Disclosures: The authors declared no conflict of interest.
Big breakfast rich in protein and fat improves glycemic control in type 2 diabetics
Article first published online: 6 DEC 2013
Copyright © 2013 The Obesity Society
Volume 22, Issue 5, pages E46–E54, May 2014
How to Cite
Rabinovitz, H. R., Boaz, M., Ganz, T., Jakubowicz, D., Matas, Z., Madar, Z. and Wainstein, J. (2014), Big breakfast rich in protein and fat improves glycemic control in type 2 diabetics. Obesity, 22: E46–E54. doi: 10.1002/oby.20654
- Issue published online: 1 MAY 2014
- Article first published online: 6 DEC 2013
- Accepted manuscript online: 29 OCT 2013 05:12AM EST
- Manuscript Accepted: 19 OCT 2013
- Manuscript Revised: 24 SEP 2013
- Manuscript Received: 18 JUN 2013
Our goal was to evaluate the effect of breakfast size and composition on body weight, glycemic control, and metabolic markers in adults with type 2 diabetes mellitus (T2DM).
59 overweight/obese adults with T2DM were randomized to one of two isocaloric diabetic diets for 3 months; big breakfast (BB), breakfast was rich in fat and protein and provided 33% of total daily energy or small breakfast (SB), breakfast was rich in carbohydrates and provided 12.5% of total daily energy.
Although body weight was reduced similarly in both groups, the BB group showed greater HbA1c and systolic blood pressure reductions (HbA1c: −4.62% vs. −1.46%, p = 0.047; SBP −9.58 vs. −2.43 mmHg; p = 0.04). T2DM medication dose was reduced in a greater proportion of the BB participants (31% vs. 0%; p = 0.002) while in the SB, a greater proportion of participants had a dose increases (16.7% vs. 3.4%; p = 0.002). Hunger scores were lower in the BB group and greater improvements in fasting glucose were observed in the BB group.
A simple dietary manipulation enriching breakfast with energy as protein and fat appears to confer metabolic benefits and might be a useful alternative for the management of T2DM.
Overweight and obesity are associated with increased glucose intolerance and type 2 diabetes mellitus (T2DM) risk, and obesity and T2DM prevalence have increased dramatically over the past few decades [1-3]. Both obesity and T2DM are associated with increased morbidity and mortality . Abdominal obesity is associated with insulin resistance, which plays a central role in the metabolic syndrome including dyslipidemia, hypertension, and increased risk of cardiovascular disease, in addition to T2DM [5, 6]. The mechanisms through which obesity leads to insulin resistance and to T2DM are purportedly related to abnormalities in free fatty acids, adipokines, proinflammatory cytokines, leptin, and other substances [7, 8].
Concurrent with the rise in obesity prevalence, eating breakfast has declined [9, 10]. Several cross-sectional studies have identified an inverse association between breakfast consumption and BMI or weight gain; this association between skipping breakfast and obesity has been reported in adults [11-13] as well as children and adolescents [14, 15]. One cross-sectional study in adults, reported elevated BMI in subjects who regularly skipped breakfast despite lower daily energy intake compared to other groups . Similar results were found in children and adolescents . Additionally, one of the National Weight Control Registry (NWCR) strategies for successful long-term weight loss and weight maintenance is regular breakfast consumption . Clinical trials in women  and men  have shown that eating vs. skipping breakfast reduced daily energy intake, glucose homeostasis, free fatty acid, and lipid profile in response to a test meal.
Analyses of intra-individual eating habits have demonstrated that increased calorie intake at breakfast is associated with a lower energy intake over the course of the day [21, 22]. In a large longitudinal study, increased percentage of daily energy consumed at breakfast was associated with relatively lower weight gain . A big breakfast rich in carbohydrate and protein has been shown to prevent weight regain by reducing diet-induced compensatory changes in hunger, cravings, and ghrelin suppression . However, another study found that increasing breakfast energy was associated with greater overall intake in normal weight and obese subjects .
Given these mixed results, further examination of the association between breakfast consumption and body weight and diabetes is warranted. The present study was designed to address whether a change in breakfast size and composition impacts body weight and metabolic outcomes leading to long term glycemic control in adults with T2DM.
The present randomized, treatment-controlled, open clinical trial was conducted at the Diabetes Unit, E. Wolfson Medical Center, Holon, Israel, from June 2011 to April 2012. The study compared the effects of two isocaloric dietary interventions with different breakfast size and composition on body weight, glycemic control, and its association with metabolic markers in adults with T2DM. Study duration was 3 months.
A total of 59 obese/overweight subjects (21 men) were recruited by means of advertising in the radio and the local newspaper. Participants were adults (age 45-70 years); overweight or obese (body mass index 25-40 kg/m2); T2DM (characterized by fasting glucose ≥126 mg/dl on two separate tests or glucose ≥200 mg/dl 2 h following oral administration of 75 g glucose) with normal thyroid, liver, and kidney function as assessed by standard blood tests. Excluded from participation were subjects with type I diabetes; insulin-dependent; using medications designed to treat obesity or GLP-1 agonists; and/or had a change in any medication dose within the 3 months preceding study onset. Subjects with gastrointestinal problems possibly preventing dietary adherence; pregnancy or lactation; cancer or other characteristics (psychological or physical disabilities) deemed likely to interfere with participation in or compliance with the study were also excluded.
Most subjects were sedentary at baseline and were asked to maintain their usual physical activity levels. The protocol and potential risks and benefits of the study were fully explained to each subject before he/she provided a written informed consent. All experimental procedures followed ethical standards of and were approved by the Institutional Review Board (Helsinki Committee) at the E. Wolfson Medical Center, Holon, Israel.
On day 0, participants met the project dietitian, completed questionnaires, underwent anthropometric measurements and blood samples were collected. Subjects were randomly assigned to one of two isocaloric weight loss diets using the lottery method. Diets differed primarily in the size and composition of the breakfast meal.
In order to have significant difference in breakfast size between the groups, the following distribution of energy during the day were used:
For breakfast 13% of total energy was recommended in the small breakfast diet (SB), while 33% of total energy was recommended in the big breakfast diet (BB).
For lunch and dinner, each meal has recommended 33% of total energy in the SB group, and 25% in the BB group.
The remaining of the total energy was divided evenly to two or three snacks throughout the day in the two groups.
Breakfast composition differed between the groups. In the SB group, the breakfast was a high carbohydrate meal (12-18% protein, 14-22% fat, 60-70% carbohydrates) while in the BB diet the breakfast was a fat- and protein-enriched meal (23-30% protein, 29-37% fat, 37-48% carbohydrates).
The total daily energy requirement for each subject was calculated using the Harris–Benedict formula , with energy reduced of 500 kcal to achieve a gradual decrease in weight.
The study dietitian met all participants personally at 1-3-week intervals in order to perform a comprehensive inquiry and estimate adherence to dietary regimen and caloric intake. Participants were asked to record their dietary intake three times during each 2-week period to assess compliance.
Subjects were weighed every 1-3 weeks during the study on a Detecto Physician Beam Scale (HOSPEQ, Inc Miami, Florida), before breakfast, wearing light clothes but no shoes. Standing height was measured without shoes to the nearest 0.5 cm using the height rod attached to the scale. Waist and hip circumference were measured using a stretch-resistant tape. Percent body fat (%BF) estimates were determined using the Tanita bioelectrical impedance analysis (BIA) system. Blood pressure was measured with the patient in a supine position using a standard cuff and sphygmomanometer after participants rested for at least 10 min.
All assays were performed after overnight fasting on day 0 and at the end of month 3. Venous blood samples were collected into SST tubes, centrifuged at room temperature at 1500g for 10 min. Glucose, CRP, TG, total cholesterol, and HDL-cholesterol in serum were measured on the same day by Olympus AU2700 Beckman Coulter analyzer using manufacture's kits. LDL-cholesterol was calculated for TG < 300 mg/dl by using Fricdwald formula . Cortisol was analyzed by Abbott AxSYM immunochemistry analyzer using manufacture's kit. The remaining serum was divided to 0.5 ml portions and stored at −70°C until analyzed.
Human Adiponectin, IL-6 high sensitive, TNF-α high sensitive, and Leptin were measured with commercial sandwich enzyme immunoassay technique R&D system, Minneapolis, USA. HbA1c was analyzed on whole blood samples collected into EDTA tubes by Olympus 640AU Beckman Coulter analyzer using manufacture's kit.
Blood samples for Gherlin-active were drawn into EDTA tubes contained 5000 U of approtonin/ml. Blood put on ice immediately and centrifuged within 1 h at 1500g for 10 min at 4°C. 1 ml of plasma was acidified with 100 µl 1N HCL and stored at −70°C until analyzed with two-site sandwich enzyme-linked immunosorbent assay (ELISA) kit, Simco Billerica MA kit.
H-SS questionnaire was adapted from Paul E. Garfinkel  and translated into Hebrew . Participants were asked to complete the questionnaire three times during each 2-week period for each one of the following times: before meals (breakfast, lunch, and dinner) and after breakfast. The participants chose statements, which best described how they felt at each time point. Hunger-Satiety Score (H-SSc) is a scale of descriptions that ranges from starving (1 point) to devastatingly full (10 points). High H-SSc indicates less hunger and greater satiety. Other questions analyzed dealt with “urge to eat” and “preoccupation with thoughts about food.”
Glucose Levels Monitored by Glucometer
A glucometer was provided to each participant and instructions for use were given by the Diabetes Unit nurse. For each 2-week period, participants were asked to measure blood glucose not less than three times at each of the following times: before meals (fasting glucose before lunch, and before dinner), 2 h after meals (2 h after breakfast, lunch, and dinner), and before sleep. Changes in medications dose were made as necessary by clinic physicians.
Sample Size and Study Power
A sample size of 52 participants (26 in each treatment group) provided 80% power to detect a true, between-group difference of 1 ± 1.25 kg at the end of follow-up. This calculation assumes equal variances and a two-sided alpha of 0.05. Additional subjects were recruited to retain study power assuming a 10% dropout rate.
Data were stored on spreadsheet (Excel, Microsoft Inc., USA) and analyzed on SPSS v 19.0 (IBM Inc., USA). Distributions of continuous data were tested for normality using the Kolmogorov–Smirnov test. Continuous data are described as mean ± SD. Categorical data are described as frequency counts and presented as n (%). Continuous data were compared by treatment assignment in two analyses: once in the total cohort using last observation carried forward (LOCF) to impute missing data and the intention-to-treat principle; and second using only the completer's cohort. In these analyses, data were compared by treatment assignment using the t test for independent samples or the Mann–Whitney U as appropriate. Data were also analyzed as a between-group comparison of change-from-baseline values. Data measured on several occasions were analyzed using general linear modeling repeated measures analysis including treatment assignment as the fixed factor in all cases. In another analysis, end of treatment values or change from baseline values were used as the dependent variable, treatment assignment was included as a fixed factor and the baseline value of the same variable was included as a fixed factor in a univariate general linear model. This approach was used in neutralize regression towards the mean in variables with large baseline differences. Within group changes were evaluated using the t test for paired samples or the Wilcoxon-signed ranks test as appropriate. Categorical variables were compared by treatment assignment using the chi square test, exact as appropriate. All tests were two-sided and considered significant at p < 0.05.
Out of 150 screened patients, 80 met inclusion criteria, of which 59 patients expressed their willingness to commit to the research and were randomly allocated to the experimental (BB diet; n = 29) or control (SB diet; n = 30) groups. 46 patients completed the 3-month diet regimen; the dropout rates were similar in both groups.
A flow diagram of the study and detailed patient dispensation is shown in Figure 1.
Individuals who dropped out of the study (n = 13) had significantly greater baseline BMI than those who completed the study (n = 46): 34.6 ± 3.7 vs. 31.8 ± 3.5 kg/m2, p = 0.02. Otherwise, there were no significant differences in terms of age, sex, smokers, or biochemical measures (data not shown).
As shown in Table 1, baseline characteristics were similar between the treatment groups.
|Subject characteristics||Experimental group (BB) (n = 29)||Control group (SB) (n = 30)||p value|
|Age (years)||59.8 ± 6.7||61.6 ± 6||0.29|
|Men, n (%)||11 (38%)||10 (33.3%)||0.71|
|Weight (kg)||87.05 ± 12.2||89.23 ± 14.7||0.53|
|BMI (g/m²)||31.95 ± 3.7||32.8 ± 3.7||0.38|
|Waist circumference (cm)||106.8 ± 9.2||105.8 ± 11.1||0.69|
|Hip circumference (cm)||111.6 ± 9.1||111.4 ± 7.8||0.91|
|Body fat percent (%)||38.16 ± 8.1||38.13 ± 7.3||0.99|
|Systolic BP (mmHg)||135.3 ± 20.9||134 ± 17.1||0.79|
|Diastolic BP (mmHg)||75.76 ± 8.7||75.6 ± 9.7||0.96|
|Heart rate (bpm)||74.83 ± 11.36||72.6 ± 12.14||0.47|
|Physical activity (minutes per week)||117.6 ± 118.7||116.3 ± 114.6||0.97|
|HbA1c (%)||6.9 ± 1||6.85 ± 1.1||0.84|
|Estimated average glucose (mg/dl)||151.5 ± 30.6||149.8 ± 31.5||0.83|
|Glucose (mg/dl)||139.8 ± 36.9||146.8 ± 53.6||0.56|
|Insulin (µIU/ml)||17.58 ± 9.07||16.45 ± 7.74||0.61|
|C-peptide (ng/ml)||3.4 ± 0.85||3.52 ± 1.46||0.74|
|Cholesterol total (mg/dl)||178.5 ± 29.8||183.1 ± 37.3||0.61|
|Triglycerides (mg/dl)||149 ± 56.9||188.9 ± 196.7||0.31|
|HDL-Cholesterol (mg/dl)||43.3 ± 10.97||45.1 ± 10||0.51|
|LDL-Cholesterol (mg/dl)||107.1 ± 23.2||105.1 ± 32.7||0.8|
|Cortisol total (µg/dl)||11.87 ± 2.67||13 ± 4.7||0.26|
|CRP (mg/dl)||0.47 ± 0.36||0.46 ± 0.33||0.89|
|IL-6 (pcg/ml)||3.97 ± 3.34||3.48 ± 2.1||0.58|
|TNF-α (pcg/ml)||3.28 ± 2.36||2.38 ± 0.75||0.11|
|Ghrelin (pcg/ml)||327.3 ± 266||363.3 ± 249||0.66|
|Adiponectin (ng/ml)||6,702.6 ± 2,565||7,405.4 ± 3,137||0.44|
|Leptin (pg/ml)||22,020 ± 13,617||24,150 ± 10,746||0.59|
|Average of fasting plasma glucose (mg/dl)||136.6 ± 27.1||127.5 ± 26.8||0.26|
In the analysis of the dietary records we found that the patients consumed significantly higher calories than their recommended (Table 2).
|BB (n = 23)||SB (n = 23)||p|
|Recommended kcal||1,427 ± 249||1,373 ± 258||0.467|
|Actual kcal intake||1,631 ± 179||1,586 ± 223||0.458|
|% kcal breakfast||30 ± 2.7||13.2 ± 1.5||≤0.001|
|% CHO breakfast||38.8 ± 3.6||55 ± 7.9||≤0.001|
|% Protein breakfast||21.6 ± 2||14.9 ± 2.6||≤0.001|
|% Fat breakfast||39.4 ± 4||29.6 ± 8.9||≤0.001|
|% kcal snack 1||7.2 ± 2.2||7.9 ± 2.6||0.331|
|% kcal lunch||28.9 ± 3||34.4 ± 4||≤0.001|
|% kcal snack2||7.3 ± 2.2||8.9 ± 3.4||0.059|
|% kcal dinner||21.6 ± 3.4||28 ± 4.3||≤0.001|
|% kcal snack3||5 ± 2.3||7.6 ± 3||0.002|
The overall compliance to diet was as expected and significant differences between the groups were found in breakfast size and composition.
Additionally, the SB group consumed significantly higher caloric in snack 3.
Anthropometric and Hemodynamic Measures
After 3 months, between-group differences in weight loss were not detected (Table 3). Weight loss was −2.43 ± 0.46 kg (2.75% of body weight) in the BB group vs. −1.86 ± 0.4 kg (2.22% of body weight) in the SB group, p = 0.35.
|BB (n = 29)||SB (n = 30)||p||pa|
|Weight (kg)||−2.43 ± 0.46||−1.86 ± 0.4||0.35||N.S.|
|BMI (g/m²)||−0.88 ± 0.17||−0.69 ± 0.15||0.389||N.S.|
|Waist circumference (cm)||−2.65 ± 0.66||−2.2 ± 0.47||0.575||N.S.|
|Hip circumference (cm)||−2.28 ± 0.46||−1.53 ± 0.43||0.245||N.S.|
|Body fat percent (%)||−0.49 ± 0.32||−0.675 ± 0.23||0.656||N.S.|
|Systolic bp (mmHg)||−9.58 ± 2.23||−2.43 ± 3.06||0.066||0.04|
|Diastolic bp (mmHg)||−3.14 ± 1.5||−0.6 ± 1.46||0.231||N.S.|
|Heart rate (bpm)||−3.9 ± 1.37||−0.06 ± 1.62||0.072||0.94|
|HbA1c (%)||−0.46 ± 0.15||−0.146 ± 0.07||0.065||0.047|
|Estimated average glucose (mg/dl)||−13.2 ± 4.3||−4 ± 2.25||0.061||0.044|
|Glucose (mg/dl)||−9.27 ± 5.08||−4.8 ± 3.5||0.47||N.S.|
|Insulin (µIU/ml)||−4.22 ± 1.37||−3.3 ± 1.13||0.61||N.S.|
|C-peptide (ng/ml)||−0.32 ± 0.1||−0.3 ± 0.15||0.93||N.S.|
|Cholesterol total (mg/dl)||−1.48 ± 5.1||3.56 ± 4.09||0.44||N.S.|
|Triglycerides (mg/dl)||−2.92 ± 12.5||−14 ± 8.1||0.45||N.S.|
|HDL-choesterol (mg/dl)||1.22 ± 0.92||0.166 ± 0.8||0.39||N.S.|
|LDL-choelestrol (mg/dl)||2.04 ±4.86||6.85 ± 4.5||0.19||N.S.|
|Cortisol total (µg/dl)||0.85 ± 0.68||0.18 ± 1.01||0.58||N.S.|
|CRP (mg/dl)||−0.09 ± 0.06||0.016 ± 0.02||0.1||N.S.|
Changes in BMI, waist circumference, hip circumference, and body fat percent were also similar between the groups (Table 3).
Hemodynamic changes are also presented in Table 3. Between-group differences in systolic blood pressure and heart rate were marginal. After adjusting for differences in baseline values, significantly greater systolic blood pressure decreases were observed in the BB group vs. the SB group (−9.58 vs. −2.43 mmHg, p = 0.04).
Serum Biochemical Parameters
At the end of the study, HbA1c had declined marginally more in the BB than the SB group (−0.462 vs. −0.146 %, p = 0.065). Marginally greater reductions in estimated average glucose were also observed in the BB vs. SB group (−13.2 vs. −4 mg/dl, p = 0.061). After adjusting for differences in baseline values, there were significantly greater decreases in HbA1c and estimated average glucose in the BB group compared to the SB group (p = 0.047 and p = 0.044 respectively, Table 3).
Changes in glucose, C-peptide, lipid profile, inflammatory, and hormonal parameters were similar between the groups (Table 3).
Changes in Medications Dose
A significantly greater proportion of subjects in the BB group underwent T2DM medication dose reductions compared to the SB group (31% vs. 0%; p = 0.002, Table 4).
|% Within group||65.50%||83.30%|
|% Within group||31.00%||0.00%||0.002|
|% Within group||3.40%||16.70%||0.002|
However, a significantly greater proportion of subjects in the SB group underwent T2DM medication dose increases compared to the BB group (16.7% vs. 3.4%; p = 0.002).
Changes in doses of other medications (statins, anticoagulants, antihypertensive, heartburn drugs, SSRI, HRT, and nutrition supplements) were similar in both groups.
Glucose Levels Monitored by Glucometer
Both fasting and bedtime glucose levels decreased significantly more in the BB group compared to the SB group (fast −14.95 vs. −4.91; p = 0.001, before sleep −30.7 vs. −5; p = 0.009) from the end of the first to the end of the third month of follow-up.
|BB (n = 23)||SB (n = 23)||p|
|HbA1c (%)||−0.58 ± 0.18||−0.13 ± 0.08||0.029|
|Estimated average glucose (mg/dl)||−16.6 ± 5.2||−3.43 ± 2.4||0.026|
|Glucose (mg/dl)||−11.7 ± 6.3||−8.04 ± 4||0.63|
|Insulin (μIU/ml)||−4.96 ± 1.57||−4.2 ± 1.44||0.72|
|C-peptide (ng/ml)||−0.39 ± 0.12||−0.39 ± 0.21||1|
|Cholesterol total (mg/dl)||−1.74 ± 6||−2.17 ± 4.8||0.61|
|Triglycerides (mg/dl)||−3.4 ± 14.7||−16 ± 10.3||0.49|
|HDL-Choesterol (mg/dl)||1.43 ± 1.1||0 ± 1||0.34|
|LDL-Choelestrol (mg/dl)||−2.2 ± 5.3||−5.3 ± 4.6||0.29|
|Cortisol total (μg/dl)||1.08 ± 0.86||0.36 ± 1.22||0.63|
|CRP (mg/dl)||−0.09 ± 0.06||−0.02 ± 0.03||0.11|
|IL-6 (pcg/ml)||0.72 ± 0.71||1.16 ± 0.62||0.66|
|TNF-α (pcg/ml)||−0.2 ± 0.28||−0.17 ± 0.2||0.29|
|Ghrelin (pcg/ml)||34 ± 56.45||8.65 ± 56.3||0.75|
|Adiponectin (ng/ml)||−228.2 ± 353.2||−400 ± 321||0.72|
|Leptin (pg/ml)||3,990 ±1735||3,390 ± 1562||0.79|
|Average of fasting plasma glucose (mg/dl)||−14.51 ± 13||−4.91 ± 12.3||0.037|
Changes in H-SSc over 3 months of the study are presented in Figure 3 (a-d). Higher score in the H-SSc indicates less hunger and more satiety.
After adjusting for baseline values, H-SSc values were significantly higher in the BB than SB groups after breakfast before lunch and before dinner.
Analysis of additional questions in the questionnaire found that after adjusting for baseline values, the BB group had a significantly reduced urge to eat over time, while the SB group had an increased urge to eat over time; after breakfast (p = 0.049), before lunch (p < 0.0001) and before dinner (p < 0.0001). Additionally, the preoccupation with food decreased over time in the BB group while in the SB it was increased over time; before breakfast (p = 0.09), after breakfast (p = 0.046), before lunch (p < 0.0001), and before dinner (p = 0.001) (data not shown).
Among the 46 subjects who completers, differences in body weight and BMI were not detected. At the end of the study, in the BB group vs. the SB group a significantly greater decrease in HbA1c (−0.58% vs. −0.13%, p = 0.029) and estimated average glucose (−16.6 vs. −3.43, p = 0.026, Table 5) were detected.
These results were not significant after adjusting for differences in baseline values.
The aim of the present clinical trial was to investigate the effects of breakfast size and composition on body weight, glycemic control, and its association with metabolic markers in adults with T2DM.
Weight loss seen was small in both groups. In the analysis of the dietary records we found that the patients consumed significantly higher calories than recommended.
When we analyzed the composition of lunch, dinner and snacks (data not shown) we found no difference in the composition of carbohydrates, fats, or proteins between groups.
The actual caloric intake was not very low given the age and sedentary lifestyle of the participants. It is also possible that weight loss is hampered by metabolic dysfunction such as insulin resistance.
Weight reduction did not differ by diet type, which makes metabolic differences more compelling. There is a direct connection between weight loss and glycemic control; weight-loss improved glycemic control, and improve risk factor profile even when weight regain occurred . Metabolic improvement in our trial occurred despite the fact that the mean between-group difference in weight loss was <1 kg; therefore, the improvement in glycemic control in the BB group cannot be attributed to this factor.
Specifically, we found a significant decrease in several biochemical parameters related to glycemic control: HbA1c, average glucose levels and fasting glucose levels were reduced significantly more in the BB than SB group. Additionally, T2DM medication doses were reduced in a greater proportion of the BB participants while in the SB, a greater proportion of participants had a dose increases.
Previous studies have found that eating breakfast compared to skipping breakfast, improved the area under the curve of glucose  and insulin [19, 20] responses to a test meal. In our study, both groups consumed breakfast. The differences were the size and composition of the meal. The BB diet included a protein and fat-enriched breakfast while the SB diet included a carbohydrate-rich breakfast lower in energy content. Previous studies have shown that a carbohydrate rich breakfast (vs. protein-rich breakfast) is associated with increased fasting blood glucose and insulin levels, increased area under the curve glucose and insulin response to a test meal, and increased secretion of glucagon [31, 32].
Fasting insulin, c-peptide, adiponectin, cortisol, CRP, IL-6, or TNF-α did not differ by intervention in our study. It is possible that these variables are not influenced by the dietary changes undertaken in the present clinical trial. Alternatively, it is possible that the study was underpowered to detect differences in these outcomes. It is also possible that these variables would be altered by the result of the diet—that is, weight reduction, which was modest overall and did not differ by treatment group. Perhaps the relatively modest weight reduction did not produce an accompanying reduction in markers of inflammation.
Systolic blood pressure decreased significantly more in the BB group than the SB group. In the NHANES cross-sectional study, breakfast consumption was associated with an improved cardiometabolic risk profile including blood pressure . Hypertension is an integral part of the metabolic syndrome and improvement of systolic blood pressure can improve the outcomes of morbidity and mortality of the syndrome.
Throughout the 3-month period of the trial, we evaluated hunger and satiety. As the study progressed, we found that despite similar weight reduction in both groups, hunger scores increased significantly in the SB group while satiety scores increased in the BB group. In addition, the BB group reported a reduced urge to eat and a less preoccupation with food, while the SB group had increased preoccupation with food and a greater urge to eat over time.
Previous studies have shown that compared to a carbohydrate-rich breakfast, a protein-rich breakfast decreased postprandial ghrelin concentrations [31, 32, 34] reduced gastric emptying and increased secretion of cholecystokinin . Protein-rich breakfasts resulted in a greater feeling of satiety and reduced hunger in appetite ratings [22, 32, 34]. This was reflected by decreased calorie consumption 24 h after breakfast .
During energy restriction, protein consumed at breakfast (compared to lunch or dinner) leads to greater initial and sustained feelings of fullness . A previous study demonstrated greater post prandial ghrelin suppression following a large, high carbohydrate and protein-enriched breakfast compared to a small, low carbohydrate, high protein breakfast .
It is possible that a big breakfast rich in protein causes suppression of ghrelin secretion, which is reflected in enhanced satiety ratings. Post-prandial ghrelin and leptin levels were not examined in our study, and no differences in fasting ghrelin or leptin levels were observed.
Although dietary restriction often results in initial weight loss, the majority of dieters fail to maintain their reduced weight . These diets are typified by short-term (3-6 months) success; however, many individuals cannot maintain such weight loss strategies over time . Proposed predictors of weight regain after weight loss include increased subjective appetite scores, especially increased hunger and craving [38, 39]. Therefore, it is possible that a BB diet can be used as a strategy to maintain weight loss over time by increasing satiety feeling, but this must be directly tested in further long-term studies.
Findings of the present study must be considered in light of design limitations. It is not possible to isolate a single cause of between-group differences. Did metabolic improvement occur as a result of the timing of the large meal or its composition? Could both be correct? A drawback of many diet studies is this very inability to isolate a single causal factor. Nevertheless, our findings suggest a metabolic improvement associated with the pattern of large protein-rich breakfast.
Overall, we have demonstrated improved measures of glycemic control, reduced hunger and improved satiety using a relatively simple diet intervention. Our results suggest possible dietary alternatives which may benefit overweight/obese individuals with T2DM. Further research is required to confirm and clarify the mechanisms by which this relatively simple diet approach enhances satiety, leads to better glycemic outcomes compared to a more conventional dietary approach.
The authors thank the staff of the Diabetes Unit, Wolfson Medical Center, Holon, Israel, for their dedication. We also thank medical and laboratory staff of Wolfson Medical Center, Holon, Israel. We thank the volunteers who participated in the study.
- 10Trends in breakfast consumption for children in the United States from 1965-1991. Am J Clin Nutr 1998;67:748–756., , .
- 28Corcoran K, Fischer J (eds). Hunger Satiety Scales. Measures for Clinical Practice: Instruments for Adults. A Sourcebook, 3rd edn. Free Press: New York 2000; 343–346.. In:
- 33The relationship of breakfast skipping and type of breakfast consumed with overweight/obesity, abdominal obesity, other cardiometabolic risk factors and the metabolic syndrome in young adults. The National Health and Nutrition Examination Survey (NHANES): 1999-2006. Public Health Nutr 2012;3:1–10., , , , .