Subjects were recruited by public advertisement. Subjects over 30 years of age with a body mass index (BMI) between 27 to 40 kg/m2, and in good health by history, physical examination, and basic laboratory screening (complete blood count, serum chemistries, liver panel, and lipid panel) were selected for study. Subjects with type 2 diabetes or glucose intolerance were excluded as were individuals who regularly drank more than one alcoholic beverage daily,
One hundred men and women who met the selection criteria were randomly assigned to either the HP or SP treatment. This was a single-blinded study. The protein powder jars were labeled as either A or B, depending on their protein content. Subjects were randomized in a 1:1 manner to either HP or SP diet for 12 weeks using a computerized random proportion model. Diet plans were individualized per subject. Caloric intake to achieve weight loss was based on a 500 Kcal deficit of the participants' estimated resting metabolic rate as determined by body composition analysis by bioelectrical impedance.
Participants in the HP group received a diet plan that provided 2.2 grams of protein per kg of LBM while the diet for the SP group provided 1.1 grams of protein per kg of LBM. The meal energy macronutrient composition in the HP group was approximately 30% protein, 30% fat, and 40% carbohydrate. The macronutrient composition in the SP diet was approximately 15% protein, 30% fat, and 55% carbohydrate. Both groups received the same isocaloric MR (Formula 1, Herbalife Intl., Los Angeles) with either a protein supplement for the HP group (Performance Protein Powder, Herbalife Intl., Los Angeles) or with a similar tasting carbohydrate placebo for SP group. Two MR and two meals were eaten daily.
Instructions were provided for preparation of the MR and subjects were advised to consume one MR as a meal and the other as snack. All subjects were given individualized menu plans that incorporated the two MR (one meal and one snack) and included two all-food meals. All participants met individually with a registered dietitian at baseline for dietary instruction, and at 2, 4, and 8 weeks to assess compliance.
Participants were weighed and protein powder meal replacement products were dispensed at each visit to ensure compliance. Subjects were given general advice for increasing their activity level with a goal of 30 minutes of aerobic exercise per day, but no heavy resistance exercise.
Body weight and composition
Subjects were weighed at each visit (Detecto-Medic; Deteco-Scales; Brooklyn, NY) while wearing no shoes and after an overnight fast. Height was measured with a stadiometer (Detecto-Medic; Deteco-Scales; Brooklyn, NY) at week 0. BMI was calculated as weight (kg)/height squared (m). Body composition was determined by bioelectrical impedance analysis (BIA) (310e Bioimpedance analyzer; Biodynamics; Seattle, WA) and was performed at 0 and 12 weeks.
Biochemistry
Fasting blood samples were collected at weeks 0, 4, 8, and 12 for measurement of lipid profiles, blood glucose and liver function tests.
Statistical analysis
Weight loss was the primary outcome and the data were analyzed according to intention to treat allocation utilizing SAS version 9 (Cary, North Carolina) in the Department of Biostatistics.
Patient characteristics and baseline measurements of the two study groups were compared using t-test (for numerical variables) or Chi-square test (for categorical variables) to evaluate quality of the randomization.
Standard t-tests were used to compare weight losses between the two arms. In addition, to assess weight loss within each treatment arm, paired t-tests were conducted comparing baseline and 12 week weight for each subject. All data except baseline characteristics are presented as means +/- standard error of the mean (SEM). A univariate analysis of variance was used to assess differences between treatment and outcome variables. Since the distributions of change in fat weight and percentage change in fat weight were not normal, signed rank test was used for testing change from baseline within each group. The Wilcoxon rank sum test was used for comparing the change between the two groups. Multivariate analysis was performed to compare the difference between the two diet groups using general linear model. Square root transformation was applied before the multivariate analysis was performed.