Dietary n-3-polyunsaturated fatty acids and energy balance in overweight or moderately obese men and women: a randomized controlled trial

Subjects were 26 healthy overweight or mildly obese (body mass index 28–33 kg/m²) men and women. Subjects were weight stable [within 2.27 kg (5 pounds) in the past 6 months] and within 4.54 kg (10 pounds) of their lifetime maximum weight. Exclusion criteria included the use of tobacco or recreational drugs, alcohol abuse, a history of cardiovascular disease or diabetes mellitus, the presence of any chronic or psychiatric illness, restrictive eating behavior, pregnancy, and the intake of selective serotonin-reuptake inhibitors, lipid-lowering drugs, β-blockers, glucocorticoids, or anabolic steroids. All subjects underwent informed consent procedures. The study protocol was approved by the University of Washington Institutional Review Board (#03-9082-D05), and was in compliance with the Helsinki Declaration.

In the first two weeks of the study, all subjects consumed a control diet. Caloric content of the diet was calculated on the basis of a 3-day diet record completed by the subjects, and the Mifflin formula [11] adjusted for an activity factor based on a modified Blair physical activity questionnaire [12] to maintain each subject's weight to within 1 kg of baseline weight. Subjects were instructed to consume all food provided, not to eat any additional food, and to take a portion of their calculated daily fat intake as capsules filled with high-oleic safflower oil. They were asked to keep a daily food record and write down all the foods they had eaten, estimate the percentage eaten, and any deviations from the diet protocol. On the morning of day 14 of the control diet, subjects were admitted to the University of Washington Clinical Research Center (CRC) after an overnight fast of at least 10 hours for clinic visit #1 (CRC 1).

Upon completion of CRC 1, subjects were randomized to continue consuming the control diet or to consume a diet enriched in n-3-PUFA, for another two weeks. The n-3-PUFA diet was identical to the control diet in all respects apart from the sources of dietary fat. A portion of the augmented n-3-PUFA intake was provided by switching the subjects from capsules containing high-oleic safflower oil to an identical number of capsules containing fish oil, with the balance of the n-3-PUFA provided in prepared foods [13]. The n-3-PUFA diet contained 1.4% of energy in the form of the long-chain n-3-PUFA eicosapentaenoic acid (20:5n-3), docosapentaenoic acid (22:5n-3), and docosahexaenoic acid (22:6n-3), and 2.2% of energy in the form of α-linolenic acid (18:3n-3) from plant oils, while the control diet provided 0% of energy in the form of marine n-3-PUFA and 0.5% of energy as α-linolenic acid. Subjects continued to pick up food, turn in food records, and were weighed twice weekly. The total caloric content of dispensed food items was identical to the amount that led to weight stability in the lead-in diet period ('isoenergetic'). The subjects were again instructed to consume all food and not to eat any other food. On day 14 of this period, subjects were again admitted to the CRC for visit 2.

After CRC 2, subjects continued to consume the diets to which they had been randomized for another 12 weeks but now freely chose the amount they wished to eat (ad libitum). In this phase, the caloric content of food provided to subjects was 115% of the amount they had received in the previous isoenergetic period. Subjects continued to weigh twice weekly, pick up food and return uneaten food items, allowing an accurate assessment of the calories consumed on each day of the study. During each food pick-up visit, subjects met with a dietitian for about 15 minutes to review the food records and the returned, uneaten foods. This meeting and the comparison of the food records and returned foods were implemented as they provided valuable information about the compliance of the subject. Specifically, meeting with the subject in person twice weekly for the 16-week dietary period created a rapport with the subject and helped us respond in a timely fashion to specific requests such as switching specific meals or foods. Further, any major or repeated deviations from the study protocol would have required a substantial investment of time and effort from the subject to make sure that the amounts of foods returned were always consistent with those recorded on the food diary. To further increase compliance, subjects were allowed to drink up to 30 alcoholic drinks, and to exchange up to 10 study meals for "meals out" or meals prepared at home during the first 10 weeks of the ad libitum diet period. We asked subjects to provide detailed information about alcoholic beverages and meals out on the daily food record. Specifically, we asked them to note the amount and type of alcoholic beverage consumed, and to provide recipes for meals prepared at home, including details such as the amounts and sources of dietary fat. If they had a meal at a restaurant, we asked them to provide the names of the restaurant and the dish they had chosen. A dietitian of the CRC Nutrition Research Kitchen would then call the restaurant to obtain more detailed information about the dish. Of note, 18 out of the 26 subjects who completed the study exchanged a study meal for a "meal out" on less than the allowed 10 occasions, suggesting an excellent compliance with the dietary regimen. On day 84 of this diet period, subjects were again admitted to the CRC for visit 3.

During each of the three admissions to the CRC, subjects were weighed immediately after the first morning voiding. Breakfast, lunch, dinner and a snack were served at 0800, 1200, 1730, and 2000, respectively, and subjects were asked to complete each meal within 30 minutes. An intravenous line was used to sample 5 ml of blood every half-hour between 0800 and 2100, and every hour between 2100 and 0800 the next day. Additional fasting blood samples were drawn at 0800 on day 2 of each CRC visit and plasma prepared from all samples was kept frozen at -70°C until analyses. The 38 diurnal plasma samples were frozen individually, for measurement of plasma leptin concentrations. During CRC visits 1 and 3, a whole-body dual-energy x-ray absorptiometry scan was performed using a GE Lunar Prodigy (General Electric Healthcare, Waukesha, Wisconsin, USA) scanner to assess body fat mass. At 0700 on day 2 of each CRC admission, we performed indirect calorimetry for 30 minutes to assess resting energy expenditure using a TrueOne 2400 metabolic cart (Parvo Medics, Sandy, Utah, USA). Throughout the study, subjects completed daily visual analogue scale (VAS) questionnaires to assess hunger and fullness. They were also asked to complete a modified Blair physical activity questionnaire [12] every 2–4 weeks. Additional details about subject characteristics and study design have been published previously [13].

Plasma leptin and ghrelin were measured in duplicate by commercially available radioimmunoassays (Millipore, Billerica, Massachusetts, USA). Leptin was measured in all 38 timed diurnal samples drawn during each of the three clinic visits. Ghrelin was measured in fasting samples, and in pools of plasma created from the timed blood samples by combining 50 μL plasma from each blood sample drawn at 30-minute intervals and 100 μL plasma from each blood sample drawn at 60-minute intervals ('pool'-sample). Intra- and interassay CV were 2.4% and 3.6%, respectively, for leptin, and 4.3% and 5.3%, respectively, for ghrelin.

Sample size calculation revealed that completing 36 subjects would provide 80% power to find a difference in body weight of 2.25 kg between the n-3-PUFA and control groups. The estimate for the variability in body weight change over the 12-week ad libitum diet period was derived from two previous controlled dietary trials conducted by our group that had employed an identical design [14, 15]. We determined that we wanted to perform one interim analysis after 26 subjects had completed the study, which would allow us either to stop the study for effect or futility or to continue until the full 36 subjects had completed the study. We therefore designed this study as a sequential clinical trial, using the Power Family of Two-Sided, Inner Wedge Tests (chapter 5 in [16]). This test was chosen as it ensures specified α-error and power conditions [16].