Overweight or obese (BMI 25–40 kg/m2) men and women between the ages of 55–75 years were recruited from the local community though newspaper advertisements. For inclusion in the study, individuals were required to be weight stable (± 2 kg, >1 year) and nonsmokers. Individuals were excluded if they reported a history of depression, eating disorders, diabetes, uncontrolled hypertension (>159/99 mm Hg), heart, lung, kidney disease; cancer, food allergies/intolerances to items used in the laboratory test meals; or current use of medications known to alter food intake or body weight. Individuals were blinded to the specific purpose of the study, and were informed that the study involved examination of dietary factors believed to influence weight loss. The study protocol was approved by the Institutional Review Board of Virginia Polytechnic Institute and State University. All participants provided written informed consent prior to study enrollment.
Initial screening procedures and baseline assessments. An overview of the study protocol is depicted in Figure 1. Individuals meeting initial enrollment criteria completed baseline laboratory assessments over a series of four visits. Height was measured in meters without shoes using a wall-mounted stadiometer, and body weight was measured to the nearest 0.1 kg using a digital scale with participants wearing light street clothing and no shoes (Scale-Tronix model 5002, Wheaton, IL). Percentage body fat, absolute fat mass, fat-free mass, and total body bone mineral content were measured using dual-energy X-ray absorptiometry (GE Lunar Prodigy; GE Healthcare, Madison, WI). Waist circumference was measured to the nearest 0.5 cm at the umbilicus, using a Gulick tape measure (Gulick, Country Technology, Gays Mill, WI). Resting blood pressure was measured in the seated position using a mercury sphygmomanometer after a 15-min period of rest; the average of three measurements ±6 mm Hg was used. To assess habitual dietary intake and beverage consumption, participants were instructed in proper methods to record their food and beverage intake (including water consumption) for 4 consecutive days, which included 3 weekdays and 1 weekend day, and provided with food models to assist in portion size determination. Records were reviewed for completeness upon their return, and analyzed using diet analysis software (NDS-R 4.05; University of Minnesota, Minneapolis, MN). A second trained technician reviewed all diet analyses for data entry errors. To assess habitual beverage consumption, baseline and week 12 food intake records were manually reviewed to calculate mean daily amounts (kcal, g) of water and other beverages consumed. Dietary energy density (ED; kcal/g) was calculated from the food and beverage intake records and was expressed in four ways (22): total ED including all foods and beverages consumed; beverage ED including water; beverage ED excluding water; and ED from food only, excluding all beverages. When comparing ED (food + beverages) between individuals or over time, excluding water from the calculation could lead to higher ED values among water consumers or those increasing water intake, compared to those consuming energy-free beverages (diet sodas, coffee, and tea) (22). Thus, multiple ED calculations were performed. Participants collected urine for one 24-h period for assessment of total urine volume, and specific gravity was determined using a refractometer (Fisher UriSystem; Fisher Scientific, Hampton, NH). Blood was sampled from an antecubital vein for assessment of lipid and lipoprotein concentrations, which were performed using a SynchronLX20 (Beckman Coulter, Fullerton, CA). Total cholesterol and triglyceride concentrations were determined using the timed endpoint method, high-density lipoprotein cholesterol was determined by homogenous assay, and low-density lipoprotein cholesterol was determined by calculation. Habitual physical activity (steps/day) was measured using GT1M activity monitors for a 4-day period (ActiGraph, Pensacola, FL).
Following initial assessments, each participant underwent two laboratory test meal conditions within a 2-week period, separated by a minimum of 2 days, in a random order as follows: (i) 30-min waiting period (no preload (NP)) followed by an ad libitum breakfast meal, and (ii) preload consisting of 500 ml (∼16 fl oz) chilled bottled water followed within 30 min by an ad-lib meal. Condition 1 served as the “baseline” EI for comparison. A 30-min time interval between the preload and ad libitum meal is the most effective time interval to study EI compensation using preloads (23). Subjects were instructed not to eat or drink for at least 12 h prior to arriving for the test meal. The meal consisted of typical breakfast items (cinnamon raisin bagel, cream cheese, margarine, jelly, vanilla yogurt, banana, mozzarella cheese stick, cereal bar, orange juice, coffee, cream, and sugar) provided in excess of what would normally be consumed, from which the participants were allowed to self-select during a 20-min meal period. All foods used in the breakfast meals were evaluated for palatability prior to study initiation. Foods were presented on a meal tray and arranged in the same manner (i.e., location on tray, temperature) on both testing days, and meals were served in individuals cubicles under standardized laboratory conditions (i.e., quiet, temperature controlled). All foods were covertly weighed (±0.1 g) before being served and again after the completion of the meal to determine the amount consumed. Meal energy and nutrient intake were calculated using diet analysis software (NDS-R; University of Minnesota, Minneapolis, MN). Participants completed visual analog scales during the test meal procedure at times 0, 30, 60, 90, 120, and 150 min to subjectively rate their feelings of hunger, satiety (fullness) and thirst (24,25,26). Time 0 represented arrival for the meal and time 30 represented the time immediately prior to receiving the meal.
Intervention period. Following completion of all baseline assessments (Figure 1), participants were randomly assigned to one of two diet groups for 12 weeks: (i) hypocaloric diet + 16 fl oz (500 ml) bottled water prior to each of the three daily meals (“water group”), or (ii) hypocaloric diet alone (“nonwater group”). Individuals assigned to the water group were provided with cases of bottled water (Aquafina; Pepsico, Purchase, NY), and were instructed to consume one bottle prior to each meal (3 × 16 fl oz bottles/day). Water group participants were provided with a daily tracking form to record their premeal water consumption, which was returned to the study personnel at weekly visits for calculation of weekly water consumption (%) compliance. Nonwater group participants were offered bottled water, but were not given instructions or recommendations on water consumption. Both groups were provided with a variety of additional foods consistent with their meal plans, in order to keep participants blinded to the study purpose. Consumption of these items was not mandatory. Participants received one “provided” food per week in addition to the bottled water, and all participants received the same food item during that week (e.g., seven red delicious apples, 55 kcal each; seven navel oranges, 62 kcal each; one box of microwave popcorn, Orville Redenbacher's Smart Pop 94% Fat-Free, four Butter-Flavored 100-calorie packs; ConAgra Foods, Omaha, NE). Both groups received individualized instruction by a registered dietitian on a hypocaloric diet (women: 1,200 kcal, men: 1,500 kcal), which was developed using United States Department of Agriculture food guide pyramid guidelines (27). Consumption of fruits, vegetables, lean sources of protein, lowfat/nonfat dairy products, and whole grains was emphasized; both groups were instructed to moderate their consumption of high-fat snack foods, sweetened energy-containing beverages, and alcohol. Meal plan booklets with sample menus were also provided. Average energy and macronutrient content (% energy from fat/carbohydrate/protein, ED) of the 1,200 and 1,500 kcal sample menus, not including optional energy-free beverages (e.g., water, diet soft drinks) were as follows: 1,191 kcal (30/52/21, 1.28 kcal/g); 1,425 kcal (28/53/22, 0.93 kcal/g). Participants were instructed to maintain their current level of physical activity throughout the intervention.
Participants returned weekly to the laboratory for body weight measurement and dietary counseling, and dietary intake records were repeated at weeks 4 and 8 to encourage compliance.
Post-testing. Following the 12-week intervention, participants repeated all baseline measurements (body weight and composition, 4-day dietary intake record and activity monitoring, fasting blood draw, resting blood pressure, 24-h urine collection, two ad libitum laboratory test meal studies), completed an exit survey, and were compensated $50.
Power calculations (α = 0.05, power = 0.8) were performed based upon expected differences in weight loss between hypocaloric diets groups (2.0 ± 2.5 kg) to determine the targeted final sample size (n = 40). Baseline group demographic characteristics were assessed using independent samples t-test and Pearson's χ2-tests (SPSS vs. 12.0 for windows). To assess group difference in weight loss over 12 weeks, a random coefficients (mixed) model (i.e., growth curve analysis) was used, which includes all available data from an individual, corrects for unreliability of measurement and emphasizes individual growth trajectories rather than average values at each occasion (28,29). The growth curve model was fitted using STATA 9.1 xtmixed function. Full-information maximum likelihood estimation, which uses all available data (i.e., weekly body weight measurements) on the 48 participants enrolled into the intervention, was used to address partially observed data. To capture potential variations in the effect of increased water consumption on weight loss over the 12-week intervention, a quadratic effect of time (week-squared) was included in the model as a covariate. The intercept was specified at the first occasion of measurement (i.e., week = 0). Follow-up occasions occurred weekly for 12 weeks, and time was coded as 0–12. All main effects and their interactions with the linear and quadratic effects remained in the model regardless of the significance of the effect.
For secondary outcome variables, repeated measures ANOVA was used to assess group and time differences for subjects completing the 12-week intervention; analysis of covariance was used to adjust for baseline differences when present. When significant interactions were detected, t-tests were used for post hoc analyses. Group differences in pre-to-post change values (Δ) were analyzed using independent samples t-test. The trapezoidal model was used to calculate area under the curve (AUC) for each visual analog scale variable (30), and differences in visual analog scale ratings during the test meal period were assessed using repeated measures ANOVA. Associations among variables were assessed by simple correlational analyses (Pearson's r). The α-level was set a priori at P < 0.05. Data are expressed as mean ± s.e.m.