E., Rimm, E., Colditz, G. A.,Speizer, F. E., Hennekens, C. H. & Willett, W. C. (1999) Dietary protein and riskof ischemic heart disease in women. Am. J. Clin. Nutr. 70: 221–227.

46. German, J. B. & Dillard, C. J. (2004) Saturated fats: what dietaryintake? Am. J. Clin. Nutr. 80: 550–559.

LAYMAN ET AL.1910

The Emerging Role of Dairy Proteins and BioactivePeptides in Nutrition and Health

Dietary Protein Impact on Glycemic Control during Weight Loss1

Donald K. Layman2 and Jamie I. Baum*

Department of Food Science and Human Nutrition; *Division of Nutritional Sciences, University of IllinoisUrbana-Champaign, Urbana, IL 61801

ABSTRACT Diets with higher protein (1.5 g � kg�1 � d�1) and reduced carbohydrates (120 to 200 g/d) appear toenhance weight loss due to a higher loss of body fat and reduced loss of lean body mass. While studies ofprolonged use of moderate protein diets are not available, short-term studies report beneficial effects associatedwith increased satiety, increased thermogenesis, sparing of muscle protein loss, and enhanced glycemic control.Combined impacts of a moderate protein diet are likely derived from lower carbohydrates resulting in lowerpostprandial increase in blood glucose and lower insulin response, and higher protein providing increased BCAAleucine levels and gluconeogenic substrates. A key element in the diet appears to be the higher intake of BCAAleucine with unique regulatory actions on muscle protein synthesis, modulation of the insulin signal, and sparingof glucose use by stimulation of the glucose-alanine cycle. This review focuses on the contributions of leucine andthe BCAA to regulation of muscle protein synthesis and glycemic control. J. Nutr. 134: 968S–973S, 2004.

KEY WORDS: ● obesity ● insulin ● leucine ● BCAA

There is general consensus that the most critical factor inweight management is total energy intake. Yet the ideal bal-ance of macronutrients for weight loss and adult weight man-agement remains widely disputed. Often this debate focuses onthe relative merits or risks of carbohydrates vs. lipids. How-ever, when the energy content of the diet is equal, the relativelevels of carbohydrates and lipids in the diet appear to haveminimal affect on either weight loss or body composition(1–3). On the other hand, there is increasing evidence thatdiets with reduced levels of carbohydrates and higher levels ofprotein may be beneficial for weight loss (4–10). These studiesreport that diets with reduced carbohydrates and higher pro-tein appear to increase weight loss (4–6,8,9), increase loss ofbody fat (5,6,8), or reduce loss of lean body mass (4,6,8,10).While potential benefits for higher protein diets during weightloss are emerging, a metabolic explanation for optimal levels ofcarbohydrates and proteins remains unknown (11).

Possible explanations for the beneficial effects of diets withhigher protein and reduced levels of carbohydrates includelower energy intake associated with increased satiety(5,7,9,12), reduced energy efficiency or increased thermogen-esis (6,13), sparing of muscle protein loss (8,14), and enhancedglycemic control (8,10). Our research has focused on the role

of amino acids in regulation of muscle protein metabolism(8,15,16) and glycemic control (17,18). This presentation islimited to these topics.

The role of protein in the diet is to provide the 20 naturallyoccurring amino acids and specifically to provide the 9 indis-pensable amino acids. Each of these amino acids has a uniquerequirement as a building block for body proteins. However,the dietary requirement is not tightly linked to substrate needsfor protein synthesis. One reason for the lack of a directrelationship is the recycling of amino acids after degradation ofexisting proteins. Amino acids are efficiently reutilized forsynthesis of new proteins. Even during maximum rates ofgrowth the body deposits �10 g of protein per d (19). Hencethe dietary protein needed to maintain essential protein turn-over appears quantitatively small and with no clear metabolicrelationship to the current RDAs (recommended dietary al-lowance) of 0.8 g � kg�1 � d�1.

Beyond the needs for amino acids required for synthesis ofnew proteins, amino acids participate in numerous metabolicroles. In many cases the significance of these pathways isproportional to dietary intake, such as dietary intake of tryp-tophan or phenylalanine (i.e., tyrosine) as precursors to neu-rotransmitters with dietary intake potentially impacting appe-tite regulation (20,21), or intake of arginine altering epithelialproduction of nitrous oxide and cell signaling pathways(22,23). Another example of an amino acid with metabolicroles proportional to dietary intake is the BCAA leucine withpotential regulatory roles on skeletal muscle protein synthesisand glycemic control (16,17).

Leucine exhibits an array of metabolic roles. Like all aminoacids, leucine is essential for protein synthesis. Based on ni-trogen-balance measurements, the requirement for leucine to

1 Published in a supplement to The Journal of Nutrition. Presented as part ofthe 94th American Oil Chemists’ Annual Meeting & Expo held in Kansas City, MO,May 4–7, 2003. This symposium was sponsored by the National Dairy Council,Kraft Foods Inc., The Whey Protein Institute, and the U.S. Dairy Export Council.Guest editors for the supplement publication were Peter J. Huth, National DairyCouncil, Rosemont, IL; Donald K. Layman, University of Illinois, Urbana, IL; andPeter H. Brown, Kraft Foods Research and Development, Kraft Foods Inc.,Glenview, IL.

2 To whom correspondence should be addressed. Email: [email protected].

0022-3166/04 $8.00 © 2004 American Society for Nutritional Sciences.

968S

maintain short-term stability of body protein is �1 to 3 g perd (11,24). However, leucine participates in numerous othermetabolic processes including serving as a fuel for skeletalmuscle (25,26), modulation of the intracellular insulin/phos-phatidylinositol-3 kinase (PI3-K)3 signaling cascade (27,28), aunique regulator of muscle protein synthesis (29,30), andserving as a donor of an amino group for production of alanineor glutamine (31,32). In each of these pathways, the impact ofleucine is proportional to availability and is dependent on itsintracellular concentration. To optimize these pathways, weestimate that the leucine requirement is �8 g/d (8,33,34).Leucine is relatively abundant in the food supply, accountingfor �8% of dietary protein with dairy products being particu-larly rich in leucine and the BCAA (Table 1). This range ofleucine intake is reasonable within the guidelines of the di-etary reference intakes (11). These metabolic roles for leucineform the bases for our hypothesis for the importance of in-creased dietary protein during weight loss (8,17).

Leucine and regulation of muscle protein synthesis

The role of leucine in muscle protein synthesis is differentfrom other essential amino acids. During catabolic periodssuch as energy restriction, supplementation with leucine or acomplete mixture of the 3 BCAAs, leucine, isoleucine, andvaline, stimulates muscle protein synthesis (35–37). This re-search suggests a regulatory role of leucine that is dependenton intracellular concentration and is different from traditionalsubstrate roles for protein synthesis or nitrogen balance(36,38,39). We found that leucine supplementation stimulatesrecovery of muscle protein synthesis during food restriction orafter endurance exercise (38,39).

The molecular mechanisms for the actions of leucine inprotein synthesis are now known to involve regulation ofphosphorylation events and components of the insulin signal-ing pathway. The site for leucine action is a kinase in theinsulin signaling cascade previously identified as mTOR(mammalian target of rapamycin) (Fig. 1). This regulation wasfirst recognized associated with translational control of muscleprotein synthesis (28,38). Increases in leucine concentrationstimulate mTOR kinase activity for phosphorylation control ofthe eIF4 initiation complex and of the S6 ribosomal protein.Specifically, leucine stimulates phosphorylation of the inhib-itory binding protein (4E-BP1) causing the binding protein todissociate from the eIF4E translational initiation factor. After

dissociation, eIF4E is available to bind with eIF4G and formthe active initiation complex (Fig. 1). Leucine via mTOR alsoincreases activation of p70S6 kinase leading to phosphoryla-tion of the S6 ribosomal protein and enhanced global rates ofprotein synthesis (39). The mechanisms for translational reg-ulations by leucine have been recently reviewed (29,30). Thisunique role of leucine in regulation of muscle protein synthesisis consistent with the sparing of lean body mass seen with useof higher protein diets during weight loss (8,10,14).

Regulation of blood glucose

Before examining roles for amino acids in glucose ho-meostasis, 2 concepts for regulation of blood glucose concen-tration will be reviewed briefly. In a fasted condition, the liveris the sole source of endogenous glucose production (EGP) andmaintains a rate of production [or rate of appearance (Ra)] of�5 to 7 g of glucose per h. Hepatic glucose production isderived from a combination of glycogen breakdown and glu-coneogenesis. The relative contribution of these 2 pathwayshas been debated; however, gluconeogenesis appears to be thepredominate pathway accounting for up to 75% of EGP (40–42). The specific contribution from gluconeogenesis is influ-enced by dietary factors including the duration of the fastedperiod and previous dietary intakes of carbohydrates and pro-tein. During nonabsorptive conditions, the rate of hepaticglucose production is exactly balanced with the rate of glucoseuse by peripheral tissues under basal insulin levels ranging from5–20 �U/mL (35–140 pmol/L).

During periods of food intake and absorption of exogeneousglucose, hepatic release of glucose into the blood may exceed30 g/h (41). The rise in dietary glucose in portal circulationand the increased rate of appearance of glucose from the liverinto the blood stimulate release of insulin to accelerate pe-ripheral disposal of glucose in skeletal muscle and adiposetissue. The increase in insulin also serves to reduce EGP byinhibiting hepatic gluconeogenesis and glycogen breakdown.In nondiabetic conditions, the net balance of hepatic releaseand peripheral uptake limits the rise in postprandial bloodglucose to �7.77 mmol/L.

Critical questions in evaluating the balance of protein and

3 Abbreviations used: AUC, area under the curve; BCKAD, branched-chainketoacid dehydrogenase; 4E-BP1, inhibitory binding protein; EGP, endogenousglucose production; GLUT4, insulin dependent glucose transporter; IRS-1, insulinreceptor substrate; mTOR, mammalian target of rapamycin; PI3-K, phosphatidyl-inositol-3 kinase; Ra, rate of appearance.

FIGURE 1 Insulin signaling cascade. GLUT1, insulin independentglucose transporter; PKC, protein kinase C; eIF, translational initiationfactors.

TABLE 1

Leucine and BCAA content of foods1

Leucine BCAA

Whey protein isolate 14% 26%Milk protein 10% 21%Egg protein 8.5% 20%Muscle protein 8% 18%Soy protein isolate 8% 18%Wheat protein 7% 15%

1 Values reflect g of amino acids/100 g of protein. Source: USDAFood Composition Tables.

DIETARY PROTEIN AND WEIGHT LOSS 969S

carbohydrate in the diet concern the fundamental assumptionabout the ideal regulation of blood glucose. If the assumptionis that insulin is the primary regulator of blood glucose, thenthe experimental model is likely to be based on evaluatingfactors affecting the ability of insulin to handle an exogenousbolus of carbohydrates. However, if the assumption is that theliver is the primary regulator of blood glucose then the idealmodel is focused on hepatic control of the rate of glucoseappearance and regulation of the rate of gluconeogenesis. Thecorrect model is likely different for a 22 y old athlete with highmuscle activity consuming a 3500 kcal/d (14653 kJ/d) diet vs.a sedentary 53 y old adult attempting to restrict energy intaketo 1700 kcal/d (7117 kJ/d) to achieve weight loss. The deci-sion about the correct model underpins the dietary decisionabout the ideal mixture of protein vs. carbohydrates duringweight loss.

Figure 2 illustrates a theoretical oral glucose response curveplus a parallel insulin response curve. These curves representresponses to a meal providing �400 kcal (1675 kJ). At timezero (fasted), the Ra for glucose is low; fasting blood glucose is4.7 to 5.0 mmol/L; insulin is at basal levels; and regulation ofblood glucose is predominately hepatic. With the beginning ofa high carbohydrate meal, there is a rapid rise in blood glucoseto a maximum of 7.77 mmol/L and then a return to fastedconcentrations after 120 min. The time frame for an OGRC isdependent on the size and composition of the meal. The mealresponse for insulin follows a similar time course with fastinginsulin of �15 �U/mL (105 pmol/L) increasing rapidly to anearly peak of perhaps 60 to 80 �U/mL (430–575 pmol/L) andthen a prolonged slow return to fasting levels. These curveshighlight 2 important periods of glycemic control, fasted ornonabsorptive periods at times 0 and 120 min, and the ab-sorptive period. During the nonabsorptive periods, hepaticproduction releases glucose in proportion to peripheral use andblood glucose remains stable. During the absorptive period,there is a high rate of appearance of exogenous glucose. As theglucose appears in the blood, insulin is released to coordinatethe rate of glucose disposal with the rate of appearance. Underconditions of insulin resistance as seen in type-2 diabetes thisregulation is ineffective and blood glucose continues to riseoften well above 11.1 mmol/L.

Assuming that insulin is the primary regulator of bloodglucose, then experiments are often designed to evaluate peakinsulin response or area under the curve (AUC) of an oralglucose response curve. For this approach, 1 of the most usefultechniques has been the hyperinsulinemic, euglycemic clamp

(43). This method requires insulin to be infused into a periph-eral vein at a constant rate to stabilize blood insulin at 80–150�U/mL (575–1075 pmol/L). Then glucose is infused at ratesthat are adjusted to achieve stable concentration of bloodglucose at �7.77 mmol/L. Under these conditions, the rate ofinfusion of glucose equals the rate of glucose disposal bytissues. For type 2 diabetes, a much lower level of glucosewould be infused reflecting a lower rate of peripheral glucoseclearance and insulin insensitivity. Using the same insulinmodel and the euglycemic clamp, it is possible to test theeffects of infusion of amino acids on glycemic control. Inves-tigators using this approach demonstrate that amino acid in-fusion reduces glucose uptake leading to the conclusion thatincreased amino acid availability produces insulin insensitivityand inhibits glucose utilization (44).

Impact of amino acids on glycemic control

Interactions of amino acids with carbohydrate metabolismhave been recognized for years; however, the research litera-ture is unclear whether dietary protein has a positive or neg-ative impact on glycemic control. Amino acids directly con-tribute to de novo synthesis of glucose via gluconeogenesis andparticipate in re-cycling of glucose carbon via the glucose-alanine cycle. Amino acids including arginine and leucinestimulate insulin release from the pancreas; leucine also ap-pears to modulate the intracellular insulin signal in skeletalmuscle and adipose tissue (45). The net impact of amino acidson glucose homeostasis appears to be dependent on the exper-imental approach and the amount of amino acids used.

In 1927 Sweeney (46) reported that young adults fed dietshigh in protein displayed reduced ability to dispose of oralglucose. Specifically, subjects were fed test diets for 2 days thatwere either mostly carbohydrates (e.g., bread, potatoes, riceand oatmeal) or mostly protein (e.g., lean meats and eggwhites). On d 3, subjects were tested for their response to anoral glucose tolerance test using �100 g of glucose (1.75 g/kg)and blood was obtained at 0, 30, 60, and 120 min. Subjectspreconditioned with the carbohydrate diet exhibited peakplasma glucose concentrations of 6.66 mmol/L, while subjectspreconditioned with the protein meals had peak glucose con-centrations of �8.88 mmol/L. These data suggest that a highprotein diet decreases oral glucose tolerance.

Later studies reported that amino acids decrease glucosedisposal, induce hyperinsulinemia and hyperglycemia, and po-tentially lead to insulin resistance (44,47–49). Most of thesestudies used direct intravenous infusion of amino acids into thehuman forearm under fasted conditions and used euglycemicclamp techniques to measure glucose uptake and insulin resis-tance. Using these techniques, investigators found that acuteincreases in plasma amino acid concentrations resulted inhigher plasma glucose concentrations, lower glucose uptake,and elevated plasma insulin levels (44,47,48). Possible mech-anisms for these actions include competition between aminoacids and glucose as oxidative substrates (47,48,50) or modu-lation of the insulin response including reduced glucose uptakeor direct interaction with early steps in insulin signaling (27).These studies used supraphysiological concentrations of insu-lin and amino acids. While these acute conditions are usefulfor discerning increments in insulin sensitivity, the physiolog-ical relevance of these supraphysiological concentrations ad-ministered i.v. for predicting response to chronic dietary con-ditions is unclear.

One of the first studies of the differences in amino acidmetabolism between i.v. administration and oral intake was byFloyd et al. (51,52). These investigators evaluated the insulin

FIGURE 2 Theoretical glucose and insulin curves for responses toa 400 kcal (1675 kJ) breakfast meal.

SUPPLEMENT970S

response to i.v. infusion of amino acids or glucose (51) and alsoexamined the insulin response to oral intake of protein (52).They found that infusion of 30 g of amino acids produced a3-fold higher insulin response (�180 �U/mL) than infusion of30 g of glucose (�50 �U/mL), suggesting a dramatic hyper-insulinemic effect of amino acids. However, these investigatorsalso examined the same measurements after subjects consumeda meal of 500 g of beef liver and found that the peak insulinresponse to the protein meal was only 30 �U/mL. Assumingthat leucine is 1 of the most potent insulin secretagogues, thei.v. infusion provided �5 g of leucine while the beef mealprovide �14 g of leucine (52). These data suggest that aminoacids have minimal impact on plasma insulin concentrationswhen entering the body via the GI tract.

Nuttall and Gannon (53–55) reported similar findings.Using isoenergetic meals, they demonstrated that substitutingdietary protein for carbohydrates reduced the meal responses ofboth plasma glucose and insulin (53). Likewise, they reportedthat consumption of a test meal containing 50 g of protein(consumed as lean beef) vs. 50 g of glucose that the proteinintake alone had essentially no impact on basal blood glucoseconcentrations and the insulin response to the meal was�20% of the response with a comparable energy intake fromglucose (54). While these results are intuitively obvious, theydirectly contradict the findings that protein is hyperinsuline-mic and hyperglycemic.

Explanations for the differences in handling of i.v. vs. oralamino acids involve diverse metabolic pathways. Unlike glu-cose, which appears in the blood rapidly after a meal, aminoacids are slow to leave the gut (56), extensively modified incomposition by the gut and liver (13), and appear in the bloodslowly with metabolism over an extended postprandial period(48). Metabolism of dietary amino acids by the gut and liverhas a major impact on the amino acid profile reaching systemiccirculation. Specific examples include removal of nearly 100%of dietary glutamine and glutamate, 60% of threonine, and40% of phenylalanine during the absorption process largely byoxidative degradation (13). The primary exceptions to thispattern of modifications are the BCAA, with over 80% ofdietary content of leucine, valine, and isoleucine directlyreaching blood circulation. A second important issue in con-sidering amino acid metabolism compared with glucose han-dling is the time course. For glucose, the postprandial handlingoccurs mostly within the first 2 h (43); however for aminoacids the rate of disposal is much slower with �20% of thedietary amino acids degraded within the first 2 h (48). Thus,direct comparison of a high carbohydrate diet vs. a highprotein diet is that the carbohydrate diet requires rapid equil-ibration of the glucose and insulin metabolic system withdramatic shifts between hepatic vs. peripheral regulations,while a high protein diet serves to stabilize the glycemicenvironment with delayed metabolism and less reliance onperipheral insulin actions.

Glycemic control with moderate protein, moderatecarbohydrate weight loss diets

Diets with reduced carbohydrates and higher protein pro-duce lower meal responses for glucose and insulin (10,18,54).During studies of weight loss, we found that adult womenmaintained on a moderate protein diet for 10 wk had morestable blood glucose after an overnight fast and at 2 h after atest meal (8,18). The moderate protein diet also appeared tostabilize the insulin response to a test meal, while subjectsreceiving an isoenergetic diet high in carbohydrates increasedthe insulin needed to respond to a test meal. Similar meal

responses were reported by Farnsworth et al. (10). They foundthat after 16 wk subjects consuming a moderate protein weightloss diet displayed lower meal responses for peak glucose andinsulin concentrations and total AUC after test meals. Thesedata suggest that diets with reduced carbohydrates and higherprotein stabilize glycemic control during weight loss.

We observed an additional example of this effect withsubjects exhibiting elevated postprandial insulin responses(Layman, unpublished data). During preliminary screening ofoverweight subjects, we identified 10 subjects with abnormallyhigh insulin responses at 2 h after a test meal (Fig. 3). Normalvalues for a 2-h postprandial response are �5–15 �U/mLabove basal insulin concentrations, while these subjects aver-aged 76 �U/mL above basal levels. Subjects were paired forbody weight and insulin values and randomly assigned toeither the moderate protein or high carbohydrate diet. Afterconsuming the respective diets for 4 and 10 wk, we evaluatedinsulin responses to the test meals. As expected, as the subjectslost weight (�6.3 kg) during the 10-wk energy restriction andthey improved glycemic control as measured by reduced post-prandial insulin response to the test meal. For the CHOGroup, average values at wk 0 � 77 �U/mL and at wk 10 � 38�U/mL. On the other hand, subjects consuming the moderateprotein diet achieved normal values for 2-h insulin responseafter only 4 wk on the diet with average values at wk 0 � 75�U/mL and at wk 10 � 12 �U/mL. These changes appear tobe beneficial associated with the overall risk patterns of obesityand Metabolic Syndrome (57,58).

Reasons for enhanced glycemic control with use of moder-ate protein diets remain to be fully elucidated; however, ele-ments of possible regulations have been established. The over-all contributions of dietary amino acids to glucose homeostasiswere established by quantitative evaluations of hepatic glucoseproduction. Jungas et al. (59) reported that amino acids serveas a primary fuel for the liver and the primary carbon source forhepatic gluconeogenesis. Other investigators found that glu-coneogenesis provides �70% of fasting hepatic glucose re-lease, with amino acids serving as the principal carbon source(40,41). Estimates of the contribution of amino acid carbon tode novo glucose synthesis range from 0.6 to 0.7 g of glucosefrom 1 g of dietary protein (13,55). In addition to the directconversion of amino acid carbon to gluconeogenesis precur-

FIGURE 3 Curves represent 2-h postprandial insulin responses toa 400 kcal (1675 kJ) test meal in overweight women assigned to eithermoderate protein (Protein Group) or a high carbohydrate (CHO Group)weight loss diets. Both diets provided 1700 kcal/d (7117 kJ/d) (8). Theinsulin response was determined as the 2-h postprandial value minusthe fasting value for each subject. Values represent means � SEM, n� 5 (Layman, unpublished).


sors, there is also the contribution of the BCAA to glucoserecycling via the glucose-alanine cycle (31,32). There is acontinuous flux of BCAA from visceral tissues through theblood to skeletal muscle where transamination of the BCAAprovides the amino group for production of alanine frompyruvate with a corresponding movement of alanine frommuscle to liver to support hepatic gluconeogenesis. Althoughthe impact of the glucose-alanine cycle has been debated,Ahlborg et al. (31) reported that alanine accounted for 40% ofendogenous glucose production during prolonged exercise.Under normal conditions, alanine arising from BCAA nitro-gen likely accounts for about 25% of gluconeogenesis fromamino acids (17). These studies provide evidence for thelinkage between dietary protein and glucose homeostasis.

A focus on leucine

The interactions between amino acids and glycemic controlare influenced by total dietary protein providing substrates forgluconeogenesis and by total intake of BCAA determining thecapacity for glucose-alanine cycling. Within these require-ments, leucine serves as a selective marker for intracellularrecognition of the quantity and/or quality of protein in thediet. Peripheral tissues have a unique ability to sense theintracellular leucine concentration. Increases in leucine trig-ger an array of phosphorylation events that serve to maintainskeletal muscle mass and limit oxidative use of glucose bymuscle. The combination of circulating insulin and tissuelevels of leucine allow skeletal muscles to manage proteinmetabolism and fuel selection in relation to diet composition.

As outlined above, leucine stimulates translational regula-tion of muscle protein synthesis through modulation of down-stream elements of the insulin/PI3-k signal pathway (Fig. 1).Higher leucine stimulates mTOR kinase activity and phos-phorylation of the inhibitory binding protein 4E-BP1 andp70S6kinase. The mechanism allowing mTOR to respond toleucine concentrations likely involves a secondary proteinwith a potential candidate identified as rapTOR (60). Thisrelationship allows for skeletal muscle to sense the quantity orquality of dietary protein and to adjust the rate of muscleprotein synthesis in proportion to the availability of substrate.

Parallel with mTOR actions on translational initiationfactors, mTOR has been shown to stimulate upstream phos-phorylations of insulin receptor substrate (IRS-1) potentiallyaltering the insulin receptor signal (27,28). These findingshave been suggested as possible explanations of the aminoacid-induced insulin resistance observed with euglycemicclamps (27,44). We observed similar effects in muscle usingoral gavage of leucine. Our preliminary findings suggest thatleucine induces mTOR phosphorylation resulting in down-stream phosphorylations of 4E-BP1 with activation of theinitiation factors and upstream phosphorylations of IRS-1 withdecreased activity of the PI3-K complex (61). However, wefound no change in the rate of glucose uptake into the mus-cles. These data suggest that potential downregulation of theinsulin signal does not produce negative outcomes on eitherrates of protein synthesis or glucose transport. Possible expla-nations for the apparent disconnect of the insulin signal withglucose transport may be that under sedentary conditions basallevels of insulin dependent glucose transporter (GLUT4) onthe cell membrane are adequate to maintain glucose transport,or that levels of the noninsulin dependent GLUT1 are impor-tant for baseline levels of glucose transport. An additionalpossibility is that mTOR phosphorylation of IRS-1 is a com-ponent of the normal feedback regulation of the insulin signaland decreased activity of PI3-K represents degradation of the

signaling complex. In support of this possibility, we observedthat the decrease in PI3-K activity was greater at 60 min afteroral gavage than at 30 min suggesting an initial activationwith accelerated degradation of the IRS1-PI3-K complex (61).

Leucine also serves as a metabolic signal for fuel choices.Increases in the intracellular concentration of leucine appearto be the primary regulator of the branched-chain ketoaciddehydrogenase (BCKAD), the rate-limiting step in oxidationof the BCAA (62). An increase in leucine raises the concen-tration of its keto-analogue �-ketoisocaproate, a potent inhib-itor of the BCKAD kinase that is responsible for inactivationof the BCKAD by phosphorylation. Inhibition of the BCKADkinase leaves the BCKAD phosphatase unopposed, resultingin dephosphorylation and activation of the BCKAD. TheBCKAD stimulates decarboxylation of the 3 BCAAs andcommits them to oxidation. At the same time, the rise inleucine concentration also inhibits pyruvate dehydrogenase(50), limiting pyurvate oxidation and moderating glucose deg-radation by skeletal muscle (48). Thus when intracellularlevels of leucine are elevated, muscle has the potential to useglucose derived from either the blood or muscle glycogen as aglycolytic fuel and then trap the pyruvate carbon as alanine viatransamination with amino acid-nitrogen derived fromBCAA. These mechanisms appear to be particularly importantduring periods of low energy intake or endurance exercisewhen BCAAs are increased in muscle, insulin is low, andsparing of blood glucose is important (17,25,26).

In summary, use of diets with higher protein and reducedcarbohydrates appears to enhance weight loss with greater lossof body fat and reduced loss of lean body mass. Beneficialeffects of high protein diets may be increased satiety, increasedthermogenesis, sparing of muscle protein loss, and enhancedglycemic control. Specific mechanisms to explain each of theobserved outcomes remain to be fully elucidated. We suggestthat a key to understanding the relationship between dietaryprotein and carbohydrates is the relationship between theintakes of leucine and glucose. Leucine is now known tointeract with the insulin-signaling pathway with apparentmodulation of the downstream signal for control of proteinsynthesis resulting in maintenance of muscle protein duringperiods of restricted energy intake. Leucine also appears tomodulate glucose use by skeletal muscle. While total protein isimportant in providing substrates for gluconeogenesis, leucineappears to regulate oxidative use of glucose by skeletal musclethrough stimulation of glucose recycling via the glucose-ala-nine cycle. These mechanisms appear to provide a stableglucose environment with low insulin responses during energy-restricted periods.

LITERATURE CITED1. Yang, M. U. & Van Itallie, T. B. (1976) Composition of weight lost

during short-term weight reduction: metabolic responses of obese subjects tostarvation and low-calorie ketogenic and non-ketogenic diets. J. Clin. Invest. 58:722–730.

2. Hirsch, J., Hudgins, L. C., Leibel, R. L. & Rosenbaum, M. (1998) Dietcomposition and energy balance in humans. Am. J. Clin. Nutr. 67: 551S–555S.

3. Golay, A., Allax, A. F., Morel, Y., de Tonnac, N., Tankova, S. & Reaven,G. M. (1996) Similar weight loss with low- or high carbohydrate diets. Am. J.Clin. Nutr. 63: 174–178.

4. Piatti, P. M, Monti, L. D., Magni, F., Fermo, I., Baruffaldi, L., Nasser, R.,Santambrogio, G., Librenti, M. C., Galli-Kienle, M., Pontiroli, A. E. & Pozza, G.(1994) Hypocaloric high-protein diet improves glucose oxidation and spareslean body mass: comparison to hypocaloric high carbohydrate diet. Metab. 43:1481–1487.

5. Skov, A. R., Toubro, S., Ronn, B., Holm, L. & Astrup, A. (1999) Ran-domized trial on protein vs carbohydrate in ad libitum fat reduced diet for thetreatment of obesity. Int. J. Obes. 23: 528–536.

6. Parker, B., Noakes, M., Luscombe, N. & Clifton, P. (2002) Effect of ahigh protein, high monounsaturated fat weight loss diet on glycemic control andlipid levels in type 2 diabetes. Diabetes Care 25: 425–430.

SUPPLEMENT972S

7. Westman, E. C., Yancy, W. S., Edman, J. S., Tomlin, K. F. & Perkins, C. E.(2002) Effect of 6-month adherence to a very low carbohydrate diet program.Am. J. Med. 113: 30–36.


9. Foster, G. D., Wyatt, H. R., Hill, J. O., McGuckin, B. G., Brill, C., Moham-med, S., Szapary, P. O., Rader, D. J., Edman, J. S. & Klein, S. (2003) Arandomized trial of a low-carbohydrate diet for obesity. N. Engl. J. Med. 348:2082–2090.

10. Farnsworth, E., Luscombe, N. D., Noakes, M., Wittert, G., Argyiou, E. &Clifton, P. M. (2003) Effect of a high-protein, energy-restricted diet on bodycomposition, glycemic control, and lipid concentrations in overweight and obesehyperinsulinemic men and women. Am. J. Clin. Nutr. 78: 31–39.

11. Institute of Medicine (2002) Dietary Reference Intakes for Energy,Carbohydrate, Fiber, Fat, Fatty Acids, Cholesterol, Protein and Amino Acids.Food and Nutrition Board. Washington DC, National Academy Press.

12. Latner, J. D. & Schwartz, M. (1999) The effects of a high-carbohydrate,high-protein or balanced lunch upon later food intake and hunger ratings. Appe-tite 33: 119–128.

13. Reeds, P. J., Burrin, D. G., Davis, T. A. & Stoll, B. (1998) Amino acidmetabolism and the energetics of growth. Arch. Anim. Nutr. 51: 187–197.

14. Bistran, B. R., Winterer, J., Blackburn, G. L., Young, V. & Sherman, M.(1977) Effect of a protein-sparing diet and brief fast on nitrogen metabolism inmildly obese subjects. J. Lab. Clin. Med. 89: 1030–1035.

15. Gautsch, T. A., Anthony, J. C., Kimball, S. R., Paul, G. L., Layman, D. K.& Jefferson, L. S. (1998) Availability of eIF4E regulates skeletal muscle proteinsynthesis during recovery from exercise. Am. J. Physiol. Cell Physiol. 274: C406–C414.

16. Layman, D. K. (2002) Role of leucine in protein metabolism duringexercise and recovery. Can. J. Appl. Physiol. 27: 592–608.

17. Layman, D. K. (2003) The role of leucine in weight loss diets andglucose homeostasis. J. Nutr. 133: 261S–267S.


19. Pacy, P. J., Price, G. M., Halliday, D., Quevedo, M. R. & Millward, D. J.(1994) Nitrogen homeostasis in man: the diurnal responses of protein synthesisand degradation and amino acid oxidation to diets with increasing protein intakes.Clin. Sci. 86: 103–118.

20. Fernstrom, J. D. & Wurtman, R. J. (1972) Brain serotonin content:physiological regulation by plasma neutral amino acids. Science. 178: 414–416.

21. Fernstrom, M. H. & Fernstrom, J. D. (1995) Brain tryptophan concen-trations and serotonin synthesis remain responsive to food consumption after theingestion of sequential meals. Am. J. Clin. Nutr. 61: 312–319.

22. Moncada, S. & Higgs, E. A. (1995) Molecular mechanisms and ther-apeutic strategiesrelated to nitric oxide. FASEB J. 9: 1319–1330.

23. Wu, G. & Morris, S. M. (1998) Arginine metabolism: nitric oxide andbeyond. Biochem. J. 336: 1–17.

24. FAO/WHO/UNU (1985) Energy and protein requirements. Report ofjoint FAO/WHO/UNU expert consultation. WHO Tech. Rep. Sev. 724: 1–206.

25. Wagenmaker, A.J.M. (1998) Muscle amino acid metabolism at restand during exercise: role in human physiology and metabolism. Exerc. Sport Sci.Rev. 26: 287–314.

26. Rennie, M. J. & Tipton, K. D. (2000) Protein and amino acid metabo-lism during and after exercise and the effects of nutrition. Annu. Rev. Nutr. 20:457–483.

27. Patti, M.-E., Brambilla, E., Luzi, L., Landaker, E. J. & Kahn, C. R. (1998)Bidirectional modulation of insulin action by amino acids. J. Clin. Invest. 101:1519–1529.

28. Xu, G., Kwon, G., Marshall, C. A., Lin, T.-A. & Lawrence, J. C. (1998)Branched-chain amino acids are essential in the regulation of PHAS-I and p70 S6kinase by pancreatic �-cells. J. Biol. Chem. 273: 28178–28184.

29. Kimball, S. R. & Jefferson, L. S. (2001) Regulation of protein synthesisby branched-chain amino acids. Curr. Opin. Clin. Nutr. Metab. Care 4: 39–43.

30. Anthony, J. C., Anthony, T. G., Kimball, S. R. & Jefferson, L. S. (2001)Signaling pathways involved in translational control of protein synthesis in skeletalmuscle by leucine. J. Nutr. 131: 856S–860S.

31. Ahlborg, G., Felig, P., Hagenfeldt, R. & Wahren, J. (1974) Substrateturnover during prolonged exercise in man. J. Clin. Invest. 53: 1080–1090.

32. Ruderman, N. B. (1975) Muscle amino acid metabolism and glucone-ogenesis. Ann. Rev. Med. 26: 245–258.

33. El-Khoury, A. E., Kukagawa, N. K., Sanchez, M., Tsay, R. H., Gleason,R. E., Chapman, T. E. & Young, V. R. (1994) The 24-h pattern and rate ofleucine oxidation, with particular reference to tracer estimates of leucine require-ments in healthy adults. Am. J. Clin. Nutr. 59: 1012–1020.

34. Evans, W. J., Fisher, E. C., Hoerr, R. A. & Young, V. R. (1983) Proteinmetabolism and endurance exercise. Physician Sports Med. 11: 63–72.

35. Li, J. B. & Jefferson, L. S. (1978) Influence of amino acid availability on

protein turnover in perfused skeletal muscle. Biochim. Biophys. Acta 544: 351–359.

36. Buse, M. G. & Reid, S. S. (1975) Leucine. A possible regulator ofprotein turnover in muscle. J. Clin. Invest. 56: 1250–1261.

37. Hong, S. C. & Layman, D. K. (1984) Effects of leucine on in vitroprotein synthesis and degradation in rat skeletal muscle. J. Nutr. 114: 1204–1212.

38. Gautsch, T. A., Anthony, J. C., Kimball, S. R., Paul, G. L., Layman, D. K.& Jefferson, L. S. (1998) Availability of eIF-4E regulates skeletal muscle proteinsynthesis during recovery from exercise. Am. J. Physiol. 274: C406–C414.

39. Anthony, J. C., Yoshizawa, F., Gautsch-Anthony, T., Vary, T. C., Jeffer-son, L. S. & Kimball, S. R. (2000) Leucine stimulates translation initiation inskeletal muscle of postabsorptive rats via a rapamycin-sensitive pathway. J. Nutr.130: 2413–2419.

40. Katz, J. & Tayek, J. A. (1998) Gluconeogenesis and the Cori cycle in12-, 20- and 40-h fasted humans. Am. J. Physiol. 38: E537–E542.

41. Balasubramanyam, A., McKay, S., Nadkarni, P., Rajan, A. S., Farza, A.,Pavlik, V., Herd, J. A., Jahoor, F. & Reeds, P. J. (1999) Ethnicity affects thepostprandial regulation of glycogenolysis. Am. J. Physiol. 40: E905–E914.

42. Hellerstein, M. K., Neese, R. A., Linfoot, P., Christiansen, M., Turner, S. &Letscher, A. (1997) Hepatic gluconeogenic fluxes and glycogen turnover dur-ing fasting in humans. J. Clin. Invest. 100: 1305–1319.

43. Radziuk, J. & Pye, S. (2001) Hepatic glucose uptake, gluconeogenesisand the regulation of glycogen synthesis. Diab. Metab. Res. Rev. 17: 250–272.

44. Krebs, M., Krssak, M., Bernroider, E., Anderwald, C., Brehm, A., Meyer-speer, M., Nowotny, P., Roth, E., Waldhausl, W. & Roden, M. (2002) Mecha-nism of amino acid-induced skeletal muscle insulin resistance in humans. Diabe-tes 51: 599–605.

45. Lynch, C. H., Hutson, S. M., Patson, B. J., Vaval, A. & Vary, T. C. (2002)Tissue-specific effects of chronic dietary leucine and norleucine supplementationon protein synthesis in rats. Am. J. Physiol. 283: E824–E835.

46. Sweeney, J. S. (1927) Dietary factors that influence the dextrosetolerance test. Arch. Internal Med. 40: 818–830.

47. Schwenk, W. F. & Haymond, M. W. (1987) Decreased uptake ofglucose by human forearm during infusion of leucine, isoleucine, or threonine.Diabetes. 36: 199–204.

48. Ferrannini, E., Bevilacqua, S., Lanzone, L., Bonadonna, R., Brandi, L.,Oleggini, M., Boni, C., Buzzigoli, G., Ciociaro, D., Luzi, L. & DeFronzo, R. A.(1988) Metabolic interactions of amino acids and glucose in healthy humans.Diab. Nutr. Metab. 3: 175–186.

49. Rossetti, L., Rothman, D. L., DeFronzo, R. A. & Shulman, G. I. (1989)Effect of dietary protein on in vivo insulin action and liver glycogen repletion.Am. J. Physiol. 257: E212–E219.

50. Chang, T. W. & Goldberg, A. L. (1978) Leucine inhibits oxidation ofglucose and pyruvate in skeletal muscle during fasting. J. Biol. Chem. 253:3696–3701.

51. Floyd, J. C., Fajans, S. S., Conn, J. W., Knopf, R. F. & Rull, J. (1966)Stimulation of insulin secretion by amino acids. J. Clin. Invest. 45: 1487–1502.

52. Floyd, J. C., Fajans, S. S., Conn, J. W., Knopf, R. F. & Rull, J. (1966)Insulin secretion in response to protein ingestion. J. Clin, Invest. 45: 1479–1486.

53. Nuttall, F. Q., Gannon, M. C., Wald, J. L. & Ahmed, M. (1985) Plasmaglucose and insulin profiles in normal subjects ingesting diets of varying carbo-hydrate, fat, and protein content. J. Am. Coll. Nutr. 4: 437–450.

54. Krezowski, P. A., Nuttall, F. Q., Gannon, M. C. & Bartosh, N. H. (1986)The effect of protein ingestion on the metabolic response to oral glucose innormal individuals. Am. J. Clin. Nutr. 44: 847–856.

55. Gannon, M. C., Nuttall, J. A., Gamberg, G., Gupta, V. & Nuttall, F. Q.(2001) Effect of protein ingestion on the glucose appearance rate in people withtype 2 diabetes. J. Clin. Endocrinol. Metab. 86: 1040–1047.

56. Morens, C., Gaudichon, C., Fromentin, G. Marsset-Baglieri, A., Bensaid,A., Larue-Achagiotis, C., Leungo, C. & Tome, D. (2001) Daily delivery of dietarynitrogen to the periphery is stable in rats adapted to increased protein intake.Am. J. Physiol. 281: E826–E836.

57. Reavens, G. M. (1993) Role of insulin resistance in human disease(Syndrome X): an expanded definition. Annu. Rev. Med. 44: 121–131.

58. Hansen, B. C. (1999) The metabolic syndrome X. Ann. N.Y. Acad. Sci.892: 1–24.

59. Jungas, R. L., Halperin, M. L. & Brosnan, J. T. (1992) Quantitativeanalysis of amino acid oxidation and related gluconeogensis in humans. Physiol.Rev. 72: 419–448.

60. Kim, D. H., Sarbassov, D. D., Ali, S. M., King, J. E., Latek, R. R., Erdju-ment-Bromage, H, Tepmst, P. & Sabatini, D. M. (2002) mTOR interacts withraptor to form a nutrient-sensitive complex that signals to the cell growth ma-chinery. Cell 110: 163–175.

61. Baum, J. I., Seyler, J. E., O’Conner, J. C., Freund, G. G. & Layman, D. K.(2003) The effect of leucine on glucose homeostasis and the insulin signalingpathway. FASEB J. 17: A811.

62. Harris, R. A., Kobayashi, R., Murakami, T. & Shimomura, Y. (2001)Regulation of branched-chain �-keto acid dehydrogenase kinase expression inrat liver. J. Nutr. 131: 841S–845S.


Human Nutrition and Metabolism - Qivana Products

Documents

Transcript of Human Nutrition and Metabolism - Qivana Products