Glycemic index and load matters more than carbohydrate level

[Guest guest]

Hi All,

The below speaks well for itself. The glycemic index can be low, carbohydrates

can

be relatively high, calorie intakes can be the same and the resulting levels of

blood triacylglycerols is much lower and the plasminogen activator inhibitor

(PAI-1)

levels can be improved. " Attention has been directed toward controlling

triacylglycerol and PAI-1 concentrations in light of the direct associations

between

these variables and cardiovascular events " .

See the corrupted pdf-available below.

Jennie Brand-

Optimizing the cardiovascular outcomes of weight loss

Am J Clin Nutr 2005 81: 949-950.

See corresponding article on page 976.

The past 2 y have seen a steady stream of reports indicating that restriction or

modification of carbohydrate intakes can favorably affect weight loss and

cardiovascular disease (CVD) risk factors (1–6). The article by Ebbeling et al

(7)

in this issue of the Journal represents one more in favor of diets with a low

glycemic index (GI) or glycemic load (GL). Serum triacylglycerol concentrations

in

young overweight adults with similar weight loss fell nearly twice as far with

the

ad libitum low-GL diet as with the energy-restricted low-fat diet, whereas

concentrations of plasminogen activator inhibitor 1, an important measure of

thrombogenicity, significantly worsened (ie, rose) in subjects who were

following

the energy-restricted low-fat diet. The study was small (n = 23) and from a

group at

Harvard that had published other studies on the same topic, but it was long-term

(12

mo) and carefully carried out.

That lowering daylong glycemia, with or without weight loss, might improve CVD

risk

should not come as a surprise. Many intervention studies have tested the

hypothesis

that low-GI diets will improve not only glucose control but also lipid

metabolism.

Twenty years ago, et al (8) showed in 3-way crossover studies that

low-GI

diets improve triacylglycerol and total cholesterol concentrations in

hyperlipidemic

subjects more than do conventional low-fat diets. Recently, Patel et al (9)

showed

that women with advanced CVD who were awaiting bypass surgery spent

significantly

fewer days in the hospital than did their counterparts who were following a

conventional low-fat diet (7.1 and 9.5 d, respectively). Slowing the rate of

carbohydrate absorption per se by using the -glucosidase inhibitor acarbose was

found to reduce cardiovascular events by 50% over 3 y in a large population with

impaired glucose tolerance (IGT).

Long-term studies in animals have provided additional evidence that the GI

itself,

and not fiber intake or any other confounding factor, is important in relation

to

weight gain, body fat, and CVD risk. Animals fed diets differing only in the

type of

starch (high- or low-GI) gained body fat faster with the high-GI diet than with

the

low-GI diet (10). Even when fed to similar body weight, high-GI diet–fed rats

have

more body fat (71%), less lean body mass, and higher plasma triacylglycerol

concentrations than do low-GI diet–fed rats (11).

Large-scale observational studies show links—even in nondiabetic persons—between

the

presence of postchallenge hyperglycemia and an increased risk of chronic disease

(12). In a meta-analysis of 39 prospective studies of nondiabetic cohorts,

Levitan

et al (13) found that groups with the highest 120-min postload glucose

concentration

had a 27% greater risk of CVD than did those with the lowest glucose

concentrations,

and the relative risk was higher in women than in men (1.56 and 1.23,

respectively).

Adjustment for traditional CVD risk factors attenuated but did not abolish the

relation. Moreover, Liu et al (14) showed that average dietary GI and GL were

also

independent predictors of 10-y prospective CVD risk in US women. The latter

study is

particularly important, because it implies that postprandial glycemia induced by

carbohydrate foods in everyday settings (and not glucose tolerance testing) is

clinically relevant.

There has been fundamental progress in showing that glucose itself can directly

damage vascular cells, by a variety of mechanisms. All of these mechanisms

appear to

reflect a single hyperglycemia-induced process of overproduction of superoxide

by

the mitochondrial electron-transport chain (15). Normal concentrations of

glycemia

such as those encountered during a standard meal have been shown to acutely

decrease

plasma antioxidant capacity, which reflects a significant oxidative stress.

Moreover, the vascular endothelium is a prime target because endothelial cells,

unlike many other cells in the body, are unable to regulate glucose transport

across

the cell membrane.

Taken together, intervention, observational, and experimental studies suggest

that

postprandial glycemia plays a greater role in CVD than is generally

acknowledged,

perhaps more so in women than in men. Because decreasing the intakes of total

and

saturated fat has been the goal of efforts to reduce the incidence of obesity

and

CVD, high-carbohydrate foods have been recommended, not so much because of their

intrinsic nutritional merit, but because they fill the calorie space formerly

occupied by fat. But one of the more subtle changes in the food supply over the

past

few decades has been the replacement of traditionally processed grains by more

highly processed, high-GI cereal products. Less-processed foods are more likely

to

contain slowly digested carbohydrates because the starches and sugars remain

closely

embedded in the plant's original botanical structure, surrounded by bran and

other

barriers that inhibit starch gelatinization. In contrast, modern methods of food

production using finer flours, extrusion technology, and high temperatures and

pressures increase starch gelatinization and thus the rate of digestion in vivo.

Compared with sugars, high-GI starchy foods receive little attention, and yet

they

have a greater capacity than do sugars to increase the glycemic and

insulinotropic

potency of the whole diet.

Because the overweight are now the majority in most industrialized nations, we

can

no longer afford to direct dietary guidelines to just the " healthy " population.

Moreover, we need efficacious guidelines that work in practice, not just in

theory.

During the past 2 decades, when low-fat diets and plenty of cereal foods were

actively promoted, health trends were the opposite of those we would wish. Along

with obesity, the diagnoses of type 2 diabetes and IGT have soared,

" maturity-onset "

diabetes is being diagnosed in children, and the metabolic or insulin resistance

syndrome affects 1 in 4 adults. Even normal-weight individuals can have the

metabolic syndrome and thus a higher risk of CVD. Diseases such as polycystic

ovarian syndrome, nonalcoholic steatohepatitis, and fatty liver, which have

their

roots in insulin resistance, have also reached alarming proportions.

It must now be clear that the conventional low-fat diet (with no consideration

of

the nature of the starch) is not the ideal diet for most of the population.

Dietary

Guidelines for Americans 2005 (16) sensibly gives greater emphasis to increased

consumption of whole grains rather than to refined grains. However, this is

unlikely

to improve daylong glycemia, because many so-called whole-grain breads and

breakfast

cereals produce as much postprandial glycemia as do their white-flour

counterparts

(17). Moreover, recommending whole-grain and high-fiber cereals is nothing

new—nutritionists have been doing that for at least 50 y. A high proportion of

the

population will dismiss outright any suggestion of eating whole grains or whole

meal. We urgently need nutrition messages that fire the imagination and

encourage

even unmotivated people to adopt effective dietary strategies that reduce the

risk

of chronic disease. In Australia and the United Kingdom, the GI has become a

popular

concept in its own right. The message that slowly digested carbohydrates can

" keep

you fuller for longer " is one that the general public, young and old,

intuitively

understands. Indeed, many people warm to a plan that helps keep blood sugar

concentrations " under control. " Furthermore, as Ebbeling et al (7) point out,

their

ad libitum low-GL diet is less extreme and restrictive than is either a

low-energy,

low-fat diet or a low-carbohydrate diet, and it still produces better outcomes.

It may be argued that the evidence for a role of GI or GL in weight management

and

CVD prevention is still insufficient to justify the place of either in nutrition

advice to the general public. We need to clarify whether reducing the GL of the

diet

by changing the type of carbohydrate (substituting low-GI sources of

carbohydrate

for high-GI sources) or by substituting protein or fat for carbohydrate [or a

combination of all 3 alternatives as Ebbeling et al (7) did] will have different

metabolic consequences. Nevertheless, we must also acknowledge the shortcomings

of

the conventional low-fat (ipso facto high-GI) diet currently advocated by public

health agencies and must be prepared to entertain the idea that the GI might be

a

useful and appealing concept after all.

ACKNOWLEDGMENTS

The author is a coauthor of The New Glucose Revolution book series (New York:

Marlowe and Co).

Cara B Ebbeling, M Leidig, B Sinclair, G Seger-Shippee,

Henry A

Feldman, and S Ludwig

Effects of an ad libitum low-glycemic load diet on cardiovascular disease risk

factors in obese young adults

Am J Clin Nutr 2005 81: 976-982.

Background: The optimal nutritional approach for the prevention of

cardiovascular

disease among obese persons remains a topic of intense controversy. Available

approaches range from conventional low-fat to very-low-carbohydrate diets.

Objective: The aim of this pilot study was to evaluate the efficacy of an ad

libitum

low-glycemic load diet, without strict limitation on carbohydrate intake, as an

alternative to a conventional low-fat diet.

Design: A randomized controlled trial compared 2 dietary treatments in obese

young

adults (n = 23) over 12 mo. The experimental treatment emphasized ad libitum

consumption of low-glycemic-index foods, with 45–50% of energy from

carbohydrates

and 30–35% from fat. The conventional treatment was restricted in energy

(250–500

kcal/d deficit) and fat (<30% of energy), with 55–60% of energy from

carbohydrate.

We compared changes in study outcomes by repeated-measures analysis of

log-transformed data and expressed the results as mean percentage change.

Results: Body weight decreased significantly over a 6-mo intensive intervention

in

both the experimental and conventional diet groups (–8.4% and –7.8%,

respectively)

and remained below baseline at 12 mo (–7.8% and –6.1%, respectively). The

experimental diet group showed a significantly greater mean decline in plasma

triacylglycerols than did the conventional diet group (–37.2% and –19.1%,

respectively; P = 0.005). Mean plasminogen activator inhibitor 1 concentrations

decreased (–39.0%) in the experimental diet group but increased (33.1%) in the

conventional diet group (P = 0.004). Changes in cholesterol concentrations,

blood

pressure, and insulin sensitivity did not differ significantly between the

groups.

Conclusion: An ad libitum low-glycemic load diet may be more efficacious than a

conventional, energy-restricted, low-fat diet in reducing cardiovascular disease

risk.

INTRODUCTION

The alarming prevalence of obesity and the associated risk of cardiovascular

disease

(CVD) have been well documented (1) and extensively publicized in the United

States.

As a result, millions of obese adults are following weight-loss diets. Recently,

Atkins-type very-low-carbohydrate diets have rapidly grown in popularity (2),

although low-fat diets remain the cornerstone of conventional treatment based on

clinical practice recommendations (3, 4). Whereas a few studies have suggested

that

carbohydrate-restricted diets may have significantly greater benefits than do

low-fat diets in reducing CVD risk (5, 6), there is widespread concern regarding

the

safety and long-term efficacy of severe carbohydrate restriction (7, 8).

A low-glycemic load (GL) diet, containing unrestricted amounts of carbohydrate

from

low-glycemic index (GI) foods, represents an alternative to low-fat diets on the

one

hand and to low- carbohydrate diets on the other. The GI is defined as the

incremental area under the blood glucose response curve after consumption of 50

g of

available carbohydrate from a test food, divided by the area under the curve

after

consumption of 50 g of carbohydrate from a reference food (ie, glucose or white

bread) (9). The GL is the arithmetic product of the amount of carbohydrate

consumed

and the GI (10) and thus describes the overall effects of both quantity and

source

of carbohydrate on postprandial glycemia (11). Risk of CVD has been inversely

associated with dietary GI or GL in some (12–15) but not all (16) epidemiologic

studies. Moreover, whereas several short-term intervention studies have

described

beneficial effects of low-GI diets on blood lipids in overweight adults (17–20)

and

on the capacity for fibrinolysis in diabetic patients (21, 22), the long-term

efficacy of low-GL diets in reducing CVD risk has not previously been evaluated

(23).

The aim of this pilot study was to evaluate the efficacy of an experimental ad

libitum low-GL diet. We hypothesized that the experimental diet would have a

more

beneficial effect on CVD risk factors than would a conventional,

energy-restricted,

low-fat diet over a 12-mo intervention.

SUBJECTS AND METHODS

Screening and enrollment

The protocol was approved by the institutional review board at Children's

Hospital

Boston, and written informed consent was obtained from each subject. Inclusion

criteria included: age between 18 and 35 y, body mass index (BMI; in kg/m2) >27,

body weight <136 kg (300 lb), and absence of major medical illness as assessed

by

physical examination and laboratory screening tests (ie, kidney and liver

enzymes,

thyrotropin, glycosylated hemoglobin, fasting plasma glucose, and urinalysis).

After

being screened for eligibility, 34 obese young adults (30 females and 4 males)

were

enrolled in the study. Of these, 22 females and 1 male completed assessments at

the

end of the 12-mo intervention (Figure 1).

Study design

Subjects were randomly assigned to the experimental (low-GL diet) or

conventional

(low-fat diet) treatment group between August 2001 and July 2002. The study

comprised a 6-mo intensive intervention (12 dietary counseling sessions) and a

6-mo

follow-up (2 dietary counseling sessions). The duration of each counseling

session

was 1 h. Study outcomes were measured at 0, 6, and 12 mo.

Experimental diet

The experimental diet prescription was not energy restricted. Rather, we used an

ad

libitum approach based on previous research that suggested greater satiety and

decreased voluntary energy intake among subjects consuming low-GL diets (24).

Subjects were counseled to consume carbohydrate-containing foods with a

relatively

low GI (eg, nonstarchy vegetables, fruit, legumes, nuts, and dairy products;

24), to

consume carbohydrate with protein and healthful fat at every meal and snack, and

to

eat to satiety and snack when hungry. A low-GL food pyramid provided the basis

for

nutrition education (25). The target macronutrient composition was 45–50% of

energy

from carbohydrate, 30–35% of energy from fat, and the remainder from protein.

Conventional diet

The conventional diet prescription was based on current recommendations for

weight

loss and CVD risk reduction, with emphasis on restricting energy intake by

reducing

dietary fat (3). Meal plans were based on an exchange system (26) designed to

elicit

an energy deficit of 250–500 kcal/d. Energy requirements were estimated by using

the

-Benedict equation (27), multiplied by 1.5 to account for physical

activity

and adjusted for baseline dietary intake. The American Diabetes Association's

diabetes food pyramid provided the basis for nutrition education (28). The

target

macronutrient composition was 55–60% of energy from carbohydrate, <30% of energy

from fat, and the remainder from protein.

Behavioral therapy and physical activity recommendations

Both groups received the same behavioral therapy and physical activity

recommendations. Behavioral therapy focused on enhancing self-efficacy for

lifestyle

change by using social cognitive theory as a conceptual framework (29).

ing

behavioral capability (ie, knowledge and skill) and self-control was the primary

objective during the dietary counseling sessions. Patient expectations (ie,

anticipated outcomes), expectancies (ie, values ascribed to outcomes), and

perceptions of environmental influences were among the topics of discussion. To

operationalize the self-control construct, the study dietitian encouraged

patients

to set goals around eating behaviors, to self-monitor goal attainment, and to

explore solutions to problems. Physical activity recommendations were consistent

with public health guidelines (30).

Process evaluation

The intervention process was evaluated on the basis of attendance at the dietary

counseling sessions and adherence to respective diet prescriptions. All subjects

received extensive instruction in keeping food diaries. Three-dimensional food

models, plates, bowls, glasses, and measuring cups and spoons were used to

educate

subjects regarding accurate appraisal of portion sizes. The diaries were

reviewed

with each subject at the time of receipt to provide clarification, as necessary,

on

recorded foods and beverages. FOOD PROCESSOR PLUS software (version 8.2; ESHA

Research, Salem, OR) was used to quantify intakes of fat, carbohydrate, protein,

and

fiber. The GI of individual carbohydrate-containing foods was assigned according

to

published values based on a glucose reference (31). Daily GL was calculated by

multiplying the total amount of dietary carbohydrate (in g) by the weighted GI

for

each food and then adjusted for energy intake:

To ensure that treatments were delivered according to established procedures,

the

study dietitian completed a tracking form and progress note immediately after

each

session. Seven-day food diaries were obtained at baseline (month 0), during the

intensive intervention period (3 and 6 mo), and at the end of follow-up (12 mo)

for

evaluation of process outcomes. In addition, patients were encouraged to keep

food

diaries throughout the intervention as a self-monitoring strategy. Patients were

not

given explicit information regarding the target contributions of each

macronutrient

to total energy intake. Rather, the study dietitian reviewed the diaries after

each

counseling session and provided practical advice, as necessary, to foster eating

behaviors consistent with the diet prescriptions. The project director met with

the

study dietitian on a regular basis to review food diaries, tracking forms, and

progress notes.

Assessment of study outcomes

Weight and height were assessed by using an electronic scale (model 6702;

Scale-Tronix, White Plains, NY) and a wall-mounted stadiometer (Holtain Limited,

Crymych, United Kingdom), respectively. Body composition was measured by

dual-energy

X-ray absorptiometry (DXA) with the use of Hologic instrumentation (model QDR

4500;

Hologic Inc, Bedford, MA). Blood pressure was determined by using an automated

system (Dinamapp, Tampa, FL) while the subject sat quietly. A blood sample was

drawn

by venipuncture after a 12-h overnight fast.

Laboratory analyses

Plasma lipid concentrations were measured in a laboratory certified by the

Centers

for Disease Control and Prevention–National Heart, Lung, and Blood Institute

Lipid

Standardization Program. Total cholesterol, HDL cholesterol, and

triacylglycerols

were measured by using a Hitachi 911 analyzer (Roche Diagnostics, Indianapolis,

IN),

and LDL cholesterol was measured by using a homogenous enzymatic assay (Genzyme

Corp, Cambridge, MA) (32). Plasma concentrations of plasminogen activator

inhibitor

1 (PAI-1) were measured by using an enzyme-linked immunosorbent assay (ELISA;

Diagnostica Stago, Parsippany, NJ). Plasma glucose and serum insulin

concentrations

were measured by using a Hitachi 917 analyzer (Roche Diagnostics) and an Elecsys

2010 system (Roche Diagnostics), respectively. With the use of glucose (mg/dL)

and

insulin (µU/mL) concentrations, we calculated the quantitative insulin

sensitivity

check index: 1/(insulin + log glucose) (33).

Statistical analysis

We analyzed dietary data and study outcomes by repeated-measures analysis of

variance. We tested each variable for change over time (0, 6, and 12 mo) and for

a

difference in time course between the 2 groups (experimental and conventional

diet)

by assessing the main effect of time and the group x time interaction,

respectively.

To avoid the increased risk of type I inferential error from multiple

comparisons,

we limited statistical testing of time trends to the overall change (from 0 to 6

to

12 mo), with the exception of 2 planned comparisons for process measures, ie, GL

and

dietary fat. We accounted for within-subject correlation by using a banded

covariance structure, which allowed a lower correlation between the 0- and 12-mo

observations than between the 0- and 6-mo or the 6- and 12-mo observations.

Statistical significance was defined as P < 0.05.

The primary analysis included data from only the 23 subjects who completed the

study. Secondary analyses included available data from all 34 randomly assigned

subjects. Study outcomes were log transformed for analysis, and results are

expressed as percentage change. Dietary data were analyzed without

transformation.

We used SAS software (release 9.0; SAS Institute Inc, Cary, NC) for all

computations.

RESULTS

Subjects

Baseline data for the subjects who completed the study (n = 23; 67.6% of those

randomly assigned to a treatment group) are presented in Table 1. There were no

significant differences in baseline measures between diet groups. The male who

completed the study was in the conventional diet group.

Process data

Attendance at the 14 dietary counseling sessions approximated 100% for the 23

subjects who completed the study; 2 subjects missed just one visit each.

Nutrient

intake data derived from the food diaries are presented in Table 2. At baseline,

we

found no significant differences between groups with respect to the nutrients of

interest. GL decreased significantly in the experimental diet group (0–6 mo, P <

0.001; 0–12 mo, P < 0.001) and did not change in the conventional diet group.

Dietary fat decreased significantly in the conventional diet group (0–6 mo, P <

0.001; 0–12 mo, P = 0.004) and increased nonsignificantly in the experimental

diet

group.

TABLE 2 Dietary data

---------------------------------------------------------------

------------------------------------------------------Group----P1

---------------------------------------------------------------

Experimental diet (n = 11) Conventional diet (n = 12)----Group Time Group x time

interaction

--------------------------------------------------------------

Glycemic load (g/1000 kcal) <0.001 <0.001 <0.001

Baseline3 77.2 ± 5.72 77.8 ± 2.2

Interim4 54.4 ± 2.0 78.4 ± 1.4

12 mo 53.0 ± 2.7 77.1 ± 2.4

Glycemic index 0.006 <0.001 0.004

Baseline 56.2 ± 1.2 56.6 ± 1.0

Interim 46.2 ± 1.6 52.8 ± 0.9

12 mo 46.3 ± 2.0 52.9 ± 1.1

Carbohydrate (% of energy) <0.001 0.43 <0.001

Baseline 52.7 ± 2.2 54.8 ± 1.4

Interim 47.2 ± 1.6 59.4 ± 0.8

12 mo 45.5 ± 1.1 58.3 ± 1.9

Total fat (% of energy) <0.001 0.03 0.006

Baseline 32.6 ± 1.6 30.0 ± 1.1

Interim 33.0 ± 1.2 23.4 ± 0.9

12 mo 35.4 ± 1.2 24.3 ± 2.0

Saturated fat (% of energy) 0.04 <0.001 0.18

Baseline 11.3 ± 0.8 10.7 ± 0.7

Interim 9.1 ± 0.7 7.5 ± 0.3

12 mo 10.6 ± 0.9 7.6 ± 0.7

Protein (% of energy) 0.17 <0.001 0.07

Baseline 15.7 ± 1.0 16.1 ± 0.9

Interim 21.1 ± 1.1 18.7 ± 0.4

12 mo 20.5 ± 0.9 18.1 ± 1.0

Fiber (g/1000 kcal) 0.22 <0.001 0.53

Baseline 9.6 ± 1.0 8.2 ± 0.5

Interim 14.9 ± 1.3 12.6 ± 1.1

12 mo 13.5 ± 1.1 12.8 ± 1.2

Energy (kcal) 0.84 <0.001 0.76

Baseline 1860 ± 72 1802 ± 116

Interim 1391 ± 79 1409 ± 46

12 mo 1494 ± 82 1472 ± 85

----------------------------------------

1 Testing for overall difference in level between experimental and conventional

diet

groups (main effect of group), change over time (main effect of time), and

difference in time course between groups (group x time interaction).

Repeated-measures ANOVA was used to account for within-subject correlations.

2 ± SEM (all such values).

3 Testing for equal means in experimental and conventional diet groups by

independent-sample t test found no significant baseline differences for any of

the

listed nutrients.

4 Means of data collected at 3 and 6 mo.

Outcomes

Study outcomes are presented in Table 3. Body weight decreased significantly

over

the 6-mo intensive intervention in the experimental and conventional diet

groups,

and it remained below baseline at 12 mo. Mean weight loss did not differ

significantly between the groups, and there were no significant differences

between

the experimental and conventional diet groups in the mean percentage change in

fat

mass (–16.5 compared with –15.7; P = 0.97) and lean mass (–1.1 compared with

–1.5; P

= 0.92). Nevertheless, the experimental diet group showed greater mean declines

in

plasma triacylglycerols. Mean changes in plasma PAI-1 concentrations also

differed

between the groups, decreasing in the experimental diet group and increasing in

the

conventional diet group. Decreases in total cholesterol and increases in HDL

cholesterol were marginally nonsignificant and did not differ significantly

between

groups. There were no significant changes in LDL cholesterol or blood pressure

in

either group throughout the study. The insulin sensitivity index increased

significantly in both groups. Results were materially unchanged in the secondary

analyses that included data from all randomly assigned subjects (data not

presented).

TABLE 3 Study outcomes1

---------------------------------------------------------------

------------------------------------------------------Group----P1

---------------------------------------------------------------

Experimental diet (n = 11) Conventional diet (n = 12)----Group Time Group x time

interaction

--------------------------------------------------------------

Weight 0.18 <0.001 0.89

Interim3 –8.4 (–11.4, –5.3) –7.8 (–10.7, –4.9)

12 mo –7.8 (–13.0, –2.2) –6.1 (–11.2, –0.7)

Total cholesterol 0.90 0.06 0.22

Interim –9.9 (–16.7, –2.5) –2.1 (–9.2, 5.5)

12 mo –8.5 (–17.4, 1.5) –6.2 (–15.0, 3.5)

LDL cholesterol 0.85 0.17 0.59

Interim –9.1 (–18.6, 1.4) –2.6 (–12.3, 8.2)

12 mo –9.7 (–21.6, 3.9) –7.4 (–19.1, 6.0)

HDL cholesterol 0.41 0.08 0.20

Interim 2.3 (–6.0, 11.3) –0.3 (–8.1, 8.2)

12 mo 12.2 (2.9, 22.3) 1.1 (–6.9, 9.8)

Triacylglycerols 0.96 <0.001 0.005

Interim –35.4 (–44.6, –24.7) –7.1 (–19.8, 7.6)

12 mo –37.2 (–47.7, –24.5) –19.1 (–32.2, –3.6)

PAI-1 0.78 0.11 0.004

Interim –58.3 (–74.7, –31.3) 30.4 (–19.2, 110.4)

12 mo –39.0 (–70.2, 24.9) 33.1 (–32.9, 164.3)

Systolic blood pressure 0.78 0.81 0.99

Interim –0.9 (–5.9, 4.2) –0.5 (–5.3, 4.4)

12 mo 0.2 (–4.7, 5.3) 0.6 (–4.1, 5.5)

Diastolic blood pressure 0.84 0.72 0.82

Interim –2.0 (–7.2, 3.4) 0.3 (–4.8, 5.6)

12 mo –0.3 (–6.2, 6.0) 1.4 (–4.4, 7.6)

Insulin sensitivity index 0.32 <0.001 0.94

Interim 6.4 (1.5, 11.5) 5.8 (1.1, 10.7)

12 mo 10.4 (3.6, 17.6) 8.7 (2.3, 15.5)

1 Mean change in log-transformed variable at 6 and 12 mo (, retransformed to

percentage change [100% x (exp( –1)], with 95% confidence limits.

Repeated-measures ANOVA was used to account for within-subject correlations.

2 Testing for overall difference in level between experimental and conventional

groups (main effect of group), change over time (main effect of time, 2 df), and

difference in time course between groups (group x time interaction, 2 df).

3 Data collected at 6 mo.

DISCUSSION

In light of widespread concern regarding the high toll of the obesity epidemic

on

human suffering (34) and health care costs (35), development of effective

weight-management strategies is a public health priority (36). Debate about the

appropriate diets for promoting weight loss and decreasing CVD risk has focused

largely on the metabolic effect of dietary carbohydrate and fat (37–41). Obesity

has

become increasingly prevalent over the last 2 decades (42), and the contribution

of

carbohydrate to total energy intake has increased in tandem with a reduction in

the

contribution of fat (43). The increase in carbohydrate intake can be largely

attributed to consumption of high-GI foods (44). Taken together, these

observations

suggest that both the quantity and the source of carbohydrate may be important

considerations.

To our knowledge, the randomized controlled trial presented herein is the first

long-term study comparing a low-GL diet, with emphasis on consumption of low-GI

sources of carbohydrate, with a low-fat diet for decreasing CVD risk in obese

young

adults. We hypothesized that less hunger or greater satiety in response to an ad

libitum low-GL diet may facilitate a decrease in energy intake (45), without the

need for externally imposed energy restriction. Our hypothesis is supported, in

that

mean weight loss among persons following the ad libitum low-GL diet and mean

weight

loss among persons following the energy-restricted low-fat diet did not differ

significantly during the intensive 6-mo intervention (–8.4% and –7.8%,

respectively), and there was no significant weight rebound during the follow-up.

Moreover, we previously observed greater decreases in BMI and fat mass among

adolescents in response to a low-GL diet than in response to a low-fat diet

(46).

Differences in the response to an energy-restricted diet between adolescents and

adults, particularly women, may partially explain the varied patterns of weight

loss

between our 2 studies. Adolescents have a strong desire for autonomy and seem to

resist the use of an exchange system that imposes energy restriction; for this

reason, the flexibility of an ad libitum approach may be especially beneficial

in

this age group. In contrast, many young women are accustomed to following

conventional energy-restricted diets, which may limit the likelihood of seeing a

group effect over a 12-mo period in this patient population. Additional studies

are

needed to examine group effects with longer-term follow-up. Nevertheless, our

findings compare favorably with those of studies evaluating the effect of severe

carbohydrate restriction on weight loss (5, 6). et al (5) observed

decreases

in body weight of 4.4% and 2.5% at 12 mo among patients prescribed

carbohydrate-restricted and low-fat diets, respectively. In a similar 12-mo

study,

Stern et al (6) observed decreases of 3.5% and 2.4%.

A low-GL diet, such as that used in the present study, may represent an optimal

compromise between low-fat diets at one end of the spectrum and

carbohydrate-restricted diets at the other. Although changes in body weight did

not

differ between the 2 groups, the metabolic benefits of a low-GL diet in

decreasing

CVD risk may be significantly greater than those achieved with either of the

more

restrictive approaches. The low-fat diet had a significantly less favorable

effect

on circulating triacylglycerol and PAI-1 concentrations than did the low-GL

diet.

Indeed, low-fat diets typically have a high carbohydrate content, which causes

postprandial hyperglycemia and hyperinsulinemia (47). In turn, these episodes

may

enhance hepatic triacylglycerol production or reduce peripheral clearance (39,

48)

and also promote the synthesis and secretion of PAI-1 (49) via plausible

physiologic

and molecular mechanisms. Attention has been directed toward controlling

triacylglycerol and PAI-1 concentrations in light of the direct associations

between

these variables and cardiovascular events (50, 51). Whereas

very-low-carbohydrate

diets have beneficial effects on triacylglycerol concentrations (perhaps as a

result

of their low GL), the sustainability of such highly restrictive diets over the

long

term is questionable (5, 6). A low-GL diet,containing moderate amounts of

carbohydrate and fat, offers a potentially more flexible approach. In contrast

to

very-low-carbohydrate diets, the reduction in GL in the experimental group was

achieved by a relatively small decrease in carbohydrate intake that was

accompanied

by a substantial reduction in GI. Nevertheless, the mean decrease in

triacylglycerol

concentration in the experimental group over 12 mo (37.2%) compares favorably

with

decreases of 17.0% (5) and 28.6% (6) in previous studies of

very-low-carbohydrate

diets. Data from metabolic studies, epidemiologic investigations, and clinical

trials lend support to the efficacy of eating patterns that are consistent with

a

low-GL diet (38), including consumption of vegetables, fruit, and whole grains

as

primary sources of carbohydrate. Moreover, our findings extend data from

previous

short-term studies showing beneficial reductions in triacylglycerol

concentrations

with low-GI diets (52–55).

Several issues pertaining to study design warrant consideration. Strengths of

the

study include the use of treatments of equal intensity in both experimental and

control diet groups, which would eliminate this factor as a source of

confounding;

excellent attendance at counseling sessions among those who completed the study;

a

longer follow-up than in previous studies of GL or GI and CVD risk factors (17,

18,

20–22); and careful attention to process evaluation. Moreover, changes in

dietary

fiber, a frequently cited confounder in evaluations of GI or GL (56, 57), were

similar between groups. Limitations include the self-reporting of dietary

intakes

for process evaluation, reliance on published GI values for calculating dietary

GL

(31), and a small, predominately female sample from which there was some

attrition.

Underreporting of dietary intake is a well-recognized phenomenon in all

outpatient

studies aiming to assess the effects of diet composition on weight loss,

although

adjustment of other dietary variables for energy intake may partially correct

for

underreporting. When calculating GL from self-report data, we relied on

published GI

values (31), many which were derived from studies conducted in countries where

foods

may differ from those consumed in the United States. An attrition rate of 32.4%,

although considered problematic in terms of drawing unbiased conclusions (58),

is

similar to rates observed in previous long-term dietary intervention studies (5,

6).

In conclusion, a low-GL diet containing moderate amounts of carbohydrate from

low-GI

sources may be more efficacious than a conventional low-fat diet in reducing CVD

risk. The greater benefits in response to an ad libitum diet, compared with an

energy-restricted diet, are particularly noteworthy. This pilot study provides a

rationale for conducting long-term, larger-scale studies comparing the effects

of

low-GL, low-fat, and very-low-carbohydrate diets on CVD risk among obese

persons.

Al Pater, PhD; email: old542000@...

__________________________________

Mobile

Take with you! Check email on your mobile phone.

http://mobile./learn/mail