[Skip to Navigation]
Sign In
August 2003

A Reduced–Glycemic Load Diet in the Treatment of Adolescent Obesity

Author Affiliations

From the Division of Endocrinology (Drs Ebbeling and Ludwig, Mr Leidig, and Ms Sinclair), Department of Medicine (Drs Ebbeling and Ludwig, Mr Leidig, and Mss Sinclair and Hangen), and the Optimal Weight for Life Program (Mss Sinclair and Hangen and Dr Ludwig), Children's Hospital Boston, and the Department of Pediatrics (Drs Ebbeling and Ludwig), Harvard Medical School, Boston, Mass.

Arch Pediatr Adolesc Med. 2003;157(8):773-779. doi:10.1001/archpedi.157.8.773

Background  The incidence of type 2 diabetes increases markedly for obese children after puberty. However, the effect of dietary composition on body weight and diabetes risk factors has not been studied in adolescents.

Objective  To compare the effects of an ad libitum, reduced–glycemic load (GL) diet with those of an energy-restricted, reduced-fat diet in obese adolescents.

Design  Randomized control trial consisting of a 6-month intervention and a 6-month follow-up.

Main Outcome Measures  Body composition (body mass index [BMI; calculated as weight in kilograms divided by the square of height in meters] and fat mass) and insulin resistance (homeostasis model assessment) were measured at 0, 6, and 12 months. Seven-day food diaries were used as a process measure.

Subjects  Sixteen obese adolescents aged 13 to 21 years.

Intervention  Experimental (reduced-GL) treatment emphasized selection of foods characterized by a low to moderate glycemic index, with 45% to 50% of energy from carbohydrates and 30% to 35% from fat. In contrast, conventional (reduced-fat) treatment emphasized selection of low-fat products, with 55% to 60% of energy from carbohydrates and 25% to 30% from fat.

Results  Fourteen subjects completed the study (7 per group). The GL decreased significantly in the experimental group, and dietary fat decreased significantly in the conventional group (P<.05 for both). At 12 months, mean ± SEM BMI (−1.3 ± 0.7 vs 0.7 ± 0.5; P = .02) and fat mass (−3.0 ± 1.6 vs 1.8 ± 1.0 kg; P = .01) had decreased more in the experimental compared with the conventional group, differences that were materially unchanged in an intention-to-treat model (n = 16) (BMI, P = .02; fat mass, P = .01). Insulin resistance as measured by means of homeostasis model assessment increased less in the experimental group during the intervention period (−0.4 ± 0.9 vs 2.6 ± 1.2; P = .02). In post hoc analyses, GL was a significant predictor of treatment response among both groups (R2 = 0.51; P = .006), whereas dietary fat was not (R2 = 0.14; P = .22).

Conclusions  An ad libitum reduced-GL diet appears to be a promising alternative to a conventional diet in obese adolescents. Large-scale randomized controlled trials are needed to further evaluate the effectiveness of reduced-GL and –glycemic index diets in the treatment of obesity and prevention of type 2 diabetes.

TYPE 2 DIABETES mellitus, previously characterized as an adult-onset disease, now accounts for as many as 33% to 45% of new diabetes cases among adolescents.1 Most adolescents with a diagnosis of type 2 diabetes mellitus are obese,2,3 and increased incidence of the disease has been largely attributed to the pediatric obesity epidemic.4 Although there is widespread agreement that dietary intervention is an important component of treatment for obesity, optimal prescriptions for lasting weight loss remain elusive.

Although many studies have evaluated obesity treatments in children,5,6 relatively little attention has been directed toward adolescents. Treatments evaluated in a few adolescent studies have incorporated a reduced-fat diet in combination with varying exercise prescriptions and behavioral strategies.7-10 However, these studies were not designed to evaluate the independent effects of diet, precluding conclusions regarding the influence of diet composition on treatment outcomes.

Recently, the novel dietary factors glycemic index (GI) and glycemic load (GL) have been linked to the risk for obesity or type 2 diabetes mellitus on experimental and theoretical grounds.11,12 The GI is defined as the incremental area under the glucose response curve after consuming 50 g of available carbohydrates from a test food, divided by the area under the curve after consuming 50 g of carbohydrates from a reference food, usually white bread or glucose.13 The GL of a food, meal, or diet is calculated by multiplying the GI by the amount of carbohydrates consumed.11,14

One-day feeding studies have consistently reported increased hunger or voluntary food intake in subjects after eating highcompared with low-GI foods.15 Consumption of high-GI and -GL meals appears to elicit a sequence of hormonal events that limit availability of metabolic fuels in the late postprandial period, leading to excessive hunger and overeating.16 Short-term treatment studies have described beneficial effects of low-GI or -GL diets on body weight and composition. Slabber et al17 reported greater weight loss after 3 months among obese women who were counseled to eat low-GI foods compared with those who did not receive this advice. Bouche et al18 found that fat mass decreased significantly more in overweight men after 5 weeks of a lowcompared with a high-GI diet. Spieth et al19 retrospectively observed a greater reduction in body mass index (BMI; calculated as weight in kilograms divided by the square of height in meters) with a reduced-GL, compared with a reduced-fat diet prescription in children and adolescents who were followed up for an average of 4 months. In addition, 2 large-scale, prospective observational studies found a direct association between GL and risk for type 2 diabetes mellitus, after controlling for the effects of BMI and other potential confounders.11,20 However, the long-term efficacy of reduced-GL diets in the treatment of obesity and reduction of diabetes risk factors has not been previously evaluated.

The aims of this work were (1) to develop a reduced-GL diet for use in an adolescent population; (2) to determine whether adolescents following this diet will successfully achieve long-term reduction of GL; and (3) to compare the long-term effects of a reduced-GL diet with those of a conventional reduced-fat diet in a pilot study involving obese adolescents. Development and testing of novel dietary treatment for obesity is arguably of major public health significance, in view of the continuing debate regarding the efficacy of conventional reduced-fat diets for weight control.21,22


Overview of study design

Subjects were randomly assigned to experimental (reduced GL) or conventional (reduced fat) dietary treatment. Both groups received similar behavioral therapy, physical activity recommendations, and treatment intensity. The study, conducted in the General Clinical Research Center of Children's Hospital Boston, Boston, Mass, consisted of a 6-month intensive intervention (12 dietary counseling sessions) and a 6-month follow-up (2 dietary counseling sessions). Body composition and insulin resistance were measured at 0, 6, and 12 months.


The protocol was approved by the Institutional Review Board at Children's Hospital Boston. Written informed consent was obtained from subjects 18 years or older and from the parents of minors; assent was obtained from minors. After screening for study eligibility, 16 obese patients (5 male and 11 female patients; 13 white and 3 nonwhite) aged 13 to 21 years were enrolled and randomized between December 1, 2000, and September 30, 2001 (Figure 1). Obesity was defined as a BMI that exceeded sex- and age-specific 95th percentiles.23 All subjects were free of major medical illness, as assessed by means of physical examination and screening laboratory tests (measurement of kidney and liver enzymes, thyrotropin, glycosylated hemoglobin, and fasting plasma glucose levels and urinalysis).

Figure 1. 
Trial design. GL indicates glycemic load.

Trial design. GL indicates glycemic load.

Diet prescriptions

The reduced-GL prescription emphasized selection of carbohydrate-containing foods (eg, nonstarchy vegetables, fruits, legumes, nuts, and dairy) that are characterized by a low to moderate GI.14 Patients were instructed to balance consumption of carbohydrates with protein and fat at every meal and snack. The prescription was not energy restricted. Rather, subjects were advised to eat to satiety and to snack when hungry. The targeted proportion of energy from carbohydrate and fat were 45% to 50% and 30% to 35%, respectively, with the remainder from protein. An ad libitum approach was used in light of preliminary evidence that suggests greater satiety and decreased voluntary energy intake occurs among children and adolescents consuming reduced-GL diets.24

The reduced-fat prescription was based on current recommendations for weight loss and diabetes prevention,25,26 with emphasis on limiting dietary fat intake and increasing the intake of grains, vegetables, and fruits. Meal plans were designed to elicit a negative energy balance of 250 to 500 kcal/d. Energy requirements were estimated using the Harris-Benedict equation,27 with an activity factor of 1.5 and consideration of baseline dietary intake. Subjects were counseled to obtain 55% to 60% of energy from carbohydrates, 25% to 30% from fat, and the remainder from protein.

Nutrition education and behavioral therapy

Social cognitive theory provided a conceptual framework for the educational and behavioral components of treatment that was consistent between intervention groups. Counseling focused on enhancing self-efficacy for dietary change using the concepts of behavioral capability (knowledge and skill) and self-control.28 Patient expectations (anticipated outcomes), expectancies (values ascribed to outcomes), and perceptions of environmental influences were discussed during treatment sessions. Written materials included topic modules, food choice lists, and a select-a-meal menu. The topic modules were the primary mechanism for presenting nutrition intervention messages and facilitating self-assessment, goal setting, and problem solving. These modules were designed to promote dialogue between the patient and study dietitian. Food choice lists were used to enhance practical application of intervention messages presented in the topic modules. For the experimental group, lists corresponded to food groups delineated by a reduced-GL food pyramid.15 In contrast, for the conventional group, the lists corresponded to the diabetes food pyramid and were presented as an exchange system.29,30 The select-a-meal menu contained recipes and ideas for meal planning to complement the food choice lists. One topic module was devoted to physical activity, with subjects in both groups receiving information based on current recommendations.31,32 Nutrition bars were offered to subjects in the experimental (Balance Bar; kindly provided by Kraft Foods, Inc, Northfield, Ill) and conventional groups (Nature Valley Granola Bar; General Mills, Inc, Minneapolis, Minn) for occasional use as snacks.

Process evaluation

The treatment process was evaluated from the perspectives of interventionist adherence to nutrition education and counseling protocols, and subject participation and adherence to diet prescriptions.

Interventionist adherence may be conceptualized as treatment fidelity, a term encompassing integrity and differentiation.33 Integrity is the degree to which treatment is implemented according to established procedures, and differentiation is the extent to which interventions are distinct from one another. Several strategies were used to maximize treatment fidelity in the present study. First, written materials for the experimental and conventional groups were developed specifically for the pilot study such that the format and quality were completely parallel. The materials differed only with regard to the specific intervention messages pertaining to GL and dietary fat. Second, the interventionist was a dietitian (M.M.L.) trained in the science underlying each intervention and given precise instructions regarding use of written materials to ensure differentiation of the intervention messages. Third, to prevent shifts in implementation, the dietitian completed a tracking form and progress note immediately after each session. The project director (C.B.E.) met with the study dietitian on a regular basis to review tracking forms and progress notes and to discuss treatment of individual subjects.

Subject participation was evaluated on the basis of session attendance, and adherence was assessed by means of self-report of dietary intake. All subjects received extensive instruction and practice in keeping food diaries. Three-dimensional food models, plates, bowls, glasses, and measuring cups and spoons were used to educate them regarding accurate appraisal of portion sizes. The diaries were reviewed with the subject at the time of receipt to obtain clarification, as necessary, on recorded foods and beverages. Food Processor Plus software (Version 7.9; ESHA Research, Salem, Ore) was used to quantify intakes of fat, carbohydrate, protein, and fiber. The GI of individual foods was assigned according to published values.14 We calculated the GL by multiplying the total gram amount of dietary carbohydrate by the weighted GI for each food and then adjusted for energy intake,11,14,34 using the following equations: weighted GI = Σ(GI for food item × proportion of total carbohydrate contributed by item), and GL = (weighted GI × grams of carbohydrate)/1000 kcal.

Data from 7-day food diaries were used to evaluate process outcomes at baseline, during the intervention period (months 3 and 6), and at the end of the follow-up (month 12). In addition, self-monitoring of food intake was encouraged throughout the intervention period to enhance self-control and facilitate problem solving. The study dietitian reviewed self-reported intake after each individual counseling session to identify deviations from diet prescriptions and made corrective recommendations to the subject when necessary.

Assessment of study outcomes

Total body mass and fat mass were measured by dual-energy x-ray absorptiometry using Hologic instrumentation (Model QDR 4500; Hologic, Inc, Bedford, Mass). Height was measured using a wall-mounted stadiometer (Holtain Limited, Crymych, Wales).35 A blood sample was drawn by means of venipuncture after a 12-hour overnight fast. Plasma glucose level was measured using a Hitachi analyzer (Model 917; Roche Diagnostics, Indianapolis, Ind), and serum insulin level was measured using an Elecsys system (Model 2010; Roche Diagnostics). We used homeostasis model assessment (HOMA) to estimate insulin resistance by means of the following equation: (fasting serum insulin level [in microunits per milliliter] × fasting plasma glucose level [in millimoles per liter])/22.5.36 This index has been used previously for assessing insulin resistance in youth, with higher values indicating lower insulin sensitivity.37-39

Statistical methods

We conducted statistical analyses using SAS software (Release 8.2; SAS Institute Inc, Cary, NC). Repeated-measures analysis of variance was performed using the mixed linear model procedure. Component contrasts were estimated from the fitted model for preplanned comparisons (within treatment group changes over time and differences between groups for changes over time). The time intervals of interest were 0 to 6 months and 0 to 12 months. Data from all subjects who completed the study (n = 14) were included, irrespective of their level of adherence. Change in insulin resistance was adjusted for change in BMI. Secondary intention-to-treat analyses of body composition end points were performed for all randomized subjects (n = 16), assuming no change from baseline for the 2 subjects who were lost to follow-up (1 refused to return; 1 began medication known to affect body mass, an exclusionary criterion). Results are presented as mean ± SEM.

We conducted simple linear regression (using the general linear models procedure), pooling data from both groups, to explore whether changes in dietary GL or fat intake (average of values obtained at 3 and 6 months) independently predicted change in body fat. When exploring relationships between dietary variables and body fat, we eliminated 1 outlier from regression analyses, although data from this outlier are displayed. The outlier was a young woman who matriculated in medical school during participation in the study. On the basis of previous research, we speculated that apparent underestimation of dietary intake by this subject may be attributed, in part, to expression of a social desirability bias.40


Process data and baseline characteristics

Fourteen subjects finished the study (7 per group), yielding a completion rate of 87.5%. There were no group differences in session attendance, with subjects in the experimental and conventional groups completing 9.4 ± 0.6 of the 12 planned sessions during the intervention period. All 14 subjects attended the 2 scheduled sessions during the follow-up.

At baseline, we found no differences between the experimental and conventional groups for age (16.9 ± 1.3 vs 15.3 ± 0.9 years; P = .33), body mass (103.5 ± 6.0 vs 104.7 ± 4.8 kg; P = .88), height (171.6 ± 3.3 vs 167.9 ± 2.1 cm; P = .37), BMI (34.9 ± 1.0 vs 37.1 ± 1.2; P = .21), and HOMA estimation of insulin resistance (3.5 ± 0.7 vs 4.3 ± 0.7; P = .43). However, fat mass was lower for the experimental group compared with the conventional group (38.8 ± 2.6 vs 48.5 ± 3.0 kg; P = .03).

Changes in dietary variables are presented in the Table 1. The GL decreased for the experimental group (0-6 months, P = .005; 0-12 months, P = .007) and did not change significantly for the conventional group. Dietary fat decreased for the conventional group (0-6 months, P = .01; 0-12 months, P = .03) and tended to increase for the experimental group (0-6 months, P = .06). Carbohydrate intake decreased for the experimental group (0-6 months, P = .03; 0-12 months, P = .07).

Study outcomes

As shown in Figure 2, BMI (P = .03) and fat mass (P = .02) decreased in the experimental group from 0 to 12 months, and neither outcome changed significantly in the conventional group. At 12 months, BMI (−1.3 ± 0.7 vs 0.7 ± 0.5; P = .02) and fat mass (−3.0 ± 1.6 vs 1.8 ± 1.0 kg; P = .01) had decreased significantly more in the experimental compared with the conventional group. These results were materially unchanged in intention-to-treat analyses (BMI, −1.2 ± 0.7 vs 0.6 ± 0.5 [P = .02]; fat mass, −2.6 ± 1.5 vs 1.6 ± 0.9 kg [P = .01]). Of interest, there was no weight regain between 6 and 12 months for the experimental group. Insulin resistance, as assessed by means of HOMA, increased significantly less in the experimental compared with the conventional group during the intervention period (−0.4 ± 0.9 vs 2.6 ± 1.2; P = .02), and statistical adjustment for BMI did not materially alter this result (P = .03).

Figure 2. 
Changes over time for study outcomes. BMI indicates body mass index (calculated as weight in kilograms divided by the square of height in meters); black circles, means for the experimental group; white circles, means for the conventional group; limit lines, SEM; and asterisk, P<.05, significant change from baseline.

Changes over time for study outcomes. BMI indicates body mass index (calculated as weight in kilograms divided by the square of height in meters); black circles, means for the experimental group; white circles, means for the conventional group; limit lines, SEM; and asterisk, P<.05, significant change from baseline.

Dietary variables as predictors of change in body fat

Figure 3 depicts results of bivariate linear regression analysis, using dietary GL or fat intake during the intervention period as the independent variable and change in body fat from 0 to 6 months as the dependent variable. Change in GL was a strong predictor of this study outcome, explaining about half of the variance in both groups combined (R2 = 0.51; P = .006). In contrast, change in dietary fat was not significantly associated with change in body fat (R2 = 0.14; P = .22).

Figure 3. 
Changes in dietary glycemic load (GL) or fat intake as predictors of change in body fat. One outlier (indicated by the circled data point) was excluded when calculating regression lines (solid) and 95% confidence intervals (dashed). Black circles indicate subjects in the experimental group; white circles, subjects in the conventional group.

Changes in dietary glycemic load (GL) or fat intake as predictors of change in body fat. One outlier (indicated by the circled data point) was excluded when calculating regression lines (solid) and 95% confidence intervals (dashed). Black circles indicate subjects in the experimental group; white circles, subjects in the conventional group.


Treatment of obesity in adolescents is characterized by modest weight loss and substantial relapse.5 Virtually no studies in this age range have sought to examine the influence of dietary composition per se on body weight. In this pilot study, we investigated the independent effects of an experimental reduced-GL diet vs a conventional reduced-fat diet, using similar behavioral strategies, physical activity prescriptions, and treatment intensity in both groups. Although both groups showed the intended changes in targeted dietary factors, measures of adiposity decreased significantly more in the reduced-GL group. This result is of particular interest, in that the reduced-GL diet was prescribed in an ad libitum fashion, whereas the reduced-fat diet was energy restricted, consistent with conventional practices.5

Subjects receiving the reduced-GL diet may have felt less hunger and/or greater satiety, facilitating reduced energy intake and allowing for weight loss to occur without an externally imposed energy restriction. The potential flexibility of such a diet may have particular behavioral benefits for adolescents who have a strong desire for autonomy.41,42 During focus groups, teenagers mentioned not wanting to be told what to do as a barrier to participating in family meals.41 The lack of a strict energy restriction may help to prevent tension between adolescents who are ambivalent regarding dietary change and adults who want to support change. In a previous study, children with type 1 diabetes mellitus were more easily able to select their own foods when prescribed a low-GI diet compared with a regimented meal plan based on an exchange system without consideration for GI.43 They also achieved better glycemic control, while experiencing less family conflict, with the low-GI diet. At the end of the study, children expressed an overall preference for the low-GI diet compared with the regimented plan.

We also examined insulin resistance, as prevention of type 2 diabetes mellitus is a primary goal of obesity treatment in adolescents.2,44 Insulin resistance, as assessed by means of HOMA, increased significantly in the reduced-fat group, possibly owing to the increase in insulin resistance that occurs with progression through puberty. In contrast, insulin resistance did not change in the reduced-GL group. Moreover, the group effect remained significant after adjustment for change in BMI. These findings are consistent with epidemiological data11,20 that show lower risk for diabetes among individuals consuming a low-GL diet, after controlling for BMI. Thus, reducing dietary GL may protect against diabetes through weight-dependent and weight-independent mechanisms, as previously hypothesized.12

Adherence to diet prescriptions, regardless of group assignment, was likely affected by intrapersonal, interpersonal, and environmental influences.45 Not surprisingly, motivation varied widely, with some adolescents showing a strong desire to change their eating behaviors to lose weight and others demonstrating very little concern about their weight status. Although most participating parents were well intentioned, not all provided adequate support with regard to provision of recommended foods, transportation to counseling sessions, and encouragement. Adolescents who participated in the pilot study often enjoyed socializing with peers at food courts and fast-food restaurants, where they were challenged to make reduced-GL or reduced-fat food choices. Research is needed to develop novel strategies for motivating obese adolescents to change their eating behaviors and for motivating parents to provide appropriate support in environments that pose challenges to adherence.

Several issues pertaining to study design should be noted. Strengths of the study include careful attention to treatment fidelity, a well-defined conceptual framework to promote behavior change, a relatively long follow-up, and high subject retention. Study limitations include a small sample size and reliance on self-report for dietary assessment. Underreporting of energy intake is a recognized source of measurement error when assessing adolescent diets.46 Adjusting other dietary variables for energy intake may partially correct for underreporting, although differences in the accuracy of self-report among foods is not well understood.46 Study findings may be confounded by the baseline group difference in fat mass despite randomization; however, there is no a priori reason to believe that subjects in the experimental group, who had lower fat mass at baseline than their counterparts in the conventional group, would lose more absolute body fat. Furthermore, we cannot definitely attribute treatment effects exclusively to changes in GL, as other dietary factors such as fiber or palatability may mediate or confound the relationship between changes in GL and changes in adiposity to some degree (an issue in all long-term outpatient dietary studies).


Our results suggest that reducing dietary GL may have greater benefits than reducing dietary fat when treating adolescent obesity to lower the risk for type 2 diabetes mellitus. Although findings must be considered preliminary, this study provides relevant pilot data to inform future research. Large-scale randomized controlled trials are needed to evaluate the effectiveness and public health applications of reduced-GL and -GI diets.


Corresponding author: David S. Ludwig, MD, PhD, Division of Endocrinology, 300 Longwood Ave, Children's Hospital Boston, Boston, MA 02115 (e-mail: david.ludwig@tch .harvard.edu).

Accepted for publication March 20, 2003.

This study was supported by grants 1R01DK59240 and 1K01DK62237, from the National Insitute of Diabetes and Digestive Kidney Diseases (Bethesda, Md); the Charles H. Hood Foundation (Boston, Mass); pilot and feasibility project grant DK46200 from the Boston Obesity and Nutrition Research Center (Boston); and grant M01 RR02172 awarded by the National Institutes of Health (Bethesda) to support the General Clinical Research Center at Children's Hospital Boston, Boston.

We thank Janet Washington, MPH, RD, for her contribution to intervention development; Suzanne Muggeo and Eve Callahan, MS, RD, for performing the dual-energy x-ray absorptiometry scans; Catherine Murphy, RN, and Linda Lynch, RN, for help with patient care; and Irena Clark, MHP, Erica Garcia-Lago, and Gina Masse for technical assistance.

What This Study Adds

Adolescent obesity increases the risk for a wide range of serious complications, including type 2 diabetes mellitus. Given the current obesity epidemic, novel treatment strategies are needed urgently. No previous studies have evaluated the effects of diet composition per se on treatment outcomes in obese adolescents. The results of this study indicate that a reduced-GL diet may yield greater benefits than a conventional reduced-fat diet. These pilot data, derived from a relatively long-term intervention study, provide strong rationale for conducting large-scale randomized controlled trials to evaluate the effectiveness of a reduced-GL diet in the treatment of obesity and prevention of diabetes mellitus. This line of investigation has the potential to change clinical practice and public health guidelines.

Fagot-Campagna  APettitt  DJEngelgau  MM  et al.  Type 2 diabetes among North American children and adolescents: an epidemiologic review and a public health perspective.  J Pediatr. 2000;136664- 672PubMedGoogle ScholarCrossref
American Diabetes Association, Type 2 diabetes in children and adolescents.  Diabetes Care. 2000;23381- 389PubMedGoogle ScholarCrossref
Pinhas-Hamiel  OStandiford  DHamiel  DDolan  LMCohen  RZeitler  PS The type 2 family: a setting for development and treatment of adolescent type 2 diabetes mellitus.  Arch Pediatr Adolesc Med. 1999;1531063- 1067PubMedGoogle ScholarCrossref
Rosenbloom  ALJoe  JRYoung  RSWinter  WE Emerging epidemic of type 2 diabetes in youth.  Diabetes Care. 1999;22345- 354PubMedGoogle ScholarCrossref
Epstein  LHMyers  MDRaynor  HASaelens  BE Treatment of pediatric obesity.  Pediatrics. 1998;101554- 570PubMedGoogle Scholar
Jelalian  ESaelens  BE Empirically supported treatments in pediatric psychology: pediatric obesity.  J Pediatr Psychol. 1999;24223- 248PubMedGoogle ScholarCrossref
Brownell  KDKelman  JHStunkard  AJ Treatment of obese children with and without their mothers: changes in weight and blood pressure.  Pediatrics. 1983;71515- 523PubMedGoogle Scholar
Johnson  WGHinkle  LKCarr  RE  et al.  Dietary and exercise interventions for juvenile obesity: long-term effect of behavioral and public health models.  Obes Res. 1997;5257- 261PubMedGoogle ScholarCrossref
Wabitsch  MHauner  HHeinze  E  et al.  Body-fat distribution and changes in the atherogenic risk-factor profile in obese adolescent girls during weight reduction.  Am J Clin Nutr. 1994;6054- 60PubMedGoogle Scholar
Wadden  TAStunkard  AJRich  LRubin  CJSweidel  GMcKinney  S Obesity in black adolescent girls: a controlled clinical trial of treatment by diet, behavior modification, and parental support.  Pediatrics. 1990;85345- 352PubMedGoogle Scholar
Salmeron  JManson  JEStampfer  MJColditz  GAWing  ALWillett  WC Dietary fiber, glycemic load, and risk of non–insulin-dependent diabetes mellitus in women.  JAMA. 1997;277472- 477PubMedGoogle ScholarCrossref
Ludwig  DS The glycemic index: physiological mechanisms relating to obesity, diabetes, and cardiovascular disease.  JAMA. 2002;2872414- 2423PubMedGoogle ScholarCrossref
Wolever  TMJenkins  DJJenkins  ALJosse  RG The glycemic index: methodology and clinical implications.  Am J Clin Nutr. 1991;54846- 854PubMedGoogle Scholar
Foster-Powell  KHolt  SHABrand-Miller  JC International table of glycemic index and glycemic load values: 2002.  Am J Clin Nutr. 2002;765- 56PubMedGoogle Scholar
Ludwig  DS Dietary glycemic index and obesity.  J Nutr. 2000;130(2S, suppl)280S- 283SPubMedGoogle Scholar
Ludwig  DSMajzoub  JAAl-Zahrani  ADallal  GEBlanco  IRoberts  SB High glycemic index foods, overeating, and obesity.  Pediatrics [serial online]. 1999;103E26Accessibility verified May 29, 2003PubMedGoogle Scholar
Slabber  MBarnard  HCKuyl  JMDannhauser  ASchall  R Effects of a low-insulin-response, energy-restricted diet on weight loss and plasma insulin concentrations in hyperinsulinemic obese females.  Am J Clin Nutr. 1994;6048- 53PubMedGoogle Scholar
Bouche  CRizkalla  SWLuo  J  et al.  Five-week, low–glycemic index diet decreases total fat mass and improves plasma lipid profile in moderately overweight nondiabetic men.  Diabetes Care. 2002;25822- 828PubMedGoogle ScholarCrossref
Spieth  LEHarnish  JDLenders  CM  et al.  A low–glycemic index diet in the treatment of pediatric obesity.  Arch Pediatr Adolesc Med. 2000;154947- 951PubMedGoogle ScholarCrossref
Salmeron  JAscherio  ARimm  EB  et al.  Dietary fiber, glycemic load, and risk of NIDDM in men.  Diabetes Care. 1997;20545- 550PubMedGoogle ScholarCrossref
Willett  WC Is dietary fat a major determinant of body fat?  Am J Clin Nutr. 1998;67(3, suppl)556S- 562SPubMedGoogle Scholar
Bray  GAPopkin  BM Dietary fat intake does affect obesity!  Am J Clin Nutr. 1998;681157- 1173PubMedGoogle Scholar
Kuczmarski  RJOgden  CLGrummer-Strawn  LM  et al.  CDC Growth Charts: United States.  Hyattsville, Md National Center for Health Statistics2000;Advance Data from Vital and Health Statistics, No. 314
Ebbeling  CBLudwig  LS Treating obesity in youth: should dietary glycemic load be a consideration?  Adv Pediatr. 2001;48179- 212PubMedGoogle Scholar
Franz  MJBantle  JPBeebe  CA  et al.  Evidenced-based nutrition principles and recommendations for the treatment and prevention of diabetes and related complications.  Diabetes Care. 2002;25148- 198PubMedGoogle ScholarCrossref
Diabetes Prevention Program Research Group, The Diabetes Prevention Program (DPP): description of lifestyle intervention.  Diabetes Care. 2002;252165- 2171PubMedGoogle ScholarCrossref
Harris  JABenedict  FG A Biometric Study of Basal Metabolism in Man.  Washington, DC Carnegie Institution1919;Publication 279
Baranowski  TPerry  CLParcel  GSGlanz  KedLewis  FMedRimer  BKed How individuals, environments, and health behavior interact.  Health Behavior and Health Education. San Francisco, Calif Jossey-Bass/Pfeiffer1997;153- 178Google Scholar
Warshaw  H Diabetes Meal Planning Made Easy: How to Put the Food Pyramid to Work for Your Busy Lifestyle. 2nd ed. Alexandria, Va American Diabetes Association2000;
Holler  HJBarrier  PCronmiller  NDelahanty  LFranz  MWheeler  M Exchange Lists for Weight Management.  Alexandria, Va American Diabetes Association, American Dietetic Association1995;
Sallis  JFPatrick  K Physical activity guidelines for adolescents: consensus statement.  Pediatr Exerc Sci. 1994;6302- 314Google Scholar
Committee on Public Education, American Academy of Pediatrics, Children, adolescents, and television.  Pediatrics. 2001;107423- 426PubMedGoogle ScholarCrossref
Moncher  FJPrinz  RJ Treatment fidelity in outcome studies.  Clin Psychol Rev. 1991;11247- 266Google ScholarCrossref
Wolever  TMSNguyen  P-MChiasson  J-L  et al.  Determinants of diet glycemic index calculated retrospectively from diet records of 342 individuals with non–insulin-dependent diabetes mellitus.  Am J Clin Nutr. 1994;591265- 1269PubMedGoogle Scholar
Lohman  TGRoche  AFMartorell  R Anthropometric Standardization Reference Manual.  Champaign, Ill Human Kinetics Books1988;
Bonora  ETargher  GAlberiche  M  et al.  Homeostasis model assessment closely mirrors the glucose clamp technique in the assessment of insulin sensitivity: studies in subjects with various degrees of glucose tolerance and insulin sensitivity.  Diabetes Care. 2000;2357- 63PubMedGoogle ScholarCrossref
Bavdekar  AYajnik  CSFall  CHD  et al.  Insulin resistance syndrome in 8-year-old Indian children: small at birth, big at 8 years, or both?  Diabetes. 1999;482422- 2429PubMedGoogle ScholarCrossref
Dabelea  DPettitt  DJHanson  RLImperatore  GBennett  PHKnowler  WC Birth weight, type 2 diabetes, and insulin resistance in Pima Indian children and young adults.  Diabetes Care. 1999;22944- 950PubMedGoogle ScholarCrossref
Sinha  RFisch  GTeague  B  et al.  Prevalence of impaired glucose tolerance among children and adolescents with marked obesity.  N Engl J Med. 2002;346802- 810PubMedGoogle ScholarCrossref
Hebert  JREbbeling  CBMatthews  CE  et al.  Systematic errors in middle-aged women's estimates of energy intake: comparing three self-report measures to total energy expenditure from doubly labeled water.  Ann Epidemiol. 2002;12577- 586PubMedGoogle ScholarCrossref
Neumark-Sztainer  DStory  MAckard  DMoe  JPerry  C The "family meal": views of adolescents.  J Nutr Educ. 2000;32329- 334Google ScholarCrossref
Kelsey  KSCampbell  MKVanata  DF Parent and adolescent girls' preferences for parental involvement in adolescent health promotion programs.  J Am Diet Assoc. 1998;98906- 907PubMedGoogle ScholarCrossref
Gilbertson  HRBrand-Miller  JCThorburn  AWEvans  SChondros  PWerther  GA The effect of flexible low glycemic index dietary advice versus measured carbohydrate exchange diets on glycemic control in children with type 1 diabetes.  Diabetes Care. 2001;241137- 1143PubMedGoogle ScholarCrossref
Ludwig  DSEbbeling  CB Type 2 diabetes mellitus in children: primary care and public health considerations.  JAMA. 2001;2861427- 1430PubMedGoogle ScholarCrossref
Davison  KKBirch  LL Childhood overweight: a contextual model and recommendations for future research.  Obes Rev. 2001;2159- 171PubMedGoogle ScholarCrossref
Livingstone  MBRobson  PJ Measurement of dietary intake in children.  Proc Nutr Soc. 2000;59279- 293PubMedGoogle ScholarCrossref