Author Affiliations: Division of Endocrinology and Metabolism, Department of Medicine (Drs Pedersen and Kline), and Department of Community Health Sciences (Ms Kang), University of Calgary, Calgary, Alberta.
Portion size is an important determinant of energy intake. To our knowledge, no randomized controlled trial has evaluated the efficacy of portion control tools to induce weight loss. In patients with type 2 diabetes mellitus, weight reduction improves glycemic control.
We randomly assigned 130 obese patients with type 2 diabetes mellitus (including 55 patients taking insulin) to the daily use of a commercially available portion control plate for 6 months (intervention group) vs to usual care in the form of dietary teaching (usual care control group).
Follow-up was 93.8%. Patients in the intervention group lost significantly more weight than control subjects (mean ± SD, 1.8% ± 3.9% vs 0.1% ± 3.0%, P = .006). Compared with controls, more patients in the intervention group required a decrease in their diabetes medications at 6 months (26.2% vs 10.8%, P = .04).
Compared with usual care, the portion control tool studied was effective in inducing weight loss. The portion control plate also enabled patients with diabetes mellitus to decrease their hypoglycemic medications without sacrificing glycemic control.
clinicaltrials.gov Identifier: NCT00254124
Obesity has become a worldwide epidemic, and its prevalence is increasing. Between 1960 and 2000, the prevalence of obesity among American adults increased from 13.4% to 30.9%.1,2
Most cases of type 2 diabetes mellitus (DM) can be attributed directly to obesity. The risk of DM varies with the degree, duration, and distribution of obesity.3,4 Dietary caloric restriction has been shown to improve glycemic control by virtue of weight loss,5 with an additional benefit independent of weight loss.6
The increasing prevalence of obesity is paralleled by increasing portion sizes in the marketplace.7 Portion sizes are an important determinant of energy intake; the number of calories ingested by subjects at a meal has been directly correlated with the serving size offered.7- 9 Despite the plethora of diet strategies available, it is primarily the net caloric intake that determines the net change in weight.10,11 The principles of portion control are commonly used by diabetes educators to effect a hypocaloric diet; patients are often taught to use reference sizes relative to their hand size to guide protein, carbohydrate, and fat portions.12 However, to our knowledge, no clinical trial has been published that examined the efficacy of a food portion control tool to control caloric intake and thereby induce weight loss. We conducted a 6-month randomized controlled trial designed to determine the efficacy of a portion control plate to induce weight loss among a population with obesity and type 2 DM.
Patients were recruited from April 1, 2004, through June 30, 2004, from an outpatient diabetes center in Calgary, Alberta. Patients were recruited via fliers posted at the diabetes center or by referral from their diabetes health care providers (ie, physicians, nurses, or dieticians). All included subjects had type 2 DM and a body mass index (calculated as weight in kilograms divided by height in meters squared) of at least 30 and had to be clients of the diabetes center, with at least 6 months' prestudy teaching and management by diabetes nurse educators and dieticians. Exclusion criteria were as follows: presence of cancer, current weight loss medication, history of bulimia or anorexia nervosa, psychiatric illness under the care of a psychiatrist, surgery in the 3 months before enrollment or planned during the study period, weight loss exceeding 10 lb (4.5 kg) in the 2 months preceding study enrollment, and consumption of dinner at a restaurant more than twice weekly or consumption of more than 30% of all meals at restaurants. All participants provided written informed consent, and the local institutional ethical review board approved the protocol. The study was funded by the Stewart Diabetes Education Fund, a non–industry-based trust fund for diabetes research.
A detailed medical and demographic history was recorded at the baseline appointment. Sealed envelopes containing group designations were used to randomly assign subjects to 1 of 2 groups. The intervention group received a calibrated dinner plate and cereal bowl as a means of dietary portion control. A 10- to 15-minute individual counseling session and a booklet outlining the use of these tools were provided. The control group did not receive any treatment or counseling outside of their usual clinical care in the form of dietary assessment and teaching by their dieticians. During the study, patients had access to their dieticians on an as-needed basis as per usual care.
For the 6-month duration of the study, participants agreed not to start any weight loss medication or special diets during the trial period. Participants were contacted by telephone on 4 occasions during the study to remind them of study participation. Study investigators were not involved in any medication adjustments that were made; patients were instructed to consult with their usual physician to make any necessary adjustments.
Patients were seen in follow-up at 6 months, at which time changes in medications and dosages were assessed on an individual basis. Any new hypoglycemic medications or change from one medication to another was assumed to represent an increase in intensity of hypoglycemic medical therapy. If the number of units of insulin was increased in conjunction with a decrease or discontinuation of an oral hypoglycemic agent, the patient was classified as having an increase in intensity of medical therapy. The same principles were applied to any changes made in lipid-lowering or antihypertensive agents.
A commercially available calibrated dinner plate and breakfast bowl constituted the intervention (Figure 1).13 The plate sizes are sex specific; the male plate is calibrated for an approximate 800-cal (3347-J) meal, whereas the female plate is calibrated for a 650-cal (2720-J) meal. This caloric goal is accomplished by dividing the plate area into sections designed to contain predetermined volumes of carbohydrates, proteins, cheese, and sauces, with the remainder of the plate open for vegetables. For consumption of mixed meals that cannot be readily divided into the respective macronutrient sections, food is portioned into the section that represents that meal's dominant macronutrient.
Diet plate and breakfast bowl (The Diet Plate, Glossop, England).13
The cereal bowl is calibrated to portion-control breakfast cereal and is not sex specific. This bowl is designed to allow a 200-cal (837-J) meal consisting of cereal and cup (118 mL) of milk; this goal is accomplished by calibration at 4 different levels to accommodate cereals of differing caloric density. Subjects were provided with a list of commonly consumed cereals with their corresponding calibration levels in the bowl.
Subjects were instructed to use the calibrated plate for their largest meal of the day and to use the cereal bowl on days when cereal was consumed as their breakfast meal. Patients were counseled not to compensate for portion control at mealtime by eating more at other times of day. To document adherence, participants were asked to keep a daily log documenting their use.
Outcome measures were assessed at baseline and at 6 months. Body weight of the participants, while wearing a hospital gown and no shoes, was measured using a single calibrated scale (Health O Meter; Continental Scale Corporation, Chicago, Ill). Blood pressure was measured using a validated automated machine (BP Tru; VSM Med Tech Ltd, Vancouver, British Columbia). Glycosylated hemoglobin level was measured at baseline and at 6 months by immunoassay (Roche Integra 700; Roche Diagnostics, Indianapolis, Ind), and serum lipid levels were measured following an overnight fast (Roche Hitachi Modular P800 Analyzer; Roche Diagnostics). Low-density lipoprotein levels were calculated according to the Friedewald equation.14
The primary end point was defined as percentage change from baseline weight. A sample size of 98 individuals would have an 80% power to detect a significant 5% weight loss in the intervention group, assuming a 1% reduction in body weight in the control group and a common SD of 7%. Given an anticipated dropout rate of 25%, the enrollment target was set at 130 patients. A prespecified analysis examining the proportion of patients in each group who achieved at least 5% weight loss was performed. Analyses of change in weight were conducted by intent to treat. For the 8 patients who did not complete the study, we assumed their weight to be unchanged from baseline.
Independent sample t tests were performed to compare continuous variables between the 2 groups. The groups were compared using Pearson χ2 test on categorical variables, and the Fisher exact test was used where appropriate. The percentage change in weight was assessed using a 2-sided t test. The Fisher exact test was used to compare propartional weight loss and change in diabetes medication requirements between groups. The level of significance was set at P<.05 for testing the effects of the intervention.
Secondary outcomes were defined as changes in glycosylated hemoglobin level, blood pressure, or cholesterol profile. Changes in these variables were assessed using unpaired t tests, with no adjustment for multiple testing. Subgroup analyses were conducted using linear and logistic regression analyses. Adjusted odds ratios and confidence intervals were calculated using normal approximation with small sample adjustments.15 Commercially available software (R 2.0.0; http://www.R-project.org)16 was used for all analyses.
Two hundred forty-five patients were screened to obtain 130 participants for the study (Figure 2). Sixty-five were excluded because of failure to meet enrollment criteria; 50 declined participation for unspecified reasons. One hundred twenty-two patients were seen at the 6-month visit, for a follow up of 93.8%.
Flow of participants through the trial.
The 2 groups were well matched for baseline characteristics (Table 1). The mean age of participants at baseline was 56 years (age range, 31-77 years), 77 (59.2%) were women, and 55 (42.3%) were taking insulin. Dropouts did not differ from participants who completed the study with respect to any baseline variable.
Two protocol violations were observed during the trial period. Specifically, 1 patient in each group started an unrelated commercial weight loss diet during the study period. Both patients were included in all analyses.
Seven participants in the intervention group were unable to provide logbook documentation of portion control tool use at the 6-month follow-up visit. Among the 53 patients who kept a logbook, the median compliance was 70.8%.
Patients in the intervention group lost significantly more weight than control subjects (mean ± SD, 1.8% ± 3.9% vs0.1% ± 3.0%, P = .006; absolute values, 2.1 ± 4.9 kg vs 0.1 ± 3.5 kg, P = .01) (Table 2). Compared with controls, a significantly greater proportion of the intervention group achieved at least 5% weight loss (11/65 [16.9%] vs 3/65 [4.6%], P = .048).
In contrast to patients taking insulin in the control group, patients taking insulin in the intervention group experienced weight loss (−2.6% ± 4.3% vs 0.1% ± 3.0%, P = .01) (Table 2). The difference in proportions of patients taking insulin achieving at least 5% weight loss showed a trend toward significance (6/26 [23.1%] in the intervention group vs 1/25 [4.0%] in the control group, P = .10).
In the subgroup of patients not taking insulin, there was no significant difference in change in weight between the intervention group and the control group (−1.3% ± 3.7% vs −0.2% ± 3.2%, P = .16) (Table 2). Similarly, there was no significant difference between the 2 groups in the percentage of patients achieving at least 5% weight loss (5/34 [14.7%] in the intervention group vs 2/37 [5.4%] in the control group, P = .25).
Glycosylated hemoglobin levels did not differ between the groups at baseline (Table 1) or at follow-up (Table 3). The intrapatient change in glycosylated hemoglobin level was not different between the groups. The mean within-group change did not differ significantly from zero. However, a significant difference was seen in terms of change in hypoglycemic medications. Compared with controls, a significantly greater proportion of the intervention group experienced a decrease in hypoglycemic medication use (17/65 [26.2%] vs 7/65 [10.8%], P = .04).
Conversely, compared with the intervention group, a greater proportion of the control group required an increase in hypoglycemic medication use at 6 months (22/65 [33.8%] vs 9/65 [13.8%], P = .01) (Table 3). Among insulin users, compared with increased requirements in the control group, patients in the intervention group required a significant decrease in daily insulin (mean ± SD, −8.4 ± 19.3 vs 7.2 ± 24.7 U, P = .02).
Compared with control subjects, patients in the intervention group had a significantly greater decrease in non–high-density lipoprotein cholesterol levels (13.6% ± 20.2% vs 4.8% ± 23.4%, P = .04). There was no difference between groups in terms of proportions of each group that experienced an increase in lipid-lowering medication use. No other differences were seen in the cholesterol profiles between groups.
Change in systolic blood pressure was significantly higher in the intervention group at 6 months compared with that in the control group (5.5% ± 15.4% vs −4.2% ± 11.0%, P<.001) (Table 3). No significant difference was seen between groups with respect to change in diastolic blood pressure. No significant difference was seen between groups in terms of changes in antihypertensive medication use.
Consistent results, as defined by achievement of at least 5% weight loss, were observed regardless of age, sex, education level, baseline body mass index, obesigenic hypoglycemic medication use, or insulin use at baseline (Figure 3). Likewise, no variables were found to be predictive of compliance with the intervention.
Clinically significant weight loss by subgroup analysis. Unless otherwise indicated, data are given as percentage of patients (No.) in the subgroups who lost at least 5% of body weight. BMI indicates body mass index (calculated as weight in kilograms divided by height in meters squared); OR, odds ratio.
One hypoglycemic reaction was reported in the intervention group. It necessitated a visit to the emergency department but had no serious sequelae.
In this first randomized trial (to our knowledge) evaluating the use of a portion control tool as a treatment for obesity, we demonstrated that this simple inexpensive intervention is able to induce weight loss in an obese population with type 2 DM. The proportion of patients who achieved at least 5% weight loss in the intervention group was significantly higher than that in the control group receiving usual care. This is important, as a 5% weight loss has been shown to be clinically significant in terms of decreasing morbidity and mortality associated with obesity-linked disorders such as cancer and myocardial infarction.17- 19 We demonstrated that this portion control system is effective in decreasing hypoglycemic medication requirements and in preventing much of the progressive increase in medication requirements typical of patients with type 2 DM. The decrease in hypoglycemic medication use was made without sacrificing glycemic control.
These results are comparable to results of several studies using pharmaceutical intervention for weight loss in type 2 DM. In a recent 6-month trial, patients with type 2 DM were randomized to receive sibutramine hydrochloride vs placebo.20 The study excluded patients taking insulin, thus selecting a subgroup in whom it may be easier to achieve weight loss. Although the study found a significantly greater weight loss compared with that in control subjects (4.5 vs 1.7 kg, P<.001), a significant difference was not seen between groups in the proportions of patients losing at least 5% of body weight. Furthermore, all patients in the sibutramine trial received dietary counseling geared toward moderate caloric restriction, which was reinforced at monthly clinic visits. Such counseling was not reinforced during our study, which emphasizes the broader generalizability of our results to routine clinical care. From an economic perspective, the intervention in our study is a 1-time investment that costs approximately one sixth of the cost of a 6-month prescription of sibutramine in Canada.
Similar comparisons can be made with trials of orlistat in type 2 DM. In a 1-year randomized trial that excluded patients taking insulin and placed all subjects on a calculated hypocaloric diet, 39% of patients taking orlistat lost at least 5% of baseline weight compared with 15.7% receiving placebo (odds ratio, 2.48),21 compared with 11 (16.9%) of 65 patients in our intervention group and 3 (4.6%) of 65 in our control group (odds ratio, 3.67). Although the study was a 1-year trial, the change in weight was stable from 6 months to 1 year. Similar changes in hypoglycemic medication use were seen as were observed in our study. The orlistat trial had a high attrition rate of 40%, which is typical of studies of pharmacotherapy for weight loss. The 1-time cost of our portion control intervention is one fourth of the cost of a 6-month prescription of orlistat.
Although we observed no difference between groups in terms of glycosylated hemoglobin levels, this may be because the patients' primary physicians were free to decrease the number or dosage of hypoglycemic medications in response to lower glycemia. The observed change in hypoglycemic medications may be of great clinical importance by decreasing cost and potential for serious adverse effects. It is plausible that a component of the decrease in hypoglycemic medication use seen in this study was by virtue of decreasing carbohydrate intake alone. The cause of increased blood pressure seen in the intervention group is unclear; this outcome may have been a random play of chance.
It is interesting that the significant weight loss using the intervention in this study was primarily seen in patients taking insulin, based on post hoc analyses, as it is notoriously difficult to achieve weight loss in patients using this anabolic hormone. Further prospective studies are required to explore the success of this tool in combination with various types of hypoglycemic agents.
The successful weight loss seen using this intervention in patients with type 2 DM lends optimism toward its potential success in populations without DM, as investigations of weight loss interventions are typically less successful among obese patients with DM compared with those with obesity without DM.22,23 This portion control intervention was designed to target an overweight adult population and is not specific for use in patients with DM. It should be studied in other obese populations.
This study is subject to some limitations. Overall compliance with the intervention was poor, thereby underestimating the potential success. Having a support system in the form of group meetings or more frequent patient contact may have improved compliance with the portion control tools. Strategies of obesity treatment are more successful when support strategies are in place for patient assistance and for ongoing encouragement.24,25 Also, patients who eat frequent meals outside of the home were excluded from the study, as the use of this intervention would be difficult in restaurants. Proving that this tool induces weight loss among patients who take most of their meals at home may serve as motivation for future patients using this tool to take their meals at home as well, in an environment that is more amenable to control of caloric intake.
As a trial of a meal control tool, this study could not be conducted in double-blind fashion. However, the outcomes in question were objective and were not prone to interpretation bias. This may be a strength of the study in that it describes positive results in a real-life application of this intervention. Finally, data were not collected regarding quantitative dietary composition changes or caloric expenditure; the exact mechanism of the intervention may be more complex than portion control alone.
In conclusion, the portion control tool studied in this trial was effective in inducing weight loss in obese persons with type 2 DM comparable to that seen in investigations of weight loss pharmacotherapy. This simple inexpensive tool also enabled obese patients with DM to decrease their hypoglycemic medication requirements. This intervention holds promise for use in overweight populations with and without DM.
Correspondence: Sue D. Pedersen, MD, FRCPC, Division of Endocrinology and Metabolism, Department of Medicine, University of Calgary, Foothills Medical Centre, 1441 29th St NW, North Tower, Room 302, Calgary, Alberta, Canada T2N 4J8 (email@example.com).
Accepted for Publication: February 21, 2007.
Author Contributions: Dr Pedersen had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis. Study concept and design: Pedersen and Kline. Acquisition of data: Pedersen. Analysis and interpretation of data: Pedersen, Kang, and Kline. Drafting of the manuscript: Pedersen. Critical revision of the manuscript for important intellectual content: Pedersen, Kang, and Kline. Statistical analysis: Kang. Obtained funding: Pedersen and Kline. Administrative, technical, and material support: Pedersen. Study supervision: Kline.
Financial Disclosure: None reported.
Funding/Support: This study was sponsored by the Stewart Diabetes Education Fund. Portion control tools were donated for use in the study by The Diet Plate.
Role of the Sponsors: Neither the Stewart Diabetes Education Fund board nor The Diet Plate had any role in the design or conduct of the study; collection, management, analysis, or interpretation of the data; or preparation, review, or approval of the manuscript.
Previous Presentations: This study was presented at the 87th Meeting of The Endocrine Society; June 6, 2005; San Diego, Calif; and at the 9th Annual Meeting of the Canadian Diabetes Association; October 21, 2005; Edmonton, Alberta.
Acknowledgment: We thank Tammy MacLeod, RD, Krista Engstrom, RD, and Katie Couzens, RD, for their assistance throughout the study. We thank Rollin Brant, PhD, for his statistical guidance.
Thank you for submitting a comment on this article. It will be reviewed by JAMA Internal Medicine editors. You will be notified when your comment has been published. Comments should not exceed 500 words of text and 10 references.
Do not submit personal medical questions or information that could identify a specific patient, questions about a particular case, or general inquiries to an author. Only content that has not been published, posted, or submitted elsewhere should be submitted. By submitting this Comment, you and any coauthors transfer copyright to the journal if your Comment is posted.
* = Required Field
Disclosure of Any Conflicts of Interest*
Indicate all relevant conflicts of interest of each author below, including all relevant financial interests, activities, and relationships within the past 3 years including, but not limited to, employment, affiliation, grants or funding, consultancies, honoraria or payment, speakers’ bureaus, stock ownership or options, expert testimony, royalties, donation of medical equipment, or patents planned, pending, or issued. If all authors have none, check "No potential conflicts or relevant financial interests" in the box below. Please also indicate any funding received in support of this work. The information will be posted with your response.
Some tools below are only available to our subscribers or users with an online account.
Download citation file:
Web of Science® Times Cited: 18
Customize your page view by dragging & repositioning the boxes below.
The Rational Clinical Examination EDUCATION GUIDES
Abdominal Aortic Aneurysm
All results at
Enter your username and email address. We'll send you a link to reset your password.
Enter your username and email address. We'll send instructions on how to reset your password to the email address we have on record.
Athens and Shibboleth are access management services that provide single sign-on to protected resources. They replace the multiple user names and passwords necessary to access subscription-based content with a single user name and password that can be entered once per session. It operates independently of a user's location or IP address. If your institution uses Athens or Shibboleth authentication, please contact your site administrator to receive your user name and password.