|Home | About | Journals | Submit | Contact Us | Français|
The use of meal replacements (MR) in lifestyle modification programs (LMP) for obese adults significantly increases weight loss, compared with the prescription of an isocaloric conventional diet (CD). The present 12-month randomized trial examined 113 obese adolescents (mean [SD] age of 15.0 [1.3] yr and body mass index [BMI] of 37.1 [5.1] kg/m2) who were assigned to a LMP, combined with CD or MR (randomized in 1:2 ratio). For months 1–4, all participants were prescribed 1,300–1500 kcal/d, consumed as a CD or as 3 Slim-Fast® shakes/d and one pre-packaged meal/d (with 5 fruit/vegetable servings/d). After month 4, participants were unmasked to their random assignment of continued use of MR (i.e., MR+MR) or transitioned to CD (i.e., MR+CD). Those initially treated by CD continued with it (i.e., CD). Regression models were used to evaluate percentage change in BMI from baseline to month 4, months 5 to 12, and baseline to month 12. At month 4, participants assigned to MR (N=65) achieved a mean (+SE) 6.3±0.6% reduction in BMI, compared to a significantly (P=0.01) smaller 3.8±0.8% for CD participants (N = 37). From months 5–12, BMI increased significantly (P<0.001) in all three treatment conditions (i.e., MR+MR, MR+CD, and CD), resulting in no significant (P=0.39) differences between groups in percentage change in BMI at month 12. Across groups, mean reduction in BMI from baseline to month 12 was 3.4±0.7% (p<0.01). Use of MR significantly improved short-term weight loss, compared with CD, but its continued use did not improve the maintenance of lost weight.
The current treatment of choice for adolescent obesity is a comprehensive program of lifestyle modification that includes parental participation (1). This approach provides adolescents a set of principles and techniques to modify their food intake, physical activity, and cognitions to induce a modest negative energy balance (e.g., 500 kcal/d) required for weight loss. Comprehensive programs typically induce a loss of 2 to 4 kg in 6 months, equal to about a 2 to 4% reduction in initial BMI (1–4). However, long-term maintenance of weight loss is poor (1, 4). Thus, innovative interventions are needed which both induce greater weight loss and facilitate the maintenance of lost weight.
Meal replacements have addressed both of these needs in obese adults. These products, which include liquid shakes, meal bars, and frozen food entrees, provide dieters a fixed amount of food, with a known calorie content, which appears to facilitate adherence to calorie goals. When added to a lifestyle modification program, meal replacements have reliably increased weight losses by 2.5 kg and 2.4 kg, at 3 and 12 months, respectively, compared with isocaloric diets of conventional foods (10). In the most successful study to date, Ditschuneit et al. (5) found that adults who were prescribed a 1200–1500 kcal/d diet of conventional foods lost 1.3 kg in 3 months, compared with a significantly greater loss of 7.1 kg in those who were prescribed the same calorie goal but replaced two meals per day with shakes (i.e., SlimFast ) and two snacks with bars. Dieters know precisely how many calories they have consumed when they eat meal replacements. By contrast, they typically underestimate their calorie intake by 40 to 50% when consuming a diet of conventional foods (6). Furthermore, meal replacements often reduce dieters’ contact with problem foods and the difficulties they may experience in deciding what to eat (which may lead to overeating) (7).
The present study examined the use of meal replacements in the treatment of obese adolescents. All adolescents and their parents (guardians) received a family-based lifestyle modification program and were randomized to: 1) 12 months of a 1300–1500 kcal/d conventional diet (CD); 2) 4 months of an isocaloric meal replacement (MR) approach, followed by consumption of an isocaloric CD from months 5–12 (MR/CD); or 3) 4 months of MR with continued use of MR from months 5–12 (MR/MR) (Figure 1). In order to minimize potential intervention and compliance biases, participants in the MR condition were informed of their maintenance diet (for months 5 – 12) at the end of month 4, as were the interventionists and other research staff. We hypothesized that participants who received MR would lose significantly more weight at month 4 than those treated with CD. We also hypothesized that participants in the MR/MR condition would maintain their weight losses significantly better at month 12 than would the two groups that were prescribed a CD from months 5–12.
Participants were 13 to 17 year old males and postmenarcheal females who had a body mass index (BMI, calculated as weight in kg divided height in m2) of 28 kg/m2 to 50 kg/m2. Contraindications to participation included cardiovascular disease (including arrhythmias); types 1 or 2 diabetes mellitus; major psychiatric disorders; pregnancy; use of a weight-loss medication or a weight loss of 5 kg or more in the prior 6 months; use of medications promoting weight gain (e.g., oral steroids); or cigarette smoking. Adolescents, accompanied by a parent or guardian, completed a behavioral assessment, conducted by a staff psychologist or psychiatrist. Written informed consent was obtained from the parent and assent from the adolescent. The adolescent’s primary care physician performed a history and physical examination to exclude the noted contraindications. This study was approved by the Institutional Review Boards of the Children’s Hospital of Philadelphia and of the University of Pennsylvania (ClinicalTrials.Gov Identifier: NCT0021217).
Adolescents in all treatment conditions received the same comprehensive family-based lifestyle modification program (LMP), which was delivered following detailed treatment manuals (8, 9). The program included self-monitoring of eating habits and physical activity, stress management, stimulus control, problem solving, contingency management, cognitive restructuring, and social support. Counseling at each visit promoted increased physical activity (with a goal of ≥30 minutes a day) and reduced sedentary behavior. Parents met separately in group sessions held on the same schedule as the adolescents’ meetings. Groups were led by dietitians, master’s level therapists, psychologists, or psychiatrists.
Adolescents and their parents attended weekly group LMP meetings. After 2 weeks of orientation and learning how to self-monitor dietary intake, for the next 14 weeks adolescents were prescribed a 1300–1500 kcal/d CD, comprised of conventional foods, or an isocaloric MR diet. The daily MR plan consisted of 3 Slim-Fast® shakes that were provided free of charge (Unilever Corporation; Englewood, NJ), combined with 1 pre-packaged meal of the adolescents’ choice, 2 servings of fruit, and 3 servings of vegetables (all purchased by participants). Pre-packaged meals were selected from a list of frozen food entrées, all of which provided about 225–300 kcal per serving, with 20 to 25 g of protein. Participants who consumed the CD were instructed to consume a diet consistent with recommendations of the U.S. Dietary guidelines (10). Participants were encouraged to consume ≤ 30% of kcal/d from fat, approximately 15% from protein, and the remainder from carbohydrates.
After month 4 (Phase 1), all 3 groups were treated for an additional 8 months (Phase 2). Participants who originally received MR for 4 months either continued use of MR for weight maintenance (i.e., MR/MR, which included the use of 2 liquid MR/d and 1 prepackaged meal/d, with a breakfast of conventional foods, and 5 servings/d of fruits and vegetables) or transitioned to a CD for weight maintenance (i.e., MR/CD). Participants treated with CD during months 1–4, continued with CD (i.e., CD). From months 5–7, all participants received twice monthly group LMP meetings, followed by monthly group meetings from months 8–12. During the maintenance phase, all participants were instructed to consume 1300–1500 kcal/d and to be active for at least 30 minutes or more per day.
Weight was measured at each treatment session and at the major assessment visits (i.e., baseline, month 4, and month 12). Height and waist circumference were measured at the major assessment visits following standard techniques (11).
Serum lipids, glucose, and insulin levels were measured at the major assessment visits following an overnight fast. Systolic and diastolic blood pressure, as well as pulse rate, also were assessed at this time, using a Dinamap monitor (DINAMAP PRO 100, GE Healthcare Worldwide, UK). These latter assessments were obtained after the adolescent sat still for at least 5 minutes. Three readings were taken at 1-minute intervals, and the second and third values were averaged. Hunger, disinhibition of eating, and cognitive restraint were measured by the Eating Inventory (12).
The study’s primary outcome, change in degree of obesity, was evaluated by percentage change in initial BMI. This measure controls for changes in height over time, as well as for differences among participants in baseline BMI. Secondary outcome measures included changes in degree of obesity, as evaluated by changes in weight, BMI, BMI z-scores (13), and waist circumference. A total of 120 participants were randomized to the three treatment groups (CD, MR/CD, MR/MR). Regression analyses were utilized with the 113 adolescents who attended at least the baseline treatment session to assess changes in degree of obesity in each phase. In Phase 1, the regression models included change in obesity from baseline to month 4 as the outcome and group (i.e., CD or MR) as the independent variable. Similar models were fit for Phase 2 with the outcome being change in obesity from months 5 to 12, adjusted for phase 1 changes in degree of obesity. To assess changes throughout the entire study, repeated measures mixed effects models were fitted that included group (CD, MR/CD, MR/MR) as the between subjects factor and time (changes from baseline to months 4, 12) as the within subjects factor. Secondary analyses included a baseline carried forward (BCF) approach in which missing data were replaced by the participant’s baseline body weight. Correlation analyses were used to examine the relation between percentage change in initial BMI and changes in physiological and behavioral measures at month 12. For all analyses, the significance level (α) was equal to 0.05. Analyses were conducted using SAS, version 9.2.
Characteristics of the 113 participants who attended at least one treatment session are shown in Table 1. There were no significant differences between treatment groups on any of the baseline measures. Eighty-one percent of participants were females; 62% were African American, 26% were Caucasian, and 12% of other race/ethnicity. Participants had a mean±SD age of 15.0±1.3 yr and BMI of 37.1±5.1 kg/m2. At month 4 (Phase 1), retention was 90%; there was no significant difference in retention between groups (MR 92%, CD 88%; P=0.53) (see Figure 1). At month 12 (Phase 2), overall retention was 67% (CD: 62%, MR/CD: 70%, MR/MR 70%). There were no significant differences between groups in attrition at month 12 (P=0.68).
At month 4, participants who received MR (N=65) achieved a significantly greater percentage reduction in initial BMI than the CD group (N = 37), with decreases of (mean ± SE) −6.3 %±0.6 and −3.8%±0.8, respectively (P =0.01), resulting in an effect size of 0.52. As shown in Table 2, there were also significant differences among groups in changes in BMI (kg/m2) and weight (kg) (P’s=0.007 and 0.02. respectively), while differences in BMI z-score and waist circumference approached significance (P’s=0.07 and 0.09, respectively).
The BCF analysis revealed similar statistically significant results, with a 5.8±0.56% reduction in BMI in the MR group, compared with 3.4±0.73% in the CD group. Using the BCF data, the percentage of participants who lost 5% or more of their initial BMI at month 4 was 51% in the MR group compared to 29% in CD participants (P=0.02). Corresponding values for the achievement of a 10% weight loss were 15% and 12%, respectively (P=0.60).
From months 5–12, there were no significant differences between groups in any of the measures of change in obesity (P > 0.38). All groups (CD, MR/CD and MR/MR) gained weight from months 5–12 (means presented in Table 3: left column under each group). Percentage changes in BMI from months 5–12, adjusted for change in Phase 1, in the MR conditions were +2.9% (MR/CD) and +2.7% (MR/MR), compared to +0.9% for the CD condition (see Figure 2 and Table 3).
Based on the mixed model analyses, there were no significant differences between groups in changes in degree of obesity from baseline to month 12 (see Table 3 and Figure 2). At month 12, the CD, MR/CD, and MR/MR groups had lost −2.8±1.3%, −3.9±1.4%, and −3.4±1.3% of their initial BMI, respectively (“time” p-value=0.0003, “group” p-value=0.37, “time*group” p-value=0.40). Collapsed across the three groups, mean reduction in BMI at month 12, as measured from baseline, was −3.4±0.7% (p<0.01).
The BCF analysis revealed similar results as the mixed-effects model at month 12, with reductions of −3.1±0.97%, −4.1±1.1%, and −3.8±1.0%, respectively. At month 12, using the BCF data, 31% of the CD group, 45% of the MR/CD group, and 32% of the MR/MR participants lost 5 % or greater of their initial BMI (P=0.47). Corresponding values for the achievement of a 10% loss were 10%, 18%, and 16%, respectively (P=0.69).
There were no statistically significant differences between the three groups at the end of month 4 or month 12 in changes in blood pressure, pulse, lipid levels, glucose, insulin, HOMA, disinhibition, restraint, or depression (Table 4) (P’s > 0.08). At month 4, the −1.88±1.0 reduction in hunger in the MR participants was significantly (P=0.01) different from the 0.1±0.6 increase in the CD group. Collapsed across conditions, there were clinically significant improvements from baseline to month 12 in systolic blood pressure, total cholesterol, HDL cholesterol, LDL cholesterol, insulin, glucose, and HOMA (see Table 4). In addition, there was a clinically significant improvement from baseline to month 12 in cognitive restraint of eating (see Table 4). Based on correlation analyses, percentage change in initial BMI was significantly related to changes in systolic blood pressure (r=0.255, p=0.04), total cholesterol (r=0.418, p=0.001), triglycerides (r=0.467, p=0.0003), and cognitive restraint (r=−0.395, p=0.004).
The principal finding of this study was that the addition of meal replacements to a program of lifestyle modification significantly increased weight loss during the first 4 months, compared to the same program provided with an isocaloric diet of conventional foods. However, the benefits of the MR plan were not maintained at month 12.
The difference in weight loss between the MR and CD groups at month 4 in the present study is similar to that found in studies of obese adults (7, 14). In a meta-analysis of six randomized controlled trials (14), MR was found to increase mean weight loss by 2.5 kg at month 3, compared to conventional reduced calorie diets. The mean percentage weight loss at 3 months with the conventional reducing diet was approximately 4%, compared with 7% in the MR group. These results are very similar to those obtained in the present study of obese adolescents and confirm the benefits of MR in increasing initial weight loss.
However, the potential benefit of MR in maintaining weight loss in adolescents was not supported. All groups gained weight from months 5–12. Although the CD participants lost the least amount of weight at month 4, this group did not gain as much weight during months 5–12 as participants in the two original MR conditions. Moreover, there was no evidence that long-term use of MR facilitated the maintenance of lost weight, as revealed by comparing the group that remained on MR from months 5 to 12 with the one that was switched to a CD. Overall percentage change in BMI at month 12 from baseline was −2.8% in CD, −3.9% in MR/CD, and −3.4% in the MR/MR groups.
Long-term results for MR with our adolescents are not consistent with those reported in adults. In the previously described meta-analysis, MR increased mean weight loss by 2.4 kg at month 12 compared to a conventional reduced calorie diet (14). Overall, weight loss with the conventional reducing diet was approximately 5%, compared with 7.5% for MR (14). The discrepancy between the result of the adult and adolescent studies may be attributable to decreased adherence, in the adolescents, to self-monitoring and to consumption of MR (which we observed clinically between months 5–12). Unfortunately, we do not have adequate data to test these hypotheses. Future studies should systematically assess adherence to MR in adolescents.
Two previous studies of the treatment of adolescent obesity with a comprehensive LMP combined with sibutramine, a serotonin-norepinephrine re-uptake inhibitor (SNRI), suggest that this combination is associated with greater long-term weight loss than LMP plus MR. In both studies, sibutramine was associated with an 8.0–8.5% reduction in initial BMI at 1 year (15, 16). However, results of the current trial suggest that there may be a role for MR in the induction of weight loss in a stepwise approach to treatment. Adding a medication, such as sibutramine, after 4–6 months of treatment with MR may facilitate further weight loss and maintenance of weight loss in the long term. In addition, using a combination of MR and sibutramine with LMP may provide the most potent nonsurgical treatment. This hypothesis has been tested in obese adults (17, 18). Wadden et al. (17) found that the addition of MR to sibutramine and LMP substantially increased both short- and long-term weight losses compared to a 1200–1500 kcal/day diet of conventional foods with sibutramine and LMP. At the end of the first 4 months, participants prescribed MR and sibutramine lost 15.2% of initial weight and maintained a loss of 16.5% at 1 year. In contrast, those who were prescribed a conventional diet, sibutramine and LMP lost a significantly smaller 9.6% at month 4 and maintained a loss of 10.4% at 1 year. This combination therapy has yet to be tested in obese adolescents. (We note that sibutramine may no longer be an appropriate medication for adults or adolescents, given its risk of increasing cardiovascular events in patients with a pre-existing history of heart disease ).
In the present study, all three treatment groups maintained a loss of 2.8 to 3.8% of initial BMI at month 12. These reductions resulted in clinically significant changes in several physiological variables, including reductions in systolic blood pressure, total and LDL cholesterol, insulin, glucose, and HOMA. Significant increases in HDL cholesterol were also observed at month 12. There were not significant differences between groups on any of these measures. Weight loss in adolescents has been associated with improvements in cardiometabolic variables in previous studies (20, 21). In addition, there was an increase in cognitive restraint, which has also been found to be associated with weight reduction (22).
Limitations of the current study include the lack of dietary adherence data and the relatively small sample size. However, the weight losses of the three groups at month 12 were so similar that sample size most likely does not explain the lack of differences. The study’s strengths include the treatment of a demographically and ethnically diverse sample by experienced clinicians who followed detailed treatment protocols.
In summary, the use of meal replacements significantly increased weight loss during the first 4 months but did not facilitate the maintenance of lost weight at month 12. Further research is needed on behavioral and pharmacological treatments, whether used alone or together, that will improve the induction and maintenance of weight loss in obese adolescents.
Funding/Support: This study was supported by grants DK054713 and K24-DK065018 from the National Institutes of Health and grant UL1RR024134 from the National Center for Research Resources. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Center for Research Resources or the National Institutes of Health. Unilever provided SlimFast for the study.
None of the authors has any conflicts of interest to declare.