Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 14-Day Trial for You or Your Team.

Learn More →

Effect of a Low-Fat Vegan Diet on Body Weight, Insulin Sensitivity, Postprandial Metabolism, and Intramyocellular and Hepatocellular Lipid Levels in Overweight Adults

Effect of a Low-Fat Vegan Diet on Body Weight, Insulin Sensitivity, Postprandial Metabolism, and... Key Points Question What are the effects of a IMPORTANCE Excess body weight and insulin resistance lead to type 2 diabetes and other major low-fat vegan diet on body weight, health problems. There is an urgent need for dietary interventions to address these conditions. insulin resistance, postprandial metabolism, and intramyocellular and OBJECTIVE To measure the effects of a low-fat vegan diet on body weight, insulin resistance, hepatocellular lipid levels in postprandial metabolism, and intramyocellular and hepatocellular lipid levels in overweight adults. overweight adults? Findings In this 16-week randomized DESIGN, SETTING, AND PARTICIPANTS This 16-week randomized clinical trial was conducted clinical trial, a low-fat plant-based between January 2017 and February 2019 in Washington, DC. Of 3115 people who responded to dietary intervention reduced body flyers in medical offices and newspaper and radio advertisements, 244 met the participation criteria weight by reducing energy intake and (age 25 to 75 years; body mass index of 28 to 40) after having been screened by telephone. increasing postprandial metabolism, which was associated with reductions in INTERVENTIONS Participants were randomized in a 1:1 ratio. The intervention group (n = 122) was hepatocellular and intramyocellular fat asked to follow a low-fat vegan diet and the control group (n = 122) to make no diet changes for and increased insulin sensitivity. 16 weeks. Meaning A low-fat plant-based diet is MAIN OUTCOMES AND MEASURES At weeks 0 and 16, body weight was assessed using a an effective tool for reducing body calibrated scale. Body composition and visceral fat were measured by dual x-ray absorptiometry. weight and increasing insulin sensitivity Insulin resistance was assessed with the homeostasis model assessment index and the predicted and postprandial metabolism. insulin sensitivity index (PREDIM). Thermic effect of food was measured by indirect calorimetry over 3 hours after a standard liquid breakfast (720 kcal). In a subset of participants (n = 44), Supplemental content hepatocellular and intramyocellular lipids were quantified by proton magnetic resonance spectroscopy. Repeated measure analysis of variance was used for statistical analysis. Author affiliations and article information are listed at the end of this article. RESULTS Among the 244 participants in the study, 211 (87%) were female, 117 (48%) were White, and the mean (SD) age was 54.4 (11.6) years. Over the 16 weeks, body weight decreased in the intervention group by 5.9 kg (95% CI, 5.0-6.7 kg; P < .001). Thermic effect of food increased in the intervention group from baseline to 16 weeks and did not change significantly in the control group (between-group difference in effect size, 14.1%; 95% CI, 6.5-20.4; P < .001). The homeostasis model assessment index decreased (−1.3; 95% CI, −2.2 to −0.3; P < .001) and PREDIM increased (0.9; 95% CI, 0.5-1.2; P < .001) in the intervention group. Hepatocellular lipid levels decreased in the intervention group by 34.4%, from a mean (SD) of 3.2% (2.9%) to 2.4% (2.2%) (P = .002), and intramyocellular lipid levels decreased by 10.4%, from a mean (SD) of 1.6 (1.1) to 1.5 (1.0) (P = .03). None of these variables changed significantly in the control group over the 16 weeks. The change in PREDIM correlated negatively with the change in body weight (r = −0.43; P < .001). Changes in (continued) Open Access. This is an open access article distributed under the terms of the CC-BY License. JAMA Network Open. 2020;3(11):e2025454. doi:10.1001/jamanetworkopen.2020.25454 (Reprinted) November 30, 2020 1/14 JAMA Network Open | Nutrition, Obesity, and Exercise Effect of a Low-Fat Vegan Diet on Metabolic Measures in Overweight Adults Abstract (continued) hepatocellular and intramyocellular lipid levels correlated with changes in insulin resistance (both r = 0.51; P = .01). CONCLUSIONS AND RELEVANCE A low-fat plant-based dietary intervention reduces body weight by reducing energy intake and increasing postprandial metabolism. The changes are associated with reductions in hepatocellular and intramyocellular fat and increased insulin sensitivity. TRIAL REGISTRATION ClinicalTrials.gov Identifier: NCT02939638 JAMA Network Open. 2020;3(11):e2025454. Corrected on May 27, 2021. doi:10.1001/jamanetworkopen.2020.25454 Introduction Overweight and associated diseases, particularly type 2 diabetes and metabolic syndrome, remain worldwide challenges. There is an urgent need for dietary interventions to address these problems and for a better understanding of how different dietary interventions work. 1,2 Obesity is uncommon in individuals whose diets are based on plant-derived foods. In clinical trials, such diets caused weight loss, for which 2 explanations have been offered. First, a high-fiber, low-fat diet has a low energy density, which reduces energy intake. Second, a low-fat, vegan diet increases the thermic effect of food, which accounts for approximately 10% of total energy expenditure. However, in the latter randomized clinical trial, the control group was following an active diet based on National Cholesterol Education Program guidelines. Because there was no untreated control group, the effect of a low-fat vegan diet on thermogenesis remains unclear. Studies have reported that people following a vegan diet have lower concentrations of intramyocellular lipids compared with those following omnivorous diets, suggesting that by reducing intramyocellular or hepatocellular lipid levels, a plant-based diet may lead to increased mitochondrial 6,7 activity and postprandial metabolism. This is particularly important because the accumulation of 8-10 lipids in muscle and liver cells may also be associated with insulin resistance and type 2 diabetes. The aim of this study was to measure the effects of a low-fat vegan diet on body weight, insulin resistance, postprandial metabolism, and intramyocellular and hepatocellular lipid levels in overweight adults. Methods Study Design and Eligibility This randomized clinical trial using a single-center, open parallel design was conducted between January 2017 and February 2019 in Washington, DC, in 4 replications (the trial protocol is given in Supplement 1). Adults aged 25 to 75 years with a body mass index (BMI) (calculated as weight in kilograms divided by height in meters squared) of 28 to 40 were enrolled. Exclusion criteria were diabetes, smoking, alcohol or drug use, pregnancy or lactation, and current use of a vegan diet. The additional exclusion criteria for the subset of participants undergoing the proton magnetic resonance spectroscopy were the presence of any metal implant, claustrophobia, BMI higher than 38, and waist circumference of more than 102 cm. The study protocol was approved by the Chesapeake Institutional Review Board. All participants gave written informed consent. This study followed the Consolidated Standards of Reporting Trials (CONSORT) reporting guideline. JAMA Network Open. 2020;3(11):e2025454. doi:10.1001/jamanetworkopen.2020.25454 (Reprinted) November 30, 2020 2/14 JAMA Network Open | Nutrition, Obesity, and Exercise Effect of a Low-Fat Vegan Diet on Metabolic Measures in Overweight Adults Randomization and Study Groups With use of a computer-generated system, participants were randomly assigned (in a 1:1 ratio) to an intervention group, which was asked to follow a low-fat vegan diet, or a control group, which was asked to make no diet changes. The randomization protocol could not be accessed by the participants or the staff allocating the participants into groups beforehand. Because assignment was done simultaneously, allocation concealment was unnecessary. The participants were not blinded to their group assignment. The intervention diet (approximately 75% of energy from carbohydrates, 15% protein, and 10% fat) consisted of vegetables, grains, legumes, and fruits without animal products or added fats. Vitamin B was supplemented (500 μg/d). The intervention group attended weekly classes for detailed instruction and cooking demonstrations and received printed materials and small food samples. No meals were provided. For both groups, alcoholic beverages were limited to 1 per day for women and 2 per day for men. All participants were asked to maintain their customary exercise habits and medications unless modified by their personal physicians. Outcomes All measurements were performed at baseline and 16 weeks. The outcome assessors (K.F.P., G.I.S., and A.T.) were blinded to group assignment. The primary outcomes were body weight, insulin resistance, postprandial metabolism, and the concentrations of intramyocellular and hepatocellular lipids. At baseline and at 16 weeks, dietary intake data over 3 consecutive days were collected and analyzed by staff members certified in the Nutrition Data System for Research, version 2016, developed by the Nutrition Coordinating Center, University of Minnesota, Minneapolis. In addition, study dietitians made unannounced telephone calls to assess participants’ dietary adherence. All study participants were asked not to alter their exercise habits and to continue their preexisting medication regimens for the duration of the study. Physical activity was assessed by the International Physical Activity Questionnaire. Laboratory assessments were made after an overnight fast. Height (baseline only) and weight were measured using a stadiometer and a calibrated digital scale, respectively. Body composition and visceral fat volume were assessed using dual energy x-ray absorptiometry (iDXA; GE Healthcare), 14 15 which has been validated against computed tomography and magnetic resonance imaging. The measurement of total body fat and visceral fat had a coefficient of variation (CV) of 1.0% and 5.4%, 16,17 respectively. Insulin secretion was assessed after a standardized liquid breakfast (Boost Plus, Nestle) (720 kcal, 34% of energy from fat, 16% protein, and 50% carbohydrate). Plasma glucose, immunoreactive insulin, and C-peptide concentrations were measured at 0, 30, 60, 120, and 180 minutes. Plasma glucose concentration was analyzed using the Hexokinase UV end point method (the intra-assay CV was 1.4%, and the inter-assay CV was 1.9%), and immunoreactive insulin and C-peptide concentrations were determined using insulin and C-peptide electro-chemiluminescence immunoassay (the intra-assay CVs were 5.1% and 3.8%, respectively, and the inter-assay CVs were 5.7% and 3.9%, respectively). Glycated hemoglobin level was measured by turbidimetric inhibition immunoassay (the intra-assay CV was 3.7%, and the inter-assay CV was 3.5%), and lipid concentrations were measured by enzymatic colorimetric methods (intra-assay CV: total cholesterol, 2.1%; high-density lipoprotein cholesterol, 2.4%; low-density lipoprotein cholesterol, 2.0%; and triglycerides 2.2%; inter-assay CV: total cholesterol, 2.7%; high-density lipoprotein cholesterol, 3.8%; low-density lipoprotein cholesterol, 3.0%; and triglycerides 3.2%). All test kits were made by Roche. Insulin resistance was calculated using the homeostasis model assessment index. The predicted insulin sensitivity index (PREDIM) provided a validated measure of dynamic insulin sensitivity. Resting energy expenditure and postprandial metabolism were measured by indirect calorimetry (Cosmed Quark CPET) using a ventilated hood system (accuracy of measurement with a JAMA Network Open. 2020;3(11):e2025454. doi:10.1001/jamanetworkopen.2020.25454 (Reprinted) November 30, 2020 3/14 JAMA Network Open | Nutrition, Obesity, and Exercise Effect of a Low-Fat Vegan Diet on Metabolic Measures in Overweight Adults 20,21 CV<1% and repeatability of measurement with a CV of 1.2%). Each measurement was performed for 15 minutes after an overnight fast and 30, 60, 120, and 180 minutes after the standard breakfast. In a subset of 44 participants (23 in the intervention group and 21 in the control group), proton magnetic resonance spectroscopy was performed at the Magnetic Resonance Research Center, Yale School of Medicine. Hepatocellular and intramyocellular lipids were quantified by proton magnetic resonance spectroscopy at 4T (Bruker). This method has been shown to provide a precise quantification of fat fractions, with a mean bias of −1.1.% to 0.5%. Hepatocellular lipid content was measured by H respiratory-gated stimulated echo acquisition mode spectroscopy in a 15 × 15 × 15-mm voxel. Acquisition was synchronized to the respiratory cycle and triggered at the end of expiration. A water-suppressed lipid spectrum and a lipid-suppressed water spectrum were acquired in 3 locations of the liver to account for liver inhomogeneity, and the total lipid content was averaged and calculated. In addition, hepatocellular lipid content was corrected for transverse relaxation using the transverse relaxation times of 22 ms for water and 44 ms for lipid. Intramyocellular lipid content was measured in the soleus muscle using an 8.5-cm diameter circular 13 1 C surface coil with twin, orthogonal circular 13-cm H quadrature coils. Scout images of the lower leg were obtained to ensure correct positioning of the participant and to define an adequate volume for localized shimming using the FastMap procedure. Power Analysis Sample size was based on the change in body weight, insulin resistance, and postprandial metabolism previously observed with a plant-based diet, with an α level of 0.05. The assumed change for body weight was a mean (SD) of 5.8 (3.2) kg in the intervention arm and 1 (3.2) kg in the control arm; for insulin sensitivity, the assumed change was 1.1 (2.1) in the intervention arm and 0.1 (2.1) in the control arm; and for the thermic effect of food, the assumed change was 4.7 (12) (area under the curve) in the intervention arm and 0.3 (9.4) in the control arm. For the primary efficacy comparison, a total of 22 participants (11 per arm) were required for 90% power to detect a significant treatment effect on body weight between the 2 study arms. For insulin sensitivity, a total of 142 participants (71 in each arm) were required for 90% power. Assuming that the treatment effect for postprandial metabolism was of the same magnitude at each of the 5 evaluation points used in metabolic assessment and that the SD was approximately 10.85 points for all observations, with 5 observations per participant correlated at a magnitude of 0.7 with each other, and assuming an attrition of 10%, the required sample size was 81 per group (162 total) for 80% power and 108 per group (216 total) for 90% power. For the substudy assessing the role of intramyocellular and hepatocellular lipids in insulin sensitivity, a study from 2012 provided a basis for a power analysis. In that study, 7 lean individuals with insulin resistance followed a hypocaloric (1200 kcal/d) diet for 9 weeks. The mean (SD) intramyocellular lipid level decreased from 1.1% (0.2%) to 0.8% (0.1%). Assuming a mean (SD) change in the intramyocellular lipid level of 0.3% (0.2%) in the intervention arm and a mean change of 0 with a similar SD in the control arm, to have 90% power to detect a difference of this magnitude between, the 2 arms would each require 11 individuals (22 total). Because this was an exploratory substudy and variability in response to the diet was largely unknown, 20 participants were recruited per arm (a total of 40 participants). Statistical Analysis For baseline characteristics, between-group t tests were performed for continuous variables and χ or Fisher exact test for categorical variables. A repeated measure analysis of variance (ANOVA) model was used with between-person and within-person factors and interactions, including group, person, and time. The interaction between group and time was calculated for each variable. For thermic effect of food, minutes were included in the ANOVA model. Data from only individuals with measurements at both time points were included in the ANOVA model. Within each group, paired JAMA Network Open. 2020;3(11):e2025454. doi:10.1001/jamanetworkopen.2020.25454 (Reprinted) November 30, 2020 4/14 JAMA Network Open | Nutrition, Obesity, and Exercise Effect of a Low-Fat Vegan Diet on Metabolic Measures in Overweight Adults comparison t tests were calculated to test whether the changes from baseline to 16 weeks were statistically significant. To eliminate skewed data distribution and heteroscedasticity, data were transformed to a gaussian distribution before further processing by a power transformation using the statistical software Statgraphics Centurion, version XV (Statpoint Inc). The transformed data underwent multivariable regression using the method of orthogonal projections to latent structure. This method is effective in addressing severe multicollinearity within the matrix of independent variables. In our model, changes in thermic effect of food and in hepatocellular lipid levels were chosen as the dependent variables and the metabolic variables (body weight, fat mass, visceral fat, and insulin resistance) represented the independent variables. The variability was separated into 2 independent components. The predictive component contained the variability in the metabolic variables, which was shared with changes in dependent variables, and the orthogonal component contained the variability shared within the metabolic variables. A detailed description of the orthogonal projections to latent structure model is available elsewhere. The statistical software SIMCA-P, version 11.5 (Umetrics AB) identified the number of relevant components using the prediction error sum of squares and also allowed the detection of multivariable nonhomogeneities and testing for multivariable normal distribution and homoscedasticity (constant variance). The statisticians (M.H, R.H.) were blinded to the interventions and group assignment. Results are presented as means with 95% CIs. Two-tailed tests were used to determine significance at the 5% level. Results Participant Characteristics Of 3115 people screened by telephone, 244 met the participation criteria, signed the consent form and were randomly assigned to the intervention (n = 122) or control (n = 122) groups in a 1:1 ratio (Figure 1). The mean (SD) age of the intervention group was 53 (10) years compared with 57 (13) years in the control group (P = .01) (eTable 1 in Supplement 2). There were no other significant differences between the groups. Five intervention group and 16 control group participants dropped out, mostly for reasons unrelated to the study, leaving 223 (91.0%) individuals who completed the study. eTable 2 in Supplement 2 shows the baseline characteristics of those who completed the study and those who dropped out. There were no significant differences between these groups. The main outcomes are reported in Table 1. The treatment effects were largely unaffected by the adjustment for age and race/ethnicity (eTable 4 in Supplement 2). eTable 3 in Supplement 2 shows the characteristics of the subgroup that underwent magnetic resonance spectroscopy. This group had a lower BMI compared with the rest of the study population. The model adjusted for baseline BMI for magnetic resonance spectroscopy is presented in eFigure 2 in Supplement 2. Dietary Intake and Physical Activity Self-reported energy intake decreased in both groups but more so in the intervention group (treatment effect, −354.9 kcal/d; 95% CI, −519.0 to −190.8 kcal/d; P < .001) (Table 2). In the intervention group, mean intakes of carbohydrate and fiber increased, whereas mean fat, protein, and cholesterol intake decreased. These values did not change significantly in the control group. Physical activity decreased slightly in both groups (−709.8 metabolic equivalents [95% CI, −1346 to −73.9 metabolic equivalents] in the control group and −604.8 metabolic equivalents [95% CI, −1388 to −178.6 metabolic equivalents] in the intervention group; between-group P = .84). Body Weight, Body Composition, and Blood Lipid Levels Mean body weight decreased by 6.4 kg in the intervention group compared with 0.5 kg in the control group (treatment effect, −5.9 kg; 95% CI, −6.7 to −5.0; interaction between group and time, P < .001). This difference was largely attributable to a reduction in body fat, as noted by significant decreases in fat mass and visceral fat volume in the intervention group participants. Total and JAMA Network Open. 2020;3(11):e2025454. doi:10.1001/jamanetworkopen.2020.25454 (Reprinted) November 30, 2020 5/14 JAMA Network Open | Nutrition, Obesity, and Exercise Effect of a Low-Fat Vegan Diet on Metabolic Measures in Overweight Adults low-density lipoprotein cholesterol levels decreased by 0.5 mmol/L and 0.4 mmol/L (to convert to milligrams per deciliter, divide by 0.0259), respectively, in the intervention group, with no significant changes in the control group (0.1 mmol/L and 0.07 mmol/L, respectively) (P < .001 for both). Insulin Sensitivity Fasting plasma insulin concentration decreased by 21.6 pmol/L (to convert to micro-IU per milliliter, divide by 6.945) in the intervention group, with no significant change in the control group (23.6 pmol/L; 95% CI, −5.0 to 54.3; between-group P = .006). The homeostasis model assessment index (a measure of insulin resistance) decreased significantly (−1.3; 95% CI, −2.2 to −0.3; P < .001), and PREDIM (a measure of insulin sensitivity) increased significantly in the intervention group (0.9; 95% CI, 0.5-1.2; P < .001); neither changed significantly in the control group (Table 2). Within the intervention group, the change in PREDIM correlated negatively with the change in body weight (r = −0.43; P < .001). Figure 1. CONSORT Diagram of Participant Flow Through Trial 3115 Participants screened over phone 413 In-person meetings 81 Excluded 36 Outside BMI range 3 Diabetes diagnosis 7 Medical exclusion 31 Not willing to be in the vegan or control group 4 Unable to attend weekly classes 332 Signed the consent form 88 Excluded 72 Did not turn in their diet record or did not come to their baseline assessment 1 Family emergency 1 Had medication changes and decided not to participate 1 Had already started a vegan diet 6 Unable to attend weekly classes 7 Withdrew 244 Randomized 122 Randomized to vegan diet 122 Randomized to control 16 Dropped out 5 Dropped out 3 Withdrew owing to personal reasons 1 Withdrew owing to family reasons 2 Not willing to be in the control or 1 Withdrew owing to health reasons vegan group 1 Unable to contact 2 Due to health reasons 1 Unable to follow diet 8 Unable to contact 1 Unable to attend classes 1 Unable to participate in all aspects of the study 117 Completed final assessment 106 Completed final assessment 117 Were included in the analysis 106 Were included in the analysis JAMA Network Open. 2020;3(11):e2025454. doi:10.1001/jamanetworkopen.2020.25454 (Reprinted) November 30, 2020 6/14 JAMA Network Open | Nutrition, Obesity, and Exercise Effect of a Low-Fat Vegan Diet on Metabolic Measures in Overweight Adults Table 1. Changes in Outcomes During the Study in the Low-Fat Vegan Dietary Intervention Group vs the Control Group Value, Mean (95% CI) Control group Intervention group Outcome Baseline Week 16 Change Baseline Week 16 Change Effect Size P value Dietary intake Energy intake, kcal/d 1793 (1670 to 1657 (1548 to −135.8 (−250.7 to 1834 (1729 to 1344 (1260 to −490.7 (−607.9 to −354.9 (−519.0 to <.001 b c 1915) 1766) −20.8) 1940) 1428) −373.5) −190.8) Fiber intake, g/d 23.9 (21.9 to 23.3 (21.4 to −0.56 (−2.6 to 24.1 (22.1 to 34.6 (32.1 to 10.6 (7.8 to 11.1 (7.8 to <.001 25.9) 25.3) 1.5) 26.0) 37.2) 13.3) 14.5) Cholesterol intake, 244.5 (211.4 to 230.5 (196.1 to −14.0 (−51.7 to 238.6 (212.3 to 5.5 (3.8 to −233.1 (−259.4 to −219.1 (−264.9 to <.001 mg/d 277.6) 264.9) 23.7) 265.0) 7.3) −206.8) −173.3) Saturated fatty 22.9 (20.5 to 20.5 (18.1 to −2.4 (−4.9 to 23.6 (21.4 to 5.1 (4.5 to −18.6 (−20.7 to −16.2 (−19.4 to <.001 acids, g/d 25.3) 23.0) 0.1) 25.8) 5.6) −16.5) −13.0) Monounsaturated fatty 27.9 (25.2 to 25.4 (23.0 to −2.5 (−4.9 to 27.2 (25.3 to 8.4 (7.6 to −18.8 (−20.7 to −16.3 (−19.3 to <.001 b c acids, g/d 30.7) 27.8) −0.1) 29.1) 9.2) −16.9) −13.3) Polyunsaturated fatty 19.1 (17.0 to 18.0 (16.3 to −1.1 (−3.0 to 18.4 (16.8 to 9.5 (8.6 to −8.9 (−10.6 to −7.8 (−10.4 to <.001 acids, g/d 21.1) 19.7) 0.9) 19.9) 10.4) −7.1) −5.2) Physical activity, METs 2863 (2224 to 2153 (1605 to −709.8 (−1346 to 2719 (1805 to 2114 (1619 to −604.8 (−1388 to 105 (−898 to .84 3502) 2702) −73.9) 3633) 2609) 178.6) 1108) Anthropometric variables and body composition Weight, kg 92.7 (90.0 to 92.2 (89.4 to −0.5 (−1.0 to 93.6 (91.0 to 87.2 (84.9 to −6.4 (−7.0 to −5.9 (−6.7 to <.001 95.3) 94.9) 0.1) 96.1) 89.6) −5.7) −5.0) BMI 33.6 (32.9 to 33.9 (32.6 to 0.3 (−0.7 to 33.3 (32.6 to 31.4 (30.5 to −1.9 (−2.5 to −2.2 (−3.3 to <.001 34.3) 35.2) 1.3) 34.0) 32.4) −1.3) −1.1) Fat mass, kg 40.9 (39.1 to 41.0 (39.0 to 0.01 (−0.3 to 40.6 (38.9 to 36.5 (34.9 to −4.1 (−4.6 to −4.1 (−4.7 to <.001 42.8) 42.9) 0.4) 42.2) 38.1) −3.6) −3.5) Lean mass, kg 49.5 (47.9 to 48.9 (47.4 to −0.6 (−0.9 to 50.5 (49.0 to 48.4 (47.1 to −2.1 (−2.4 to −1.5 (−1.9 to <.001 c c 51.1) 50.5) −0.3) 51.9) 49.8) −1.8) −1.1) VAT volume, cm 1517 (1339 to 1510 (1324 to −7.7 (−78.5 to 1459 (1286 to 1243 (1096 to −216.5 (−280.9 to −208.8 (−303.7 to <.001 1695) 1695) 63.0) 1632) 1390) −152.2) −113.7) Hepatocellular 3.3 (3.1 to 3.6 (3.5 to 0.3 (−0.5 to 3.2 (3.0 to 2.4 (2.3 to −0.8 (−1.5 to −1.2 (−2.2 to .002 lipids, % 3.5) 3.8) 1.2) 3.4) 2.5) −0.1) −0.1) Intramyocellular 1.5 (1.4 to 1.7 (1.5 to 0.13 (−0.05 to 1.6 (1.5 to 1.5 (1.4 to −0.1 (−0.2 to −0.3 (−0.4 to .03 lipids, % 1.6) 1.8) 0.21) 1.7) 1.6) 0.05) −0.1) Parameters of glucose control and insulin resistance HbA , DCCT, % 5.7 (5.6 to 5.7 (5.6 to 0.01 (−0.04 to 5.6 (5.6 to 5.6 (5.5 to −0.06 (−0.12 to −0.07 (−0.1 to .07 1c 5.8) 5.8) 0.05) 5.7) 5.7) −0.002) 0.01) Fasting plasma insulin 78.9 (68.3 to 103.5 (71.4 to 23.6 (−5.0 to 91.2 (79.9 to 69.6 (56.9 to −21.6 (−35.9 to −46.2 (−79.0 to .006 level, pmol/L 89.4) 135.6) 54.3) 102.5) 82.3) −7.3) −13.4) Fasting plasma glucose 5.0 (4.7 to 5.5. (5.4 to 0.5 (0.2 to 5.2 (5.1 to 5.1 (5.0 to −0.1 (−0.2 to −0.6 (−0.2 to .001 level, mmol/L 5.4) 5.7) 0.8) 5.3) 5.2) 0.02) −1.0) PREDIM 4.4 (4.1 to 4.2 (3.9 to −0.2 (−0.4 to 4.1 (3.8 to 4.7 (4.4 to 0.7 (0.5 to 0.9 (0.5 to <.001 4.7) 4.5) 0.04) 4.3) 5.0) 0.9) 1.2) HOMA 2.7 (2.3 to 3.2 (2.4 to 0.5 (−0.3 to 3.2 (2.7 to 2.3 (1.9 to −0.8 (−1.3 to −1.3 (−2.2 to <.001 3.2) 4.0) 1.2) 3.6) 2.8) −0.3) −0.3) Lipid levels, mmol/L Total cholesterol 5.0 (4.7 to 5.1 (4.9 to 0.1 (−0.1 to 5.2 (5.0 to 4.7 (4.5 to −0.5 (−0.7 to −0.6 (−0.9 to <.001 5.2) 5.3) 0.4) 5.4) 4.9) −0.4) −0.4) Triglycerides 1.3 (1.2 to 1.3 (1.1 to −0.01 (−0.14 to 1.2 (1.1 to 1.4 (1.3 to 0.2 (0.08 to 0.2 (0.03 to .02 1.4) 1.5) 0.12) 1.3) 1.5) 0.3) 0.4) HDL cholesterol 1.7 (1.5 to 1.5 (1.4 to −0.2 (−0.4 to 1.6 (1.5 to 1.4 (1.3 to −0.2 (−0.3 to 0.01 (−0.2 to .93 b c 1.9) 1.6) −0.02) 1.6) 1.4) −0.1) 0.2) LDL cholesterol 2.9 (2.6 to 3.0 (2.9 to 0.07 (−0.02 to 3.1 (3.0 to 2.7 (2.5 to −0.4 (−1.0 to −0.5 (−0.8 to <.001 3.1) 3.2) 0.2) 3.3) 2.9) −0.3) −0.3) Abbreviations: BMI, body mass index (calculated as weight in kilograms divided by P values are for the interaction between group and time assessed by repeated height in meters squared); DCCT, Diabetes Control and Complications Trial; HbA , measures analysis of variance. 1c glycated hemoglobin; HDL, high-density lipoprotein; HOMA, homeostasis model b P < .05 for within-group changes from baseline assessed by paired comparison t tests. assessment; LDL, low-density lipoprotein; METs, metabolic equivalents; PREDIM, P < .001 for within-group changes from baseline assessed by paired comparison t tests. predicted insulin sensitivity index; VAT, visceral adipose tissue. SI conversion factors: To convert plasma insulin level to μIU/mL, divide by 6.945; plasma glucose level to mg/dL, divide by 0.0555; and lipid levels to mg/dL, divide by 0.0259. JAMA Network Open. 2020;3(11):e2025454. doi:10.1001/jamanetworkopen.2020.25454 (Reprinted) November 30, 2020 7/14 JAMA Network Open | Nutrition, Obesity, and Exercise Effect of a Low-Fat Vegan Diet on Metabolic Measures in Overweight Adults Postprandial Metabolism Postprandial energy expenditure (the thermic effect of food) increased in the intervention group from baseline to 16 weeks and did not change significantly in the control group (between-group difference effect size, 14.1%; 95% CI, 6.5%-20.4%; interaction between group and time, P < .001) (Figure 2A). The F values were as follows: group, F =1.7 (P = .19); week, F =15.4 (P < .001); time, F = 122.4 (P < .001); group × week, F =11.9 (P < .001); group × time, F =1.1 (P = .35); week × time, F = 1.38 (P = .25). The results were similar in models adjusted for age and race/ethnicity (eFigure 1 in Supplement 2). Within the intervention group, the change in thermic effect of food did not correlate significantly with changes in body weight (r = −0.15; P = .09), PREDIM (r = 0.06; P = .54), energy intake (r = 0.01; P = .90), or fiber consumption (r = 0.07; P = .48). In both groups combined, changes in thermic effect of food correlated negatively with changes in fat mass (r = −0.30; P < .05) and positively with changes in PREDIM (r = 0.36; P < .05). That is, as fat mass decreased and insulin sensitivity improved, postprandial metabolism increased (Table 2). A linear regression model for changes in reported energy intake and body weight showed that every 100 kcal/d change in energy intake was associated with a 0.15 kg change in body weight (eFigure 3 in Supplement 2). The mean (SD) reported energy reduction of 355 (617) kcal in the intervention group compared with the control group would therefore be associated with a mean (SD) weight loss of 0.53 (4.4) kg. For changes in postprandial energy expenditure and body weight, every change in postprandial energy expenditure of 10 000 U in area under the curve was associated with a change in body weight of 0.48 kg (eFigure 3 in Supplement 2). The mean (SD) decrease in postprandial energy expenditure of 8588 (34 020) U of area under the curve was associated with an mean (SD) weight loss of 0.41 (2.8) kg. Hepatocellular and Intramyocellular Lipid Levels In the 44 participants for whom hepatocellular and intramyocellular lipid levels were quantified, 29,30 baseline hepatocellular lipid content was generally in the normal range. Nonetheless, hepatocellular lipid content decreased in the intervention group by 34.4% (from a mean [SD] of 3.2% [2.9%] to 2.4% [2.2%]; P = .03) and remained unchanged in the control group (from a mean [SD] of 3.3% [4.3%] to 3.6% [4.7%]) (group, F =3.1 [P = .09]; week, F =1.27 [P = .27]; group × week, F =10.8 [P = .002]) (Figure 2B). Results were similar in models adjusted for age and race/ethnicity (eFigure 1 in Supplement 2) and for baseline BMI (eFigure 2 in Supplement 2). Within the intervention group, the decrease in hepatocellular lipid levels was significantly associated with change in body weight (r = 0.42; P = .04) but not with changes in reported energy intake (r = 0.24; P = .27) or fiber consumption (r = 0.07; P = .76). In both groups combined, changes in hepatocellular lipid levels correlated negatively with changes in PREDIM (r = −0.47; P < .05). That is, as hepatocellular lipid level decreased, insulin sensitivity increased. Changes in hepatocellular lipid Table 2. Relationship Between Changes in Thermic Effect of Food and the First Predictive Component as Evaluated by the OPLS Model OPLS predictive component Multiple regression a b Variable Component loading t Statistic R P value for R Regression coefficient t Statistic Matrix X Baseline BMI 0.191 2.46 0.209 <.05 −0.015 −0.33 Baseline fat mass 0.256 2.89 0.283 <.05 −0.014 −0.28 Baseline TEF −0.850 −11.96 −0.938 .005 −0.505 −5.69 Change in PREDIM 0.324 2.41 0.359 <.05 0.105 1.37 Change in fat mass −0.271 −2.59 −0.301 <.05 −0.122 −1.55 Matrix Y Change in TEF 1.000 5.27 0.540 .003 NA NA Abbreviations: BMI, body mass index; NA, not applicable; OPLS, orthogonal projections Component loadings expressed as a correlation coefficients with predictive to latent structure; PREDIM, predicted insulin sensitivity index; TEF, thermic effect component. of food. The explained variability was 29.2% (24.3% after cross-validation). JAMA Network Open. 2020;3(11):e2025454. doi:10.1001/jamanetworkopen.2020.25454 (Reprinted) November 30, 2020 8/14 JAMA Network Open | Nutrition, Obesity, and Exercise Effect of a Low-Fat Vegan Diet on Metabolic Measures in Overweight Adults levels correlated positively with changes in body weight (r = 0.91; P < .01), BMI (r = 0.90; P < .01), fat mass (r = 0.91; P < .01), and visceral fat (r = 0.80; P <.01)(Table 3). Changes in intramyocellular lipid levels were not statistically significant in within-group comparisons, but owing to the opposite trends, the treatment effect was significantly decreased in the intervention group by 10.4%, from a mean (SD) of 1.6 (1.1) to 1.5 (1.0) (P = .03) (group, F =4.7 [P = .04]; week, F = 0.02 [P = .88]; group × week, F =5.1 [P = .03]) (Figure 1C). Within the intervention group (n = 23), changes in both hepatocellular and intramyocellular lipid levels correlated with changes in insulin resistance, as measured by the homeostasis model assessment index (both r = 0.51; P = .01). In both groups combined, changes in intramyocellular lipid levels correlated positively with changes in fat mass (r = 0.51; P < .05) and homeostasis model assessment index score (r = 0.52; P < .05). That is, as fat mass decreased, intramyocellular lipid levels and insulin resistance decreased. Discussion In this trial, the dietary intervention reduced body weight, apparently owing to its tendency to reduce energy intake and increase postprandial energy expenditure. The intervention also improved glycemic control and reduced insulin concentrations, owing in part to reduced lipid accumulation in liver and muscle cells and thus reduced insulin resistance in these organs. Figure 2. Changes in the Thermic Effect of Food, Liver Fat, and Intramyocellular Lipid Levels in the Intervention and Control Groups A Thermic effect of food in the control group B Thermic effect of food in the intervention group 4.5 4.5 Week 0 4.0 4.0 Week 16 3.5 3.5 Week 0 3.0 3.0 Week 16 2.5 2.5 2.0 2.0 1.5 1.5 1.0 1.0 30 60 120 180 30 60 120 180 Time, min Time, min C Liver fat D Intramyocellular lipids 2.6 1.8 Control group Control group 2.4 1.6 2.2 2.0 1.5 1.8 1.4 Intervention group 1.6 Intervention group 1.4 1.3 0 16 0 16 Week Week Whiskers represent 95% CIs. JAMA Network Open. 2020;3(11):e2025454. doi:10.1001/jamanetworkopen.2020.25454 (Reprinted) November 30, 2020 9/14 Liver fat, % Thermic effect of food, kcal/kg Intramyocellular lipids, % Thermic effect of food, kcal/kg JAMA Network Open | Nutrition, Obesity, and Exercise Effect of a Low-Fat Vegan Diet on Metabolic Measures in Overweight Adults The intervention diet’s effect on weight and insulin action are clinically important. Hepatocellular and intramyocellular lipids play central roles in hepatic and muscle insulin resistance, respectively, and in type 2 diabetes. A 16-week diet of 1200 kcal per day resulted in a moderate weight loss of approximately 8 kg, which was sufficient to normalize liver lipid content and fasting plasma glucose concentrations as well as reverse hepatic insulin resistance in patients with obesity and type 2 diabetess. A potential mechanism explaining the improvement in insulin sensitivity is the reduction in intracellular diacylglycerol levels, which reduce insulin signaling in liver and muscle, 22,32,33 leading to tissue-specific insulin resistance. The effects of the dietary intervention on hepatocellular and intramyocellular lipid levels and insulin sensitivity—the presumed basis for the improved glycemic control—has not previously been quantified in clinical trials. Energy restriction has been shown to reduce intramyocellular and hepatocellular lipid levels and improve glycemic control in healthy young individuals without 26,34,35 diabetes. In young, lean individuals with insulin resistance, a hypocaloric diet (approximately 1200 kcal) led to a mean weight loss of 4.1 kg and a 30% reduction of intramyocellular lipids during a 9-week intervention. In contrast, the intervention diet in the present study did not restrict energy intake but nonetheless led to 34% and 10% reductions in hepatocellular and intramyocellular lipid levels, respectively. The reductions in hepatocellular and intramyocellular lipid levels correlated with 26,36,37 the reduction in fat mass, consistent with prior studies. The present finding that the increase in thermic effect of food was associated with decreased 38,39 fat mass and increased insulin sensitivity confirm the findings of previous research. The increased insulin sensitivity may have contributed to the increased postprandial metabolism. In addition, increased postprandial metabolism may have promoted further reduction in fat mass and an increase in insulin sensitivity. Despite the ad libitum diet, the participants in the intervention group reduced their energy intake, consistent with many previous trials using vegan diets. This not only contributes to weight loss but also may have contributed to the decrease in hepatocellular triglyceride content. 40-43 Postprandial metabolism is influenced by meal composition. In the present study, however, the test meal was identical for all study phases. These results suggest that the increased postprandial thermogenesis was attributable to improved insulin sensitivity. Table 3. Relationship Between Changes in Liver Fat and the First Predictive Component as Evaluated by OPLS Model OPLS predictive component Multiple regression a b Variable Component loading t Statistic R P value for R Regression coefficient t Statistic P value for t Matrix X Control group 0.339 9.88 0.795 .004 0.069 3.69 .007 Intervention group −0.339 −9.88 −0.795 .004 −0.069 −3.69 .007 Baseline PREDIM 0.214 8.89 0.498 .003 0.038 4.99 .005 Baseline HOMA −0.218 −2.71 −0.509 <.05 −0.060 −2.07 <.05 Baseline weight −0.228 −2.18 −0.535 <.05 −0.065 −2.19 <.05 Baseline fat mass −0.221 −2.18 −0.518 <.05 −0.058 −2.42 <.05 Change in PREDIM −0.199 −2.35 −0.468 <.05 −0.021 −2.54 <.05 Change in weight 0.388 14.13 0.910 .005 0.079 5.47 .005 Change in BMI 0.384 13.92 0.901 .005 0.077 5.64 .003 Change in fat mass 0.389 21.06 0.911 .002 0.078 5.87 .006 Change in visceral fat 0.341 8.23 0.798 .007 0.060 2.63 <.05 Matrix Y Change in liver fat 1.000 4.66 0.495 .009 NA NA NA Abbreviations: BMI, body mass index; HOMA, homeostasis model assessment; OPLS, Component loadings expressed as a correlation coefficients with predictive orthogonal projections to latent structure; NA, not applicable; PREDIM, predicted insulin component. sensitivity index. Explained variability was 24.5% (20.8% after cross-validation). JAMA Network Open. 2020;3(11):e2025454. doi:10.1001/jamanetworkopen.2020.25454 (Reprinted) November 30, 2020 10/14 JAMA Network Open | Nutrition, Obesity, and Exercise Effect of a Low-Fat Vegan Diet on Metabolic Measures in Overweight Adults Strengths and Limitations This study has several strengths. The randomized parallel design in which all participants within each cohort began the study simultaneously controlled for seasonal diet fluctuations. The study duration provided sufficient time for adaptation to the diet. Physiologic stimulation by a standard mixed meal permitted quantification of insulin sensitivity and insulin secretion during a physiologic perturbation. Measurement of visceral, hepatocellular, and intramyocellular lipid levels, in addition to the detailed assessment of the thermic effect of food, are also strengths. The low attrition suggests that the intervention was acceptable. The study also has limitations. Self-reports of dietary intake have well-known limitations. However, it is reassuring that the reported diet changes were paralleled by changes in weight and plasma lipid levels. Health-conscious participants may not be representative of the general population but may be representative of a clinical population seeking help for weight problems or type 2 diabetes. We followed the participants for 16 weeks and were not able to estimate the effects of the diet over a longer period. In addition, the study design did not allow separation of the specific effects of the low-fat vegan diet from the weight loss it causes. Conclusions This randomized clinical trial found that a low-fat plant-based dietary intervention reduces body weight by reducing energy intake and increasing postprandial metabolism, apparently owing to increased insulin sensitivity resulting from reduced hepatocellular and intramyocellular fat. This intervention may be an effective treatment for overweight adults. ARTICLE INFORMATION Accepted for Publication: September 17, 2020. Published: November 30, 2020. doi:10.1001/jamanetworkopen.2020.25454 Correction: This article was corrected on January 7, 2021, to edit the Role of the Funder/Sponsor section; on February 1, 2021, to fix errors in the abstract and main Results sections; and on May 27, 2021, to fix an error in the Results section and eTable 2 in Supplement 2. Open Access: This is an open access article distributed under the terms of the CC-BY License. © 2020 Kahleova H et al. JAMA Network Open. Corresponding Author: Hana Kahleova, MD, PhD, Physicians Committee for Responsible Medicine, 5100 Wisconsin Ave NW, Ste 400, Washington, DC 20016 (hkahleova@pcrm.org). Author Affiliations: Physicians Committee for Responsible Medicine, Washington, DC (Kahleova, Alwarith, Rembert, Barnard); Department of Internal Medicine, Yale School of Medicine, New Haven, Connecticut (Petersen, Shulman); Department of Cellular & Molecular Physiology, Yale School of Medicine, New Haven, Connecticut (Shulman); Metabolic Unit, CNR Institute of Neuroscience, Padua, Italy (Tura); Institute of Endocrinology, Prague, Czech Republic (Hill); School of Medicine, University of Utah, Salt Lake City (Holubkov); George Washington University School of Medicine and Health Sciences, Washington, DC (Barnard). Author Contributions: Drs Kahleova and Barnard had full access to all of the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis. Concept and design: Kahleova, Petersen, Shulman, Barnard. Acquisition, analysis, or interpretation of data: All authors. Drafting of the manuscript: Kahleova, Shulman, Alwarith, Rembert, Tura, Barnard. Critical revision of the manuscript for important intellectual content: Kahleova, Petersen, Shulman, Hill, Holubkov. Statistical analysis: Hill, Holubkov. Obtained funding: Petersen, Shulman. Administrative, technical, or material support: Kahleova, Petersen, Alwarith, Rembert. Supervision: Kahleova, Petersen, Shulman, Barnard. JAMA Network Open. 2020;3(11):e2025454. doi:10.1001/jamanetworkopen.2020.25454 (Reprinted) November 30, 2020 11/14 JAMA Network Open | Nutrition, Obesity, and Exercise Effect of a Low-Fat Vegan Diet on Metabolic Measures in Overweight Adults Conflict of Interest Disclosures: Dr Kahleova reported being director of clinical research at the Physicians Committee, a nonprofit organization that provides nutrition education and research. Dr Rembert reported compensation from the Physicians Committee for Responsible Medicine outside the submitted work. Dr Holubkov reported receiving personal fees from the Physicians Committee for Responsible Medicine during the conduct of the study. Dr Barnard reported to serving as president of the Physicians Committee for Responsible Medicine and Barnard Medical Center; receiving royalties from Hachette Book Group, Penguin Random House, Rodale, and Da Capo publishers; and receiving honoraria from Yale, Rush, George Washington, Loma Linda, Rockford Universities, Montefiore Medical Center, the Mayo Clinic, Northwell Health, Christiana Care, Oticon, and the National Organization of Professional Athletes. No other disclosures were reported. Funding/Support: This work was funded by the Physicians Committee for Responsible Medicine and grants P30 DK-045735 and R01 DK-113984 from the Yale Diabetes Center (Drs Shulman and Petersen). Role of the Funder/Sponsor: The Yale Diabetes Center had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication. Drs Kahleova, Alwarith, and Rembert, as employees of Physicians Committee for Responsible Medicine, were involved in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication. The research team of the Physicians Committee for Responsible Medicine had full autonomy in all aspects of the study. Data Sharing Statement: See Supplement 3. REFERENCES 1. Qian F, Liu G, Hu FB, Bhupathiraju SN, Sun Q. Association between plant-based dietary patterns and risk of type 2 diabetes: a systematic review and meta-analysis. JAMA Intern Med. 2019. doi:10.1001/jamainternmed.2019.2195 2. Tonstad S, Butler T, Yan R, Fraser GE. Type of vegetarian diet, body weight, and prevalence of type 2 diabetes. Diabetes Care. 2009;32(5):791-796. doi:10.2337/dc08-1886 3. Barnard ND, Levin SM, Yokoyama Y. A systematic review and meta-analysis of changes in body weight in clinical trials of vegetarian diets. J Acad Nutr Diet. 2015;115(6):954-969. doi:10.1016/j.jand.2014.11.016 4. Barnard ND, Scialli AR, Turner-McGrievy G, Lanou AJ, Glass J. The effects of a low-fat, plant-based dietary intervention on body weight, metabolism, and insulin sensitivity. Am J Med. 2005;118(9):991-997. doi:10.1016/j. amjmed.2005.03.039 5. National Cholesterol Education Program. Summary of the second report of the National Cholesterol Education Program (NCEP) Expert Panel on Detection, Evaluation, and Treatment of High Blood Cholesterol in Adults (Adult Treatment Panel II). JAMA. 1993;269(23):3015-3023. doi:10.1001/jama.1993.03500230097036 6. Goff LM, Bell JD, So P-W, Dornhorst A, Frost GS. Veganism and its relationship with insulin resistance and intramyocellular lipid. Eur J Clin Nutr. 2005;59(2):291-298. doi:10.1038/sj.ejcn.1602076 7. Gojda J, Patková J, Jaček M, et al. Higher insulin sensitivity in vegans is not associated with higher mitochondrial density. Eur J Clin Nutr. 2013;67(12):1310-1315. doi:10.1038/ejcn.2013.202 8. Itani SI, Ruderman NB, Schmieder F, Boden G. Lipid-induced insulin resistance in human muscle is associated with changes in diacylglycerol, protein kinase C, and IkappaB-alpha. Diabetes. 2002;51(7):2005-2011. doi:10. 2337/diabetes.51.7.2005 9. Goodpaster BH, Thaete FL, Kelley DE. Thigh adipose tissue distribution is associated with insulin resistance in obesity and in type 2 diabetes mellitus. Am J Clin Nutr. 2000;71(4):885-892. doi:10.1093/ajcn/71.4.885 10. Yatsuya H, Nihashi T, Li Y, et al. Independent association of liver fat accumulation with insulin resistance. Obes Res Clin Pract. 2014;8(4):e350-e355. doi:10.1016/j.orcp.2013.08.002 11. Schulz KF, Altman DG, Moher D; CONSORT Group. CONSORT 2010 statement: updated guidelines for reporting parallel group randomised trials. BMC Med. 2010;8(1):18. doi:10.1186/1741-7015-8-18 12. Schakel SF, Sievert YA, Buzzard IM. Sources of data for developing and maintaining a nutrient database. JAm Diet Assoc. 1988;88(10):1268-1271. 13. Hagströmer M, Oja P, Sjöström M. The International Physical Activity Questionnaire (IPAQ): a study of concurrent and construct validity. Public Health Nutr. 2006;9(6):755-762. doi:10.1079/PHN2005898 14. Kaul S, Rothney MP, Peters DM, et al. Dual-energy X-ray absorptiometry for quantification of visceral fat. Obesity (Silver Spring). 2012;20(6):1313-1318. doi:10.1038/oby.2011.393 15. Neeland IJ, Grundy SM, Li X, Adams-Huet B, Vega GL. Comparison of visceral fat mass measurement by dual- X-ray absorptiometry and magnetic resonance imaging in a multiethnic cohort: the Dallas Heart Study. Nutr Diabetes. 2016;6(7):e221. doi:10.1038/nutd.2016.28 JAMA Network Open. 2020;3(11):e2025454. doi:10.1001/jamanetworkopen.2020.25454 (Reprinted) November 30, 2020 12/14 JAMA Network Open | Nutrition, Obesity, and Exercise Effect of a Low-Fat Vegan Diet on Metabolic Measures in Overweight Adults 16. Mellis MG, Oldroyd B, Hind K. In vivo precision of the GE Lunar iDXA for the measurement of visceral adipose tissue in adults: the influence of body mass index. Eur J Clin Nutr. 2014;68(12):1365-1367. doi:10.1038/ejcn. 2014.213 17. Carver TE, Christou NV, Andersen RE. In vivo precision of the GE iDXA for the assessment of total body composition and fat distribution in severely obese patients. Obesity (Silver Spring). 2013;21(7):1367-1369. doi:10. 1002/oby.20323 18. Matthews DR, Hosker JP, Rudenski AS, Naylor BA, Treacher DF, Turner RC. Homeostasis model assessment: insulin resistance and beta-cell function from fasting plasma glucose and insulin concentrations in man. Diabetologia. 1985;28(7):412-419. doi:10.1007/BF00280883 19. Tura A, Chemello G, Szendroedi J, et al. Prediction of clamp-derived insulin sensitivity from the oral glucose insulin sensitivity index. Diabetologia. 2018;61(5):1135-1141. doi:10.1007/s00125-018-4568-4 20. Kaviani S, Schoeller DA, Ravussin E, et al. Determining the accuracy and reliability of indirect calorimeters utilizing the methanol combustion technique. Nutr Clin Pract. 2018;33(2):206-216. doi:10.1002/ncp.10070 21. Blond E, Maitrepierre C, Normand S, et al. A new indirect calorimeter is accurate and reliable for measuring basal energy expenditure, thermic effect of food and substrate oxidation in obese and healthy subjects. e-SPEN, the European e-Journal of Clinical Nutrition and Metabolism 2011;6(1):e7–e15. doi:10.1016/j.eclnm.2010.12.001 22. Petersen KF, Dufour S, Feng J, et al. Increased prevalence of insulin resistance and nonalcoholic fatty liver disease in Asian-Indian men. Proc Natl Acad SciUSA. 2006;103(48):18273-18277. doi:10.1073/pnas.0608537103 23. Lee SS, Lee Y, Kim N, et al. Hepatic fat quantification using chemical shift MR imaging and MR spectroscopy in the presence of hepatic iron deposition: validation in phantoms and in patients with chronic liver disease. J Magn Reson Imaging. 2011;33(6):1390-1398. doi:10.1002/jmri.22583 24. Rabøl R, Petersen KF, Dufour S, Flannery C, Shulman GI. Reversal of muscle insulin resistance with exercise reduces postprandial hepatic de novo lipogenesis in insulin resistant individuals. Proc Natl Acad SciUSA. 2011;108 (33):13705-13709. doi:10.1073/pnas.1110105108 25. Gruetter R. Automatic, localized in vivo adjustment of all first- and second-order shim coils. Magn Reson Med. 1993;29(6):804-811. doi:10.1002/mrm.1910290613 26. Petersen KF, Dufour S, Morino K, Yoo PS, Cline GW, Shulman GI. Reversal of muscle insulin resistance by weight reduction in young, lean, insulin-resistant offspring of parents with type 2 diabetes. Proc Natl Acad Sci U S A. 2012;109(21):8236-8240. doi:10.1073/pnas.1205675109 27. Trygg J, Holmes E, Lundstedt T. Chemometrics in metabonomics. J Proteome Res. 2007;6(2):469-479. doi:10. 1021/pr060594q 28. Sosvorova L, Hill M, Mohapl M, Vitku J, Hampl R. Steroid hormones in prediction of normal pressure hydrocephalus. J Steroid Biochem Mol Biol. 2015;152:124-132. doi:10.1016/j.jsbmb.2015.05.004 29. Romeo S, Kozlitina J, Xing C, et al. Genetic variation in PNPLA3 confers susceptibility to nonalcoholic fatty liver disease. Nat Genet. 2008;40(12):1461-1465. doi:10.1038/ng.257 30. Szczepaniak LS, Nurenberg P, Leonard D, et al. Magnetic resonance spectroscopy to measure hepatic triglyceride content: prevalence of hepatic steatosis in the general population. Am J Physiol Endocrinol Metab. 2005;288(2):E462-E468. doi:10.1152/ajpendo.00064.2004 31. Petersen KF, Dufour S, Befroy D, Lehrke M, Hendler RE, Shulman GI. Reversal of nonalcoholic hepatic steatosis, hepatic insulin resistance, and hyperglycemia by moderate weight reduction in patients with type 2 diabetes. Diabetes. 2005;54(3):603-608. doi:10.2337/diabetes.54.3.603 32. Petersen MC, Shulman GI. Mechanisms of insulin action and insulin resistance. Physiol Rev. 2018;98(4): 2133-2223. doi:10.1152/physrev.00063.2017 33. Samuel VT, Shulman GI. Mechanisms for insulin resistance: common threads and missing links. Cell. 2012;148 (5):852-871. doi:10.1016/j.cell.2012.02.017 34. Lara-Castro C, Newcomer BR, Rowell J, et al. Effects of short-term very low-calorie diet on intramyocellular lipid and insulin sensitivity in nondiabetic and type 2 diabetic subjects. Metabolism. 2008;57(1):1-8. doi:10.1016/j. metabol.2007.05.008 35. Jazet IM, Schaart G, Gastaldelli A, et al. Loss of 50% of excess weight using a very low energy diet improves insulin-stimulated glucose disposal and skeletal muscle insulin signalling in obese insulin-treated type 2 diabetic patients. Diabetologia. 2008;51(2):309-319. doi:10.1007/s00125-007-0862-2 36. Sinha R, Dufour S, Petersen KF, et al. Assessment of skeletal muscle triglyceride content by (1)H nuclear magnetic resonance spectroscopy in lean and obese adolescents: relationships to insulin sensitivity, total body fat, and central adiposity. Diabetes. 2002;51(4):1022-1027. doi:10.2337/diabetes.51.4.1022 JAMA Network Open. 2020;3(11):e2025454. doi:10.1001/jamanetworkopen.2020.25454 (Reprinted) November 30, 2020 13/14 JAMA Network Open | Nutrition, Obesity, and Exercise Effect of a Low-Fat Vegan Diet on Metabolic Measures in Overweight Adults 37. Moro C, Galgani JE, Luu L, et al. Influence of gender, obesity, and muscle lipase activity on intramyocellular lipids in sedentary individuals. J Clin Endocrinol Metab. 2009;94(9):3440-3447. doi:10.1210/jc.2009-0053 38. Camastra S, Bonora E, Del Prato S, Rett K, Weck M, Ferrannini E; EGIR (European Group for the Study of Insulin Resistance). Effect of obesity and insulin resistance on resting and glucose-induced thermogenesis in man. Int J Obes Relat Metab Disord. 1999;23(12):1307-1313. doi:10.1038/sj.ijo.0801072 39. Ravussin E, Acheson KJ, Vernet O, Danforth E, Jéquier E. Evidence that insulin resistance is responsible for the decreased thermic effect of glucose in human obesity. J Clin Invest. 1985;76(3):1268-1273. doi:10.1172/JCI112083 40. Bowden VL, McMurray RG. Effects of training status on the metabolic responses to high carbohydrate and high fat meals. Int J Sport Nutr Exerc Metab. 2000;10(1):16-27. doi:10.1123/ijsnem.10.1.16 41. Nagai N, Sakane N, Moritani T. Metabolic responses to high-fat or low-fat meals and association with sympathetic nervous system activity in healthy young men. J Nutr Sci Vitaminol (Tokyo). 2005;51(5):355-360. doi: 10.3177/jnsv.51.355 42. Thyfault JP, Richmond SR, Carper MJ, Potteiger JA, Hulver MW. Postprandial metabolism in resistance-trained versus sedentary males. Med Sci Sports Exerc. 2004;36(4):709-716. doi:10.1249/01.MSS.0000121946.98885.F5 43. Barr SB, Wright JC. Postprandial energy expenditure in whole-food and processed-food meals: implications for daily energy expenditure. Food Nutr Res. 2010;54:54. doi:10.3402/fnr.v54i0.5144 44. Yuan C, Spiegelman D, Rimm EB, et al. Relative validity of nutrient intakes assessed by questionnaire, 24-hour recalls, and diet records compared with urinary recovery and plasma concentration biomarkers: findings for women. Am J Epidemiol. 2017. doi:10.1093/aje/kww104 SUPPLEMENT 1. Trial Protocol SUPPLEMENT 2. eTable 1. Baseline characteristics of the study population eTable 2. Baseline characteristics of the study population, comparing study completers and drop-outs eTable 3. Baseline characteristics of the study population, comparing the subsample undergoing magnetic resonance spectroscopy (MRS) with the rest of the study population eTable 4. Treatment effects for the main outcomes, adjusted for age and race eFigure 1. Changes in the thermic effect of food, liver fat, and intramyocellular lipids after adjustment for race and age eFigure 2. Changes in liver fat and intramyocellular lipids after adjustment for baseline BMI eFigure 3. Linear regression model for changes in energy and body weight and postprandial energy expenditure and body weight SUPPLEMENT 3. Data Sharing Statement JAMA Network Open. 2020;3(11):e2025454. doi:10.1001/jamanetworkopen.2020.25454 (Reprinted) November 30, 2020 14/14 1 A Randomized, Controlled Trial on Diet, Insulin Sensitivity, 2 and Postprandial Metabolism 6 Summary 8 This randomized, controlled trial aims to elucidate the mechanisms by which a plant- 9 based dietary intervention causes weight loss. Using a low-fat, plant-based diet for 16 10 weeks, along with an untreated control for comparison, the study will measure changes 11 in insulin sensitivity, postprandial metabolism, and intracellular lipid, and assess their 12 associations with changes in body weight. 14 1. SPECIFIC AIMS AND OVERVIEW 15 1.1. Specific Aims 16 Specific Aim 1. This study tests the hypothesis that weight changes associated with a 17 low-fat plant-based diet are, in part, the result of increased postprandial metabolism 18 (thermic effect of food). 19 Specific Aim 2. This study tests the hypothesis that increased postprandial metabolism 20 in response to a diet intervention is the result of increased insulin sensitivity. 21 Specific Aim 3. This study conducts a pilot substudy to test the hypothesis that 22 changes in insulin sensitivity observed in response to a diet intervention correlate with 23 changes in intramyocellular and/or intrahepatocellular lipid. 25 1.2. Protocol Overview 27 In a 16-week trial, overweight adults will be randomly assigned to two groups. Changes 28 in insulin sensitivity, postprandial metabolism, and body weight will be the primary 29 dependent variables. 31 The Diet Group will be asked to follow a low-fat, vegan diet and will receive weekly 32 classes and support. 34 The Control Group will be asked to make no changes in diet or exercise for 16 weeks, 35 but will be instruct 37 1.3. Investigative Team 39 The project will be conducted by investigators from the Physicians Committee for 40 Responsible Medicine (PCRM), a nonprofit 501(c)(3) organization located at 5100 41 Wisconsin Avenue, NW, Washington DC 20016, which conducts nutrition-related 42 research. Its medical, nutrition, and research staff will oversee participant recruitment, 43 screening, group assignment, nutrition teaching and monitoring, and data collection and 44 analysis. Recruitment interviews, the dietary intervention, and most assessments will 45 take place at its offices. 47 Researchers from the Department of Internal Medicine, Yale University School of 48 Medicine, will conduct MR spectroscopy studies for intracellular lipid. 50 2. BACKGROUND AND SIGNIFICANCE 52 Excess body weight is a major contributor to many health problems, including diabetes, 53 cardiovascular disease, orthopedic problems, and certain forms of cancer. In 54 epidemiologic studies, individuals following vegan diets tend to have significantly lower 55 body weights, compared with individuals following other dietary patterns. In studies of 56 overweight individuals, the adoption of a low-fat plant-based diet predictably reduces 57 body weight, even in the absence of any specified limitation on energy intake. 58 The mechanisms by which plant-based diets reduce body weight are not entirely clear. 59 Previous studies have identified two possible explanations: First, to the extent that 60 vegan diets are low in fat and high in fiber, they have a relatively low energy density, 61 which reduces energy intake. Second, a low-fat vegan diet may increase postprandial 62 metabolism (the thermic effect of food). These observations suggest that the diet leads 63 to weight loss by (1) reducing energy intake and (2) increasing postprandial energy 64 output. 65 A prior study including 64 overweight postmenopausal women randomly assigned to a 66 low-fat vegan diet or a comparison diet based on the guidelines of the National 67 Cholesterol Education Program for 14 weeks found that the vegan diet led to 68 significantly greater weight loss (5.8 kg for the vegan group, compared with 3.8 kg for 69 the comparison group). The vegan diet group also had a 16% increase in postprandial 70 metabolism and an increase in insulin sensitivity that was significant within group, 71 although not between groups. However, because the comparison group used an active 72 diet and there was no untreated control group, that study was not able to show the 73 degree to which a plant-based diet influences energy expenditure, compared with 74 untreated participants. 75 Insulin resistance has been shown to be related to fat accumulation within muscle cells 76 (intramyocellular lipid) and liver cells (intrahepatocellular lipid). The above findings 77 suggest the possibility that low-fat, vegan diets reduce the quantity of lipid stored within 78 these cells, which, in turn, improves insulin sensitivity. 79 Some evidence suggests that the accumulation of intracellular fat may be responsive to 80 diet. In a 2012 study at Yale University, 7 lean, young individuals who were known to be 81 insulin-resistant and who had parents with type 2 diabetes underwent a hypocaloric 82 (1200 kcal/d) diet for 9 weeks, leading to an average weight loss of 4.1 ± 0.6 kg. During 83 this intervention period, average intramyocellular lipid fell approximately 30%, from 1.1 ± 84 0.2% to 0.8 ± 0.1%. 85 High-fat diets appear to downregulate the genes required for mitochondrial oxidative 86 phosphorylation in skeletal muscle and increase intramyocellular lipid. In contrast, a 87 case-control study found that soleus muscle intramyocellular lipid concentrations were 88 significantly lower in a group of 21 vegans, compared with 25 omnivores. 89 In research studies, the acceptability of plant-based diets appears to be similar to that of 90 other therapeutic diets over both the short and long term, as indicated by rates of 8-11 91 retention, diet adherence, and diet acceptance questionnaires. If a plant-based diet 92 increases postprandial metabolism, its use for the prevention and management of 93 weight disorders and related health problems will have a more solid rationale. This is 94 especially important given that weight problems are widespread and there is a great 95 deal of confusion among the public regarding which diet method to select. 97 3. RESEARCH DESIGN, RECRUITMENT, AND ASSESSMENTS 99 3.1. Overview of Research Design 101 In a randomized, controlled trial, we will test the effects of a low-fat, plant-based diet on 102 insulin sensitivity, postprandial metabolism, and body weight in overweight adults over a 103 16-week period, using for comparison an untreated control. A substudy will examine 104 effects on intracellular lipid. 106 3.2. Key Personnel 108 Key personnel include: 109 Neal D. Barnard, MD, FACC, Principal Investigator, is an Adjunct Associate Professor 110 of Medicine at the George Washington University and President of PCRM. He has been 111 the Principal Investigator of several clinical trials, as noted herein. 112 Hana Kahleova, MD, PhD, is an endocrinologist and Director of Clinical Research at 113 PCRM. She has been involved in several clinical trials in diabetes and insulin 114 resistance. 115 Susan Levin, MS, RD; Karen Smith, RD, and Maggie Neola, RD, are Registered 116 Dietitians at PCRM who provide nutrition instruction and participate in clinical 117 assessments. 118 Francesca Valente and Rosendo Flores coordinate clinical research studies at 119 PCRM. 120 Richard Holubkov, PhD, is a biostatistician with the University of Utah, who works with 121 PCRM on contract. 123 Gerald I. Shulman, MD, PhD, FACP, MACE, is a Professor of Medicine, Cellular and 124 Molecular Physiology and The Howard Hughes Medical Institutes at Yale University 125 School of Medicine. 126 Kitt Falk Petersen, MD, is a Professor of Medicine at Yale University School of 127 Medicine. 129 3.3. Recruitment and Screening Procedures 132 offices, letters sent to patients of medical practitioners, and advertisements placed in 133 newspapers, on radio, and in buses in the Washington, DC, area, as well as social 134 media postings. (Appendix 1) 137 participants using a telephone screening script (Appendix 2). Research staff will 138 explain the study, review participation criteria, and inquire about other motivations for 139 volunteering, filling out a paper interview screening form for each person who calls. 140 Volunteers who satisfy the participation criteria will be scheduled for group and/or 141 individual information sessions. The names/identifiable information of volunteers who do 142 not satisfy participation criteria will be destroyed (shredded) immediately. For these 143 individuals, the research team will retain only de-identified demographic information and 144 the reason for exclusion, for purposes of evaluating participation statistics. 146 At the group and/or individual information sessions (some volunteers may attend 147 individually; others may be seen in groups), the investigators and research staff will 148 explain the study and its scientific basis in detail and review participation criteria in 149 simple, nontechnical terms. They will also provide instruction on filling out a diet record. 150 Additional content will be determined by questions raised by volunteers, and may relate 151 to study logistics, the recruitment process, the content of the vegan diet and ease of 152 following it, clinical assessments, or the weekly classes. To protect patient privacy, 153 volunteer names will not be used at these meetings. Volunteers will have a chance to 154 ask questions about the study and the informed consent process in private, and each 155 volunteer will meet in private with study personnel at the conclusion of the group 156 session, even if he or she has no questions. Volunteers who choose to complete the 157 informed consent document (Appendix 3) will also be asked to complete a contact 158 information form (Appendix 4) and a general medical history form (Appendix 5). 159 Volunteers will be assigned identification numbers in the order in which they complete 160 the informed consent document. These numbers will be used in place of identifying 161 information for purposes of data collection, assessment, and analysis. During the 162 screening process, prospective volunteer 163 phone number and contact information in order to be able to reschedule and cancel 164 their appointments, if necessary. 166 Prospective volunteers will then be asked to complete a practice 2-day dietary record to 167 demonstrate their ability to track nutrient intake for research purposes. A letter will be 168 sent to the participants to remind them to complete their 2-day diet record with 169 instructions (Appendix 6). When completed, these records will be reviewed for 170 completeness by a registered dietitian. 172 Volunteers who have completed the informed consent process and practice dietary 173 records and meet the study participation criteria will be asked to schedule individual 174 appointments for baseline assessments. There, they will be asked to fill out new 3-day 175 dietary records. Volunteers can either submit questionnaires online, print, scan and 176 email them, or send them through regular mail. Those who satisfactorily complete the 177 baseline assessments and 3-day dietary records will be enrolled in the study. 179 The cost of all tests and procedures will be covered by PCRM. If any examination or 180 test reveals that a participant has a medical condition that requires additional diagnostic 181 tests or treatment, research staff will advise the participant of that fact, but will not 182 provide such additional diagnostic tests or treatment. There is no cost for the weekly 184 assessments, meetings, and group sessions. 187 3.4. Inclusion and Exclusion Criteria 189 Inclusion criteria are as follows: 191 1. Men and women age 18 years of age 192 2. Body mass index 28-40 kg/m 194 Exclusion criteria are as follows: 196 1. Diabetes mellitus, type 1 or 2, history of diabetes mellitus or of any endocrine 197 condition that would affect body weight, such as thyroid disease, pituitary 198 abnormality, or Cushing's syndrome 199 2. Smoking during the past six months 200 3. Alcohol consumption of more than 2 drinks per day or the equivalent, episodic 201 increased drinking (e.g., more than 2 drinks per day on weekends), or a history of 202 alcohol abuse or dependency followed by any current use 203 4. Use of recreational drugs in the past 6 months 204 5. Use within the preceding six months of medications that affect appetite or body 205 weight, such as estrogens or other hormones, thyroid medications (unstable 206 dose within the preceding 6 months), systemic steroids, antidepressants 207 (tricyclics, MAOIs, SSRIs), antipsychotics, lithium, anticonvulsants, appetite 208 suppressants or other weight-loss drugs, herbs for weight loss or mood, St. 209 John's wort, ephedra, beta blockers 210 6. Pregnancy or intention to become pregnant during the study period 211 7. Unstable medical or psychiatric illness 212 8. Evidence of an eating disorder 213 9. Likely to be disruptive in group sessions 214 10. Already following a low-fat, vegan diet 215 11. Lack of English fluency 216 12. Inability to maintain current medication regimen 217 13. Inability or unwillingness to participate in all components of the study 218 14. Intention to follow another weight-loss method during the trial 221 3.5. Group Assignment 223 Participants will be told that, if accepted, they will be assigned either to a Diet Group or 224 a Control Group. Accepted volunteers will be assigned to these groups using a 225 computer-generated random-number table. Because assignment will be done 226 simultaneously within each replication, allocation concealment is unnecessary. 228 3.6. Clinical Assessments 230 The following determinations will be made at baseline and 16 weeks, except as noted: 232 Assessments of Dietary Intake and Physical Activity 234 3-day dietary record. A 3-day dietary record will be used to assess macro- and 235 micronutrient intakes. Records will be analyzed using Nutrition Data System for 236 Research software version 2016, developed by the Nutrition Coordinating Center 237 (NCC), University of Minnesota, Minneapolis, MN, US, by a registered dietitian certified 238 by the NCC. Random 24-hour dietary recalls will be conducted by registered dietitians 239 to determine compliance but will not be part of the final nutrient analysis. 241 The International Physical Activity Questionnaire short form assesses recent 242 physical activity patterns. The method is highly reliable; an assessment of test-retest 243 repeatability produced a correlation of 0.8. (Appendix 7) 245 Assessments of Physical Health, Weight, and Metabolism 247 General status, symptoms, and medication accounting. Participants will be asked to 248 report changes in their health and medication use. 250 Height. Height will be measured at baseline (only) with participants standing barefoot 251 with their backs to a wall-mounted stadiometer and heels against the wall, recorded to 252 the nearest 0.5 cm. 254 Body weight. With participants wearing light, indoor clothing but without shoes, body 255 weight will be measured to the nearest 0.1 kg, using a digital scale. Body weight will 256 also be assessed at each weekly group session for the Diet Group, but only data from 257 weeks 0, 8, and 16 will be included in the analysis. 259 Comprehensive Metabolic Panel. These values will be evaluated at baseline only. 261 Serum cholesterol and triacylglycerol concentrations and hemoglobin A1c will be 262 measured using standard methods. 264 The following measures will be assessed at weeks 0 and 16: 266 Glucose Tolerance and Insulin Sensitivity. An oral glucose tolerance test will be 267 performed for three hours after an overnight fast. (Matsuda 1999) 268 Resting Energy Expenditure (REE). Participants will be asked to report to the 269 laboratory within 60 minutes of waking and after a 12-hour fast. Following 30 minutes of 270 quiet rest in a dimly lit room, pulse, respiratory rate, and body temperature will be 271 measured. REE will be measured for 20 minutes through indirect calorimetry (Cosmed 272 Quark RMR, Chicago, IL) utilizing a ventilated hood system. The laboratory temperature 273 will be maintained at 23 degrees C throughout, and precautions will be taken to 274 minimize any disturbances that could affect the metabolic rate. 275 For premenopausal women, measures will be timed so as to occur in the luteal phase of 276 the menstrual cycle. 277 Postprandial metabolism (thermic effect of food, TEF). After the REE determination, 278 participants will be given a 720-kilocalorie test meal (Sustacal, Mead Johnson, 279 Evansville, IN) to be ingested within 10 minutes. Metabolic rate will be measured in the 280 same manner as above for 30 minutes at 2 and 4 hours postingestion. 281 Body Composition. Body composition will be measured by dual energy x-ray 282 absorptometry (Lunar iDXA, GE Healthcare; Madison, WI) with Encore® 2005 283 v.9.15.010 software. The iDXA can measure body composition with low X ray exposure 284 and short scanning time. The iDXA unit will be calibrated daily using the GE Lunar 285 calibration phantom, and a trained operator will perform all scans following standard 286 protocol for participant positioning. The iDXA is equipped with the CoreScan module 287 (GE Healthcare, Madison, WI), which can also provide an estimate of visceral adipose 288 tissue volume and mass. 289 Intramyocellular and Hepatic Lipid Content. A subset of participants will be selected 290 for MR spectroscopy studies quantifying hepatic lipid and/or intramyocellular and/or 291 contents in order to provide data regarding possible causal relationships between 292 dietary changes, ectopic lipid, and insulin sensitivity. Selected individuals with varying 293 degrees of insulin-resistance in both groups will be assessed before and after the 294 intervention period. These MRS studies will take place at the Magnetic Research Center 295 at Yale University School of Medicine, New Haven, CT. 297 Intramyocellular and hepatic lipid contents will be measured using H MRS at 4T 298 (Bruker). After safety procedures including completion of the Yale Magnetic Research 299 Center Safety Questionnaire, changing into scrubs and passing through the metal 300 detector, the participants will positioned on their back on the bed, which slides into the 301 MRS instrument. For the leg lipid measurements the right leg will be positioned in a 302 holder with the calf muscle over a receiver coil. 304 Muscle lipid content will be measured in the soleus muscle using an 8.5-cm diameter 13 1 305 circular C surface coil with twin, orthogonal circular 13-cm H quadrature coils. The 306 probe will be as tuned and matched and scout images of the lower leg will be obtained 307 to ensure correct positioning of the participant and to define an adequate volume for 308 localized shimming using the FASTMAP procedure. The measurement will take 309 approximately 30 minutes. 311 After the lipid measurements in the leg a receiver coil embedded in a plastic plate will 312 be positioned on the side of the abdomen over the liver and strapped in place with velco 313 straps. The position of the coil will be confirmed with MR images and the location of the 314 lipid measurement within the coil will be determined from these MR images.Liver 315 triglyceride content will be measured by H respiratory-gated STEAM spectroscopy in a 316 15 × 15 × 15-mm3 voxel. Acquisition will be synchronized to the respiratory cycle and 317 triggered at the end of expiration. A water-suppressed lipid spectrum and a lipid- 318 suppressed water spectrum will be acquired in three different locations of the liver to 319 account for liver inhomogeneity. A minimum of three spectra will be acquired for each 320 participant and the total lipid content will be averaged and calculated. In addition, 321 hepatic lipid content will be corrected for transverse relaxation, using the transverse 322 relaxation times of 22 ms for water and 44 ms for lipid. These MRS measurements will 323 take approximately 30 minutes. 326 Table 1: Study Procedures Schedule Week 0 16 3-day diet record International Physical Activity Questionnaire Clinical status and symptoms Medication use Height Body weight* Comprehensive Metabolic Panel (CMP) Plasma lipids and lipoproteins A1c Glucose tolerance testing REE Postprandial metabolism (TEF) Body composition IMCL and hepatic lipid (pilot) 330 4. INTERVENTION PROCEDURES 332 4.1. Intervention Diet 334 The interventions for the Diet and Control Groups are described below. 336 The Diet Group will be asked to follow a low-fat, vegan diet. According to the Academy 337 of Nutrition and Dietetics, vegan and vegetarian diets meet all nutritional requirements 338 when appropriately planned. The diet consists of whole grains, vegetables, legumes, 339 and fruits, with no restriction on energy intake. Animal products and added oils will be 340 excluded. In choosing grain products and starchy vegetables (e.g., bread, potatoes), 341 participants will be encouraged to select those retaining their natural fiber and having a 342 glycemic index <70, using tables standardized to a value of 100 for glucose. No meals 343 will be provided. Participants will handle their own food preparation and purchases, with 344 guidance from the research team. 346 The diet derives approximately 10% of energy from fat, approximately 10-15% of energy 347 from protein, and the remainder from complex carbohydrates. The diet will provide 348 approximately 30-40 grams of fiber per day. It is generally adequate in all nutrients 349 except vitamin B . 351 Participants will be provided with a commercially available supplement containing 100 352 micrograms of vitamin B and asked to take it daily during the study. Should they wish 353 to continue the diet thereafter, they will be counseled to use any standard multivitamin 354 or other reliable source of vitamin B . 356 An advantage of studies such as this one, which include volunteers who are not 357 confined to a metabolic ward or otherwise restricted, is that they can readily translate to 358 nonclinical settings. A disadvantage is that they include a degree of uncertainty as to 359 the extent to which participants have adhered to their prescribed diets. While this 360 uncertainty cannot be entirely eliminated, several measures will be taken to maximize 362 associated with dietary compliance in clinical trials. Stricter limits on fat intake, frequent 363 monitoring of reported dietary intake, family involvement, group support, and the use of 364 vegetarian diets are associated with a greater degree of dietary change. 366 Control Group members will be asked to continue their usual diets for the 16-week 367 study period. Those who wish to try the intervention diet will be given instruction in the 370 Both groups: For both groups, alcoholic beverages will be limited to one per day for 371 women, and two for men. 373 4.2. Dietary Instruction and Group Meetings 375 Diet Group participants will be asked to attend weekly, one-hour group sessions for 376 support and education. (Class Curriculum, Appendix 8). No weekly support or 377 education will be provided to the participants in the Control Group. 379 All group sessions will be conducted by a registered dietitian, nurse, physician, cooking 380 instructor, or research staff and will include information on nutrition, meal planning, 381 shopping, food preparation techniques, recipes, and everyday dietary challenges, such 382 as dining out and healthful snacking. The classes will also include education on topics 383 such as maintaining a healthy weight, cholesterol, hypertension, diabetes, and other 384 health issues. 386 For some sessions, participants will be encouraged to bring a spouse, partner, family 387 member, or friend. To facilitate interaction between diet instructors and participants, 388 classes will be conducted in sections of approximately 15 participants. 391 Health Belief Model developed by researchers with the Public Health Service and 392 adapted by others. This model describes constructs that predict health-related 393 behaviors and should be considered when planning behavioral change strategies. 394 These include perceived susceptibility, severity, benefits, and barriers, as well as cues 395 to action, and self-efficacy. Our participants are already aware that they are overweight 396 and may benefit from diet changes. Nonetheless, they need help in overcoming barriers 397 and gaining confidence in their ability to implement new dietary habits. We have 398 therefore focused the content of the weekly support group sessions on integrating 399 practical skills (e.g., menu planning, food preparation, dining out, healthful snacking) 400 with their growing understanding of how dietary choices affect health. In order to 401 facilitate individual experience with the prescribed diet, practical skills are presented 402 early, while intellectual understanding of more complex health issues (e.g., how diet 403 affects heart disease risk) is presented later. Each group session includes time for 404 participants to discuss their successes and challenges, and group problem-solving is 405 encouraged. 407 The study does not seek to separate the effects of the diet from those of regular group 408 support. Rather, group support is a means of facilitating adherence. It should also be 409 emphasized that the goal of this study is not to construct an intervention diet that is 410 isocaloric compared with diets followed by the Control Group participants. Because the 411 intervention diet is low in fat and high in fiber, self-selected energy intake is likely to fall 412 as the diet period begins, and weight loss is likely. 414 4.3. Exercise and Medication Use 416 Participants in both groups will be asked to keep their level of physical exercise and use 417 of medications constant and to add no new nutritional supplements to their current 418 medication regimens, except as recommended by their personal physicians. 420 4.4. Intervention Fidelity and Dietary Adherence 422 Individual meetings. During the initial individual meal-planning meetings with the Diet 423 Group participants, dietitians will follow a set agenda which will cover the use of vegan 424 foods, methods for reducing dietary fat, and how to avoid proscribed foods. 426 Group meetings. To maintain intervention fidelity, the group leaders will follow a set 427 course curriculum, using an agenda for each session and keeping a checklist of major 428 content items to be covered at each meeting. 430 Dietary Adherence. Each participant will complete diet records at regular intervals 431 using the methods described above. In addition, for the Diet Group, 24-hour multi-pass 432 dietary recalls will be used to assess dietary adherence to assist study personnel in 433 working with individuals who need additional teaching or support. The 24-hour recalls 434 will be performed either by telephone or in person at weeks 3 and 8. These recalls will 435 not be subjected to statistical analysis, but will allow the investigators to check for poor 436 adherence. Such recalls have the advantage that they can be conducted at 437 unscheduled times and over the telephone, and so are not subject to the planning and 438 preparation required for food records. In cases where participants appear to be 439 deviating from the prescribed diet, additional dietary counseling will be provided. 441 4.5. Participant Retention 444 reasons for volunteering other than a desire to improve their health or to advance 445 scientific understanding may be rejected. The exclusion criteria also eliminate 446 individuals with a history of unresolved substance abuse, which may influence retention. 448 Participants will be instructed that attendance at meetings is essential to study 449 participation. The research team will take attendance at each meeting. The research 450 staff will make phone calls to participants who do not attend. 452 In the weekly meetings, group support will be facilitated through group discussions and 453 encouragement to share successes and difficulties with the prescribed diet. Meeting 454 content will remain varied, including nutrition lectures, health education, cooking 455 demonstrations, and opportunities to taste food. Family members will be invited to 456 certain support group sessions. A voluntary listserv will allow Diet Group members to 457 exchange information, recipes and ideas between meetings. Only participants, study 458 coordinators, and the PI will be allowed to post on the list serve and the content of the 459 list serve will be accessible only by them. 461 Participants who complete all assessments at weeks 0 and 16 will be paid $100 at 462 completion of their final assessments. 464 4.6. Biological Specimen Handling Procedures 466 Samples for the study endpoints will be drawn by LabCorp and will be processed using 467 standard procedures. 470 5. STATISTICAL PROCEDURES 472 5.1. Power Analysis 474 Power Analysis for Overall Study 476 Sample size will be based on the change in postprandial metabolism (thermic effect of 477 food) previously observed with a plant-based diet, compared with an active dietary 478 control. The current power analysis assumes there will be a single t-test for the 479 comparison of the changes in thermic effect of food observed in the two study groups, 480 with an alpha level of 0.05. 481 In the prior study, the change at 14 weeks for the thermal effect of food was 4.7 with an 482 SD of 12 in the intervention arm, and 0.3 with an SD of 9.4 in the control arm. Assuming 483 that the true treatment effect is 4.4 kcal/170 min and that the SD of the change will be 484 12.0 units in the Diet Group and 9.4 units in the Control Group, as previously observed, 485 the sample size required is 96 per arm (192 total) for 80% power, 109 per arm for 85% 486 power, 128 per arm for 90% power. 487 However, if this magnitude of treatment effect is assumed to be present at the same 488 magnitude of effect at each of the 5 evaluation points used in TEF assessment and 489 assuming that the standard deviation is about 10.85 points for all observations, with 5 490 observations per participant, correlated at a magnitude of 0.7 with each other, the 491 sample size required is 73 participants per arm (146 total) with 80% power, 83 per arm 492 with 85% power, 98 per arm for 90% power. 493 Assuming an attrition of 10%, the required sample size is 81 per group, or 162 total for 494 80% power. 496 Power Analysis for Intracellular Lipid Substudy 498 Two studies, cited above, provide a basis for a power analysis for the substudy 499 assessing the role of intramyocellular and hepatic lipid on insulin sensitivity and, 500 ultimately, postprandial metabolism. In the 2012 Yale study, 7 lean, young insulin- 501 resistant individuals whose parents had type 2 diabetes followed a hypocaloric (1200 502 kcal/d) diet for 9 weeks, leading to an average weight loss of 4.1 ± 0.6 kg. During this 503 intervention period, average intramyocellular lipid fell approximately 30%, from 1.1 ± 504 0.2% to 0.8 ± 0.1%. In an observational study including 21 individuals following vegan 505 diets and 25 following omnivorous diets, soleus muscle intramyocellular lipid for the 506 vegan participants was found to be 11.7 (6.1 24.6), compared with 16.9 (2.7 44.7) for 507 the omnivorous participants (P = 0.01). The 95% confidence interval for the difference 508 was reported to be -13.2 to -3.3). 510 Based on the Yale study, assuming a change in IMCL of 0.3 percentage points with a 511 standard deviation of 0.2 and, in the control arm, a mean change of zero with a similar 512 standard deviation, to have 90% power to detect a difference of this magnitude between 513 the two arms would require 11 subjects per arm. Ten per arm would yield 88% power. 514 Because this is an exploratory substudy and variability in response to the diet is largely 515 unknown, we aim to include 20 participants per arm in the substudy, for a total of 40 516 participants. 518 5.2. Data Management 520 All laboratory samples, reports, questionnaires, and data sheets will be coded with 521 participant identification numbers, rather than names. Laboratory reports will be 522 delivered to the PCRM office at 5100 Wisconsin Avenue, Washington, D.C., where they, 523 along with all other history and data forms, will be maintained in individual participant 524 files in a locked cabinet. 526 Data will be promptly entered into the data tables at PCRM using Microsoft Excel. Two 527 research staff members will check the tables for accuracy against the original 528 documents. Data tables will be routinely copied onto back-up files and stored for safety 529 on an off-site, passcode-protected, secure server. Data grids will be sent electronically 530 to the biostatistician for analysis. Registered dietitians will be provided with information 531 on usual ranges for nutrients or intakes of interest and asked to check their original data 532 and analysis for errors if they fall outside of these ranges. 534 5.3. Statistical Analysis 536 Descriptive statistics for all demographic variables and clinical measures will be 537 calculated for each group. To determine if there are statistically significant differences 538 between the 2 groups at baseline, t-tests will be calculated for continuous measures 539 and chi squares will be calculated for categorical measures. Regardless of any 540 differences, baseline values for key outcome variables will be included as covariates in 541 the main assessments of the effect of diet in the multivariate analysis of covariance. An 542 alpha of 0.05 will be used for all statistical tests. 544 For nutrient intake and physical measures, descriptive statistics (means, standard 545 deviations, tests for normality) will be calculated. If data are normally distributed, 546 parametric tests for significant effects will be used; for non-normally distributed 547 variables, non-parametric tests will be used. 549 The initial test of the hypotheses will be examined by performing t-tests for independent 550 samples on the difference score denoting the change from baseline to the reporting 551 period. For missing data in a reporting period, values from the previous period will be 552 brought forward. For body weight, drop-outs will be considered to have returned to 553 baseline weights. 555 5.4. Assessment of Diet Adherence. Diet-and-supplement group participants will be 556 described as adherent or non-adherent based on whether they met the following 557 criteria: absence of proscribed foods reported on 24-hour recalls and diet records, 558 saturated fat <5% and total fat <25% of energy, and average daily cholesterol intake 559 <50 mg on 3-day dietary records. 561 For drop-out rates, we will determine if there are between-group differences, using chi- 562 square. 564 5.5. Assessment of Medication use. Any changes to lipid-lowering medications will be 565 classified as a net increase, net decrease, or mixed (changes in opposing directions for 566 2 or more medications). Using chi-square, we will determine whether there are 567 differences in medication changes between the 2 groups. 570 6. TIME LINE AND PARTICIPANT FLOW 572 6.1. Time Line 573 The study will be conducted in six cohorts, each including 40 participants, over a 3-year 574 period. 576 For the first cohort, recruitment will take place October 2016 January 2017. The 577 intervention, including weekly meetings, will take place between February and May 578 2017 for a total of 16 weeks. For the second cohort, recruitment will take place in 579 December 2016 and January, 2017. The intervention, including weekly meetings, will 580 take place between March and July 2017. Similar time frames will apply in the two 581 subsequent years. 583 6.2. Participant Flow Based on Power Analysis 585 In order to accommodate this number of participants, baseline metabolic assessments 586 will occur as follows for the first cohort: 588 January 6: 2 participants 589 January 9-13: 10 participants 590 January 17-20: 8 participants 591 January 23-27: 10 participants 592 January 30-Feb 3: 10 participants 594 The MR spectroscopy studies will be limited to 40 participants (20 per study arm) total. 595 These evaluations will be scheduled during the above scheduled days. 597 The Diet Group sessions for the first cohort will be held for 16 weeks, from February 1 598 through May 17, 2017. 600 For the first cohort in 2017, the 16-week metabolic assessments will occur as follows: 602 May 18-19: 4 participants 603 May 22-26: 10 participants 604 May 30-June 2: 8 participants 605 June 5-9: 10 participants 606 June 12-15: 8 participants 608 For the second cohort, the assessment and intervention dates will be approximately one 609 month later than those for the first cohort. 611 7. PROTECTION OF HUMAN RESEARCH PARTICIPANTS 613 7.1. Risks to the Subjects 615 Sources of Materials: Participants will be asked to complete questionnaires, provide 616 blood samples, and have several physical assessments. 618 Human Subjects Involvement and Characteristics: The proposed research will 619 include participants at least 18 years of age. 621 Potential Risks: Participation in the study entails the following risks: 623 1. Blood draws can cause transient pain, occasionally cause bruising, and may 624 cause bleeding. 626 2. A well-planned vegan diet provides all the nutrients people need except for 627 vitamin B12. People with a vitamin B12 deficiency may suffer from anemia and 628 neurologic damage. 630 3. Loss of confidential information. 632 7.2. Adequacy of Protection against Risks 634 Recruitment and Informed Consent: 635 and procedures and review the inclusion and exclusion criteria. Volunteers who appear 636 to meet the criteria for participation will be invited to a group or individual interview with 637 the principal investigator and the study coordinator, who will explain the study in detail, 638 answer questions, and provide a written consent form, as approved by the IRB. 639 Participants will have the opportunity to ask any questions individually in a private 640 setting and may take as much time as they would like to review the informed consent 641 document. The consent form will be signed by the volunteer participant and study 642 coordinator. The principal investigator will certify that the research study has been 643 explained to the volunteer, including the purpose, procedures, possible risks, and 644 potential benefits associated with participation and that any questions have been 647 he study and 648 that the investigators will not manage any aspects of their medical care. 650 To maintain confidentiality, all laboratory specimens, questionnaires, forms, and data 651 sheets will identify participants by their assigned numbers only. Data and safety 652 monitoring are described below. 654 Phlebotomy risks. All blood draws will be carried out by experienced phlebotomists at 655 Quest Diagnostics. 657 Vitamin B12 deficiency. All Diet Group participants will be given a supply of vitamin 658 B12, 100 micrograms, and will be asked to take it daily. Participants will also be 659 counseled to continue B12 supplementation if they plan to continue following a vegan 660 diet. 662 Loss of confidential information. We will make every effort to keep all research 663 records private to the extent allowed by law. We will use an identification number on 664 forms, instead of identifiable information. All study documents will be kept in locked filing 666 learn from this study may be shared at scientific or medical meetings and may be 667 published, but participants will not be personally identified. 669 Our protocol also includes the following safeguards: 671 1. All participants will remain under the care of their personal healthcare providers. 673 2. All participants will continue on the medications they were using at study entry, 674 unless modified by their personal physician(s). 676 We therefore believe that the risks to participants in a dietary intervention trial are 677 minimal, while the scientific and public health merit of such an investigation is high. By 678 studying the benefits of a dietary intervention, we hope to obtain valuable research data. 680 Confidentiality. To maintain confidentiality, all laboratory specimens, questionnaires, 681 forms, and data sheets will identify participants by their assigned numbers only. Data 682 and safety monitoring are described below. 684 7.3. Potential Benefits of the Proposed Research to the Participants and Others 686 Given that, over the long run, excess body weight contributes to morbidity and mortality, 687 the dietary instruction and consistent support provided may be of substantial benefits for 688 the Diet Group. The Control Group will be given detailed information on how to follow 691 7.4. Importance of the Knowledge to Be Gained 693 Weight problems are extremely common, and gaps remain in our understanding of how 694 intervention diets work. This study is founded on clear theoretical constructs and 695 compelling previous data on both the efficacy and acceptability of the experimental 696 intervention, as well as preliminary findings on its mechanisms of action. It investigates 697 what may be a major advance in the understanding of the role of diet in weight control. 698 The risks to participants are small, and the potential benefits are significant. 700 7.5. Assessment and Reporting of Adverse Events 702 An adverse event is any adverse physical or clinical change experienced by a 703 participant. This includes the onset of new symptoms and the exacerbation of pre- 704 existing conditions. In order to avoid bias in eliciting reports of adverse events, 705 participants will be asked, during assessments at the 706 you had any new symptoms, injuries, illness or side effects or worsening of pre-existing 709 All adverse events will be recorded in the participant's record and on the IRB continuing 710 review form. The severity of the adverse event will be assessed, and actions/outcomes 711 (e.g., hospitalization, discontinuation of therapy, etc.) will also be recorded. 713 Any actions taken and follow-up results will also be recorded on the appropriate page of 714 the IRB continuing review form, as well as in the participant's record. Follow-up 716 at a site will be reported by the investigator to the IRB according to the Data and Safety 717 Monitoring Plan, described below. 719 The following definitions will be used: Minimally serious: Awareness of sign, symptom, or event, but easily tolerated. Somewhat Serious: Discomfort enough to cause interference with usual activity and may warrant investigation. Very Serious: Incapacitating, with inability to do usual activities, or significantly affects clinical status, and warrants intervention. Life-threatening: Immediate risk of death. 721 The research team will also assess the relationship of any adverse event to the study 722 intervention, based on available information, using the following guidelines: 0 = No temporal association, or the cause of the event has been identified, or Unlikely the study interventions cannot be implicated. 1 = Temporal association, but other etiologies are likely the cause; however, Possibly involvement of the study interventions cannot be excluded. 2 = Temporal association or other etiologies are possible, but unlikely. Probably 724 7.6. Serious Adverse Events (SAEs) 726 All serious adverse events, whether or not deemed intervention-related or expected, will 727 be reported by telephone to the Safety Officer within 24 hours (one working day) of the 728 time they become known. A written report will follow as soon as possible, including a full 729 description of the event and any sequelae. This includes serious events that occur any 730 time after the inclusion of the patient in the study until completion of the last visit. A 731 serious adverse event report will also be sent via fax to the IRB chair. 733 A serious adverse event is any event that falls in any of the following categories: 734 Fatal 735 Life-threatening (the patient was at immediate risk of death from the AE as it 736 occurred) 737 Significantly or permanently disabling 738 Requires hospitalization or prolongs hospitalization 740 Important medical events that may not result in death, be life-threatening, or require 741 hospitalization may be considered serious adverse events when, upon appropriate 742 medical judgment, they may jeopardize the patient and may require medical or surgical 743 intervention to prevent one of the outcomes listed in the definition. The death of any 744 patient during the study, regardless of the cause, will be reported within 24 hours by 745 telephone to the Safety Officer and IRB. A full written report will follow as soon as 746 possible. If an autopsy is performed, a copy will be provided to the Safety Officer and 747 IRB. 749 Reports of all serious adverse events, including deaths, will be communicated to the 750 IRB in accordance with local laws and regulations. 752 7.7. Action plan if a subject becomes severely depressed or suicidal during the 753 course of the study 755 Participants with a history of severe mental illness (with current unstable status), such 756 as severe depression or suicidality, will not be enrolled in the study as indicated in the 757 exclusion criteria. If a participant becomes severely depressed during the course of the 758 study, he or she will be referred to see his or her primary care physician or psychiatrist 759 and to seek medical care. The event will be recorded and reported to the PI and Safety 760 Officer immediately. The PI will notify IRB within 24 hours of recognition of the event by 761 study personnel. All non-serious events will be reported and reviewed by the PI within 762 one week. Study personnel will inform the primary care physician of all events occurring 763 in his/her patients within 48 hours of recognition of the event. 765 If a participant becomes suicidal during the course of the study, he or she will be 766 instructed to call 911 and seek emergency medical care. The incident will be recorded 767 and reported to the PI and Safety Officer immediately. The PI will notify the IRB within 768 24 hours of recognition of the event by study personnel. All non-serious events will be 769 reported and reviewed by the PI within one week. Study personnel will inform the 770 primary care physician of all events occurring in his/her patients within 48 hours of 771 recognition of the event. 773 7.8. Data and Safety Monitoring Plan 775 Data and Safety Monitoring functions will be performed by the principal investigator (PI, 776 Neal D. Barnard, M.D.), study coordinator (Francesca Valente), study statistician, and a 777 Safety Officer who is a physician who is not part of the research staff and has no role in 778 care of the participants. The Safety Officer will have no scientific, financial, or other 779 conflict of interest related to the trial. Prior to the study onset, the study statistician and 780 Safety Officer will review the research protocol, informed consent documents, and plans 781 for data and safety monitoring. 783 During the recruitment phase, the study coordinator and PI will review enrollment 785 including accrual, demographics, thoroughness of baseline data, subject status 786 (reporting concurrent illnesses, withdrawal of consent, or loss to follow-up), and 787 adherence to participation criteria, informed consent procedures, and the study protocol. 788 The reports will be submitted to the study statistician and Safety Officer. 790 -existing 791 symptoms and medical problems will be recorded. During each monitoring visit, study 792 participants will be asked if any medical symptom, problem, or event has occurred or if 793 there has been any change in pre-existing symptoms. 795 All serious events (hospitalization, serious illness, or disability) will be recorded and 796 reported to the PI and Safety Officer. The PI will notify the IRB within 24 hours of 797 recognition of the event by study personnel. All non-serious events will be reported and 798 reviewed by the PI within one week. Study personnel will inform the primary care 799 physician of all events occurring in his/her patients within 48 hours of recognition of the 800 event. 802 At monthly intervals, the study coordinator and PI will prepare and submit to the study 803 statistician and Safety Officer a report covering each of the following areas: (1) 804 performance (including adherence to the study protocol and maintenance of data 805 integrity and confidentiality), (2) safety (including abnormal laboratory values, adverse 806 events, serious adverse events, deaths, and disease- or treatment-specific events), and 807 (3) treatment effects, including medication changes. 809 The study statistician and Safety Officer will review each safety report within one week 810 of receipt. The study statistician will review these reports to assess whether event rates 811 are of statistical concern and, if so, will alert the Safety Officer, the PI, and the IRB. The 812 study statistician and Safety Officer will also consider factors external to the study, e.g., 813 new scientific developments, that may affect the safety of participants or the conduct of 814 the trial. 816 The Safety Officer will make recommendations as necessary to the PI. If the Safety 817 Officer recommends a study change for patient safety or for ethical reasons, or if the 818 study is closed early due to slow accrual, the PI will be responsible for implementing the 819 recommendations as expeditiously as possible. If the PI does not concur with any 820 recommendation of the Safety Officer, both will be responsible for reaching a mutually 821 acceptable decision. 823 7.9. Stopping Rules 825 At the conclusion of the 16-week intervention period for the first cohort, the study 826 statistician will prepare a report on clinical changes and adverse events for presentation 827 to the PI and the Safety Officer. If evidence available at that point clearly shows either 828 (1) an effect of the intervention diet on postprandial metabolism or (2) harm associated 829 with the intervention diet, the Safety Officer may recommend early termination of the 830 study. 832 8. INCLUSION OF WOMEN, MINORITIES, AND CHILDREN 834 8.1. Inclusion of Women 836 The participation criteria, cited above, include both men and women. Recruitment 837 procedures are expected to yield roughly equal numbers of men and women. 839 8.2. Inclusion of Minorities 841 The U.S. Census Bureau reports both race and ethnicity, the latter term used primarily 842 to denote self-identification as Hispanic or non-Hispanic. According to the 2015 Census 843 Bureau, races were represented in Washington, DC, as follows: 48.3% black, 44.1% 844 white, 4.2% Asian, 0.6% American Indian/Native American, 0.2% Native Hawaiian or 845 other Pacific Islander; 2.7% 2 or more races. In addition, 10.6% of the population 846 identified themselves as Hispanic. In our prior studies, the respondent populations 847 have been demographically diverse, reflecting the profile of the greater Washington, 848 D.C area. 850 8.3. Inclusion of Children. 852 Persons less than 18 years of age will not be included in the study because they have 853 insufficient control over the dietary choices that are essential to meaningful participation. 856 9. BRIEF STATEMENT OF ANTICIPATED OUTCOMES 858 This study aims to test hypotheses that are potentially important for individual and public 859 health. It will improve our understanding of the treatment of weight problems and will 860 also have practical implications for reducing the medical, personal, and economic costs 861 associated with obesity. Anticipated outcomes for Diet Group participants include 862 beneficial changes in body weight, insulin sensitivity, and serum lipid concentrations, all 863 of which are also possible for the Control Group participants who choose to take 864 advantage of instruction in . 869 LITERATURE CITED Tonstad, S, Butler T, Yan R, Fraser GE.Type of vegetarian diet, body weight and prevalence of type 2 diabetes. Diabetes Care. 2009;32:791-6. Barnard ND, Levin SM, Yokoyama Y. A systematic review and meta-analysis of changes in body weight in clinical trials of vegetarian diets. J Acad Nutr Diet. 2015 Jun;115(6):954-69. Barnard ND, Scialli AR, Turner-McGrievy G, Lanou AJ, Glass J. The effects of a low- fat, plant-based dietary intervention on body weight, metabolism, and insulin sensitivity. Am J Med 2005;118:991-997. Shulman GI. Ectopic fat in insulin resistance, dyslipidemia, and cardiometabolic disease. N Engl J Med 2014;371:1131-1141 Petersen KF, Dufour S, Morino K, Yoo PS, Cline GW, Shulman GL. Reversal of muscle insulin resistance by weight reduction in young, lean, insulin-resistant offspring of parents with type 2 diabetes. PNAS. 2012;109:8236-40. Sparks LM, Xie H, Koza RA, et al. A high-fat diet coordinately downregulates genes required for mitochondrial oxidative phosphorylation in skeletal muscle. Diabetes. 2005;54:1926 33. Goff LM, Bell JD, So PW, Dornhorst A, Frost GS. Veganism and its relationship with insulin resistance and intramyocellular lipid. Eur J Clin Nutr. 2005;59:291 298. Barnard N, Scherwitz L, Ornish D. Adherence and acceptability of a lowfat vegetarian diet among patients with cardiac disease. J Cardiopulmonary Rehabil 1992;12:423-31. Barnard N, Scialli A, Bertron P, Hurlock D, Edmonds K. Acceptability of a therapeutic low-fat, vegan diet in premenopausal women. J Nutr Educ 2000;32:314-9. Barnard ND, Scialli AR, Turner-McGrievy G, Lanou AJ. Acceptability of a low-fat vegan diet compares favorably to a step II diet in a randomized, controlled trial. Journal of cardiopulmonary rehabilitation 2004;24(4):229-35. Barnard ND, Gloede L, Cohen J, et al. A low-fat vegan diet elicits greater macronutrient changes, but is comparable in adherence and acceptability, compared with a more conventional diabetes diet among individuals with type 2 diabetes. J Am Diet Assoc 2009;109(2):263-72. Craig CL, Marshall AL, Sjostrom M, et al. International physical activity questionnaire: 12-country reliability and validity. Medicine and science in sports and exercise 2003;35(8):1381-95. Petersen KF, et al. (2006) Increased prevalence of insulin resistance and nonalcoholic fatty liver disease in Asian-Indian men. Proc Natl Acad Sci USA 103:18273 18277. Gruetter R (1993) Automatic, localized in vivo adjustment of all first- and second- order shim coils. Magn Reson Med 29:804 811. Rabøl R, Petersen KF, Dufour S, Flannery C, Shulman GI (2011) Reversal of muscle insulin resistance with exercise reduces postprandial hepatic de novo lipogenesis in insulin resistant individuals. Proc Natl Acad Sci USA 108:13705 13709. Position of the American Dietetic Association and Dietitians of Canada: Vegetarian diets. J Am Diet Assoc 2003;103(6):748-65. Barnard N, Akhtar A, Nicholson A. Factors that facilitate dietary change. Arch Fam Med 1995;4:153-8. Becker M. The health belief model and personal health behavior. Health Education Monographs 1974;2:324-473. Buzzard I, Faucett C, Jeffery R, et al. Monitoring dietary change in a low-fat diet intervention study: advantages of using 24-hour dietary recalls vs food records. J Am Diet Assoc 1996;96:574. U.S. Census Bureau. Quick Facts. District of Columbia. Internet: http://www.census.gov/quickfacts/table/RHI125215/11, accessed August 22, 2016. Supplemental Online Content Kahleova H, Petersen KF, Shulman GI, et al. Effect of a low-fat vegan diet on body weight, insulin sensitivity, postprandial metabolism, and intramyocellular and hepatocellular lipid levels in overweight adults: a randomized clinical trial. JAMA Netw Open. 2020;3(11):e2025454. doi:10.1001/jamanetworkopen.2020.25454 eTable 1. Baseline characteristics of the study population eTable 2. Baseline characteristics of the study population, comparing study completers and drop- outs eTable 3. Baseline characteristics of the study population, comparing the subsample undergoing magnetic resonance spectroscopy (MRS) with the rest of the study population eTable 4. Treatment effects for the main outcomes, adjusted for age and race eFigure 1. Changes in the thermic effect of food, liver fat, and intramyocellular lipids after adjustment for race and age eFigure 2. Changes in liver fat and intramyocellular lipids after adjustment for baseline BMI eFigure 3. Linear regression model for changes in energy and body weight and postprandial energy expenditure and body weight This supplemental material has been provided by the authors to give readers additional information about their work. © 2020 Kahleova H et al. JAMA Network Open. Characteristic Intervention Control group P Value group (n=122) (n=122) Age (years, SD) 53 (±10) 57 (±13) 0.01 Sex (number, %) 0.85 Female 105 (86.0) 106 (86.8) Male 17 (14.0) 16 (13.1) Race, (number, %) 0.06 White 57 (46.7) 60 (49.2) Black 60 (49.2) 53 (43.4) Asian, Pacific Islander 1 (0.8) 7 (5.7) American Indian, Eskimo, Aleut 2 (1.6) 0 (0.0) Did not disclose 2 (1.6) 2 (1.6) Ethnicity, (number, %) 0.75 Non-Hispanic 97 (79.5) 101 (82.8) Hispanic 8 (6.6) 7 (5.7) Did not disclose 17 (13.9) 14 (11.5) Marital status, (number, %) 0.87 Not married 66 (54.1) 61 (50.0) Married 55 (45.1) 53 (43.4) Did not disclose 1 (0.8) 8 (6.6) Education, (number, %) 0.15 High school 8 (6.6) 11 (9.0) Associates 35 (28.7) 43 (35.2) College 72 (59.0) 63 (51.6) Graduate degree 7 (5.7) 5 (4.1) Occupation, (number, %) 0.23 Service occupation 27 (22.1) 15 (12.3) Technical, sales, administrative 34 (27.9) 32 (26.2) Professional or managerial 33 (27.0) 39 (32.0) Retired 16 (13.1) 24 (19.7) Other 12 (9.8) 12 (9.8) Medications, (number, %) Lipid-lowering therapy (%) 22 (18.0) 21 (17.2) 0.87 Antihypertensive therapy (%) 33 (27.0) 31 (25.4) 0.77 Thyroid medications (%) 16 (13.1) 12 (9.8) 0.42 Total physical activity (number, SD) 2719 (±4701) 2863 (±3085) 0.80 Dietary Intake (number, SD) Caloric intake (kcal/day) 1834 (±574) 1793 (±628) 0.61 © 2020 Kahleova H et al. JAMA Network Open. Total fiber intake (g/day) 24.1 (±10.6) 23.9 (±10.1) 0.89 Total cholesterol intake (mg/day) 238.6 (±143.9) 244.5 (±169.3) 0.78 Total saturated fatty acid intake 23.6 (±12.0) 22.9 (±12.2) 0.65 (g/day,) Monounsaturated fatty acids (g/day) 27.2 (±10.4) 27.9 (±13.9) 0.67 Polyunsaturated fatty acids (g/day) 18.4 (±8.2) 19.1 (±10.5) 0.59 Anthropometric variables, (number, SD) Body weight (kg) 93.6 (±13.8) 92.7 (±13.7) 0.62 BMI (kg/m ) 33.3 (±3.8) 33.6 (±3.7) 0.57 Lean mass (kg) 50.5 (±7.9) 49.5 (±8.1) 0.35 Fat mass (kg) 40.6 (±9.2) 40.9 (±9.6) 0.76 Visceral fat volume (cm ) 1459 (±944.2) 1517 (±907.0) 0.64 Laboratory variables (number, SD) Total cholesterol (mmol/l) 5.2 (±1.1) 5.0 (±1.3) 0.11 HDL-cholesterol (mmol/l) 1.6 (±0.4) 1.7 (±0.9) 0.16 LDL-cholesterol (mmol/l) 3.0 (±0.9) 2.9 (±1.1) 0.16 Triglycerides (mmol/l) 1.2 (±0.5) 1.3 (±0.6) 0.10 Fasting plasma glucose (mmol/l) 5.2 (±0.2) 5.0 (±0.3) 0.18 Fasting plasma insulin (pmol/l) 91.2 (±59.6) 78.9 (±51.1) 0.12 HbA1c (DCCT, %) 5.6 (±0.4) 5.7 (±0.4) 0.30 Insulin sensitivity /resistance (number, SD) PREDIM (mg/min/kg) 4.1 (±1.3) 4.4 (±1.5) 0.11 HOMA-IR (dimensionless) 3.2 (±2.2) 2.7 (±1.9) 0.17 eTable 1. Baseline characteristics of the study population. Data are means ± SD (standard deviation), or number (%). P-values refer to t- (chi-squared) categorical variables. The P-value calculated for ethnicity distribution is for the comparison between Hispanic vs. non-Hispanic categories (and all other comparisons also exclude datapoints that were not available). © 2020 Kahleova H et al. JAMA Network Open. Characteristic Drop-outs (n=22) Study completers P Value (n=223) Age (years, SD) 57.8 (±12.5) 54.4 (±11.6) 0.20 Sex (number, %) Female 19 (86.4) 192 (86.5) 1.0 Male 3 (13.6) 30 (13.5) Race, (number, %) 0.30 White 7 (31.8) 110 (49.6) Black 14 (63.6) 99 (44.6) Asian, Pacific Islander 1 (4.5) 7 (3.2) American Indian, Eskimo, Aleut 0 (0) 2 (0.9) Did not disclose 0 (0) 4 (1.8) Ethnicity, (number, %) 1.0 Non-Hispanic 18 (81.8) 180 (81.1) Hispanic 1 (4.5) 14 (6.3) Did not disclose 3 (13.6) 28 (12.6) Marital status, (number, %) 0.09 Not married 15 (68.2) 112 (50.5) Married 6 (27.3) 102 (45.9) Did not disclose 1 (4.5) 8 (3.6) Education, (number, %) 0.58 High school 3 (13.6) 16 (7.2) Associates 0 (0) 1 (0.5) © 2020 Kahleova H et al. JAMA Network Open. College 8 (36.4) 81 (36.5) Graduate degree 11 (50.0) 123 (55.4) Did not disclose 0 (0) 1 (0.5) Occupation, (number, %) 0.49 Service occupation 4 (18.2) 38 (17.1) Technical, sales, administrative 7 (31.8) 59 (26.6) Professional or managerial 3 (13.6) 69 (31.1) Retired 5 (22.7) 35 (15.8) Other 2 (9.1) 21 (9.5) Did not disclose 1 (4.5) 0 (0) Medications, (number, %) Lipid-lowering therapy (%) 6 (27.3) 39 (17.6) 0.24 Antihypertensive therapy (%) 6 (27.3) 59 (26.6) 0.84 Thyroid medications (%) 1 (4.5) 27 (12.2) 0.48 Total physical activity (number, 2863 (±3144) 2780 (±4006) 0.93 SD) Dietary Intake (number, SD) Caloric intake (kcal/day) 1589 (±442) 1812 (±599) 0.13 Total fiber intake (g/day) 20.3 (±6.8) 23.9 (±10.3) 0.16 Total cholesterol intake (mg/day) 197 (±126) 242 (±155) 0.25 Total saturated fatty acid intake 18.8 (±10.5) 23.2 (±12.1) 0.14 (g/day,) Monounsaturated fatty acids (g/day) 23.8 (±11.0) 27.5 (±12.1) 0.22 © 2020 Kahleova H et al. JAMA Network Open. Polyunsaturated fatty acids (g/day) 14.9 (±7.0) 18.6 (±9.3) 0.11 Anthropometric variables, (number, SD) Body weight (kg) 92.3 (±9.02 93.0 (±13.7) 0.74 BMI (kg/m ) 33.0 (±2.7) 33.4 (±3.7) 0.56 Lean mass (kg) 50.0 (±7.5) 49.9 (±8.0) 0.98 Fat mass (kg) 40.5 (±4.9) 40.7 (±9.3) 0.88 Visceral fat volume (cm ) 1355 (±575) 1496 (±914) 0.34 Laboratory variables (number, SD) Total cholesterol (mmol/l) 5.2 (±0.9) 5.2 (±1.1) 0.78 HDL-cholesterol (mmol/l) 1.5 (±0.4) 1.6 (±0.4) 0.55 LDL-cholesterol (mmol/l) 3.1 (±0.7) 3.1 (±0.9) 0.99 Triglycerides (mmol/l) 1.2 (±0.6) 1.2 (±0.5) 0.92 Fasting plasma glucose (mmol/l) 5.5 (±0.9) 5.4 (±0.6) 0.61 Fasting plasma insulin (pmol/l) 73.2 (±31.9) 78.8 (±62.2) 0.52 HbA1c (DCCT, %) 5.7 (±0.5) 5.7 (±0.4) 0.86 Insulin sensitivity /resistance (number, SD) PREDIM (mg/min/kg) 3.9 (±1.1) 4.2 (±1.4) 0.58 HOMA-IR (dimensionless) 3.2 (±1.8) 3.0 (±2.0) 0.78 © 2020 Kahleova H et al. JAMA Network Open. eTable 2. Baseline characteristics of the study population, comparing study completers and drop-outs. Data are means ± SD (standard deviation), or number (%). P-values refer to t-tests for (chi-squared) P-value calculated for ethnicity distribution is for the comparison between Hispanic vs. non- Hispanic categories (and all other comparisons also exclude datapoints that were not available). Characteristic MRS YES (n=44) MRS NO (n=200) P-value Age (years, SD) 55.8 (±11.1) 54.5 (±11.8) 0.52 Sex (number, %) Female 35 (79.5) 176 (88.0) 0.14 Male 9 (20.5) 24 (12.0) Race, (number, %) 0.07 White 25 (56.8) 92 (46.0) Black 15 (34.1) 98 (49.0) Asian, Pacific Islander 3 (6.8) 5 (2.5) American Indian, Eskimo, Aleut 1 (2.3) 1 (0.5) Did not disclose 0 4 (2.0) Ethnicity, (number, %) 0.50 Non-Hispanic 37 (84.1) 161 (80.5) Hispanic 4 (9.1) 11 (5.5) Did not disclose 3 (6.8) 28 (14.0) Marital status, (number, %) 0.81 Not married 22 (50.0) 105 (52.5) Married 20 (45.5) 88 (44.0) © 2020 Kahleova H et al. JAMA Network Open. Did not disclose 2 (4.5) 7 (3.5) Education, (number, %) High school 1 (2.3) 18 (9.0) Associates 0 (0) 1 (0.5) College 24 (54.5) 65 (32.5) Graduate degree 19 (43.2) 115 (57.5) Did not disclose 0 (0) 1 (0.5) Occupation, (number, %) 0.30 Service occupation 5 (11.4) 37 (18.5) Technical, sales, administrative 17 (38.6) 49 (24.5) Professional or managerial 12 (27.3) 60 (30.0) Retired 5 (11.4) 35 (17.5) Other 5 (11.4) 18 (9.0) Did not disclose 0 (0) 1 (0.5) Medications, (number, %) Lipid-lowering therapy (%) 7 (15.9) 38 (19.0) 0.62 Antihypertensive therapy (%) 10 (22.7) 55 (27.5) 0.51 Thyroid medications (%) 4 (9.1) 24 (12.0) 0.58 Total physical activity (number, 2711 (±3222) 2804 (±4080) 0.89 SD) Dietary Intake (number, SD) Caloric intake (kcal/day) 1886 (±629) 1776 (±582) 0.27 Total fiber intake (g/day) 25.7 (±11.9) 23.2 (±9.7) 0.14 © 2020 Kahleova H et al. JAMA Network Open. Total cholesterol intake (mg/day) 255.8 (±188) 234.6 (±145.0) 0.49 Total saturated fatty acid intake 24.3 (±12.0) 22.6 (±12.0) 0.39 (g/day,) Monounsaturated fatty acids (g/day) 29.2 (±14.9) 26.8 (±11.3) 0.34 Polyunsaturated fatty acids (g/day) 19.8 (±12.4) 18.0 (±8.3) 0.36 Anthropometric variables, (number, SD) Body weight (kg) 89.8 (±11.7) 93.6 (±13.7) 0.08 BMI (kg/m ) 32.1 (±2.3) 33.7 (±3.8) <0.001 Lean mass (kg) 50.5 (±8.5) 49.8 (±7.8) 0.60 Fat mass (kg) 37.3 (±7.0) 41.4 (±9.3) 0.002 Visceral fat volume (cm ) 1559 (±975) 1468 (±874) 0.55 Laboratory variables (number, SD) Total cholesterol (mmol/l) 5.4 (±1.2) 5.2 (±1.0) 0.27 HDL-cholesterol (mmol/l) 1.6 (±0.6) 1.6 (±0.4) 0.41 LDL-cholesterol (mmol/l) 3.2 (±0.9) 3.1 (±0.9) 0.42 Triglycerides (mmol/l) 1.2 (±0.6) 1.2 (±0.5) 0.91 Fasting plasma glucose (mmol/l) 5.5 (±0.7) 5.4 (±0.6) 0.35 Fasting plasma insulin (pmol/l) 66.4 (±53.3) 81.1 (±72.0) 0.07 HbA1c (DCCT, %) 5.9 (±0.5) 5.6 (±0.4) <0.001 Insulin sensitivity /resistance (number, SD) © 2020 Kahleova H et al. JAMA Network Open. PREDIM (mg/min/kg) 4.2 (±1.3) 4.2 (±1.4) 0.78 HOMA-IR (dimensionless) 2.8 (±1.9) 3.0 (±2.0) 0.58 eTable 3. Baseline characteristics of the study population, comparing the subsample undergoing magnetic resonance spectroscopy (MRS) with the rest of the study population. Data are means ± SD (standard deviation), or number (%). P-values refer to t-tests for (chi-squared) - value calculated for ethnicity distribution is for the comparison between Hispanic vs. non- Hispanic categories (and all other comparisons also exclude datapoints that were not available). Treatment effect P- Outcomes (adj.) value Weight (kg) -6.1 (-7.0 to -5.2) <.001 BMI (kg/m ) -2.4 (-3.6 to -1.2) <.001 Fat mass (kg) -4.1 (-4.7 to -3.5) <.001 Lean mass (kg) -1.5 (-2.0 to -1.1) <.001 VAT volume -217.4 (-316.2 to - <.001 (cm ) 118.7) © 2020 Kahleova H et al. JAMA Network Open. Hepatocellular -1.1 (-2.3 to -0.04) <.001 lipids (%) Intramyocellular -0.1 (-0.5 to +0.2) 0.02 lipids (%) PREDIM +0.8 (+0.5 to +1.1) <.001 HOMA -1.2 (-2.2 to -0.2) 0.01 eTable 4. Treatment effects for the main outcomes, adjusted for age and race. © 2020 Kahleova H et al. JAMA Network Open. © 2020 Kahleova H et al. JAMA Network Open. © 2020 Kahleova H et al. JAMA Network Open. © 2020 Kahleova H et al. JAMA Network Open. © 2020 Kahleova H et al. JAMA Network Open. http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png JAMA Network Open American Medical Association

Effect of a Low-Fat Vegan Diet on Body Weight, Insulin Sensitivity, Postprandial Metabolism, and Intramyocellular and Hepatocellular Lipid Levels in Overweight Adults

Loading next page...
 
/lp/american-medical-association/effect-of-a-low-fat-vegan-diet-on-body-weight-insulin-sensitivity-gysCK8ITcc

References (92)

Publisher
American Medical Association
Copyright
Copyright 2020 Kahleova H et al. JAMA Network Open.
eISSN
2574-3805
DOI
10.1001/jamanetworkopen.2020.25454
Publisher site
See Article on Publisher Site

Abstract

Key Points Question What are the effects of a IMPORTANCE Excess body weight and insulin resistance lead to type 2 diabetes and other major low-fat vegan diet on body weight, health problems. There is an urgent need for dietary interventions to address these conditions. insulin resistance, postprandial metabolism, and intramyocellular and OBJECTIVE To measure the effects of a low-fat vegan diet on body weight, insulin resistance, hepatocellular lipid levels in postprandial metabolism, and intramyocellular and hepatocellular lipid levels in overweight adults. overweight adults? Findings In this 16-week randomized DESIGN, SETTING, AND PARTICIPANTS This 16-week randomized clinical trial was conducted clinical trial, a low-fat plant-based between January 2017 and February 2019 in Washington, DC. Of 3115 people who responded to dietary intervention reduced body flyers in medical offices and newspaper and radio advertisements, 244 met the participation criteria weight by reducing energy intake and (age 25 to 75 years; body mass index of 28 to 40) after having been screened by telephone. increasing postprandial metabolism, which was associated with reductions in INTERVENTIONS Participants were randomized in a 1:1 ratio. The intervention group (n = 122) was hepatocellular and intramyocellular fat asked to follow a low-fat vegan diet and the control group (n = 122) to make no diet changes for and increased insulin sensitivity. 16 weeks. Meaning A low-fat plant-based diet is MAIN OUTCOMES AND MEASURES At weeks 0 and 16, body weight was assessed using a an effective tool for reducing body calibrated scale. Body composition and visceral fat were measured by dual x-ray absorptiometry. weight and increasing insulin sensitivity Insulin resistance was assessed with the homeostasis model assessment index and the predicted and postprandial metabolism. insulin sensitivity index (PREDIM). Thermic effect of food was measured by indirect calorimetry over 3 hours after a standard liquid breakfast (720 kcal). In a subset of participants (n = 44), Supplemental content hepatocellular and intramyocellular lipids were quantified by proton magnetic resonance spectroscopy. Repeated measure analysis of variance was used for statistical analysis. Author affiliations and article information are listed at the end of this article. RESULTS Among the 244 participants in the study, 211 (87%) were female, 117 (48%) were White, and the mean (SD) age was 54.4 (11.6) years. Over the 16 weeks, body weight decreased in the intervention group by 5.9 kg (95% CI, 5.0-6.7 kg; P < .001). Thermic effect of food increased in the intervention group from baseline to 16 weeks and did not change significantly in the control group (between-group difference in effect size, 14.1%; 95% CI, 6.5-20.4; P < .001). The homeostasis model assessment index decreased (−1.3; 95% CI, −2.2 to −0.3; P < .001) and PREDIM increased (0.9; 95% CI, 0.5-1.2; P < .001) in the intervention group. Hepatocellular lipid levels decreased in the intervention group by 34.4%, from a mean (SD) of 3.2% (2.9%) to 2.4% (2.2%) (P = .002), and intramyocellular lipid levels decreased by 10.4%, from a mean (SD) of 1.6 (1.1) to 1.5 (1.0) (P = .03). None of these variables changed significantly in the control group over the 16 weeks. The change in PREDIM correlated negatively with the change in body weight (r = −0.43; P < .001). Changes in (continued) Open Access. This is an open access article distributed under the terms of the CC-BY License. JAMA Network Open. 2020;3(11):e2025454. doi:10.1001/jamanetworkopen.2020.25454 (Reprinted) November 30, 2020 1/14 JAMA Network Open | Nutrition, Obesity, and Exercise Effect of a Low-Fat Vegan Diet on Metabolic Measures in Overweight Adults Abstract (continued) hepatocellular and intramyocellular lipid levels correlated with changes in insulin resistance (both r = 0.51; P = .01). CONCLUSIONS AND RELEVANCE A low-fat plant-based dietary intervention reduces body weight by reducing energy intake and increasing postprandial metabolism. The changes are associated with reductions in hepatocellular and intramyocellular fat and increased insulin sensitivity. TRIAL REGISTRATION ClinicalTrials.gov Identifier: NCT02939638 JAMA Network Open. 2020;3(11):e2025454. Corrected on May 27, 2021. doi:10.1001/jamanetworkopen.2020.25454 Introduction Overweight and associated diseases, particularly type 2 diabetes and metabolic syndrome, remain worldwide challenges. There is an urgent need for dietary interventions to address these problems and for a better understanding of how different dietary interventions work. 1,2 Obesity is uncommon in individuals whose diets are based on plant-derived foods. In clinical trials, such diets caused weight loss, for which 2 explanations have been offered. First, a high-fiber, low-fat diet has a low energy density, which reduces energy intake. Second, a low-fat, vegan diet increases the thermic effect of food, which accounts for approximately 10% of total energy expenditure. However, in the latter randomized clinical trial, the control group was following an active diet based on National Cholesterol Education Program guidelines. Because there was no untreated control group, the effect of a low-fat vegan diet on thermogenesis remains unclear. Studies have reported that people following a vegan diet have lower concentrations of intramyocellular lipids compared with those following omnivorous diets, suggesting that by reducing intramyocellular or hepatocellular lipid levels, a plant-based diet may lead to increased mitochondrial 6,7 activity and postprandial metabolism. This is particularly important because the accumulation of 8-10 lipids in muscle and liver cells may also be associated with insulin resistance and type 2 diabetes. The aim of this study was to measure the effects of a low-fat vegan diet on body weight, insulin resistance, postprandial metabolism, and intramyocellular and hepatocellular lipid levels in overweight adults. Methods Study Design and Eligibility This randomized clinical trial using a single-center, open parallel design was conducted between January 2017 and February 2019 in Washington, DC, in 4 replications (the trial protocol is given in Supplement 1). Adults aged 25 to 75 years with a body mass index (BMI) (calculated as weight in kilograms divided by height in meters squared) of 28 to 40 were enrolled. Exclusion criteria were diabetes, smoking, alcohol or drug use, pregnancy or lactation, and current use of a vegan diet. The additional exclusion criteria for the subset of participants undergoing the proton magnetic resonance spectroscopy were the presence of any metal implant, claustrophobia, BMI higher than 38, and waist circumference of more than 102 cm. The study protocol was approved by the Chesapeake Institutional Review Board. All participants gave written informed consent. This study followed the Consolidated Standards of Reporting Trials (CONSORT) reporting guideline. JAMA Network Open. 2020;3(11):e2025454. doi:10.1001/jamanetworkopen.2020.25454 (Reprinted) November 30, 2020 2/14 JAMA Network Open | Nutrition, Obesity, and Exercise Effect of a Low-Fat Vegan Diet on Metabolic Measures in Overweight Adults Randomization and Study Groups With use of a computer-generated system, participants were randomly assigned (in a 1:1 ratio) to an intervention group, which was asked to follow a low-fat vegan diet, or a control group, which was asked to make no diet changes. The randomization protocol could not be accessed by the participants or the staff allocating the participants into groups beforehand. Because assignment was done simultaneously, allocation concealment was unnecessary. The participants were not blinded to their group assignment. The intervention diet (approximately 75% of energy from carbohydrates, 15% protein, and 10% fat) consisted of vegetables, grains, legumes, and fruits without animal products or added fats. Vitamin B was supplemented (500 μg/d). The intervention group attended weekly classes for detailed instruction and cooking demonstrations and received printed materials and small food samples. No meals were provided. For both groups, alcoholic beverages were limited to 1 per day for women and 2 per day for men. All participants were asked to maintain their customary exercise habits and medications unless modified by their personal physicians. Outcomes All measurements were performed at baseline and 16 weeks. The outcome assessors (K.F.P., G.I.S., and A.T.) were blinded to group assignment. The primary outcomes were body weight, insulin resistance, postprandial metabolism, and the concentrations of intramyocellular and hepatocellular lipids. At baseline and at 16 weeks, dietary intake data over 3 consecutive days were collected and analyzed by staff members certified in the Nutrition Data System for Research, version 2016, developed by the Nutrition Coordinating Center, University of Minnesota, Minneapolis. In addition, study dietitians made unannounced telephone calls to assess participants’ dietary adherence. All study participants were asked not to alter their exercise habits and to continue their preexisting medication regimens for the duration of the study. Physical activity was assessed by the International Physical Activity Questionnaire. Laboratory assessments were made after an overnight fast. Height (baseline only) and weight were measured using a stadiometer and a calibrated digital scale, respectively. Body composition and visceral fat volume were assessed using dual energy x-ray absorptiometry (iDXA; GE Healthcare), 14 15 which has been validated against computed tomography and magnetic resonance imaging. The measurement of total body fat and visceral fat had a coefficient of variation (CV) of 1.0% and 5.4%, 16,17 respectively. Insulin secretion was assessed after a standardized liquid breakfast (Boost Plus, Nestle) (720 kcal, 34% of energy from fat, 16% protein, and 50% carbohydrate). Plasma glucose, immunoreactive insulin, and C-peptide concentrations were measured at 0, 30, 60, 120, and 180 minutes. Plasma glucose concentration was analyzed using the Hexokinase UV end point method (the intra-assay CV was 1.4%, and the inter-assay CV was 1.9%), and immunoreactive insulin and C-peptide concentrations were determined using insulin and C-peptide electro-chemiluminescence immunoassay (the intra-assay CVs were 5.1% and 3.8%, respectively, and the inter-assay CVs were 5.7% and 3.9%, respectively). Glycated hemoglobin level was measured by turbidimetric inhibition immunoassay (the intra-assay CV was 3.7%, and the inter-assay CV was 3.5%), and lipid concentrations were measured by enzymatic colorimetric methods (intra-assay CV: total cholesterol, 2.1%; high-density lipoprotein cholesterol, 2.4%; low-density lipoprotein cholesterol, 2.0%; and triglycerides 2.2%; inter-assay CV: total cholesterol, 2.7%; high-density lipoprotein cholesterol, 3.8%; low-density lipoprotein cholesterol, 3.0%; and triglycerides 3.2%). All test kits were made by Roche. Insulin resistance was calculated using the homeostasis model assessment index. The predicted insulin sensitivity index (PREDIM) provided a validated measure of dynamic insulin sensitivity. Resting energy expenditure and postprandial metabolism were measured by indirect calorimetry (Cosmed Quark CPET) using a ventilated hood system (accuracy of measurement with a JAMA Network Open. 2020;3(11):e2025454. doi:10.1001/jamanetworkopen.2020.25454 (Reprinted) November 30, 2020 3/14 JAMA Network Open | Nutrition, Obesity, and Exercise Effect of a Low-Fat Vegan Diet on Metabolic Measures in Overweight Adults 20,21 CV<1% and repeatability of measurement with a CV of 1.2%). Each measurement was performed for 15 minutes after an overnight fast and 30, 60, 120, and 180 minutes after the standard breakfast. In a subset of 44 participants (23 in the intervention group and 21 in the control group), proton magnetic resonance spectroscopy was performed at the Magnetic Resonance Research Center, Yale School of Medicine. Hepatocellular and intramyocellular lipids were quantified by proton magnetic resonance spectroscopy at 4T (Bruker). This method has been shown to provide a precise quantification of fat fractions, with a mean bias of −1.1.% to 0.5%. Hepatocellular lipid content was measured by H respiratory-gated stimulated echo acquisition mode spectroscopy in a 15 × 15 × 15-mm voxel. Acquisition was synchronized to the respiratory cycle and triggered at the end of expiration. A water-suppressed lipid spectrum and a lipid-suppressed water spectrum were acquired in 3 locations of the liver to account for liver inhomogeneity, and the total lipid content was averaged and calculated. In addition, hepatocellular lipid content was corrected for transverse relaxation using the transverse relaxation times of 22 ms for water and 44 ms for lipid. Intramyocellular lipid content was measured in the soleus muscle using an 8.5-cm diameter circular 13 1 C surface coil with twin, orthogonal circular 13-cm H quadrature coils. Scout images of the lower leg were obtained to ensure correct positioning of the participant and to define an adequate volume for localized shimming using the FastMap procedure. Power Analysis Sample size was based on the change in body weight, insulin resistance, and postprandial metabolism previously observed with a plant-based diet, with an α level of 0.05. The assumed change for body weight was a mean (SD) of 5.8 (3.2) kg in the intervention arm and 1 (3.2) kg in the control arm; for insulin sensitivity, the assumed change was 1.1 (2.1) in the intervention arm and 0.1 (2.1) in the control arm; and for the thermic effect of food, the assumed change was 4.7 (12) (area under the curve) in the intervention arm and 0.3 (9.4) in the control arm. For the primary efficacy comparison, a total of 22 participants (11 per arm) were required for 90% power to detect a significant treatment effect on body weight between the 2 study arms. For insulin sensitivity, a total of 142 participants (71 in each arm) were required for 90% power. Assuming that the treatment effect for postprandial metabolism was of the same magnitude at each of the 5 evaluation points used in metabolic assessment and that the SD was approximately 10.85 points for all observations, with 5 observations per participant correlated at a magnitude of 0.7 with each other, and assuming an attrition of 10%, the required sample size was 81 per group (162 total) for 80% power and 108 per group (216 total) for 90% power. For the substudy assessing the role of intramyocellular and hepatocellular lipids in insulin sensitivity, a study from 2012 provided a basis for a power analysis. In that study, 7 lean individuals with insulin resistance followed a hypocaloric (1200 kcal/d) diet for 9 weeks. The mean (SD) intramyocellular lipid level decreased from 1.1% (0.2%) to 0.8% (0.1%). Assuming a mean (SD) change in the intramyocellular lipid level of 0.3% (0.2%) in the intervention arm and a mean change of 0 with a similar SD in the control arm, to have 90% power to detect a difference of this magnitude between, the 2 arms would each require 11 individuals (22 total). Because this was an exploratory substudy and variability in response to the diet was largely unknown, 20 participants were recruited per arm (a total of 40 participants). Statistical Analysis For baseline characteristics, between-group t tests were performed for continuous variables and χ or Fisher exact test for categorical variables. A repeated measure analysis of variance (ANOVA) model was used with between-person and within-person factors and interactions, including group, person, and time. The interaction between group and time was calculated for each variable. For thermic effect of food, minutes were included in the ANOVA model. Data from only individuals with measurements at both time points were included in the ANOVA model. Within each group, paired JAMA Network Open. 2020;3(11):e2025454. doi:10.1001/jamanetworkopen.2020.25454 (Reprinted) November 30, 2020 4/14 JAMA Network Open | Nutrition, Obesity, and Exercise Effect of a Low-Fat Vegan Diet on Metabolic Measures in Overweight Adults comparison t tests were calculated to test whether the changes from baseline to 16 weeks were statistically significant. To eliminate skewed data distribution and heteroscedasticity, data were transformed to a gaussian distribution before further processing by a power transformation using the statistical software Statgraphics Centurion, version XV (Statpoint Inc). The transformed data underwent multivariable regression using the method of orthogonal projections to latent structure. This method is effective in addressing severe multicollinearity within the matrix of independent variables. In our model, changes in thermic effect of food and in hepatocellular lipid levels were chosen as the dependent variables and the metabolic variables (body weight, fat mass, visceral fat, and insulin resistance) represented the independent variables. The variability was separated into 2 independent components. The predictive component contained the variability in the metabolic variables, which was shared with changes in dependent variables, and the orthogonal component contained the variability shared within the metabolic variables. A detailed description of the orthogonal projections to latent structure model is available elsewhere. The statistical software SIMCA-P, version 11.5 (Umetrics AB) identified the number of relevant components using the prediction error sum of squares and also allowed the detection of multivariable nonhomogeneities and testing for multivariable normal distribution and homoscedasticity (constant variance). The statisticians (M.H, R.H.) were blinded to the interventions and group assignment. Results are presented as means with 95% CIs. Two-tailed tests were used to determine significance at the 5% level. Results Participant Characteristics Of 3115 people screened by telephone, 244 met the participation criteria, signed the consent form and were randomly assigned to the intervention (n = 122) or control (n = 122) groups in a 1:1 ratio (Figure 1). The mean (SD) age of the intervention group was 53 (10) years compared with 57 (13) years in the control group (P = .01) (eTable 1 in Supplement 2). There were no other significant differences between the groups. Five intervention group and 16 control group participants dropped out, mostly for reasons unrelated to the study, leaving 223 (91.0%) individuals who completed the study. eTable 2 in Supplement 2 shows the baseline characteristics of those who completed the study and those who dropped out. There were no significant differences between these groups. The main outcomes are reported in Table 1. The treatment effects were largely unaffected by the adjustment for age and race/ethnicity (eTable 4 in Supplement 2). eTable 3 in Supplement 2 shows the characteristics of the subgroup that underwent magnetic resonance spectroscopy. This group had a lower BMI compared with the rest of the study population. The model adjusted for baseline BMI for magnetic resonance spectroscopy is presented in eFigure 2 in Supplement 2. Dietary Intake and Physical Activity Self-reported energy intake decreased in both groups but more so in the intervention group (treatment effect, −354.9 kcal/d; 95% CI, −519.0 to −190.8 kcal/d; P < .001) (Table 2). In the intervention group, mean intakes of carbohydrate and fiber increased, whereas mean fat, protein, and cholesterol intake decreased. These values did not change significantly in the control group. Physical activity decreased slightly in both groups (−709.8 metabolic equivalents [95% CI, −1346 to −73.9 metabolic equivalents] in the control group and −604.8 metabolic equivalents [95% CI, −1388 to −178.6 metabolic equivalents] in the intervention group; between-group P = .84). Body Weight, Body Composition, and Blood Lipid Levels Mean body weight decreased by 6.4 kg in the intervention group compared with 0.5 kg in the control group (treatment effect, −5.9 kg; 95% CI, −6.7 to −5.0; interaction between group and time, P < .001). This difference was largely attributable to a reduction in body fat, as noted by significant decreases in fat mass and visceral fat volume in the intervention group participants. Total and JAMA Network Open. 2020;3(11):e2025454. doi:10.1001/jamanetworkopen.2020.25454 (Reprinted) November 30, 2020 5/14 JAMA Network Open | Nutrition, Obesity, and Exercise Effect of a Low-Fat Vegan Diet on Metabolic Measures in Overweight Adults low-density lipoprotein cholesterol levels decreased by 0.5 mmol/L and 0.4 mmol/L (to convert to milligrams per deciliter, divide by 0.0259), respectively, in the intervention group, with no significant changes in the control group (0.1 mmol/L and 0.07 mmol/L, respectively) (P < .001 for both). Insulin Sensitivity Fasting plasma insulin concentration decreased by 21.6 pmol/L (to convert to micro-IU per milliliter, divide by 6.945) in the intervention group, with no significant change in the control group (23.6 pmol/L; 95% CI, −5.0 to 54.3; between-group P = .006). The homeostasis model assessment index (a measure of insulin resistance) decreased significantly (−1.3; 95% CI, −2.2 to −0.3; P < .001), and PREDIM (a measure of insulin sensitivity) increased significantly in the intervention group (0.9; 95% CI, 0.5-1.2; P < .001); neither changed significantly in the control group (Table 2). Within the intervention group, the change in PREDIM correlated negatively with the change in body weight (r = −0.43; P < .001). Figure 1. CONSORT Diagram of Participant Flow Through Trial 3115 Participants screened over phone 413 In-person meetings 81 Excluded 36 Outside BMI range 3 Diabetes diagnosis 7 Medical exclusion 31 Not willing to be in the vegan or control group 4 Unable to attend weekly classes 332 Signed the consent form 88 Excluded 72 Did not turn in their diet record or did not come to their baseline assessment 1 Family emergency 1 Had medication changes and decided not to participate 1 Had already started a vegan diet 6 Unable to attend weekly classes 7 Withdrew 244 Randomized 122 Randomized to vegan diet 122 Randomized to control 16 Dropped out 5 Dropped out 3 Withdrew owing to personal reasons 1 Withdrew owing to family reasons 2 Not willing to be in the control or 1 Withdrew owing to health reasons vegan group 1 Unable to contact 2 Due to health reasons 1 Unable to follow diet 8 Unable to contact 1 Unable to attend classes 1 Unable to participate in all aspects of the study 117 Completed final assessment 106 Completed final assessment 117 Were included in the analysis 106 Were included in the analysis JAMA Network Open. 2020;3(11):e2025454. doi:10.1001/jamanetworkopen.2020.25454 (Reprinted) November 30, 2020 6/14 JAMA Network Open | Nutrition, Obesity, and Exercise Effect of a Low-Fat Vegan Diet on Metabolic Measures in Overweight Adults Table 1. Changes in Outcomes During the Study in the Low-Fat Vegan Dietary Intervention Group vs the Control Group Value, Mean (95% CI) Control group Intervention group Outcome Baseline Week 16 Change Baseline Week 16 Change Effect Size P value Dietary intake Energy intake, kcal/d 1793 (1670 to 1657 (1548 to −135.8 (−250.7 to 1834 (1729 to 1344 (1260 to −490.7 (−607.9 to −354.9 (−519.0 to <.001 b c 1915) 1766) −20.8) 1940) 1428) −373.5) −190.8) Fiber intake, g/d 23.9 (21.9 to 23.3 (21.4 to −0.56 (−2.6 to 24.1 (22.1 to 34.6 (32.1 to 10.6 (7.8 to 11.1 (7.8 to <.001 25.9) 25.3) 1.5) 26.0) 37.2) 13.3) 14.5) Cholesterol intake, 244.5 (211.4 to 230.5 (196.1 to −14.0 (−51.7 to 238.6 (212.3 to 5.5 (3.8 to −233.1 (−259.4 to −219.1 (−264.9 to <.001 mg/d 277.6) 264.9) 23.7) 265.0) 7.3) −206.8) −173.3) Saturated fatty 22.9 (20.5 to 20.5 (18.1 to −2.4 (−4.9 to 23.6 (21.4 to 5.1 (4.5 to −18.6 (−20.7 to −16.2 (−19.4 to <.001 acids, g/d 25.3) 23.0) 0.1) 25.8) 5.6) −16.5) −13.0) Monounsaturated fatty 27.9 (25.2 to 25.4 (23.0 to −2.5 (−4.9 to 27.2 (25.3 to 8.4 (7.6 to −18.8 (−20.7 to −16.3 (−19.3 to <.001 b c acids, g/d 30.7) 27.8) −0.1) 29.1) 9.2) −16.9) −13.3) Polyunsaturated fatty 19.1 (17.0 to 18.0 (16.3 to −1.1 (−3.0 to 18.4 (16.8 to 9.5 (8.6 to −8.9 (−10.6 to −7.8 (−10.4 to <.001 acids, g/d 21.1) 19.7) 0.9) 19.9) 10.4) −7.1) −5.2) Physical activity, METs 2863 (2224 to 2153 (1605 to −709.8 (−1346 to 2719 (1805 to 2114 (1619 to −604.8 (−1388 to 105 (−898 to .84 3502) 2702) −73.9) 3633) 2609) 178.6) 1108) Anthropometric variables and body composition Weight, kg 92.7 (90.0 to 92.2 (89.4 to −0.5 (−1.0 to 93.6 (91.0 to 87.2 (84.9 to −6.4 (−7.0 to −5.9 (−6.7 to <.001 95.3) 94.9) 0.1) 96.1) 89.6) −5.7) −5.0) BMI 33.6 (32.9 to 33.9 (32.6 to 0.3 (−0.7 to 33.3 (32.6 to 31.4 (30.5 to −1.9 (−2.5 to −2.2 (−3.3 to <.001 34.3) 35.2) 1.3) 34.0) 32.4) −1.3) −1.1) Fat mass, kg 40.9 (39.1 to 41.0 (39.0 to 0.01 (−0.3 to 40.6 (38.9 to 36.5 (34.9 to −4.1 (−4.6 to −4.1 (−4.7 to <.001 42.8) 42.9) 0.4) 42.2) 38.1) −3.6) −3.5) Lean mass, kg 49.5 (47.9 to 48.9 (47.4 to −0.6 (−0.9 to 50.5 (49.0 to 48.4 (47.1 to −2.1 (−2.4 to −1.5 (−1.9 to <.001 c c 51.1) 50.5) −0.3) 51.9) 49.8) −1.8) −1.1) VAT volume, cm 1517 (1339 to 1510 (1324 to −7.7 (−78.5 to 1459 (1286 to 1243 (1096 to −216.5 (−280.9 to −208.8 (−303.7 to <.001 1695) 1695) 63.0) 1632) 1390) −152.2) −113.7) Hepatocellular 3.3 (3.1 to 3.6 (3.5 to 0.3 (−0.5 to 3.2 (3.0 to 2.4 (2.3 to −0.8 (−1.5 to −1.2 (−2.2 to .002 lipids, % 3.5) 3.8) 1.2) 3.4) 2.5) −0.1) −0.1) Intramyocellular 1.5 (1.4 to 1.7 (1.5 to 0.13 (−0.05 to 1.6 (1.5 to 1.5 (1.4 to −0.1 (−0.2 to −0.3 (−0.4 to .03 lipids, % 1.6) 1.8) 0.21) 1.7) 1.6) 0.05) −0.1) Parameters of glucose control and insulin resistance HbA , DCCT, % 5.7 (5.6 to 5.7 (5.6 to 0.01 (−0.04 to 5.6 (5.6 to 5.6 (5.5 to −0.06 (−0.12 to −0.07 (−0.1 to .07 1c 5.8) 5.8) 0.05) 5.7) 5.7) −0.002) 0.01) Fasting plasma insulin 78.9 (68.3 to 103.5 (71.4 to 23.6 (−5.0 to 91.2 (79.9 to 69.6 (56.9 to −21.6 (−35.9 to −46.2 (−79.0 to .006 level, pmol/L 89.4) 135.6) 54.3) 102.5) 82.3) −7.3) −13.4) Fasting plasma glucose 5.0 (4.7 to 5.5. (5.4 to 0.5 (0.2 to 5.2 (5.1 to 5.1 (5.0 to −0.1 (−0.2 to −0.6 (−0.2 to .001 level, mmol/L 5.4) 5.7) 0.8) 5.3) 5.2) 0.02) −1.0) PREDIM 4.4 (4.1 to 4.2 (3.9 to −0.2 (−0.4 to 4.1 (3.8 to 4.7 (4.4 to 0.7 (0.5 to 0.9 (0.5 to <.001 4.7) 4.5) 0.04) 4.3) 5.0) 0.9) 1.2) HOMA 2.7 (2.3 to 3.2 (2.4 to 0.5 (−0.3 to 3.2 (2.7 to 2.3 (1.9 to −0.8 (−1.3 to −1.3 (−2.2 to <.001 3.2) 4.0) 1.2) 3.6) 2.8) −0.3) −0.3) Lipid levels, mmol/L Total cholesterol 5.0 (4.7 to 5.1 (4.9 to 0.1 (−0.1 to 5.2 (5.0 to 4.7 (4.5 to −0.5 (−0.7 to −0.6 (−0.9 to <.001 5.2) 5.3) 0.4) 5.4) 4.9) −0.4) −0.4) Triglycerides 1.3 (1.2 to 1.3 (1.1 to −0.01 (−0.14 to 1.2 (1.1 to 1.4 (1.3 to 0.2 (0.08 to 0.2 (0.03 to .02 1.4) 1.5) 0.12) 1.3) 1.5) 0.3) 0.4) HDL cholesterol 1.7 (1.5 to 1.5 (1.4 to −0.2 (−0.4 to 1.6 (1.5 to 1.4 (1.3 to −0.2 (−0.3 to 0.01 (−0.2 to .93 b c 1.9) 1.6) −0.02) 1.6) 1.4) −0.1) 0.2) LDL cholesterol 2.9 (2.6 to 3.0 (2.9 to 0.07 (−0.02 to 3.1 (3.0 to 2.7 (2.5 to −0.4 (−1.0 to −0.5 (−0.8 to <.001 3.1) 3.2) 0.2) 3.3) 2.9) −0.3) −0.3) Abbreviations: BMI, body mass index (calculated as weight in kilograms divided by P values are for the interaction between group and time assessed by repeated height in meters squared); DCCT, Diabetes Control and Complications Trial; HbA , measures analysis of variance. 1c glycated hemoglobin; HDL, high-density lipoprotein; HOMA, homeostasis model b P < .05 for within-group changes from baseline assessed by paired comparison t tests. assessment; LDL, low-density lipoprotein; METs, metabolic equivalents; PREDIM, P < .001 for within-group changes from baseline assessed by paired comparison t tests. predicted insulin sensitivity index; VAT, visceral adipose tissue. SI conversion factors: To convert plasma insulin level to μIU/mL, divide by 6.945; plasma glucose level to mg/dL, divide by 0.0555; and lipid levels to mg/dL, divide by 0.0259. JAMA Network Open. 2020;3(11):e2025454. doi:10.1001/jamanetworkopen.2020.25454 (Reprinted) November 30, 2020 7/14 JAMA Network Open | Nutrition, Obesity, and Exercise Effect of a Low-Fat Vegan Diet on Metabolic Measures in Overweight Adults Postprandial Metabolism Postprandial energy expenditure (the thermic effect of food) increased in the intervention group from baseline to 16 weeks and did not change significantly in the control group (between-group difference effect size, 14.1%; 95% CI, 6.5%-20.4%; interaction between group and time, P < .001) (Figure 2A). The F values were as follows: group, F =1.7 (P = .19); week, F =15.4 (P < .001); time, F = 122.4 (P < .001); group × week, F =11.9 (P < .001); group × time, F =1.1 (P = .35); week × time, F = 1.38 (P = .25). The results were similar in models adjusted for age and race/ethnicity (eFigure 1 in Supplement 2). Within the intervention group, the change in thermic effect of food did not correlate significantly with changes in body weight (r = −0.15; P = .09), PREDIM (r = 0.06; P = .54), energy intake (r = 0.01; P = .90), or fiber consumption (r = 0.07; P = .48). In both groups combined, changes in thermic effect of food correlated negatively with changes in fat mass (r = −0.30; P < .05) and positively with changes in PREDIM (r = 0.36; P < .05). That is, as fat mass decreased and insulin sensitivity improved, postprandial metabolism increased (Table 2). A linear regression model for changes in reported energy intake and body weight showed that every 100 kcal/d change in energy intake was associated with a 0.15 kg change in body weight (eFigure 3 in Supplement 2). The mean (SD) reported energy reduction of 355 (617) kcal in the intervention group compared with the control group would therefore be associated with a mean (SD) weight loss of 0.53 (4.4) kg. For changes in postprandial energy expenditure and body weight, every change in postprandial energy expenditure of 10 000 U in area under the curve was associated with a change in body weight of 0.48 kg (eFigure 3 in Supplement 2). The mean (SD) decrease in postprandial energy expenditure of 8588 (34 020) U of area under the curve was associated with an mean (SD) weight loss of 0.41 (2.8) kg. Hepatocellular and Intramyocellular Lipid Levels In the 44 participants for whom hepatocellular and intramyocellular lipid levels were quantified, 29,30 baseline hepatocellular lipid content was generally in the normal range. Nonetheless, hepatocellular lipid content decreased in the intervention group by 34.4% (from a mean [SD] of 3.2% [2.9%] to 2.4% [2.2%]; P = .03) and remained unchanged in the control group (from a mean [SD] of 3.3% [4.3%] to 3.6% [4.7%]) (group, F =3.1 [P = .09]; week, F =1.27 [P = .27]; group × week, F =10.8 [P = .002]) (Figure 2B). Results were similar in models adjusted for age and race/ethnicity (eFigure 1 in Supplement 2) and for baseline BMI (eFigure 2 in Supplement 2). Within the intervention group, the decrease in hepatocellular lipid levels was significantly associated with change in body weight (r = 0.42; P = .04) but not with changes in reported energy intake (r = 0.24; P = .27) or fiber consumption (r = 0.07; P = .76). In both groups combined, changes in hepatocellular lipid levels correlated negatively with changes in PREDIM (r = −0.47; P < .05). That is, as hepatocellular lipid level decreased, insulin sensitivity increased. Changes in hepatocellular lipid Table 2. Relationship Between Changes in Thermic Effect of Food and the First Predictive Component as Evaluated by the OPLS Model OPLS predictive component Multiple regression a b Variable Component loading t Statistic R P value for R Regression coefficient t Statistic Matrix X Baseline BMI 0.191 2.46 0.209 <.05 −0.015 −0.33 Baseline fat mass 0.256 2.89 0.283 <.05 −0.014 −0.28 Baseline TEF −0.850 −11.96 −0.938 .005 −0.505 −5.69 Change in PREDIM 0.324 2.41 0.359 <.05 0.105 1.37 Change in fat mass −0.271 −2.59 −0.301 <.05 −0.122 −1.55 Matrix Y Change in TEF 1.000 5.27 0.540 .003 NA NA Abbreviations: BMI, body mass index; NA, not applicable; OPLS, orthogonal projections Component loadings expressed as a correlation coefficients with predictive to latent structure; PREDIM, predicted insulin sensitivity index; TEF, thermic effect component. of food. The explained variability was 29.2% (24.3% after cross-validation). JAMA Network Open. 2020;3(11):e2025454. doi:10.1001/jamanetworkopen.2020.25454 (Reprinted) November 30, 2020 8/14 JAMA Network Open | Nutrition, Obesity, and Exercise Effect of a Low-Fat Vegan Diet on Metabolic Measures in Overweight Adults levels correlated positively with changes in body weight (r = 0.91; P < .01), BMI (r = 0.90; P < .01), fat mass (r = 0.91; P < .01), and visceral fat (r = 0.80; P <.01)(Table 3). Changes in intramyocellular lipid levels were not statistically significant in within-group comparisons, but owing to the opposite trends, the treatment effect was significantly decreased in the intervention group by 10.4%, from a mean (SD) of 1.6 (1.1) to 1.5 (1.0) (P = .03) (group, F =4.7 [P = .04]; week, F = 0.02 [P = .88]; group × week, F =5.1 [P = .03]) (Figure 1C). Within the intervention group (n = 23), changes in both hepatocellular and intramyocellular lipid levels correlated with changes in insulin resistance, as measured by the homeostasis model assessment index (both r = 0.51; P = .01). In both groups combined, changes in intramyocellular lipid levels correlated positively with changes in fat mass (r = 0.51; P < .05) and homeostasis model assessment index score (r = 0.52; P < .05). That is, as fat mass decreased, intramyocellular lipid levels and insulin resistance decreased. Discussion In this trial, the dietary intervention reduced body weight, apparently owing to its tendency to reduce energy intake and increase postprandial energy expenditure. The intervention also improved glycemic control and reduced insulin concentrations, owing in part to reduced lipid accumulation in liver and muscle cells and thus reduced insulin resistance in these organs. Figure 2. Changes in the Thermic Effect of Food, Liver Fat, and Intramyocellular Lipid Levels in the Intervention and Control Groups A Thermic effect of food in the control group B Thermic effect of food in the intervention group 4.5 4.5 Week 0 4.0 4.0 Week 16 3.5 3.5 Week 0 3.0 3.0 Week 16 2.5 2.5 2.0 2.0 1.5 1.5 1.0 1.0 30 60 120 180 30 60 120 180 Time, min Time, min C Liver fat D Intramyocellular lipids 2.6 1.8 Control group Control group 2.4 1.6 2.2 2.0 1.5 1.8 1.4 Intervention group 1.6 Intervention group 1.4 1.3 0 16 0 16 Week Week Whiskers represent 95% CIs. JAMA Network Open. 2020;3(11):e2025454. doi:10.1001/jamanetworkopen.2020.25454 (Reprinted) November 30, 2020 9/14 Liver fat, % Thermic effect of food, kcal/kg Intramyocellular lipids, % Thermic effect of food, kcal/kg JAMA Network Open | Nutrition, Obesity, and Exercise Effect of a Low-Fat Vegan Diet on Metabolic Measures in Overweight Adults The intervention diet’s effect on weight and insulin action are clinically important. Hepatocellular and intramyocellular lipids play central roles in hepatic and muscle insulin resistance, respectively, and in type 2 diabetes. A 16-week diet of 1200 kcal per day resulted in a moderate weight loss of approximately 8 kg, which was sufficient to normalize liver lipid content and fasting plasma glucose concentrations as well as reverse hepatic insulin resistance in patients with obesity and type 2 diabetess. A potential mechanism explaining the improvement in insulin sensitivity is the reduction in intracellular diacylglycerol levels, which reduce insulin signaling in liver and muscle, 22,32,33 leading to tissue-specific insulin resistance. The effects of the dietary intervention on hepatocellular and intramyocellular lipid levels and insulin sensitivity—the presumed basis for the improved glycemic control—has not previously been quantified in clinical trials. Energy restriction has been shown to reduce intramyocellular and hepatocellular lipid levels and improve glycemic control in healthy young individuals without 26,34,35 diabetes. In young, lean individuals with insulin resistance, a hypocaloric diet (approximately 1200 kcal) led to a mean weight loss of 4.1 kg and a 30% reduction of intramyocellular lipids during a 9-week intervention. In contrast, the intervention diet in the present study did not restrict energy intake but nonetheless led to 34% and 10% reductions in hepatocellular and intramyocellular lipid levels, respectively. The reductions in hepatocellular and intramyocellular lipid levels correlated with 26,36,37 the reduction in fat mass, consistent with prior studies. The present finding that the increase in thermic effect of food was associated with decreased 38,39 fat mass and increased insulin sensitivity confirm the findings of previous research. The increased insulin sensitivity may have contributed to the increased postprandial metabolism. In addition, increased postprandial metabolism may have promoted further reduction in fat mass and an increase in insulin sensitivity. Despite the ad libitum diet, the participants in the intervention group reduced their energy intake, consistent with many previous trials using vegan diets. This not only contributes to weight loss but also may have contributed to the decrease in hepatocellular triglyceride content. 40-43 Postprandial metabolism is influenced by meal composition. In the present study, however, the test meal was identical for all study phases. These results suggest that the increased postprandial thermogenesis was attributable to improved insulin sensitivity. Table 3. Relationship Between Changes in Liver Fat and the First Predictive Component as Evaluated by OPLS Model OPLS predictive component Multiple regression a b Variable Component loading t Statistic R P value for R Regression coefficient t Statistic P value for t Matrix X Control group 0.339 9.88 0.795 .004 0.069 3.69 .007 Intervention group −0.339 −9.88 −0.795 .004 −0.069 −3.69 .007 Baseline PREDIM 0.214 8.89 0.498 .003 0.038 4.99 .005 Baseline HOMA −0.218 −2.71 −0.509 <.05 −0.060 −2.07 <.05 Baseline weight −0.228 −2.18 −0.535 <.05 −0.065 −2.19 <.05 Baseline fat mass −0.221 −2.18 −0.518 <.05 −0.058 −2.42 <.05 Change in PREDIM −0.199 −2.35 −0.468 <.05 −0.021 −2.54 <.05 Change in weight 0.388 14.13 0.910 .005 0.079 5.47 .005 Change in BMI 0.384 13.92 0.901 .005 0.077 5.64 .003 Change in fat mass 0.389 21.06 0.911 .002 0.078 5.87 .006 Change in visceral fat 0.341 8.23 0.798 .007 0.060 2.63 <.05 Matrix Y Change in liver fat 1.000 4.66 0.495 .009 NA NA NA Abbreviations: BMI, body mass index; HOMA, homeostasis model assessment; OPLS, Component loadings expressed as a correlation coefficients with predictive orthogonal projections to latent structure; NA, not applicable; PREDIM, predicted insulin component. sensitivity index. Explained variability was 24.5% (20.8% after cross-validation). JAMA Network Open. 2020;3(11):e2025454. doi:10.1001/jamanetworkopen.2020.25454 (Reprinted) November 30, 2020 10/14 JAMA Network Open | Nutrition, Obesity, and Exercise Effect of a Low-Fat Vegan Diet on Metabolic Measures in Overweight Adults Strengths and Limitations This study has several strengths. The randomized parallel design in which all participants within each cohort began the study simultaneously controlled for seasonal diet fluctuations. The study duration provided sufficient time for adaptation to the diet. Physiologic stimulation by a standard mixed meal permitted quantification of insulin sensitivity and insulin secretion during a physiologic perturbation. Measurement of visceral, hepatocellular, and intramyocellular lipid levels, in addition to the detailed assessment of the thermic effect of food, are also strengths. The low attrition suggests that the intervention was acceptable. The study also has limitations. Self-reports of dietary intake have well-known limitations. However, it is reassuring that the reported diet changes were paralleled by changes in weight and plasma lipid levels. Health-conscious participants may not be representative of the general population but may be representative of a clinical population seeking help for weight problems or type 2 diabetes. We followed the participants for 16 weeks and were not able to estimate the effects of the diet over a longer period. In addition, the study design did not allow separation of the specific effects of the low-fat vegan diet from the weight loss it causes. Conclusions This randomized clinical trial found that a low-fat plant-based dietary intervention reduces body weight by reducing energy intake and increasing postprandial metabolism, apparently owing to increased insulin sensitivity resulting from reduced hepatocellular and intramyocellular fat. This intervention may be an effective treatment for overweight adults. ARTICLE INFORMATION Accepted for Publication: September 17, 2020. Published: November 30, 2020. doi:10.1001/jamanetworkopen.2020.25454 Correction: This article was corrected on January 7, 2021, to edit the Role of the Funder/Sponsor section; on February 1, 2021, to fix errors in the abstract and main Results sections; and on May 27, 2021, to fix an error in the Results section and eTable 2 in Supplement 2. Open Access: This is an open access article distributed under the terms of the CC-BY License. © 2020 Kahleova H et al. JAMA Network Open. Corresponding Author: Hana Kahleova, MD, PhD, Physicians Committee for Responsible Medicine, 5100 Wisconsin Ave NW, Ste 400, Washington, DC 20016 (hkahleova@pcrm.org). Author Affiliations: Physicians Committee for Responsible Medicine, Washington, DC (Kahleova, Alwarith, Rembert, Barnard); Department of Internal Medicine, Yale School of Medicine, New Haven, Connecticut (Petersen, Shulman); Department of Cellular & Molecular Physiology, Yale School of Medicine, New Haven, Connecticut (Shulman); Metabolic Unit, CNR Institute of Neuroscience, Padua, Italy (Tura); Institute of Endocrinology, Prague, Czech Republic (Hill); School of Medicine, University of Utah, Salt Lake City (Holubkov); George Washington University School of Medicine and Health Sciences, Washington, DC (Barnard). Author Contributions: Drs Kahleova and Barnard had full access to all of the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis. Concept and design: Kahleova, Petersen, Shulman, Barnard. Acquisition, analysis, or interpretation of data: All authors. Drafting of the manuscript: Kahleova, Shulman, Alwarith, Rembert, Tura, Barnard. Critical revision of the manuscript for important intellectual content: Kahleova, Petersen, Shulman, Hill, Holubkov. Statistical analysis: Hill, Holubkov. Obtained funding: Petersen, Shulman. Administrative, technical, or material support: Kahleova, Petersen, Alwarith, Rembert. Supervision: Kahleova, Petersen, Shulman, Barnard. JAMA Network Open. 2020;3(11):e2025454. doi:10.1001/jamanetworkopen.2020.25454 (Reprinted) November 30, 2020 11/14 JAMA Network Open | Nutrition, Obesity, and Exercise Effect of a Low-Fat Vegan Diet on Metabolic Measures in Overweight Adults Conflict of Interest Disclosures: Dr Kahleova reported being director of clinical research at the Physicians Committee, a nonprofit organization that provides nutrition education and research. Dr Rembert reported compensation from the Physicians Committee for Responsible Medicine outside the submitted work. Dr Holubkov reported receiving personal fees from the Physicians Committee for Responsible Medicine during the conduct of the study. Dr Barnard reported to serving as president of the Physicians Committee for Responsible Medicine and Barnard Medical Center; receiving royalties from Hachette Book Group, Penguin Random House, Rodale, and Da Capo publishers; and receiving honoraria from Yale, Rush, George Washington, Loma Linda, Rockford Universities, Montefiore Medical Center, the Mayo Clinic, Northwell Health, Christiana Care, Oticon, and the National Organization of Professional Athletes. No other disclosures were reported. Funding/Support: This work was funded by the Physicians Committee for Responsible Medicine and grants P30 DK-045735 and R01 DK-113984 from the Yale Diabetes Center (Drs Shulman and Petersen). Role of the Funder/Sponsor: The Yale Diabetes Center had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication. Drs Kahleova, Alwarith, and Rembert, as employees of Physicians Committee for Responsible Medicine, were involved in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication. The research team of the Physicians Committee for Responsible Medicine had full autonomy in all aspects of the study. Data Sharing Statement: See Supplement 3. REFERENCES 1. Qian F, Liu G, Hu FB, Bhupathiraju SN, Sun Q. Association between plant-based dietary patterns and risk of type 2 diabetes: a systematic review and meta-analysis. JAMA Intern Med. 2019. doi:10.1001/jamainternmed.2019.2195 2. Tonstad S, Butler T, Yan R, Fraser GE. Type of vegetarian diet, body weight, and prevalence of type 2 diabetes. Diabetes Care. 2009;32(5):791-796. doi:10.2337/dc08-1886 3. Barnard ND, Levin SM, Yokoyama Y. A systematic review and meta-analysis of changes in body weight in clinical trials of vegetarian diets. J Acad Nutr Diet. 2015;115(6):954-969. doi:10.1016/j.jand.2014.11.016 4. Barnard ND, Scialli AR, Turner-McGrievy G, Lanou AJ, Glass J. The effects of a low-fat, plant-based dietary intervention on body weight, metabolism, and insulin sensitivity. Am J Med. 2005;118(9):991-997. doi:10.1016/j. amjmed.2005.03.039 5. National Cholesterol Education Program. Summary of the second report of the National Cholesterol Education Program (NCEP) Expert Panel on Detection, Evaluation, and Treatment of High Blood Cholesterol in Adults (Adult Treatment Panel II). JAMA. 1993;269(23):3015-3023. doi:10.1001/jama.1993.03500230097036 6. Goff LM, Bell JD, So P-W, Dornhorst A, Frost GS. Veganism and its relationship with insulin resistance and intramyocellular lipid. Eur J Clin Nutr. 2005;59(2):291-298. doi:10.1038/sj.ejcn.1602076 7. Gojda J, Patková J, Jaček M, et al. Higher insulin sensitivity in vegans is not associated with higher mitochondrial density. Eur J Clin Nutr. 2013;67(12):1310-1315. doi:10.1038/ejcn.2013.202 8. Itani SI, Ruderman NB, Schmieder F, Boden G. Lipid-induced insulin resistance in human muscle is associated with changes in diacylglycerol, protein kinase C, and IkappaB-alpha. Diabetes. 2002;51(7):2005-2011. doi:10. 2337/diabetes.51.7.2005 9. Goodpaster BH, Thaete FL, Kelley DE. Thigh adipose tissue distribution is associated with insulin resistance in obesity and in type 2 diabetes mellitus. Am J Clin Nutr. 2000;71(4):885-892. doi:10.1093/ajcn/71.4.885 10. Yatsuya H, Nihashi T, Li Y, et al. Independent association of liver fat accumulation with insulin resistance. Obes Res Clin Pract. 2014;8(4):e350-e355. doi:10.1016/j.orcp.2013.08.002 11. Schulz KF, Altman DG, Moher D; CONSORT Group. CONSORT 2010 statement: updated guidelines for reporting parallel group randomised trials. BMC Med. 2010;8(1):18. doi:10.1186/1741-7015-8-18 12. Schakel SF, Sievert YA, Buzzard IM. Sources of data for developing and maintaining a nutrient database. JAm Diet Assoc. 1988;88(10):1268-1271. 13. Hagströmer M, Oja P, Sjöström M. The International Physical Activity Questionnaire (IPAQ): a study of concurrent and construct validity. Public Health Nutr. 2006;9(6):755-762. doi:10.1079/PHN2005898 14. Kaul S, Rothney MP, Peters DM, et al. Dual-energy X-ray absorptiometry for quantification of visceral fat. Obesity (Silver Spring). 2012;20(6):1313-1318. doi:10.1038/oby.2011.393 15. Neeland IJ, Grundy SM, Li X, Adams-Huet B, Vega GL. Comparison of visceral fat mass measurement by dual- X-ray absorptiometry and magnetic resonance imaging in a multiethnic cohort: the Dallas Heart Study. Nutr Diabetes. 2016;6(7):e221. doi:10.1038/nutd.2016.28 JAMA Network Open. 2020;3(11):e2025454. doi:10.1001/jamanetworkopen.2020.25454 (Reprinted) November 30, 2020 12/14 JAMA Network Open | Nutrition, Obesity, and Exercise Effect of a Low-Fat Vegan Diet on Metabolic Measures in Overweight Adults 16. Mellis MG, Oldroyd B, Hind K. In vivo precision of the GE Lunar iDXA for the measurement of visceral adipose tissue in adults: the influence of body mass index. Eur J Clin Nutr. 2014;68(12):1365-1367. doi:10.1038/ejcn. 2014.213 17. Carver TE, Christou NV, Andersen RE. In vivo precision of the GE iDXA for the assessment of total body composition and fat distribution in severely obese patients. Obesity (Silver Spring). 2013;21(7):1367-1369. doi:10. 1002/oby.20323 18. Matthews DR, Hosker JP, Rudenski AS, Naylor BA, Treacher DF, Turner RC. Homeostasis model assessment: insulin resistance and beta-cell function from fasting plasma glucose and insulin concentrations in man. Diabetologia. 1985;28(7):412-419. doi:10.1007/BF00280883 19. Tura A, Chemello G, Szendroedi J, et al. Prediction of clamp-derived insulin sensitivity from the oral glucose insulin sensitivity index. Diabetologia. 2018;61(5):1135-1141. doi:10.1007/s00125-018-4568-4 20. Kaviani S, Schoeller DA, Ravussin E, et al. Determining the accuracy and reliability of indirect calorimeters utilizing the methanol combustion technique. Nutr Clin Pract. 2018;33(2):206-216. doi:10.1002/ncp.10070 21. Blond E, Maitrepierre C, Normand S, et al. A new indirect calorimeter is accurate and reliable for measuring basal energy expenditure, thermic effect of food and substrate oxidation in obese and healthy subjects. e-SPEN, the European e-Journal of Clinical Nutrition and Metabolism 2011;6(1):e7–e15. doi:10.1016/j.eclnm.2010.12.001 22. Petersen KF, Dufour S, Feng J, et al. Increased prevalence of insulin resistance and nonalcoholic fatty liver disease in Asian-Indian men. Proc Natl Acad SciUSA. 2006;103(48):18273-18277. doi:10.1073/pnas.0608537103 23. Lee SS, Lee Y, Kim N, et al. Hepatic fat quantification using chemical shift MR imaging and MR spectroscopy in the presence of hepatic iron deposition: validation in phantoms and in patients with chronic liver disease. J Magn Reson Imaging. 2011;33(6):1390-1398. doi:10.1002/jmri.22583 24. Rabøl R, Petersen KF, Dufour S, Flannery C, Shulman GI. Reversal of muscle insulin resistance with exercise reduces postprandial hepatic de novo lipogenesis in insulin resistant individuals. Proc Natl Acad SciUSA. 2011;108 (33):13705-13709. doi:10.1073/pnas.1110105108 25. Gruetter R. Automatic, localized in vivo adjustment of all first- and second-order shim coils. Magn Reson Med. 1993;29(6):804-811. doi:10.1002/mrm.1910290613 26. Petersen KF, Dufour S, Morino K, Yoo PS, Cline GW, Shulman GI. Reversal of muscle insulin resistance by weight reduction in young, lean, insulin-resistant offspring of parents with type 2 diabetes. Proc Natl Acad Sci U S A. 2012;109(21):8236-8240. doi:10.1073/pnas.1205675109 27. Trygg J, Holmes E, Lundstedt T. Chemometrics in metabonomics. J Proteome Res. 2007;6(2):469-479. doi:10. 1021/pr060594q 28. Sosvorova L, Hill M, Mohapl M, Vitku J, Hampl R. Steroid hormones in prediction of normal pressure hydrocephalus. J Steroid Biochem Mol Biol. 2015;152:124-132. doi:10.1016/j.jsbmb.2015.05.004 29. Romeo S, Kozlitina J, Xing C, et al. Genetic variation in PNPLA3 confers susceptibility to nonalcoholic fatty liver disease. Nat Genet. 2008;40(12):1461-1465. doi:10.1038/ng.257 30. Szczepaniak LS, Nurenberg P, Leonard D, et al. Magnetic resonance spectroscopy to measure hepatic triglyceride content: prevalence of hepatic steatosis in the general population. Am J Physiol Endocrinol Metab. 2005;288(2):E462-E468. doi:10.1152/ajpendo.00064.2004 31. Petersen KF, Dufour S, Befroy D, Lehrke M, Hendler RE, Shulman GI. Reversal of nonalcoholic hepatic steatosis, hepatic insulin resistance, and hyperglycemia by moderate weight reduction in patients with type 2 diabetes. Diabetes. 2005;54(3):603-608. doi:10.2337/diabetes.54.3.603 32. Petersen MC, Shulman GI. Mechanisms of insulin action and insulin resistance. Physiol Rev. 2018;98(4): 2133-2223. doi:10.1152/physrev.00063.2017 33. Samuel VT, Shulman GI. Mechanisms for insulin resistance: common threads and missing links. Cell. 2012;148 (5):852-871. doi:10.1016/j.cell.2012.02.017 34. Lara-Castro C, Newcomer BR, Rowell J, et al. Effects of short-term very low-calorie diet on intramyocellular lipid and insulin sensitivity in nondiabetic and type 2 diabetic subjects. Metabolism. 2008;57(1):1-8. doi:10.1016/j. metabol.2007.05.008 35. Jazet IM, Schaart G, Gastaldelli A, et al. Loss of 50% of excess weight using a very low energy diet improves insulin-stimulated glucose disposal and skeletal muscle insulin signalling in obese insulin-treated type 2 diabetic patients. Diabetologia. 2008;51(2):309-319. doi:10.1007/s00125-007-0862-2 36. Sinha R, Dufour S, Petersen KF, et al. Assessment of skeletal muscle triglyceride content by (1)H nuclear magnetic resonance spectroscopy in lean and obese adolescents: relationships to insulin sensitivity, total body fat, and central adiposity. Diabetes. 2002;51(4):1022-1027. doi:10.2337/diabetes.51.4.1022 JAMA Network Open. 2020;3(11):e2025454. doi:10.1001/jamanetworkopen.2020.25454 (Reprinted) November 30, 2020 13/14 JAMA Network Open | Nutrition, Obesity, and Exercise Effect of a Low-Fat Vegan Diet on Metabolic Measures in Overweight Adults 37. Moro C, Galgani JE, Luu L, et al. Influence of gender, obesity, and muscle lipase activity on intramyocellular lipids in sedentary individuals. J Clin Endocrinol Metab. 2009;94(9):3440-3447. doi:10.1210/jc.2009-0053 38. Camastra S, Bonora E, Del Prato S, Rett K, Weck M, Ferrannini E; EGIR (European Group for the Study of Insulin Resistance). Effect of obesity and insulin resistance on resting and glucose-induced thermogenesis in man. Int J Obes Relat Metab Disord. 1999;23(12):1307-1313. doi:10.1038/sj.ijo.0801072 39. Ravussin E, Acheson KJ, Vernet O, Danforth E, Jéquier E. Evidence that insulin resistance is responsible for the decreased thermic effect of glucose in human obesity. J Clin Invest. 1985;76(3):1268-1273. doi:10.1172/JCI112083 40. Bowden VL, McMurray RG. Effects of training status on the metabolic responses to high carbohydrate and high fat meals. Int J Sport Nutr Exerc Metab. 2000;10(1):16-27. doi:10.1123/ijsnem.10.1.16 41. Nagai N, Sakane N, Moritani T. Metabolic responses to high-fat or low-fat meals and association with sympathetic nervous system activity in healthy young men. J Nutr Sci Vitaminol (Tokyo). 2005;51(5):355-360. doi: 10.3177/jnsv.51.355 42. Thyfault JP, Richmond SR, Carper MJ, Potteiger JA, Hulver MW. Postprandial metabolism in resistance-trained versus sedentary males. Med Sci Sports Exerc. 2004;36(4):709-716. doi:10.1249/01.MSS.0000121946.98885.F5 43. Barr SB, Wright JC. Postprandial energy expenditure in whole-food and processed-food meals: implications for daily energy expenditure. Food Nutr Res. 2010;54:54. doi:10.3402/fnr.v54i0.5144 44. Yuan C, Spiegelman D, Rimm EB, et al. Relative validity of nutrient intakes assessed by questionnaire, 24-hour recalls, and diet records compared with urinary recovery and plasma concentration biomarkers: findings for women. Am J Epidemiol. 2017. doi:10.1093/aje/kww104 SUPPLEMENT 1. Trial Protocol SUPPLEMENT 2. eTable 1. Baseline characteristics of the study population eTable 2. Baseline characteristics of the study population, comparing study completers and drop-outs eTable 3. Baseline characteristics of the study population, comparing the subsample undergoing magnetic resonance spectroscopy (MRS) with the rest of the study population eTable 4. Treatment effects for the main outcomes, adjusted for age and race eFigure 1. Changes in the thermic effect of food, liver fat, and intramyocellular lipids after adjustment for race and age eFigure 2. Changes in liver fat and intramyocellular lipids after adjustment for baseline BMI eFigure 3. Linear regression model for changes in energy and body weight and postprandial energy expenditure and body weight SUPPLEMENT 3. Data Sharing Statement JAMA Network Open. 2020;3(11):e2025454. doi:10.1001/jamanetworkopen.2020.25454 (Reprinted) November 30, 2020 14/14 1 A Randomized, Controlled Trial on Diet, Insulin Sensitivity, 2 and Postprandial Metabolism 6 Summary 8 This randomized, controlled trial aims to elucidate the mechanisms by which a plant- 9 based dietary intervention causes weight loss. Using a low-fat, plant-based diet for 16 10 weeks, along with an untreated control for comparison, the study will measure changes 11 in insulin sensitivity, postprandial metabolism, and intracellular lipid, and assess their 12 associations with changes in body weight. 14 1. SPECIFIC AIMS AND OVERVIEW 15 1.1. Specific Aims 16 Specific Aim 1. This study tests the hypothesis that weight changes associated with a 17 low-fat plant-based diet are, in part, the result of increased postprandial metabolism 18 (thermic effect of food). 19 Specific Aim 2. This study tests the hypothesis that increased postprandial metabolism 20 in response to a diet intervention is the result of increased insulin sensitivity. 21 Specific Aim 3. This study conducts a pilot substudy to test the hypothesis that 22 changes in insulin sensitivity observed in response to a diet intervention correlate with 23 changes in intramyocellular and/or intrahepatocellular lipid. 25 1.2. Protocol Overview 27 In a 16-week trial, overweight adults will be randomly assigned to two groups. Changes 28 in insulin sensitivity, postprandial metabolism, and body weight will be the primary 29 dependent variables. 31 The Diet Group will be asked to follow a low-fat, vegan diet and will receive weekly 32 classes and support. 34 The Control Group will be asked to make no changes in diet or exercise for 16 weeks, 35 but will be instruct 37 1.3. Investigative Team 39 The project will be conducted by investigators from the Physicians Committee for 40 Responsible Medicine (PCRM), a nonprofit 501(c)(3) organization located at 5100 41 Wisconsin Avenue, NW, Washington DC 20016, which conducts nutrition-related 42 research. Its medical, nutrition, and research staff will oversee participant recruitment, 43 screening, group assignment, nutrition teaching and monitoring, and data collection and 44 analysis. Recruitment interviews, the dietary intervention, and most assessments will 45 take place at its offices. 47 Researchers from the Department of Internal Medicine, Yale University School of 48 Medicine, will conduct MR spectroscopy studies for intracellular lipid. 50 2. BACKGROUND AND SIGNIFICANCE 52 Excess body weight is a major contributor to many health problems, including diabetes, 53 cardiovascular disease, orthopedic problems, and certain forms of cancer. In 54 epidemiologic studies, individuals following vegan diets tend to have significantly lower 55 body weights, compared with individuals following other dietary patterns. In studies of 56 overweight individuals, the adoption of a low-fat plant-based diet predictably reduces 57 body weight, even in the absence of any specified limitation on energy intake. 58 The mechanisms by which plant-based diets reduce body weight are not entirely clear. 59 Previous studies have identified two possible explanations: First, to the extent that 60 vegan diets are low in fat and high in fiber, they have a relatively low energy density, 61 which reduces energy intake. Second, a low-fat vegan diet may increase postprandial 62 metabolism (the thermic effect of food). These observations suggest that the diet leads 63 to weight loss by (1) reducing energy intake and (2) increasing postprandial energy 64 output. 65 A prior study including 64 overweight postmenopausal women randomly assigned to a 66 low-fat vegan diet or a comparison diet based on the guidelines of the National 67 Cholesterol Education Program for 14 weeks found that the vegan diet led to 68 significantly greater weight loss (5.8 kg for the vegan group, compared with 3.8 kg for 69 the comparison group). The vegan diet group also had a 16% increase in postprandial 70 metabolism and an increase in insulin sensitivity that was significant within group, 71 although not between groups. However, because the comparison group used an active 72 diet and there was no untreated control group, that study was not able to show the 73 degree to which a plant-based diet influences energy expenditure, compared with 74 untreated participants. 75 Insulin resistance has been shown to be related to fat accumulation within muscle cells 76 (intramyocellular lipid) and liver cells (intrahepatocellular lipid). The above findings 77 suggest the possibility that low-fat, vegan diets reduce the quantity of lipid stored within 78 these cells, which, in turn, improves insulin sensitivity. 79 Some evidence suggests that the accumulation of intracellular fat may be responsive to 80 diet. In a 2012 study at Yale University, 7 lean, young individuals who were known to be 81 insulin-resistant and who had parents with type 2 diabetes underwent a hypocaloric 82 (1200 kcal/d) diet for 9 weeks, leading to an average weight loss of 4.1 ± 0.6 kg. During 83 this intervention period, average intramyocellular lipid fell approximately 30%, from 1.1 ± 84 0.2% to 0.8 ± 0.1%. 85 High-fat diets appear to downregulate the genes required for mitochondrial oxidative 86 phosphorylation in skeletal muscle and increase intramyocellular lipid. In contrast, a 87 case-control study found that soleus muscle intramyocellular lipid concentrations were 88 significantly lower in a group of 21 vegans, compared with 25 omnivores. 89 In research studies, the acceptability of plant-based diets appears to be similar to that of 90 other therapeutic diets over both the short and long term, as indicated by rates of 8-11 91 retention, diet adherence, and diet acceptance questionnaires. If a plant-based diet 92 increases postprandial metabolism, its use for the prevention and management of 93 weight disorders and related health problems will have a more solid rationale. This is 94 especially important given that weight problems are widespread and there is a great 95 deal of confusion among the public regarding which diet method to select. 97 3. RESEARCH DESIGN, RECRUITMENT, AND ASSESSMENTS 99 3.1. Overview of Research Design 101 In a randomized, controlled trial, we will test the effects of a low-fat, plant-based diet on 102 insulin sensitivity, postprandial metabolism, and body weight in overweight adults over a 103 16-week period, using for comparison an untreated control. A substudy will examine 104 effects on intracellular lipid. 106 3.2. Key Personnel 108 Key personnel include: 109 Neal D. Barnard, MD, FACC, Principal Investigator, is an Adjunct Associate Professor 110 of Medicine at the George Washington University and President of PCRM. He has been 111 the Principal Investigator of several clinical trials, as noted herein. 112 Hana Kahleova, MD, PhD, is an endocrinologist and Director of Clinical Research at 113 PCRM. She has been involved in several clinical trials in diabetes and insulin 114 resistance. 115 Susan Levin, MS, RD; Karen Smith, RD, and Maggie Neola, RD, are Registered 116 Dietitians at PCRM who provide nutrition instruction and participate in clinical 117 assessments. 118 Francesca Valente and Rosendo Flores coordinate clinical research studies at 119 PCRM. 120 Richard Holubkov, PhD, is a biostatistician with the University of Utah, who works with 121 PCRM on contract. 123 Gerald I. Shulman, MD, PhD, FACP, MACE, is a Professor of Medicine, Cellular and 124 Molecular Physiology and The Howard Hughes Medical Institutes at Yale University 125 School of Medicine. 126 Kitt Falk Petersen, MD, is a Professor of Medicine at Yale University School of 127 Medicine. 129 3.3. Recruitment and Screening Procedures 132 offices, letters sent to patients of medical practitioners, and advertisements placed in 133 newspapers, on radio, and in buses in the Washington, DC, area, as well as social 134 media postings. (Appendix 1) 137 participants using a telephone screening script (Appendix 2). Research staff will 138 explain the study, review participation criteria, and inquire about other motivations for 139 volunteering, filling out a paper interview screening form for each person who calls. 140 Volunteers who satisfy the participation criteria will be scheduled for group and/or 141 individual information sessions. The names/identifiable information of volunteers who do 142 not satisfy participation criteria will be destroyed (shredded) immediately. For these 143 individuals, the research team will retain only de-identified demographic information and 144 the reason for exclusion, for purposes of evaluating participation statistics. 146 At the group and/or individual information sessions (some volunteers may attend 147 individually; others may be seen in groups), the investigators and research staff will 148 explain the study and its scientific basis in detail and review participation criteria in 149 simple, nontechnical terms. They will also provide instruction on filling out a diet record. 150 Additional content will be determined by questions raised by volunteers, and may relate 151 to study logistics, the recruitment process, the content of the vegan diet and ease of 152 following it, clinical assessments, or the weekly classes. To protect patient privacy, 153 volunteer names will not be used at these meetings. Volunteers will have a chance to 154 ask questions about the study and the informed consent process in private, and each 155 volunteer will meet in private with study personnel at the conclusion of the group 156 session, even if he or she has no questions. Volunteers who choose to complete the 157 informed consent document (Appendix 3) will also be asked to complete a contact 158 information form (Appendix 4) and a general medical history form (Appendix 5). 159 Volunteers will be assigned identification numbers in the order in which they complete 160 the informed consent document. These numbers will be used in place of identifying 161 information for purposes of data collection, assessment, and analysis. During the 162 screening process, prospective volunteer 163 phone number and contact information in order to be able to reschedule and cancel 164 their appointments, if necessary. 166 Prospective volunteers will then be asked to complete a practice 2-day dietary record to 167 demonstrate their ability to track nutrient intake for research purposes. A letter will be 168 sent to the participants to remind them to complete their 2-day diet record with 169 instructions (Appendix 6). When completed, these records will be reviewed for 170 completeness by a registered dietitian. 172 Volunteers who have completed the informed consent process and practice dietary 173 records and meet the study participation criteria will be asked to schedule individual 174 appointments for baseline assessments. There, they will be asked to fill out new 3-day 175 dietary records. Volunteers can either submit questionnaires online, print, scan and 176 email them, or send them through regular mail. Those who satisfactorily complete the 177 baseline assessments and 3-day dietary records will be enrolled in the study. 179 The cost of all tests and procedures will be covered by PCRM. If any examination or 180 test reveals that a participant has a medical condition that requires additional diagnostic 181 tests or treatment, research staff will advise the participant of that fact, but will not 182 provide such additional diagnostic tests or treatment. There is no cost for the weekly 184 assessments, meetings, and group sessions. 187 3.4. Inclusion and Exclusion Criteria 189 Inclusion criteria are as follows: 191 1. Men and women age 18 years of age 192 2. Body mass index 28-40 kg/m 194 Exclusion criteria are as follows: 196 1. Diabetes mellitus, type 1 or 2, history of diabetes mellitus or of any endocrine 197 condition that would affect body weight, such as thyroid disease, pituitary 198 abnormality, or Cushing's syndrome 199 2. Smoking during the past six months 200 3. Alcohol consumption of more than 2 drinks per day or the equivalent, episodic 201 increased drinking (e.g., more than 2 drinks per day on weekends), or a history of 202 alcohol abuse or dependency followed by any current use 203 4. Use of recreational drugs in the past 6 months 204 5. Use within the preceding six months of medications that affect appetite or body 205 weight, such as estrogens or other hormones, thyroid medications (unstable 206 dose within the preceding 6 months), systemic steroids, antidepressants 207 (tricyclics, MAOIs, SSRIs), antipsychotics, lithium, anticonvulsants, appetite 208 suppressants or other weight-loss drugs, herbs for weight loss or mood, St. 209 John's wort, ephedra, beta blockers 210 6. Pregnancy or intention to become pregnant during the study period 211 7. Unstable medical or psychiatric illness 212 8. Evidence of an eating disorder 213 9. Likely to be disruptive in group sessions 214 10. Already following a low-fat, vegan diet 215 11. Lack of English fluency 216 12. Inability to maintain current medication regimen 217 13. Inability or unwillingness to participate in all components of the study 218 14. Intention to follow another weight-loss method during the trial 221 3.5. Group Assignment 223 Participants will be told that, if accepted, they will be assigned either to a Diet Group or 224 a Control Group. Accepted volunteers will be assigned to these groups using a 225 computer-generated random-number table. Because assignment will be done 226 simultaneously within each replication, allocation concealment is unnecessary. 228 3.6. Clinical Assessments 230 The following determinations will be made at baseline and 16 weeks, except as noted: 232 Assessments of Dietary Intake and Physical Activity 234 3-day dietary record. A 3-day dietary record will be used to assess macro- and 235 micronutrient intakes. Records will be analyzed using Nutrition Data System for 236 Research software version 2016, developed by the Nutrition Coordinating Center 237 (NCC), University of Minnesota, Minneapolis, MN, US, by a registered dietitian certified 238 by the NCC. Random 24-hour dietary recalls will be conducted by registered dietitians 239 to determine compliance but will not be part of the final nutrient analysis. 241 The International Physical Activity Questionnaire short form assesses recent 242 physical activity patterns. The method is highly reliable; an assessment of test-retest 243 repeatability produced a correlation of 0.8. (Appendix 7) 245 Assessments of Physical Health, Weight, and Metabolism 247 General status, symptoms, and medication accounting. Participants will be asked to 248 report changes in their health and medication use. 250 Height. Height will be measured at baseline (only) with participants standing barefoot 251 with their backs to a wall-mounted stadiometer and heels against the wall, recorded to 252 the nearest 0.5 cm. 254 Body weight. With participants wearing light, indoor clothing but without shoes, body 255 weight will be measured to the nearest 0.1 kg, using a digital scale. Body weight will 256 also be assessed at each weekly group session for the Diet Group, but only data from 257 weeks 0, 8, and 16 will be included in the analysis. 259 Comprehensive Metabolic Panel. These values will be evaluated at baseline only. 261 Serum cholesterol and triacylglycerol concentrations and hemoglobin A1c will be 262 measured using standard methods. 264 The following measures will be assessed at weeks 0 and 16: 266 Glucose Tolerance and Insulin Sensitivity. An oral glucose tolerance test will be 267 performed for three hours after an overnight fast. (Matsuda 1999) 268 Resting Energy Expenditure (REE). Participants will be asked to report to the 269 laboratory within 60 minutes of waking and after a 12-hour fast. Following 30 minutes of 270 quiet rest in a dimly lit room, pulse, respiratory rate, and body temperature will be 271 measured. REE will be measured for 20 minutes through indirect calorimetry (Cosmed 272 Quark RMR, Chicago, IL) utilizing a ventilated hood system. The laboratory temperature 273 will be maintained at 23 degrees C throughout, and precautions will be taken to 274 minimize any disturbances that could affect the metabolic rate. 275 For premenopausal women, measures will be timed so as to occur in the luteal phase of 276 the menstrual cycle. 277 Postprandial metabolism (thermic effect of food, TEF). After the REE determination, 278 participants will be given a 720-kilocalorie test meal (Sustacal, Mead Johnson, 279 Evansville, IN) to be ingested within 10 minutes. Metabolic rate will be measured in the 280 same manner as above for 30 minutes at 2 and 4 hours postingestion. 281 Body Composition. Body composition will be measured by dual energy x-ray 282 absorptometry (Lunar iDXA, GE Healthcare; Madison, WI) with Encore® 2005 283 v.9.15.010 software. The iDXA can measure body composition with low X ray exposure 284 and short scanning time. The iDXA unit will be calibrated daily using the GE Lunar 285 calibration phantom, and a trained operator will perform all scans following standard 286 protocol for participant positioning. The iDXA is equipped with the CoreScan module 287 (GE Healthcare, Madison, WI), which can also provide an estimate of visceral adipose 288 tissue volume and mass. 289 Intramyocellular and Hepatic Lipid Content. A subset of participants will be selected 290 for MR spectroscopy studies quantifying hepatic lipid and/or intramyocellular and/or 291 contents in order to provide data regarding possible causal relationships between 292 dietary changes, ectopic lipid, and insulin sensitivity. Selected individuals with varying 293 degrees of insulin-resistance in both groups will be assessed before and after the 294 intervention period. These MRS studies will take place at the Magnetic Research Center 295 at Yale University School of Medicine, New Haven, CT. 297 Intramyocellular and hepatic lipid contents will be measured using H MRS at 4T 298 (Bruker). After safety procedures including completion of the Yale Magnetic Research 299 Center Safety Questionnaire, changing into scrubs and passing through the metal 300 detector, the participants will positioned on their back on the bed, which slides into the 301 MRS instrument. For the leg lipid measurements the right leg will be positioned in a 302 holder with the calf muscle over a receiver coil. 304 Muscle lipid content will be measured in the soleus muscle using an 8.5-cm diameter 13 1 305 circular C surface coil with twin, orthogonal circular 13-cm H quadrature coils. The 306 probe will be as tuned and matched and scout images of the lower leg will be obtained 307 to ensure correct positioning of the participant and to define an adequate volume for 308 localized shimming using the FASTMAP procedure. The measurement will take 309 approximately 30 minutes. 311 After the lipid measurements in the leg a receiver coil embedded in a plastic plate will 312 be positioned on the side of the abdomen over the liver and strapped in place with velco 313 straps. The position of the coil will be confirmed with MR images and the location of the 314 lipid measurement within the coil will be determined from these MR images.Liver 315 triglyceride content will be measured by H respiratory-gated STEAM spectroscopy in a 316 15 × 15 × 15-mm3 voxel. Acquisition will be synchronized to the respiratory cycle and 317 triggered at the end of expiration. A water-suppressed lipid spectrum and a lipid- 318 suppressed water spectrum will be acquired in three different locations of the liver to 319 account for liver inhomogeneity. A minimum of three spectra will be acquired for each 320 participant and the total lipid content will be averaged and calculated. In addition, 321 hepatic lipid content will be corrected for transverse relaxation, using the transverse 322 relaxation times of 22 ms for water and 44 ms for lipid. These MRS measurements will 323 take approximately 30 minutes. 326 Table 1: Study Procedures Schedule Week 0 16 3-day diet record International Physical Activity Questionnaire Clinical status and symptoms Medication use Height Body weight* Comprehensive Metabolic Panel (CMP) Plasma lipids and lipoproteins A1c Glucose tolerance testing REE Postprandial metabolism (TEF) Body composition IMCL and hepatic lipid (pilot) 330 4. INTERVENTION PROCEDURES 332 4.1. Intervention Diet 334 The interventions for the Diet and Control Groups are described below. 336 The Diet Group will be asked to follow a low-fat, vegan diet. According to the Academy 337 of Nutrition and Dietetics, vegan and vegetarian diets meet all nutritional requirements 338 when appropriately planned. The diet consists of whole grains, vegetables, legumes, 339 and fruits, with no restriction on energy intake. Animal products and added oils will be 340 excluded. In choosing grain products and starchy vegetables (e.g., bread, potatoes), 341 participants will be encouraged to select those retaining their natural fiber and having a 342 glycemic index <70, using tables standardized to a value of 100 for glucose. No meals 343 will be provided. Participants will handle their own food preparation and purchases, with 344 guidance from the research team. 346 The diet derives approximately 10% of energy from fat, approximately 10-15% of energy 347 from protein, and the remainder from complex carbohydrates. The diet will provide 348 approximately 30-40 grams of fiber per day. It is generally adequate in all nutrients 349 except vitamin B . 351 Participants will be provided with a commercially available supplement containing 100 352 micrograms of vitamin B and asked to take it daily during the study. Should they wish 353 to continue the diet thereafter, they will be counseled to use any standard multivitamin 354 or other reliable source of vitamin B . 356 An advantage of studies such as this one, which include volunteers who are not 357 confined to a metabolic ward or otherwise restricted, is that they can readily translate to 358 nonclinical settings. A disadvantage is that they include a degree of uncertainty as to 359 the extent to which participants have adhered to their prescribed diets. While this 360 uncertainty cannot be entirely eliminated, several measures will be taken to maximize 362 associated with dietary compliance in clinical trials. Stricter limits on fat intake, frequent 363 monitoring of reported dietary intake, family involvement, group support, and the use of 364 vegetarian diets are associated with a greater degree of dietary change. 366 Control Group members will be asked to continue their usual diets for the 16-week 367 study period. Those who wish to try the intervention diet will be given instruction in the 370 Both groups: For both groups, alcoholic beverages will be limited to one per day for 371 women, and two for men. 373 4.2. Dietary Instruction and Group Meetings 375 Diet Group participants will be asked to attend weekly, one-hour group sessions for 376 support and education. (Class Curriculum, Appendix 8). No weekly support or 377 education will be provided to the participants in the Control Group. 379 All group sessions will be conducted by a registered dietitian, nurse, physician, cooking 380 instructor, or research staff and will include information on nutrition, meal planning, 381 shopping, food preparation techniques, recipes, and everyday dietary challenges, such 382 as dining out and healthful snacking. The classes will also include education on topics 383 such as maintaining a healthy weight, cholesterol, hypertension, diabetes, and other 384 health issues. 386 For some sessions, participants will be encouraged to bring a spouse, partner, family 387 member, or friend. To facilitate interaction between diet instructors and participants, 388 classes will be conducted in sections of approximately 15 participants. 391 Health Belief Model developed by researchers with the Public Health Service and 392 adapted by others. This model describes constructs that predict health-related 393 behaviors and should be considered when planning behavioral change strategies. 394 These include perceived susceptibility, severity, benefits, and barriers, as well as cues 395 to action, and self-efficacy. Our participants are already aware that they are overweight 396 and may benefit from diet changes. Nonetheless, they need help in overcoming barriers 397 and gaining confidence in their ability to implement new dietary habits. We have 398 therefore focused the content of the weekly support group sessions on integrating 399 practical skills (e.g., menu planning, food preparation, dining out, healthful snacking) 400 with their growing understanding of how dietary choices affect health. In order to 401 facilitate individual experience with the prescribed diet, practical skills are presented 402 early, while intellectual understanding of more complex health issues (e.g., how diet 403 affects heart disease risk) is presented later. Each group session includes time for 404 participants to discuss their successes and challenges, and group problem-solving is 405 encouraged. 407 The study does not seek to separate the effects of the diet from those of regular group 408 support. Rather, group support is a means of facilitating adherence. It should also be 409 emphasized that the goal of this study is not to construct an intervention diet that is 410 isocaloric compared with diets followed by the Control Group participants. Because the 411 intervention diet is low in fat and high in fiber, self-selected energy intake is likely to fall 412 as the diet period begins, and weight loss is likely. 414 4.3. Exercise and Medication Use 416 Participants in both groups will be asked to keep their level of physical exercise and use 417 of medications constant and to add no new nutritional supplements to their current 418 medication regimens, except as recommended by their personal physicians. 420 4.4. Intervention Fidelity and Dietary Adherence 422 Individual meetings. During the initial individual meal-planning meetings with the Diet 423 Group participants, dietitians will follow a set agenda which will cover the use of vegan 424 foods, methods for reducing dietary fat, and how to avoid proscribed foods. 426 Group meetings. To maintain intervention fidelity, the group leaders will follow a set 427 course curriculum, using an agenda for each session and keeping a checklist of major 428 content items to be covered at each meeting. 430 Dietary Adherence. Each participant will complete diet records at regular intervals 431 using the methods described above. In addition, for the Diet Group, 24-hour multi-pass 432 dietary recalls will be used to assess dietary adherence to assist study personnel in 433 working with individuals who need additional teaching or support. The 24-hour recalls 434 will be performed either by telephone or in person at weeks 3 and 8. These recalls will 435 not be subjected to statistical analysis, but will allow the investigators to check for poor 436 adherence. Such recalls have the advantage that they can be conducted at 437 unscheduled times and over the telephone, and so are not subject to the planning and 438 preparation required for food records. In cases where participants appear to be 439 deviating from the prescribed diet, additional dietary counseling will be provided. 441 4.5. Participant Retention 444 reasons for volunteering other than a desire to improve their health or to advance 445 scientific understanding may be rejected. The exclusion criteria also eliminate 446 individuals with a history of unresolved substance abuse, which may influence retention. 448 Participants will be instructed that attendance at meetings is essential to study 449 participation. The research team will take attendance at each meeting. The research 450 staff will make phone calls to participants who do not attend. 452 In the weekly meetings, group support will be facilitated through group discussions and 453 encouragement to share successes and difficulties with the prescribed diet. Meeting 454 content will remain varied, including nutrition lectures, health education, cooking 455 demonstrations, and opportunities to taste food. Family members will be invited to 456 certain support group sessions. A voluntary listserv will allow Diet Group members to 457 exchange information, recipes and ideas between meetings. Only participants, study 458 coordinators, and the PI will be allowed to post on the list serve and the content of the 459 list serve will be accessible only by them. 461 Participants who complete all assessments at weeks 0 and 16 will be paid $100 at 462 completion of their final assessments. 464 4.6. Biological Specimen Handling Procedures 466 Samples for the study endpoints will be drawn by LabCorp and will be processed using 467 standard procedures. 470 5. STATISTICAL PROCEDURES 472 5.1. Power Analysis 474 Power Analysis for Overall Study 476 Sample size will be based on the change in postprandial metabolism (thermic effect of 477 food) previously observed with a plant-based diet, compared with an active dietary 478 control. The current power analysis assumes there will be a single t-test for the 479 comparison of the changes in thermic effect of food observed in the two study groups, 480 with an alpha level of 0.05. 481 In the prior study, the change at 14 weeks for the thermal effect of food was 4.7 with an 482 SD of 12 in the intervention arm, and 0.3 with an SD of 9.4 in the control arm. Assuming 483 that the true treatment effect is 4.4 kcal/170 min and that the SD of the change will be 484 12.0 units in the Diet Group and 9.4 units in the Control Group, as previously observed, 485 the sample size required is 96 per arm (192 total) for 80% power, 109 per arm for 85% 486 power, 128 per arm for 90% power. 487 However, if this magnitude of treatment effect is assumed to be present at the same 488 magnitude of effect at each of the 5 evaluation points used in TEF assessment and 489 assuming that the standard deviation is about 10.85 points for all observations, with 5 490 observations per participant, correlated at a magnitude of 0.7 with each other, the 491 sample size required is 73 participants per arm (146 total) with 80% power, 83 per arm 492 with 85% power, 98 per arm for 90% power. 493 Assuming an attrition of 10%, the required sample size is 81 per group, or 162 total for 494 80% power. 496 Power Analysis for Intracellular Lipid Substudy 498 Two studies, cited above, provide a basis for a power analysis for the substudy 499 assessing the role of intramyocellular and hepatic lipid on insulin sensitivity and, 500 ultimately, postprandial metabolism. In the 2012 Yale study, 7 lean, young insulin- 501 resistant individuals whose parents had type 2 diabetes followed a hypocaloric (1200 502 kcal/d) diet for 9 weeks, leading to an average weight loss of 4.1 ± 0.6 kg. During this 503 intervention period, average intramyocellular lipid fell approximately 30%, from 1.1 ± 504 0.2% to 0.8 ± 0.1%. In an observational study including 21 individuals following vegan 505 diets and 25 following omnivorous diets, soleus muscle intramyocellular lipid for the 506 vegan participants was found to be 11.7 (6.1 24.6), compared with 16.9 (2.7 44.7) for 507 the omnivorous participants (P = 0.01). The 95% confidence interval for the difference 508 was reported to be -13.2 to -3.3). 510 Based on the Yale study, assuming a change in IMCL of 0.3 percentage points with a 511 standard deviation of 0.2 and, in the control arm, a mean change of zero with a similar 512 standard deviation, to have 90% power to detect a difference of this magnitude between 513 the two arms would require 11 subjects per arm. Ten per arm would yield 88% power. 514 Because this is an exploratory substudy and variability in response to the diet is largely 515 unknown, we aim to include 20 participants per arm in the substudy, for a total of 40 516 participants. 518 5.2. Data Management 520 All laboratory samples, reports, questionnaires, and data sheets will be coded with 521 participant identification numbers, rather than names. Laboratory reports will be 522 delivered to the PCRM office at 5100 Wisconsin Avenue, Washington, D.C., where they, 523 along with all other history and data forms, will be maintained in individual participant 524 files in a locked cabinet. 526 Data will be promptly entered into the data tables at PCRM using Microsoft Excel. Two 527 research staff members will check the tables for accuracy against the original 528 documents. Data tables will be routinely copied onto back-up files and stored for safety 529 on an off-site, passcode-protected, secure server. Data grids will be sent electronically 530 to the biostatistician for analysis. Registered dietitians will be provided with information 531 on usual ranges for nutrients or intakes of interest and asked to check their original data 532 and analysis for errors if they fall outside of these ranges. 534 5.3. Statistical Analysis 536 Descriptive statistics for all demographic variables and clinical measures will be 537 calculated for each group. To determine if there are statistically significant differences 538 between the 2 groups at baseline, t-tests will be calculated for continuous measures 539 and chi squares will be calculated for categorical measures. Regardless of any 540 differences, baseline values for key outcome variables will be included as covariates in 541 the main assessments of the effect of diet in the multivariate analysis of covariance. An 542 alpha of 0.05 will be used for all statistical tests. 544 For nutrient intake and physical measures, descriptive statistics (means, standard 545 deviations, tests for normality) will be calculated. If data are normally distributed, 546 parametric tests for significant effects will be used; for non-normally distributed 547 variables, non-parametric tests will be used. 549 The initial test of the hypotheses will be examined by performing t-tests for independent 550 samples on the difference score denoting the change from baseline to the reporting 551 period. For missing data in a reporting period, values from the previous period will be 552 brought forward. For body weight, drop-outs will be considered to have returned to 553 baseline weights. 555 5.4. Assessment of Diet Adherence. Diet-and-supplement group participants will be 556 described as adherent or non-adherent based on whether they met the following 557 criteria: absence of proscribed foods reported on 24-hour recalls and diet records, 558 saturated fat <5% and total fat <25% of energy, and average daily cholesterol intake 559 <50 mg on 3-day dietary records. 561 For drop-out rates, we will determine if there are between-group differences, using chi- 562 square. 564 5.5. Assessment of Medication use. Any changes to lipid-lowering medications will be 565 classified as a net increase, net decrease, or mixed (changes in opposing directions for 566 2 or more medications). Using chi-square, we will determine whether there are 567 differences in medication changes between the 2 groups. 570 6. TIME LINE AND PARTICIPANT FLOW 572 6.1. Time Line 573 The study will be conducted in six cohorts, each including 40 participants, over a 3-year 574 period. 576 For the first cohort, recruitment will take place October 2016 January 2017. The 577 intervention, including weekly meetings, will take place between February and May 578 2017 for a total of 16 weeks. For the second cohort, recruitment will take place in 579 December 2016 and January, 2017. The intervention, including weekly meetings, will 580 take place between March and July 2017. Similar time frames will apply in the two 581 subsequent years. 583 6.2. Participant Flow Based on Power Analysis 585 In order to accommodate this number of participants, baseline metabolic assessments 586 will occur as follows for the first cohort: 588 January 6: 2 participants 589 January 9-13: 10 participants 590 January 17-20: 8 participants 591 January 23-27: 10 participants 592 January 30-Feb 3: 10 participants 594 The MR spectroscopy studies will be limited to 40 participants (20 per study arm) total. 595 These evaluations will be scheduled during the above scheduled days. 597 The Diet Group sessions for the first cohort will be held for 16 weeks, from February 1 598 through May 17, 2017. 600 For the first cohort in 2017, the 16-week metabolic assessments will occur as follows: 602 May 18-19: 4 participants 603 May 22-26: 10 participants 604 May 30-June 2: 8 participants 605 June 5-9: 10 participants 606 June 12-15: 8 participants 608 For the second cohort, the assessment and intervention dates will be approximately one 609 month later than those for the first cohort. 611 7. PROTECTION OF HUMAN RESEARCH PARTICIPANTS 613 7.1. Risks to the Subjects 615 Sources of Materials: Participants will be asked to complete questionnaires, provide 616 blood samples, and have several physical assessments. 618 Human Subjects Involvement and Characteristics: The proposed research will 619 include participants at least 18 years of age. 621 Potential Risks: Participation in the study entails the following risks: 623 1. Blood draws can cause transient pain, occasionally cause bruising, and may 624 cause bleeding. 626 2. A well-planned vegan diet provides all the nutrients people need except for 627 vitamin B12. People with a vitamin B12 deficiency may suffer from anemia and 628 neurologic damage. 630 3. Loss of confidential information. 632 7.2. Adequacy of Protection against Risks 634 Recruitment and Informed Consent: 635 and procedures and review the inclusion and exclusion criteria. Volunteers who appear 636 to meet the criteria for participation will be invited to a group or individual interview with 637 the principal investigator and the study coordinator, who will explain the study in detail, 638 answer questions, and provide a written consent form, as approved by the IRB. 639 Participants will have the opportunity to ask any questions individually in a private 640 setting and may take as much time as they would like to review the informed consent 641 document. The consent form will be signed by the volunteer participant and study 642 coordinator. The principal investigator will certify that the research study has been 643 explained to the volunteer, including the purpose, procedures, possible risks, and 644 potential benefits associated with participation and that any questions have been 647 he study and 648 that the investigators will not manage any aspects of their medical care. 650 To maintain confidentiality, all laboratory specimens, questionnaires, forms, and data 651 sheets will identify participants by their assigned numbers only. Data and safety 652 monitoring are described below. 654 Phlebotomy risks. All blood draws will be carried out by experienced phlebotomists at 655 Quest Diagnostics. 657 Vitamin B12 deficiency. All Diet Group participants will be given a supply of vitamin 658 B12, 100 micrograms, and will be asked to take it daily. Participants will also be 659 counseled to continue B12 supplementation if they plan to continue following a vegan 660 diet. 662 Loss of confidential information. We will make every effort to keep all research 663 records private to the extent allowed by law. We will use an identification number on 664 forms, instead of identifiable information. All study documents will be kept in locked filing 666 learn from this study may be shared at scientific or medical meetings and may be 667 published, but participants will not be personally identified. 669 Our protocol also includes the following safeguards: 671 1. All participants will remain under the care of their personal healthcare providers. 673 2. All participants will continue on the medications they were using at study entry, 674 unless modified by their personal physician(s). 676 We therefore believe that the risks to participants in a dietary intervention trial are 677 minimal, while the scientific and public health merit of such an investigation is high. By 678 studying the benefits of a dietary intervention, we hope to obtain valuable research data. 680 Confidentiality. To maintain confidentiality, all laboratory specimens, questionnaires, 681 forms, and data sheets will identify participants by their assigned numbers only. Data 682 and safety monitoring are described below. 684 7.3. Potential Benefits of the Proposed Research to the Participants and Others 686 Given that, over the long run, excess body weight contributes to morbidity and mortality, 687 the dietary instruction and consistent support provided may be of substantial benefits for 688 the Diet Group. The Control Group will be given detailed information on how to follow 691 7.4. Importance of the Knowledge to Be Gained 693 Weight problems are extremely common, and gaps remain in our understanding of how 694 intervention diets work. This study is founded on clear theoretical constructs and 695 compelling previous data on both the efficacy and acceptability of the experimental 696 intervention, as well as preliminary findings on its mechanisms of action. It investigates 697 what may be a major advance in the understanding of the role of diet in weight control. 698 The risks to participants are small, and the potential benefits are significant. 700 7.5. Assessment and Reporting of Adverse Events 702 An adverse event is any adverse physical or clinical change experienced by a 703 participant. This includes the onset of new symptoms and the exacerbation of pre- 704 existing conditions. In order to avoid bias in eliciting reports of adverse events, 705 participants will be asked, during assessments at the 706 you had any new symptoms, injuries, illness or side effects or worsening of pre-existing 709 All adverse events will be recorded in the participant's record and on the IRB continuing 710 review form. The severity of the adverse event will be assessed, and actions/outcomes 711 (e.g., hospitalization, discontinuation of therapy, etc.) will also be recorded. 713 Any actions taken and follow-up results will also be recorded on the appropriate page of 714 the IRB continuing review form, as well as in the participant's record. Follow-up 716 at a site will be reported by the investigator to the IRB according to the Data and Safety 717 Monitoring Plan, described below. 719 The following definitions will be used: Minimally serious: Awareness of sign, symptom, or event, but easily tolerated. Somewhat Serious: Discomfort enough to cause interference with usual activity and may warrant investigation. Very Serious: Incapacitating, with inability to do usual activities, or significantly affects clinical status, and warrants intervention. Life-threatening: Immediate risk of death. 721 The research team will also assess the relationship of any adverse event to the study 722 intervention, based on available information, using the following guidelines: 0 = No temporal association, or the cause of the event has been identified, or Unlikely the study interventions cannot be implicated. 1 = Temporal association, but other etiologies are likely the cause; however, Possibly involvement of the study interventions cannot be excluded. 2 = Temporal association or other etiologies are possible, but unlikely. Probably 724 7.6. Serious Adverse Events (SAEs) 726 All serious adverse events, whether or not deemed intervention-related or expected, will 727 be reported by telephone to the Safety Officer within 24 hours (one working day) of the 728 time they become known. A written report will follow as soon as possible, including a full 729 description of the event and any sequelae. This includes serious events that occur any 730 time after the inclusion of the patient in the study until completion of the last visit. A 731 serious adverse event report will also be sent via fax to the IRB chair. 733 A serious adverse event is any event that falls in any of the following categories: 734 Fatal 735 Life-threatening (the patient was at immediate risk of death from the AE as it 736 occurred) 737 Significantly or permanently disabling 738 Requires hospitalization or prolongs hospitalization 740 Important medical events that may not result in death, be life-threatening, or require 741 hospitalization may be considered serious adverse events when, upon appropriate 742 medical judgment, they may jeopardize the patient and may require medical or surgical 743 intervention to prevent one of the outcomes listed in the definition. The death of any 744 patient during the study, regardless of the cause, will be reported within 24 hours by 745 telephone to the Safety Officer and IRB. A full written report will follow as soon as 746 possible. If an autopsy is performed, a copy will be provided to the Safety Officer and 747 IRB. 749 Reports of all serious adverse events, including deaths, will be communicated to the 750 IRB in accordance with local laws and regulations. 752 7.7. Action plan if a subject becomes severely depressed or suicidal during the 753 course of the study 755 Participants with a history of severe mental illness (with current unstable status), such 756 as severe depression or suicidality, will not be enrolled in the study as indicated in the 757 exclusion criteria. If a participant becomes severely depressed during the course of the 758 study, he or she will be referred to see his or her primary care physician or psychiatrist 759 and to seek medical care. The event will be recorded and reported to the PI and Safety 760 Officer immediately. The PI will notify IRB within 24 hours of recognition of the event by 761 study personnel. All non-serious events will be reported and reviewed by the PI within 762 one week. Study personnel will inform the primary care physician of all events occurring 763 in his/her patients within 48 hours of recognition of the event. 765 If a participant becomes suicidal during the course of the study, he or she will be 766 instructed to call 911 and seek emergency medical care. The incident will be recorded 767 and reported to the PI and Safety Officer immediately. The PI will notify the IRB within 768 24 hours of recognition of the event by study personnel. All non-serious events will be 769 reported and reviewed by the PI within one week. Study personnel will inform the 770 primary care physician of all events occurring in his/her patients within 48 hours of 771 recognition of the event. 773 7.8. Data and Safety Monitoring Plan 775 Data and Safety Monitoring functions will be performed by the principal investigator (PI, 776 Neal D. Barnard, M.D.), study coordinator (Francesca Valente), study statistician, and a 777 Safety Officer who is a physician who is not part of the research staff and has no role in 778 care of the participants. The Safety Officer will have no scientific, financial, or other 779 conflict of interest related to the trial. Prior to the study onset, the study statistician and 780 Safety Officer will review the research protocol, informed consent documents, and plans 781 for data and safety monitoring. 783 During the recruitment phase, the study coordinator and PI will review enrollment 785 including accrual, demographics, thoroughness of baseline data, subject status 786 (reporting concurrent illnesses, withdrawal of consent, or loss to follow-up), and 787 adherence to participation criteria, informed consent procedures, and the study protocol. 788 The reports will be submitted to the study statistician and Safety Officer. 790 -existing 791 symptoms and medical problems will be recorded. During each monitoring visit, study 792 participants will be asked if any medical symptom, problem, or event has occurred or if 793 there has been any change in pre-existing symptoms. 795 All serious events (hospitalization, serious illness, or disability) will be recorded and 796 reported to the PI and Safety Officer. The PI will notify the IRB within 24 hours of 797 recognition of the event by study personnel. All non-serious events will be reported and 798 reviewed by the PI within one week. Study personnel will inform the primary care 799 physician of all events occurring in his/her patients within 48 hours of recognition of the 800 event. 802 At monthly intervals, the study coordinator and PI will prepare and submit to the study 803 statistician and Safety Officer a report covering each of the following areas: (1) 804 performance (including adherence to the study protocol and maintenance of data 805 integrity and confidentiality), (2) safety (including abnormal laboratory values, adverse 806 events, serious adverse events, deaths, and disease- or treatment-specific events), and 807 (3) treatment effects, including medication changes. 809 The study statistician and Safety Officer will review each safety report within one week 810 of receipt. The study statistician will review these reports to assess whether event rates 811 are of statistical concern and, if so, will alert the Safety Officer, the PI, and the IRB. The 812 study statistician and Safety Officer will also consider factors external to the study, e.g., 813 new scientific developments, that may affect the safety of participants or the conduct of 814 the trial. 816 The Safety Officer will make recommendations as necessary to the PI. If the Safety 817 Officer recommends a study change for patient safety or for ethical reasons, or if the 818 study is closed early due to slow accrual, the PI will be responsible for implementing the 819 recommendations as expeditiously as possible. If the PI does not concur with any 820 recommendation of the Safety Officer, both will be responsible for reaching a mutually 821 acceptable decision. 823 7.9. Stopping Rules 825 At the conclusion of the 16-week intervention period for the first cohort, the study 826 statistician will prepare a report on clinical changes and adverse events for presentation 827 to the PI and the Safety Officer. If evidence available at that point clearly shows either 828 (1) an effect of the intervention diet on postprandial metabolism or (2) harm associated 829 with the intervention diet, the Safety Officer may recommend early termination of the 830 study. 832 8. INCLUSION OF WOMEN, MINORITIES, AND CHILDREN 834 8.1. Inclusion of Women 836 The participation criteria, cited above, include both men and women. Recruitment 837 procedures are expected to yield roughly equal numbers of men and women. 839 8.2. Inclusion of Minorities 841 The U.S. Census Bureau reports both race and ethnicity, the latter term used primarily 842 to denote self-identification as Hispanic or non-Hispanic. According to the 2015 Census 843 Bureau, races were represented in Washington, DC, as follows: 48.3% black, 44.1% 844 white, 4.2% Asian, 0.6% American Indian/Native American, 0.2% Native Hawaiian or 845 other Pacific Islander; 2.7% 2 or more races. In addition, 10.6% of the population 846 identified themselves as Hispanic. In our prior studies, the respondent populations 847 have been demographically diverse, reflecting the profile of the greater Washington, 848 D.C area. 850 8.3. Inclusion of Children. 852 Persons less than 18 years of age will not be included in the study because they have 853 insufficient control over the dietary choices that are essential to meaningful participation. 856 9. BRIEF STATEMENT OF ANTICIPATED OUTCOMES 858 This study aims to test hypotheses that are potentially important for individual and public 859 health. It will improve our understanding of the treatment of weight problems and will 860 also have practical implications for reducing the medical, personal, and economic costs 861 associated with obesity. Anticipated outcomes for Diet Group participants include 862 beneficial changes in body weight, insulin sensitivity, and serum lipid concentrations, all 863 of which are also possible for the Control Group participants who choose to take 864 advantage of instruction in . 869 LITERATURE CITED Tonstad, S, Butler T, Yan R, Fraser GE.Type of vegetarian diet, body weight and prevalence of type 2 diabetes. Diabetes Care. 2009;32:791-6. Barnard ND, Levin SM, Yokoyama Y. A systematic review and meta-analysis of changes in body weight in clinical trials of vegetarian diets. J Acad Nutr Diet. 2015 Jun;115(6):954-69. Barnard ND, Scialli AR, Turner-McGrievy G, Lanou AJ, Glass J. The effects of a low- fat, plant-based dietary intervention on body weight, metabolism, and insulin sensitivity. Am J Med 2005;118:991-997. Shulman GI. Ectopic fat in insulin resistance, dyslipidemia, and cardiometabolic disease. N Engl J Med 2014;371:1131-1141 Petersen KF, Dufour S, Morino K, Yoo PS, Cline GW, Shulman GL. Reversal of muscle insulin resistance by weight reduction in young, lean, insulin-resistant offspring of parents with type 2 diabetes. PNAS. 2012;109:8236-40. Sparks LM, Xie H, Koza RA, et al. A high-fat diet coordinately downregulates genes required for mitochondrial oxidative phosphorylation in skeletal muscle. Diabetes. 2005;54:1926 33. Goff LM, Bell JD, So PW, Dornhorst A, Frost GS. Veganism and its relationship with insulin resistance and intramyocellular lipid. Eur J Clin Nutr. 2005;59:291 298. Barnard N, Scherwitz L, Ornish D. Adherence and acceptability of a lowfat vegetarian diet among patients with cardiac disease. J Cardiopulmonary Rehabil 1992;12:423-31. Barnard N, Scialli A, Bertron P, Hurlock D, Edmonds K. Acceptability of a therapeutic low-fat, vegan diet in premenopausal women. J Nutr Educ 2000;32:314-9. Barnard ND, Scialli AR, Turner-McGrievy G, Lanou AJ. Acceptability of a low-fat vegan diet compares favorably to a step II diet in a randomized, controlled trial. Journal of cardiopulmonary rehabilitation 2004;24(4):229-35. Barnard ND, Gloede L, Cohen J, et al. A low-fat vegan diet elicits greater macronutrient changes, but is comparable in adherence and acceptability, compared with a more conventional diabetes diet among individuals with type 2 diabetes. J Am Diet Assoc 2009;109(2):263-72. Craig CL, Marshall AL, Sjostrom M, et al. International physical activity questionnaire: 12-country reliability and validity. Medicine and science in sports and exercise 2003;35(8):1381-95. Petersen KF, et al. (2006) Increased prevalence of insulin resistance and nonalcoholic fatty liver disease in Asian-Indian men. Proc Natl Acad Sci USA 103:18273 18277. Gruetter R (1993) Automatic, localized in vivo adjustment of all first- and second- order shim coils. Magn Reson Med 29:804 811. Rabøl R, Petersen KF, Dufour S, Flannery C, Shulman GI (2011) Reversal of muscle insulin resistance with exercise reduces postprandial hepatic de novo lipogenesis in insulin resistant individuals. Proc Natl Acad Sci USA 108:13705 13709. Position of the American Dietetic Association and Dietitians of Canada: Vegetarian diets. J Am Diet Assoc 2003;103(6):748-65. Barnard N, Akhtar A, Nicholson A. Factors that facilitate dietary change. Arch Fam Med 1995;4:153-8. Becker M. The health belief model and personal health behavior. Health Education Monographs 1974;2:324-473. Buzzard I, Faucett C, Jeffery R, et al. Monitoring dietary change in a low-fat diet intervention study: advantages of using 24-hour dietary recalls vs food records. J Am Diet Assoc 1996;96:574. U.S. Census Bureau. Quick Facts. District of Columbia. Internet: http://www.census.gov/quickfacts/table/RHI125215/11, accessed August 22, 2016. Supplemental Online Content Kahleova H, Petersen KF, Shulman GI, et al. Effect of a low-fat vegan diet on body weight, insulin sensitivity, postprandial metabolism, and intramyocellular and hepatocellular lipid levels in overweight adults: a randomized clinical trial. JAMA Netw Open. 2020;3(11):e2025454. doi:10.1001/jamanetworkopen.2020.25454 eTable 1. Baseline characteristics of the study population eTable 2. Baseline characteristics of the study population, comparing study completers and drop- outs eTable 3. Baseline characteristics of the study population, comparing the subsample undergoing magnetic resonance spectroscopy (MRS) with the rest of the study population eTable 4. Treatment effects for the main outcomes, adjusted for age and race eFigure 1. Changes in the thermic effect of food, liver fat, and intramyocellular lipids after adjustment for race and age eFigure 2. Changes in liver fat and intramyocellular lipids after adjustment for baseline BMI eFigure 3. Linear regression model for changes in energy and body weight and postprandial energy expenditure and body weight This supplemental material has been provided by the authors to give readers additional information about their work. © 2020 Kahleova H et al. JAMA Network Open. Characteristic Intervention Control group P Value group (n=122) (n=122) Age (years, SD) 53 (±10) 57 (±13) 0.01 Sex (number, %) 0.85 Female 105 (86.0) 106 (86.8) Male 17 (14.0) 16 (13.1) Race, (number, %) 0.06 White 57 (46.7) 60 (49.2) Black 60 (49.2) 53 (43.4) Asian, Pacific Islander 1 (0.8) 7 (5.7) American Indian, Eskimo, Aleut 2 (1.6) 0 (0.0) Did not disclose 2 (1.6) 2 (1.6) Ethnicity, (number, %) 0.75 Non-Hispanic 97 (79.5) 101 (82.8) Hispanic 8 (6.6) 7 (5.7) Did not disclose 17 (13.9) 14 (11.5) Marital status, (number, %) 0.87 Not married 66 (54.1) 61 (50.0) Married 55 (45.1) 53 (43.4) Did not disclose 1 (0.8) 8 (6.6) Education, (number, %) 0.15 High school 8 (6.6) 11 (9.0) Associates 35 (28.7) 43 (35.2) College 72 (59.0) 63 (51.6) Graduate degree 7 (5.7) 5 (4.1) Occupation, (number, %) 0.23 Service occupation 27 (22.1) 15 (12.3) Technical, sales, administrative 34 (27.9) 32 (26.2) Professional or managerial 33 (27.0) 39 (32.0) Retired 16 (13.1) 24 (19.7) Other 12 (9.8) 12 (9.8) Medications, (number, %) Lipid-lowering therapy (%) 22 (18.0) 21 (17.2) 0.87 Antihypertensive therapy (%) 33 (27.0) 31 (25.4) 0.77 Thyroid medications (%) 16 (13.1) 12 (9.8) 0.42 Total physical activity (number, SD) 2719 (±4701) 2863 (±3085) 0.80 Dietary Intake (number, SD) Caloric intake (kcal/day) 1834 (±574) 1793 (±628) 0.61 © 2020 Kahleova H et al. JAMA Network Open. Total fiber intake (g/day) 24.1 (±10.6) 23.9 (±10.1) 0.89 Total cholesterol intake (mg/day) 238.6 (±143.9) 244.5 (±169.3) 0.78 Total saturated fatty acid intake 23.6 (±12.0) 22.9 (±12.2) 0.65 (g/day,) Monounsaturated fatty acids (g/day) 27.2 (±10.4) 27.9 (±13.9) 0.67 Polyunsaturated fatty acids (g/day) 18.4 (±8.2) 19.1 (±10.5) 0.59 Anthropometric variables, (number, SD) Body weight (kg) 93.6 (±13.8) 92.7 (±13.7) 0.62 BMI (kg/m ) 33.3 (±3.8) 33.6 (±3.7) 0.57 Lean mass (kg) 50.5 (±7.9) 49.5 (±8.1) 0.35 Fat mass (kg) 40.6 (±9.2) 40.9 (±9.6) 0.76 Visceral fat volume (cm ) 1459 (±944.2) 1517 (±907.0) 0.64 Laboratory variables (number, SD) Total cholesterol (mmol/l) 5.2 (±1.1) 5.0 (±1.3) 0.11 HDL-cholesterol (mmol/l) 1.6 (±0.4) 1.7 (±0.9) 0.16 LDL-cholesterol (mmol/l) 3.0 (±0.9) 2.9 (±1.1) 0.16 Triglycerides (mmol/l) 1.2 (±0.5) 1.3 (±0.6) 0.10 Fasting plasma glucose (mmol/l) 5.2 (±0.2) 5.0 (±0.3) 0.18 Fasting plasma insulin (pmol/l) 91.2 (±59.6) 78.9 (±51.1) 0.12 HbA1c (DCCT, %) 5.6 (±0.4) 5.7 (±0.4) 0.30 Insulin sensitivity /resistance (number, SD) PREDIM (mg/min/kg) 4.1 (±1.3) 4.4 (±1.5) 0.11 HOMA-IR (dimensionless) 3.2 (±2.2) 2.7 (±1.9) 0.17 eTable 1. Baseline characteristics of the study population. Data are means ± SD (standard deviation), or number (%). P-values refer to t- (chi-squared) categorical variables. The P-value calculated for ethnicity distribution is for the comparison between Hispanic vs. non-Hispanic categories (and all other comparisons also exclude datapoints that were not available). © 2020 Kahleova H et al. JAMA Network Open. Characteristic Drop-outs (n=22) Study completers P Value (n=223) Age (years, SD) 57.8 (±12.5) 54.4 (±11.6) 0.20 Sex (number, %) Female 19 (86.4) 192 (86.5) 1.0 Male 3 (13.6) 30 (13.5) Race, (number, %) 0.30 White 7 (31.8) 110 (49.6) Black 14 (63.6) 99 (44.6) Asian, Pacific Islander 1 (4.5) 7 (3.2) American Indian, Eskimo, Aleut 0 (0) 2 (0.9) Did not disclose 0 (0) 4 (1.8) Ethnicity, (number, %) 1.0 Non-Hispanic 18 (81.8) 180 (81.1) Hispanic 1 (4.5) 14 (6.3) Did not disclose 3 (13.6) 28 (12.6) Marital status, (number, %) 0.09 Not married 15 (68.2) 112 (50.5) Married 6 (27.3) 102 (45.9) Did not disclose 1 (4.5) 8 (3.6) Education, (number, %) 0.58 High school 3 (13.6) 16 (7.2) Associates 0 (0) 1 (0.5) © 2020 Kahleova H et al. JAMA Network Open. College 8 (36.4) 81 (36.5) Graduate degree 11 (50.0) 123 (55.4) Did not disclose 0 (0) 1 (0.5) Occupation, (number, %) 0.49 Service occupation 4 (18.2) 38 (17.1) Technical, sales, administrative 7 (31.8) 59 (26.6) Professional or managerial 3 (13.6) 69 (31.1) Retired 5 (22.7) 35 (15.8) Other 2 (9.1) 21 (9.5) Did not disclose 1 (4.5) 0 (0) Medications, (number, %) Lipid-lowering therapy (%) 6 (27.3) 39 (17.6) 0.24 Antihypertensive therapy (%) 6 (27.3) 59 (26.6) 0.84 Thyroid medications (%) 1 (4.5) 27 (12.2) 0.48 Total physical activity (number, 2863 (±3144) 2780 (±4006) 0.93 SD) Dietary Intake (number, SD) Caloric intake (kcal/day) 1589 (±442) 1812 (±599) 0.13 Total fiber intake (g/day) 20.3 (±6.8) 23.9 (±10.3) 0.16 Total cholesterol intake (mg/day) 197 (±126) 242 (±155) 0.25 Total saturated fatty acid intake 18.8 (±10.5) 23.2 (±12.1) 0.14 (g/day,) Monounsaturated fatty acids (g/day) 23.8 (±11.0) 27.5 (±12.1) 0.22 © 2020 Kahleova H et al. JAMA Network Open. Polyunsaturated fatty acids (g/day) 14.9 (±7.0) 18.6 (±9.3) 0.11 Anthropometric variables, (number, SD) Body weight (kg) 92.3 (±9.02 93.0 (±13.7) 0.74 BMI (kg/m ) 33.0 (±2.7) 33.4 (±3.7) 0.56 Lean mass (kg) 50.0 (±7.5) 49.9 (±8.0) 0.98 Fat mass (kg) 40.5 (±4.9) 40.7 (±9.3) 0.88 Visceral fat volume (cm ) 1355 (±575) 1496 (±914) 0.34 Laboratory variables (number, SD) Total cholesterol (mmol/l) 5.2 (±0.9) 5.2 (±1.1) 0.78 HDL-cholesterol (mmol/l) 1.5 (±0.4) 1.6 (±0.4) 0.55 LDL-cholesterol (mmol/l) 3.1 (±0.7) 3.1 (±0.9) 0.99 Triglycerides (mmol/l) 1.2 (±0.6) 1.2 (±0.5) 0.92 Fasting plasma glucose (mmol/l) 5.5 (±0.9) 5.4 (±0.6) 0.61 Fasting plasma insulin (pmol/l) 73.2 (±31.9) 78.8 (±62.2) 0.52 HbA1c (DCCT, %) 5.7 (±0.5) 5.7 (±0.4) 0.86 Insulin sensitivity /resistance (number, SD) PREDIM (mg/min/kg) 3.9 (±1.1) 4.2 (±1.4) 0.58 HOMA-IR (dimensionless) 3.2 (±1.8) 3.0 (±2.0) 0.78 © 2020 Kahleova H et al. JAMA Network Open. eTable 2. Baseline characteristics of the study population, comparing study completers and drop-outs. Data are means ± SD (standard deviation), or number (%). P-values refer to t-tests for (chi-squared) P-value calculated for ethnicity distribution is for the comparison between Hispanic vs. non- Hispanic categories (and all other comparisons also exclude datapoints that were not available). Characteristic MRS YES (n=44) MRS NO (n=200) P-value Age (years, SD) 55.8 (±11.1) 54.5 (±11.8) 0.52 Sex (number, %) Female 35 (79.5) 176 (88.0) 0.14 Male 9 (20.5) 24 (12.0) Race, (number, %) 0.07 White 25 (56.8) 92 (46.0) Black 15 (34.1) 98 (49.0) Asian, Pacific Islander 3 (6.8) 5 (2.5) American Indian, Eskimo, Aleut 1 (2.3) 1 (0.5) Did not disclose 0 4 (2.0) Ethnicity, (number, %) 0.50 Non-Hispanic 37 (84.1) 161 (80.5) Hispanic 4 (9.1) 11 (5.5) Did not disclose 3 (6.8) 28 (14.0) Marital status, (number, %) 0.81 Not married 22 (50.0) 105 (52.5) Married 20 (45.5) 88 (44.0) © 2020 Kahleova H et al. JAMA Network Open. Did not disclose 2 (4.5) 7 (3.5) Education, (number, %) High school 1 (2.3) 18 (9.0) Associates 0 (0) 1 (0.5) College 24 (54.5) 65 (32.5) Graduate degree 19 (43.2) 115 (57.5) Did not disclose 0 (0) 1 (0.5) Occupation, (number, %) 0.30 Service occupation 5 (11.4) 37 (18.5) Technical, sales, administrative 17 (38.6) 49 (24.5) Professional or managerial 12 (27.3) 60 (30.0) Retired 5 (11.4) 35 (17.5) Other 5 (11.4) 18 (9.0) Did not disclose 0 (0) 1 (0.5) Medications, (number, %) Lipid-lowering therapy (%) 7 (15.9) 38 (19.0) 0.62 Antihypertensive therapy (%) 10 (22.7) 55 (27.5) 0.51 Thyroid medications (%) 4 (9.1) 24 (12.0) 0.58 Total physical activity (number, 2711 (±3222) 2804 (±4080) 0.89 SD) Dietary Intake (number, SD) Caloric intake (kcal/day) 1886 (±629) 1776 (±582) 0.27 Total fiber intake (g/day) 25.7 (±11.9) 23.2 (±9.7) 0.14 © 2020 Kahleova H et al. JAMA Network Open. Total cholesterol intake (mg/day) 255.8 (±188) 234.6 (±145.0) 0.49 Total saturated fatty acid intake 24.3 (±12.0) 22.6 (±12.0) 0.39 (g/day,) Monounsaturated fatty acids (g/day) 29.2 (±14.9) 26.8 (±11.3) 0.34 Polyunsaturated fatty acids (g/day) 19.8 (±12.4) 18.0 (±8.3) 0.36 Anthropometric variables, (number, SD) Body weight (kg) 89.8 (±11.7) 93.6 (±13.7) 0.08 BMI (kg/m ) 32.1 (±2.3) 33.7 (±3.8) <0.001 Lean mass (kg) 50.5 (±8.5) 49.8 (±7.8) 0.60 Fat mass (kg) 37.3 (±7.0) 41.4 (±9.3) 0.002 Visceral fat volume (cm ) 1559 (±975) 1468 (±874) 0.55 Laboratory variables (number, SD) Total cholesterol (mmol/l) 5.4 (±1.2) 5.2 (±1.0) 0.27 HDL-cholesterol (mmol/l) 1.6 (±0.6) 1.6 (±0.4) 0.41 LDL-cholesterol (mmol/l) 3.2 (±0.9) 3.1 (±0.9) 0.42 Triglycerides (mmol/l) 1.2 (±0.6) 1.2 (±0.5) 0.91 Fasting plasma glucose (mmol/l) 5.5 (±0.7) 5.4 (±0.6) 0.35 Fasting plasma insulin (pmol/l) 66.4 (±53.3) 81.1 (±72.0) 0.07 HbA1c (DCCT, %) 5.9 (±0.5) 5.6 (±0.4) <0.001 Insulin sensitivity /resistance (number, SD) © 2020 Kahleova H et al. JAMA Network Open. PREDIM (mg/min/kg) 4.2 (±1.3) 4.2 (±1.4) 0.78 HOMA-IR (dimensionless) 2.8 (±1.9) 3.0 (±2.0) 0.58 eTable 3. Baseline characteristics of the study population, comparing the subsample undergoing magnetic resonance spectroscopy (MRS) with the rest of the study population. Data are means ± SD (standard deviation), or number (%). P-values refer to t-tests for (chi-squared) - value calculated for ethnicity distribution is for the comparison between Hispanic vs. non- Hispanic categories (and all other comparisons also exclude datapoints that were not available). Treatment effect P- Outcomes (adj.) value Weight (kg) -6.1 (-7.0 to -5.2) <.001 BMI (kg/m ) -2.4 (-3.6 to -1.2) <.001 Fat mass (kg) -4.1 (-4.7 to -3.5) <.001 Lean mass (kg) -1.5 (-2.0 to -1.1) <.001 VAT volume -217.4 (-316.2 to - <.001 (cm ) 118.7) © 2020 Kahleova H et al. JAMA Network Open. Hepatocellular -1.1 (-2.3 to -0.04) <.001 lipids (%) Intramyocellular -0.1 (-0.5 to +0.2) 0.02 lipids (%) PREDIM +0.8 (+0.5 to +1.1) <.001 HOMA -1.2 (-2.2 to -0.2) 0.01 eTable 4. Treatment effects for the main outcomes, adjusted for age and race. © 2020 Kahleova H et al. JAMA Network Open. © 2020 Kahleova H et al. JAMA Network Open. © 2020 Kahleova H et al. JAMA Network Open. © 2020 Kahleova H et al. JAMA Network Open. © 2020 Kahleova H et al. JAMA Network Open.

Journal

JAMA Network OpenAmerican Medical Association

Published: Nov 30, 2020

There are no references for this article.