If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
Portion control is a useful component of weight reduction interventions and meal replacement (MR) plans represent a promising strategy for portion control. Research performed with pooled data on the effect of MR plans according to various characteristics of MR interventions remains scarce.
Our aim was to assess the effects of MR-based diets compared with food-based diets on weight loss, according to calorie-restriction types and energy intake proportions from MR.
Electronic databases (Cochrane Central Register of Controlled Trials, PubMed, Embase, and Research Information Sharing Service) were searched for randomized controlled trials on weight loss results of MR-based calorie-restricted diets compared with food-based calorie-restricted diets from January 2000 to May 2020. Standardized mean differences (Hedges' g) from all study outcomes were calculated using a random-effects model. Heterogeneity was quantified by Q test and I2. Publication bias was assessed using a funnel plot and a trim and fill method. Both interventions (MR and control) were separated into very-low-energy diets and low-energy diets (LEDs). A meta-analysis of variance was conducted by dividing patient-related factors and treatment-related factors into subgroups. In multivariable meta-regressions, background variables were selected first, after which main independent variables were included.
Twenty-two studies involving 24 interventions and 1,982 patients who were overweight or obese were included. The effect size in which MR-based LED was compared with food-based LED for weight loss was small, favoring MR (Hedges’ g = 0.261; 95% CI 0.156 to 0.365; I2 = 21.9; 95% CI 0.0 to 53.6). Diets including ≥60% of total daily energy intake from MR had a medium effect size favoring MR with regard to weight loss among the groups (Hedges’ g = 0.545; 95% CI 0.260 to 0.830; I2 = 42.7; 95% CI 0.0 to 80.8).
The effect of MR-based LED on weight loss was superior to the effect of food-based LED, and receiving ≥60% of total daily energy intake from MR had the greatest effect on weight loss.
The Continuing Professional Education (CPE) quiz for this article is available for free to Academy members through the MyCDRGo app (available for iOS and Android devices) and through www.jandonline.org (click on “CPE” in the menu and then “Academy Journal CPE Articles”). Log in with your Academy of Nutrition and Dietetics or Commission on Dietetic Registration username and password, click “Journal Article Quiz” on the next page, then click the “Additional Journal CPE quizzes” button to view a list of available quizzes. Non-members may take CPE quizzes by sending a request to [email protected] There is a $45 fee per quiz (includes quiz and copy of article) for non-members. CPE quizzes are valid for 3 years after the issue date in which the articles are published.
Research Question: Is calorie restriction with meal replacement (MR) more effective for weight loss than conventional calorie restriction? Is there a greater effect on weight loss with an increasing percentage of total daily energy intake from MR?
Key Findings: In this systematic review and meta-analysis, the effect of an MR-based low-energy diet on weight loss was superior to the effect of food-based conventional low-energy diet (Hedges’ g = 0.261; 95% CI 0.156 to 0.365), where ≥60% of total daily energy intake from MR resulted in the greatest effect (Hedges’ g = 0.545; 95% CI 0.260 to 0.830).
The prevalence of obesity is increasing worldwide.
Calorie-restricted diets can be divided into 2 types based on the number of calories consumed per day—low-energy diets (LEDs), which are the most common and allow 1,000 to 1,500 kcal/d, and very-low-energy diets (VLED), which allow only 600 to 800 kcal/d.
Successful calorie restriction is closely related not only to the energy density of the foods consumed, but also to behavioral strategies such as self-monitoring, planning, food purchasing and preparation, and intake control. This can eventually lead to weight loss.
MR plans, which consist of commercially available products fortified with minerals and vitamins, include total diet replacement plans to replace all daily meals and partial meal replacement plans to replace 1 or 2 meals per day.
showed that, as part of energy-restriction strategies, MR plans generally promote weight loss. Therefore, MR plans have been shown to be effective for weight loss as part of energy-restriction diets; however, those meta-analyses did not compare MR plans with conventional food-based diets. Other meta-analyses by Heymsfield and colleagues,
suggested that MR plans lead to more effective weight loss than other weight loss methods, including alternative types of diets, with or without support and nominal intervention, at 3 months and 1 year, respectively. However, as Astbury and colleagues
reported, due to the small sample sizes for each subgroup, whether or not the characteristics of MR interventions impacted the observed effect was not assessed.
As described, several meta-analyses examining MR plans have addressed their effects on weight loss as a part of energy-restriction diets or compared them with conventional food-based diets. An additional question may be to assess the conditions under which MR plans might be more effective. For example, it should be investigated whether there are differences in weight loss depending on patient-related factors, such as age, sex, and initial body mass index (BMI; calculated as kg/m2), or depending on treatment-related factors, such as treatment period, number of MR meals, and percentage of total daily energy intake from MR. In several clinical trials on the effect of MR plans on weight loss, no interactions were found between sex, age, and/or BMI and treatment and extent of weight lost.
reported in a clinical trial that although total diet replacement was more effective for weight loss, partial meal replacement was more beneficial as a long-term treatment when considering individual adherence to dietary regimens. In another clinical trial, Leader and colleagues
showed that 2 MR meals per day are more effective than 1 MR meal per day for weight loss, regulation of blood glucose, and compliance with dietary prescriptions in patients with obesity and diabetes. However, no meta-analysis has analyzed patient-related and/or treatment-related factors in relation to the effect of MR plans on weight loss. Once these associations are known, MR plans can be applied more strategically to optimize weight loss outcomes.
Therefore, this meta-analysis sought to investigate the effect of an MR diet on weight loss in comparison with conventional calorie-restricted diets, according to treatment-related factors, such as treatment period, number of MR meals, and percentage of total daily energy intake from MR. These findings may have the potential to inform clinical practice.
The population, intervention, comparison, outcome and study design framework was used to formulate eligibility criteria in this study.
Published randomized controlled trials involved calorie-restricted diets using MR (which comprise commercially available products fortified with minerals and vitamins, including liquid formula, powdered mixes, energy bars, and prepackaged foods). Individuals aged between 18 and 65 years who were overweight or obese and without other chronic diseases (ie, no significant health conditions) were included. Only studies with weight loss interventions were included. There was no restriction in the duration of the interventions. For studies including 2 or more interventions, only those involving MR plans were included. Comparison groups included calorie-restricted diets without MR (ie, food-based diets). Studies needed to include weight-related outcomes (mean changes and standard deviation difference in weight or BMI) for estimating intervention effectiveness, otherwise they were excluded (insufficient data). The review was limited to English- or Korean-language articles. A nutritionist, a clinician, and a meta-analytical statistician collaborated on all aspects of this study.
The meta-analysis on weight management using MR conducted by Heymsfield and colleagues
included searches from 1960 to January 2001. Therefore, this study aimed to analyze more recent research trends regarding the effect of MR plans on weight loss by searching related studies starting from 2000. The following sources were searched for studies conducted from 2000 to May 2020: Cochrane Central Register of Controlled Trials, PubMed, Embase and Research Information Sharing Service; thesis papers were also searched. The search strategy followed the Preferred Reporting Items for Systematic reviews and Meta-Analysis guidelines.
Two co-first authors (J.M., S.-Y. K.) searched the articles independently using the following search terms: ("meal replacement"[title/abstract/keywords (tiabkw)] OR "meal supplement"[tiabkw] OR "portion control"[tiabkw]) AND ("weight loss"[tiabkw] OR "weight reduction"[tiabkw] OR "obesity management"[tiabkw]). The reference lists of the included studies were screened for potentially relevant articles, and backward and forward citation searches were performed in Google Scholar for all identified trials. In the case of any discrepancy, discussions was undertaken until agreement was reached. Figure 1 (available at www.jandonline.org) shows the complete search strategy used in the electronic database Cochrane Central Register of Controlled Trials.
Data Extraction and Quality Assessment
Once consensus was reached regarding which articles to retain, a predefined data template was prepared; data were extracted and coded based on this template. The study characteristics (publication year, journal, study design, and location), participants (sample size, age, sex, BMI, height, and weight), intervention (duration, type of calorie-restriction, frequency of MR intake, and percentage of total daily energy intake from MR), and weight-related outcomes (mean changes and standard deviation difference in weight or BMI) were included as data. When both weight and BMI were reported in a study, the BMI was used for the analyses. The data were then cross-checked by the 2 co-first authors (J.M., S.-Y. K.). Any discrepancy was further discussed until an agreement was reached. The methodological quality of the included studies was assessed using the Cochrane’s risk of bias assessment tool 2.
and the trim and fill method. Publication bias tests and funnel plots are only relevant when there are more than 10 studies; otherwise, these are underpowered for detection and tend to lead to unjustified conclusions.
The effect size, that is, d, was calculated as follows:
where and are the post-test and pretest means in the treatment group, respectively; and are the post-test and pretest means in the control group, respectively; and is the within-groups standard deviation.
where and are the sample sizes in the treatment and control groups, respectively.
The variance of was calculated using the following formula:
The standard error of is the square root of . Small-sample studies show a tendency toward increased effect size. Therefore, a small-sample study correction factor (Hedges’ g) was used.
The effect size, that is, , was calculated using a correction factor (J). Then, . For this meta-analysis, the significance of the results of each study was interpreted with the following effect sizes: 0.2, small; 0.5, medium; and 0.8, large.
Hedges’ g is a standardized mean difference. The specific effects of studies (mean differences) are also reported in Table 4 (available at www.jandonline.org) for reference.
In this study, the primary outcome was change in BMI or weight. Heterogeneity in BMI or weight among studies was assessed using Q test and I2 with CIs. The articles were thought to be sufficiently heterogeneous, given the diversity of countries in which they were conducted and the variety of races represented across participants. If homogeneity is rejected, a random-effects model is applied. The random-effects model is more conservative and provides better estimates with wider CIs than a fixed-effects model.
Therefore, although the heterogeneity was not significant (I2 = 15.364; P = .249), the random-effects model (DerSimonian-Laird method with inverse variance weighting) was used to assess overall effect sizes; each subgroup was analyzed with Comprehensive Meta-Analysis software (version 3).
The study interventions, the MR-based calorie-restricted diets, were compared with the control interventions, the food-based calorie-restricted diets, after being separated into VLED and LED groups. The standardized mean change difference was calculated first as within-group change from baseline, and then as between-group difference. The shifting unit-of-analysis method was used to prevent any unit-of-analysis error.
Meta-analysis of variance (ANOVA) and meta-regression were used to evaluate the potential clinical significance of the effects of MR diets on weight loss between groups according to the following moderators: age, sex, initial BMI, duration of treatment, frequency of MR intake, and percentage of total daily energy intake from MR. The meta-ANOVA was conducted by treating the patient-related factors as binary, and by dividing the percentage of total daily energy intake from MR, 1 of the treatment-related factors, into the following subgroups: <30%, ≥30% and <60%, and ≥60%. In a hierarchical-approach multivariable meta-regression, background variables are selected first, and the main independent variables are then included to estimate the effect of MR intake while controlling for background variables. To assess the impact of BMI and main independent variables while controlling for age and sex, they were simultaneously applied as covariates. As for the main independent variables, the percentage of total daily energy intake from MR and the frequency of MR intake were not used simultaneously because, although different from a clinical perspective, they are similar from a statistical point of view and seem to be collinear.
The initial search identified 493 potential studies for inclusion. Ultimately, 22
which included 2 interventions, compared MR-based VLED with food-based VLED. Only 2 studies consisted of all men, 6 consisted of all women, and 14 involved both sexes within a group. The country distribution was as follows: 853 patients from the United States (n = 10), 229 patients from Germany (n = 3), 232 patients from India (n = 2), 176 patients from Taiwan (n = 2), 118 patients from Australia (n = 2), 81 patients from the Republic of Korea (n = 2), 67 patients from Brazil (n = 1), and 46 patients from Singapore (n = 1). Table 1 summarizes the characteristics of the included studies. The calorie contents of the MRs ranged from 270 to 1,200 kcal/d.
Table 1Characteristics of included randomized controlled trials for the systematic review and meta-analysis of meal replacement–based, calorie-restricted diets vs conventional calorie-restricted diets
The other domains were assessed as low risk in all studies.
Table 2Evaluation of the risk of bias of included randomized controlled trials for systematic review and meta-analysis of meal replacement–based, calorie-restricted vs conventional calorie-restricted diets, using the Cochrane risk of bias tool
an effect size per study was calculated to consider the study as a unit of analysis. The overall effect size of MR diet on weight loss was small compared with conventional diets (Hedges’ g = 0.255; 95% CI 0.154 to 0.356; I2 = 15.4; 95% CI 0.0 to 48.3). Results of a meta-ANOVA analysis indicated that the effect size of MR-based LED vs food-based LED was small (Hedges’ g = 0.261; 95% CI 0.156 to 0.365; I2 = 21.9; 95% CI 0.0 to 53.6). However, the effect size of MR-based VLED vs food-based VLED (Hedges’ g = 0.166; 95% CI –0.235 to 0.567; I2 = 0) was not statistically significant (Table 3), and there was no significant difference between the LED and the VLED groups (Q = .202, degrees of freedom = 1; P = .653).
Table 3Effect size of meal replacement diets vs conventional diets on weight loss according to calorie-restriction type
Effects of MR-Based LED Compared with Food-Based LED on Weight Loss According to Percentage of Total Daily Energy Intake from MR
Figure 4 depicts the Hedges’ g values and the forest plots for MR-based LED compared with food-based LED. To examine whether the overall effects of MR-based LED differed according to the percentage of total daily energy intake from MR, 3 subgroups (<30%, ≥30% and <60%, and ≥60%) were created according to caloric intake percentage. The meta-ANOVA test showed that in those receiving ≥60% calories from MR, the effect size was medium and had the greatest effect size on weight loss (Hedges’ g = 0.545; 95% CI 0.260 to 0.830; I2 = 42.7; 95% CI 0.0 to 80.8); this was followed by the group receiving ≥30% and <60% (Hedges’ g = 0.232; 95% CI 0.114 to 0.350; I2 = 0). The group with <30% calories from MR was not statistically different (Hedges’ g = 0.055; 95% CI –0.206 to 0.316; I2 = 0); there was a significant difference between groups (Q = 9.226; degrees of freedom = 2; P = .010). In addition, to highlight the clinical significance, the results were separated by BMI and body weight to summarize unstandardized effect sizes, such as mean difference (Table 4; available at www.jandonline.org), and the results of additional meta-ANOVAs that examined whether each factor moderated the effect size of weight loss are presented in Table 5 (available at www.jandonline.org).
This multiple meta-regression model analyzed and selected independent variables using a step-by-step approach and a hierarchical regression analysis. Table 6 summarizes the results of the meta-regression. The results of the meta-regression demonstrated that the percentage of total daily energy intake from MR and the frequency of MR intake were significant moderators (Z = 2.39; P = .017 and Z = 3.15; P = .002, respectively), and none of the other factors, including age, sex, initial BMI, and treatment duration, were significant. Figure 5 depicts the results of the meta-regression on the percentage of total daily energy intake from MR (model 5).
Table 6Multiple meta-regression analysis using the hierarchical regression analysis approach, for the impact of patient- and treatment-related moderators of meal replacement–based, calorie-restricted vs conventional calorie-restricted diets on weight loss
The remaining 11 studies did not report any results for adverse events. The mild adverse events cited included constipation, diarrhea, gastrointestinal symptoms, lethargy, abnormal taste, transitory elimination of flatus, and sleep loss; in the MR-based VLED groups only, the mild adverse events were anemia, constipation, skin dryness, lack of strength, and vomiting.
To the best of the authors’ knowledge, this is the first systematic review and meta-analysis to analyze the effects of MR diets on weight loss compared with conventional food-based diets; the analysis was performed according to the characteristics of the MR plan, including the type of calorie-restriction MR plan and percentage of total daily energy intake from MR.
The results demonstrate that the effect size of MR-based LED on weight loss was small compared with that of food-based LED (Hedges’ g = 0.261); however, compared with food-based VLED, the effect of MR-based VLED on weight loss remained unclear (Hedges’ g = 0.166). The sample size from the studies was too small to allow a subgroup analysis to be performed on the effects of MR-based VLED vs food-based VLED on weight loss. In general, VLEDs are only recommended for short-term use when rapid weight loss is required clinically for treating obesity.
2013 AHA/ACC/TOS guideline for the management of overweight and obesity in adults: a report of the American College of Cardiology/American Heart Association Task Force on Practice Guidelines and The Obesity Society.
reported that VLED with behavioral programs compared with behavioral programs alone resulted in an average reduction of 3.9 kg (95% CI –6.7 to –1.1) at 1 year, with few notable adverse effects. This shows the potential of long-term VLED use as a weight loss plan. However, Bailey and colleagues
reported no significant difference between VLED and LED, and Bailey and colleagues suggested that LED is an efficacious, safe, and less burdensome diet compared with VLED. In view of these findings, the lack of research on the effect of MR-based VLED on weight loss may reflect the lack of preference for VLED in the context of long-term interventions.
A meta-ANOVA and meta-regression were conducted to examine the association between factors of interest and weight loss. First, in terms of the proportion of energy intake per day from MR, the results indicated that ≥60% of total daily energy intake from MR had the greatest effect size for weight loss (Hedges’ g = 0.545), and that ≥30% and <60% had small effect size (Hedges’ g = 0.232). However, when <30% of calories were consumed from MR, this was not statistically significant. There was also a significant difference between groups. In a similar context, when comparing the effects based on the frequency of MR intake, the results revealed that different meal frequency options had different effect sizes for weight loss; the findings also suggested that replacing 3 meals per day had a medium effect size (Hedges’ g = 0.544), whereas simply replacing 2 meals daily had a small effect size (Hedges’ g = 0.177). In the meta-regression, the last 2 variables, that is, percentage of total daily energy intake from MR and frequency of MR intake, were not used simultaneously; this is because they are similar from a statistical perspective, despite being different from the clinical one, as mentioned previously. The meta-regression also showed that only 2 factors—percentage of total daily energy intake from MR and frequency of MR intake per day—had significant effects for weight loss; this finding is similar to those of previous studies (by Guo and colleagues
dietary compliance can vary, depending on the frequency of MR intake regardless of the effect; as MR products are generally not provided for free, these factors must be considered when attempting MR plans for weight loss. The dropout rate increases otherwise. Yackobovitch-Gavan and colleagues
reported that poor weight loss during the first 2 weeks of the program was the strongest predictor of dropout. In conjunction, these results show that replacing 3 meals with MR per day, at least at the beginning of MR plans, will help patients lose weight successfully by increasing initial weight loss and motivation. Reporting the weight loss benefits of the number of meals from MR per day can help patients select a more efficient MR plan.
When comparing the effects according to duration of treatment, Table 5 (available at www.jandonline.org) shows that plans implemented for 3 or more months and fewer than 6 months were slightly effective (Hedges’ g = 0.292), whereas the others were not significantly effective. At the same time, there was no statistically significant difference between the groups. In many studies,
intervention periods for weight loss and weight maintenance were different. Because this study compared randomized controlled trials that performed the same intervention for weight loss, excluding the maintenance period after losing weight, the average treatment period lasted approximately 4 months; there were only 3 studies that assessed plans for 6 or more months (a 26-week study and two 52-week studies). Results revealed that MR-based LED for fewer than 3 months or 6 or more months was not substantially more effective than food-based LED. As it is difficult to maintain calorie-restriction diets for a long time, it can be assumed that short-term, calorie-restricted diets can be controlled by a conventional diet without MR. However, MR interventions for 6 or more months may have decreased dietary adherence as they can become monotonous.
Eleven of the 22 studies reported adverse events, 5 reported no serious adverse events, and 6 documented mild adverse events. The mild adverse events included constipation, diarrhea, gastrointestinal symptoms, lethargy, abnormal taste, transitory elimination of flatus, and decreased sleep. In previous studies on adverse effects of LED, Christensen and colleagues
reported that mild adverse events, such as abdominal and intestinal symptoms, musculoskeletal symptoms, central nervous system and psychiatric symptoms, skin and subcutaneous symptoms, and miscellaneous symptoms, lasted for 8 weeks on average, and a 1-year-long study by Moreno and colleagues
reported constipation, nausea (only for the first 4 months), and hyperuricemia (only for the first 15 days). It is difficult to assess the incidence of adverse events of MR-based diets compared with that of food-based diets, and the relationship is unclear. However, MR meals that can nutritionally enhance the diet are advantageous in terms of nutrient intake (including proteins, essential amino acids, dietary fiber, and vitamins) compared with conventional diets with similar calories.
Nevertheless, other foods without protein and/or dietary fiber, such as milk and cereal, are also used for MR plans; these may help limit caloric intake. However, nutritional adequacy cannot be ensured when consumed over a long time.
As weight loss in patients with obesity often requires calorie restriction, MR plans help ensure effective calorie-restricted diets by restricting the types of foods consumed, as well as by executing excellent portion control. This study sought to provide evidence for an effective percentage of total daily energy intake from MR by analyzing the effect of MR characteristics on weight loss. These findings may be helpful in designing more effective MR diet plans for weight loss by considering the frequency of MR intake and the percentage of total daily energy intake from MR.
There are some limitations to this study. First, as mentioned earlier, the number of included studies that explored the effect of MR-based VLED on weight loss was small. Further studies are needed to address this limitation. Second, because the studies consisting of only men or only women were limited and few studies reported results by separating sexes, the effect of MR-based calorie-restricted diets on weight loss was not analyzed according to sex. However, Miller and colleagues
reported no sex differences in terms of the effectiveness of each intervention in a meta-analysis that compared the effect of diet and/or exercise on weight loss. Third, the effect of MR plans on weight loss according to the nutritional components of MR was not analyzed. More effective MR characteristics for weight loss may be identified if studies on the benefits of MR plans are re-analyzed by systematic reviews and meta-analyses according to their components.
Fourth, meta-regression provides great advantages for investigating intervention effects, while controlling for other factors such as sex, age, and patient status of disease. However, meta-regression is usually underpowered to detect anything but massive associations. Meta-ANOVA could be a better choice for finding and interpreting the clinical meaning. Researchers should consider these merits and drawbacks together when conducting meta-regression analyses. Lastly, this systematic review study has not been registered, despite research registration being recommended by the Preferred Reporting Items for Systematic Reviews and Meta-Analyses reporting standards.
According to this systematic review and meta-analysis, the effect of MR-based LED on weight loss was superior to that of food-based LED. In addition, the findings suggest that at least 30%, and preferably 30% to 60%, of total daily energy intake should be derived from MR for optimal weight loss.
J. Min and S.-Y. Kim reviewed literature and wrote the manuscript as equally contributing authors with support from Y.-B. Park and Y.-Woo Lim. I.-S. Shin verified the analytical methods. All authors discussed the results and approved the final manuscript.
Table 4Between-group differences in body mass index or body weight of meal replacement-based low-energy diets compared with conventional low-energy diets
2013 AHA/ACC/TOS guideline for the management of overweight and obesity in adults: a report of the American College of Cardiology/American Heart Association Task Force on Practice Guidelines and The Obesity Society.