Abstract
Objective
To examine whether breakfast consumption frequency (BCF) is associated with weight loss outcomes in the Look AHEAD (Action for Health in Diabetes) trial.
Methods
Data from a subset of participants (n = 3,915) from Look AHEAD, a randomized trial comparing intensive lifestyle intervention (ILI) to diabetes support and education (DSE) in adults with overweight/obesity and type 2 diabetes, were analyzed. BCF was collected by yearly questionnaire. Multivariable linear regression models were used to estimate the association between average BCF and percentage weight change over 4 years, controlling for baseline sociodemographic, anthropometric, and diabetes-related variables. In separate models, adjustment for diet (n = 915) and physical activity level (n = 837) were performed in a subset of participants.
Results
4-year average BCF was similar in DSE (n=1916) and ILI (n=1999) arms (p=0.14). Each 1-day higher average BCF was associated with an additional 0.5% weight loss in the ILI arm (p<0.0001) but not in the DSE arm (p=0.58). This association in the ILI arm remained significant after adjustment for diet (p=0.02) but not after adjustment for physical activity (p=0.36).
Conclusions
Breakfast consumption was associated with greater weight loss in the active treatment group of an intensive lifestyle intervention, which may be mediated by increased physical activity.
Keywords: breakfast consumption, lifestyle intervention, obesity, weight loss
Introduction
Obesity remains a global epidemic. The prevalence of obesity in US adults continues to rise and was 42.4% in 2017–2018.1 Obesity-related diseases such as cardiovascular disease and type 2 diabetes mellitus (T2DM) are leading causes of preventable deaths. Despite the existence of effective lifestyle interventions, the disease burden and epidemic of obesity continues to worsen.
Recent literature on dietary interventions for weight loss indicates a conceptual shift from what we eat to when we eat. Observational studies have shown that habitual breakfast skipping is associated with obesity, cardiometabolic diseases, T2DM, and increased cardiovascular and all-cause mortality,2–5 while habitual breakfast consumption is associated with lower risk of obesity and weight gain over time.6–8 However, randomized controlled trials with a short follow-up duration have found that breakfast consumption alone is not an effective weight loss intervention,9–12 with recent systematic reviews cautiously supporting this conclusion.13,14 Data on the impact of breakfast consumption as a component of a multi-faceted lifestyle intervention for weight loss remain sparse. Additionally, previous observational and interventional studies have only included healthy individuals with overweight and obesity, while the relationship between breakfast consumption and weight outcomes in persons with underlying cardiometabolic diseases such as T2DM has not been investigated. Therefore, we sought to examine the relationship between breakfast consumption frequency (BCF) and weight loss outcomes in Look AHEAD, a large multi-center randomized controlled trial that evaluated the effects of an intensive lifestyle intervention (goal weight loss ≥7%) on cardiovascular and other health outcomes in adults with overweight/obesity and T2DM. We hypothesized that higher BCF would be associated with greater weight loss over the 4-year intervention period.
Methods
Study Participants
Details regarding the design and methods of Look AHEAD were previously published (also see Supplement).15 The study was approved by the institutional review board at each participating center. Adults aged 45–75 who met the following criteria were eligible to participate: T2DM by self-report verified by review of records or medications or glycemic testing, BMI ≥25.0 kg/m2, hemoglobin A1c ≤11%, systolic blood pressure <160 mmHg, diastolic blood pressure <100 mmHg, triglyceride level <600 mg/dL, the ability to complete a maximal exercise test, and an established relationship with a primary care provider. The current study used the Look AHEAD distributed data set, a subset of 4901 participants excluding participants from the Native American sites due to consent limitations. We examined data from baseline to year 4, the most active period of the Look AHEAD intervention. Additional inclusion criteria for this analysis were: weight measurement completed at baseline and year 4 and completion of the breakfast consumption questionnaire every year from baseline to year 4. We excluded 70 participants who had bariatric surgery during the intervention period. Our final analysis included 3,915 subjects (Figure 1).
Figure 1.
Flowchart of included and excluded subjects in this post-hoc analysis.
Abbreviations: DSE, diabetes support and education; ILI, intensive lifestyle intervention
Study Interventions
Details regarding the study interventions were previously published.16,17 In brief, participants were randomly assigned to an intensive lifestyle intervention (ILI) or diabetes support and education (DSE)/control group. As detailed below, breakfast consumption could have been implied at several points during ILI.
Intensive Lifestyle Intervention16
The ILI program aimed to achieve and maintain weight loss of ≥7%. Intervention strategies involved a calorie goal of 1200–1800 kcal/day and a physical activity goal of ≥175 minutes/week of moderate-intensity activity. Meal-replacements and weight loss medications were offered to participants who had difficulty achieving or maintaining weight loss. ILI was delivered in a combination of group and individual sessions over 4 years. In the first 6 months, participants attended weekly sessions. These sessions spaced out to 3 times per month in months 7–12. During years 2 to 4, participants had 2 individual contacts per month with their lifestyle counselor.
In the ILI group, there were 3 points during the intervention where breakfast consumption could have been implied. First, during the first 6 months, specifically for at least the first 20 weeks, participants were provided the option of using meal replacements (MR) primarily for breakfast and lunch with one meal a day (usually dinner) consisting of conventional foods. A meal plan accompanied this regimen with defined calories per meal for the day. Additional meal plans for breakfast, lunch, and dinner were made available for participants who were not willing or able to tolerate liquid MR. Starting at month 7, participants were provided the option of one MR per day and could continue this for up to 4 years. Notably, during the monthly individual sessions between the lifestyle coach and the participant, if the participants’ weight progress was not meeting expected outcomes, the use of a behavioral toolbox was used which could allow participants to stay on 2 MR per day if needed. Second, during years 2 to 4, refresher groups lasting for 6 weeks were offered every 4 months and were designed to help reverse small weight gains. Structured meal plans were again provided, with a meal-planning worksheet that included a calorie goal for breakfast. Third, throughout the 4 years, lifestyle coaches discussed the importance of balanced eating and not skipping meals as part of a healthy, weight loss strategy.
The physical activity intervention mainly involved home exercises such as brisk walking or other similar aerobic activity. Participants were initially instructed to walk for at least 50 minutes/week. They were then instructed to increase physical activity to ≥125 minutes/week by week 16 and ≥175 minutes/week by week 26. Participants were also encouraged to increase lifestyle activity such as a goal of ≥10,000 steps/day, as measured by pedometers.
Diabetes Support and Education17
The DSE (control) group was invited to 3 group sessions each year focused on diet, physical activity, and social support for years 1 through 4. Attendance at these sessions was strongly encouraged but not required. In the DSE group, one group session per year was dedicated to diet. Additional details are provided in the Supplement.
Study Assessments
Baseline demographic, clinical, and laboratory data –
Study staff collected sociodemographic (age, sex, race/ethnicity, income, education level, marital status) and baseline health data, including past medical history, diabetes medication use, and health behaviors (alcohol use and smoking) via self-report through standardized questionnaires. Glycated hemoglobin was measured at baseline and through year 4.
Anthropometric measurements –
Weight was measured in duplicate following standardized procedures at baseline and annually. Height was measured at baseline and at year 4. BMI was calculated for baseline and at year 4.
Breakfast consumption frequency (BCF) –
BCF was assessed at baseline and annually by participant self-report using a standardized questionnaire that asked, “How many days out of the 7-day week do you eat breakfast?” generating a numerical value (range 0–7). We examined BCF at study baseline and then averaged BCF across the first 4 years of the intervention to generate the 4-year average BCF continuous variable. In addition, baseline and 4-year average BCF were categorized as: 1) those who ate breakfast on average 0–3 days out of 7 days, inclusively (Never/Occasionally); 2) those who ate breakfast on average 4–6 days out of 7 days, inclusively (Most Days); 3) those who ate breakfast on average >6 days out of 7 days (Every Day).
Physical activity –
Physical activity was assessed using the Paffenbarger Activity Questionnaire, which provided an estimated weekly energy expenditure from moderate intensity physical activity. Participants at 8 of 16 centers completed the questionnaire at baseline, years 1 and 4. These 8 centers were pre-determined prior to data collection, and all participants from these centers were asked to complete this questionnaire as part of routine data collection.
Dietary Assessment –
Dietary assessment was recorded in ~50% of participants at each clinical site using a self-administered 134-line item food frequency questionnaire that assessed typical consumption during the past 6 months. Specifically, this subset included the first 157 participants enrolled in each clinical site to represent a random sample.
Statistical Methods
Main outcome –
Our main outcome was the percent (%) weight change between baseline and year 4.
Main exposure variable of interest –
The exposure of interest was the calculated average BCF across 4 years of the intervention.
We compared baseline characteristics across 4-year average BCF categories using Fisher’s exact tests for categorical variables and ANOVA for continuous variables. We performed bivariate linear regression to evaluate the association between average BCF (as a continuous variable) and % weight change in the total cohort. We then evaluated the interaction between average BCF and treatment arm using linear regression with terms for BCF, treatment arm, and their multiplicative interaction and found a significant interaction (p=0.007). Given this, we conducted our subsequent analyses stratified by treatment arm. We used multivariable linear regression with covariates chosen a priori based on previous studies on predictors of weight loss in Look AHEAD18 and factors associated with BCF. 4,6,8 Supplement provides additional details of models including analyses of physical activity and dietary assessment as mediators.
A sensitivity analysis evaluated whether BCF stability during these 4 years affected our primary findings (described in Supplement). We performed an additional analysis on the use of meal replacements in the ILI arm, as MR options were used to promote weight loss and encourage breakfast (see Supplement).
Statistical significance was defined as p-value < 0.05 using 2-sided tests. We performed analyses using Stata/SE 15.5 software (StataCorp LP, College Station, Texas, USA).
Results
Baseline characteristics based on 4-year average BCF category
Table 1 shows baseline characteristics across 4-year average BCF categories. Table S1 shows baseline characteristics stratified by intervention arm. Those who ate breakfast every day tended to be older, female, and have a lower baseline BMI. Those who ate breakfast every day were more often of non-Hispanic White race/ethnicity and less often non-Hispanic Black. Participants whose employment status was homemaker were more likely to eat breakfast every day, and participants who had full or part-time jobs were less likely to eat breakfast every day. Eating breakfast never or occasionally was associated with a higher weekly alcohol intake and current smoking status. These trends were similar for both intervention arms and the total cohort.
Table 1.
Baseline Characteristics based on 4-year Average Breakfast Consumption Frequency
Characteristics | Never/Occasionally§ N=182 | Most Days N=913 | Every Day N=2820 | p-value |
---|---|---|---|---|
| ||||
Age (yr) | 57.9 (6.5) | 57.6 (6.5) | 59.7 (6.7) | <0.001 |
Female sex, n (%) | 86 (47.2) | 499 (54.6) | 1676 (59.4) | 0.001 |
BMI (kg/m2), median (IQR) | 36.5 (32.1–39.6) | 35.5 (32.0–40.0) | 34.4 (31.2–39.0) | <0.001 |
Race, n (%) | ||||
White | 111 (61.0) | 539 (59.0) | 2031 (72.0) | <0.001 |
Non-Hispanic Black | 53 (29.1) | 224 (24.5) | 374 (13.3) | |
Hispanic | 10 (5.5) | 113 (12.4) | 323 (11.4) | |
Other/mixed | 8 (4.4) | 37 (4.1) | 92 (3.3) | |
HbA1c (%) | 7.3 (1.3) | 7.3 (1.2) | 7.2 (1.1) | 0.005 |
Insulin use, n (%) | 35 (19.2) | 162 (17.7) | 496 (17.6) | 0.83 |
Dyslipidemia, n (%) | 127 (69.8) | 642 (70.3) | 2010 (71.3) | 0.79 |
Hypertension, n (%) | 146 (80.2) | 763 (83.4) | 2366 (83.9) | 0.42 |
CVD history, n (%) | 31 (17.0) | 127 (13.9) | 382 (13.5) | 0.39 |
Education level†, n (%) | ||||
< 13 yrs | 29 (15.9) | 165 (18.1) | 483 (17.1) | 0.20 |
13–16 yrs | 75 (41.2) | 358 (39.2) | 1000 (35.5) | |
> 16 yrs | 74 (40.7) | 375 (41.1) | 1274 (45.2) | |
Employment‡, n (%) | ||||
Unemployed | 18 (11.2) | 75 (9.1) | 193 (7.8) | 0.03 |
Homemaker | 22 (14.1) | 137 (16.7) | 506 (20.6) | |
Full/part-time work | 120 (75.0) | 607 (73.9) | 1746 (71.0) | |
Marital status, n (%) | ||||
Never married | 18 (9.9) | 60 (6.6) | 212 (7.5) | 0.03 |
Married | 123 (67.6) | 603 (66.1) | 1979 (70.2) | |
Divorced/ separated | 34 (18.7) | 176 (19.3) | 449 (15.9) | |
Widowed | 7 (3.8) | 73 (8.0) | 180 (6.4) | |
Family income#, n (%) | ||||
< $20,000 | 14 (8.5) | 101 (12.2) | 231 (9.1) | 0.25 |
$20,000 – $40,000 | 33 (20.1) | 164 (19.8) | 549 (21.7) | |
$40,000 – $60,000 | 31 (18.9) | 163 (19.6) | 536 (21.1) | |
$60,000 – $80,000 | 29 (17.7) | 136 (16.4) | 448 (17.7) | |
> $80,000 | 57 (34.8) | 265 (32.0) | 771 (30.4) | |
Alcohol intake (oz/wk) | 14.0 (28.6) | 9.5 (23.3) | 8.9 (26.8) | 0.04 |
Smoking∫, n (%) | ||||
Never | 89 (48.9) | 431 (47.3) | 1423 (50.5) | 0.002 |
Former | 77 (42.3) | 436 (47.9) | 1296 (46.0) | |
Current | 16 (8.8) | 44 (4.8) | 97 (3.4) |
Continuous covariates were compared using ANOVA and categorical covariates were compared using Fisher’s exact tests. All data are mean (SD) unless stated otherwise. % were calculated as those with the baseline characteristic out of total number of subjects in that BCF category.
Never group: N=6.
Missing data for education level: N=4 for Never/Occasionally, N=15 for Most Days, N=63 for Every Day.
Employment status: N=2 were students in Most Days, N=14 were students in Every Day; missing data: N=22 for Never/Occasionally, N=92 for Most Days, N=361 for Every Day.
Missing data for family income: N=18 for Never/Occasionally, N=84 for Most Days, N=285 for Every Day.
Missing data for smoking: N=2 for Most Days, N=4 for Every Day. Yr: year; IQR: inter-quartile range; HbA1c: hemoglobin A1c; CVD: cardiovascular disease; oz/wk: ounces/week
Baseline BCF patterns and BCF trends across the 4-year intervention period
At baseline, the mean ± SD for BCF was 6.12 ± 1.75 and 6.12 ± 1.73 days/week in the ILI and DSE groups, respectively. Baseline BCF category distribution was similar between intervention arms (p=0.705; Figure 2A). During the 4 years of the intervention, the mean ± SD for average BCF was 6.28 ± 1.27 and 6.21 ± 1.36 days/week in the ILI and DSE arms, respectively. The change in BCF between baseline (pre-intervention) and year 4 was similar across treatment arms, with a mean increase of 0.125 days/week in ILI and 0.11 days/week in DSE (p=0.815 using ANOVA). The proportion of participants consuming breakfast Never/Occasionally decreased from ~11% at baseline to 4% (ILI) and 5.3% (DSE) during the intervention, while the proportion of participants consuming breakfast Most Days increased to 23.3% in both arms (Figure 2). The proportion consuming breakfast Every Day did not change substantially from baseline to the intervention period. Overall, the proportions of participants in each 4-year average BCF category did not differ significantly between treatment groups (p=0.14). The majority of participants had stable BCF during the 4 years, with 61.2% in ILI arm and 59.1% in DSE arm having the same BCF across all 4 years.
Figure 2.
Distribution of BCF categories in each treatment arm at baseline and during the 4-year intervention. Proportion of participants in each average BCF category, expressed as % of total participants in each treatment arm. Proportions of these 3 BCF categories in each treatment arm were compared using Fisher’s exact test, p=0.705 for baseline BCF and p=0.14 for 4-year average BCF.
Abbreviations: BCF, breakfast consumption frequency; DSE, diabetes support and education; ILI, intensive lifestyle intervention.
Weight outcomes based on 4-year average BCF categories
Over four years, the ILI group had a mean % weight change of −4.60% (95% CI −4.93, −4.26) and the DSE group had a mean % weight change of −0.86% (95% CI −1.17, −0.55). Figure 3 shows the mean % weight change across the 3 average BCF categories stratified by treatment arm. We found that in the ILI group, the mean % weight change differed significantly across average BCF categories (p=0.0002) such that those in the Every Day category had the largest average weight loss at −5.02% (95% CI −5.40, −4.6) and those in the Never/Occasionally category had the lowest average weight loss at −2.77% (95% CI −4.4, −1.2). In contrast, mean % weight change did not differ significantly across the 3 average BCF categories (p=0.19) in DSE.
Figure 3.
Mean percent weight change (±95% CI) across 3 categories of average BCF, stratified by intervention arm.
For DSE group, mean % weight change for “Never/Occasionally”, “Most Days” and “Every Day” BCF category were −0.045% (95% CI −1.3%, 1.4%), −1.26% (95% CI −1.9%, −0.6%) and −0.80% (95% CI −1.2%, −0.43%) respectively. For ILI group, mean % weight change for “Never/Occasionally”, “Most Days” and “Every Day” BCF category were −2.77% (95% CI −4.4%, −1.2%), −3.59% (95% CI −4.2%, −2.9%), and −5.02% (95% CI −5.4%, −4.6%) respectively. *Using ANOVA, p<0.05.
Abbreviations: DSE, diabetes support and education; ILI, intensive lifestyle intervention; BCF, breakfast consumption frequency.
Linear regression analysis revealed a significant association between average BCF and % weight change only in the ILI arm (β-coefficient −0.54), but not in the DSE arm (p for interaction between arm and BCF = 0.007; Table 2). This indicates that each one-day higher average BCF was associated with an additional 0.54% weight loss in the ILI arm. The association between BCF and % weight change in the ILI arm remained significant and largely unchanged after adjusting for baseline sociodemographic, anthropometric, and diabetes-related variables (models A, B and C). In model D, which included self-reported physical activity level at year 4 in the subset of participants with available data (n=837), the effect of average BCF on % weight change was attenuated and no longer significant (β-coefficient −0.18, p=0.36). In model E, which included self-reported dietary assessment at year 4 in the subset of participants with available data (n=915), the effect of average BCF on % weight change was attenuated but remained significant (β-coefficient −0.46, p=0.02).
Table 2.
Regression Models for 4-year Average Breakfast Consumption Frequency and 4-Year % Weight Change
Univariate | Model A | Model B | Model C | Model D | Model E | ||
---|---|---|---|---|---|---|---|
| |||||||
Treatment Arm | DSE | ILI | ILI | ILI | ILI | ILI | ILI |
| |||||||
N | 1916 | 1999 | 1999 | 1999 | 1571 | 837 | 915 |
β coefficient | −0.065 | −0.54 | −0.51 | −0.49 | −0.49 | −0.18 | −0.46 |
95% CI | −0.29, 0.16 | −0.80, −0.28 | −0.77, −0.24 | −0.75, −0.23 | −0.78, −0.19 | −0.58, 0.21 | −0.85, −0.08 |
p-value | 0.58 | <0.001 | <0.001 | <0.001 | 0.001 | 0.36 | 0.02 |
β coefficient is for average breakfast consumption frequency in days per week (range 0 to 7)
Model A: average BCF, age, sex, race, BMI
Model B: model A + baseline HbA1C, any insulin use
Model C: model B + education level, employment status, family income, marital status, smoking, alcohol
Model D: model C + Paffenbarger physical activity at year 4 (subset of participants with available data)
Model E: model C + dietary assessment at year 4 (subset of participants with available data)
DSE= Diabetes Support & Education. ILI=Intensive Lifestyle Intervention
Dietary assessment and physical activity based on 4-year BCF categories
The mean estimated daily caloric intake at year 4 was similar across average BCF categories in both treatment arms (Table 3). However, those who consumed breakfast every day reported lowest percentage of fat intake and highest percentage of carbohydrate intake in both treatment arms. In the ILI group, participants who consumed breakfast every day expended on average 351 kcal more per week than those who consumed breakfast never/occasionally (Table 2, Figure 4B). In the DSE group, physical activity at year 4 was similar across categories of BCF. Given these findings, we explored whether there was a relationship between physical activity and breakfast consumption at baseline. Indeed, Figure 4A demonstrated that at baseline, there was a positive relationship between BCF and physical activity level (p=0.03) such that those who ate breakfast every day at baseline had the highest level of physical activity.
Table 3.
Physical Activity and Dietary Assessment at Year 4 among Categories of 4-year Average Breakfast Consumption Frequency
Diabetes Support & Education | Intensive Lifestyle Intervention | |||||||
---|---|---|---|---|---|---|---|---|
| ||||||||
Characteristics | Never/Occasionally | Most Days | Every Day | p-value | Never/Occasionally | Most Days | Every Day | p-value |
| ||||||||
Physical Activity ‡ | N=63 | N=266 | N=697 | N=47 | N=252 | N=753 | ||
kcal/week | 957.3 (1556) | 962.2 (1218) | 1016.5 (1225) | 0.80 | 1005.6 (1174) | 1086.1 (1346) | 1356.5 (1562) | 0.02 |
| ||||||||
Estimated Daily Caloric Intake † | N=55 | N=253 | N=762 | N=46 | N=284 | N=823 | ||
kcal | 1612 (806) | 1655 (762) | 1590 (738) | 0.48 | 1667 (684) | 1645 (731) | 1608 (659) | 0.65 |
% Fat | 41.2 (6.7) | 40.8 (7.0) | 39.5 (7.4) | 0.02 | 37.8 (7.5) | 37.8 (7.2) | 35.9 (7.0) | 0.002 |
% Protein | 17.0 (2.9) | 17.7 (3.1) | 17.9 (3.2) | 0.10 | 17.1 (2.8) | 17.7 (2.8) | 17.8 (2.9) | 0.25 |
% Carbohydrate | 42.4 (8.0) | 42.3 (8.1) | 43.7 (8.1) | 0.04 | 45.1 (9.4) | 46.2 (8.6) | 48.5 (8.4) | <0.001 |
All data are mean (standard deviation) unless stated otherwise. Continuous covariates were compared using ANOVA and categorical covariates were compared using Fisher’s exact tests.
Paffenbarger physical activity questionnaire at year-4 obtained from total of 2078 subjects (n=1026 in DSE; n=1052 in ILI).
Dietary assessment at year-4 obtained from 2223 subjects (n=1070 in DSE; n=1153 in ILI).
Figure 4.
Physical activity level at baseline and at year 4 (±95% CI) across 3 BCF categories. A. At baseline, mean physical activity for “Never/Occasionally” (n=221), “Most Days” (n=339) and “Every Day” (n=1356) BCF category were 734.9 kcal/wk (95% CI 588, 882), 828.6 kcal/wk (95% CI 723, 934), and 939.5 kcal/wk (95% CI 873, 1005) respectively. B. At year 4, for DSE group, mean physical activity for “Never/Occasionally” (n=63), “Most Days” (n=266) and “Every Day” (n=697) 4-year average BCF category were 957.3 kcal/wk (95% CI 565, 1349), 962.2 kcal/wk (95% CI 815, 1109), and 1016.5 kcal/wk (95% CI 925, 1108) respectively. For ILI group, mean physical activity for “Never/Occasionally” (n=47), “Most Days” (n=252) and “Every Day” (n=753) BCF category were 1005.6 kcal/wk (95% CI 661, 1350), 1086.1 kcal/wk (95% CI 919, 1253), and 1356.5 kcal/wk (95% CI 1245, 1468) respectively. *Using ANOVA, p<0.05
Abbreviations: DSE, diabetes support and education; ILI, intensive lifestyle intervention; BCF, breakfast consumption frequency.
Additional Analyses
The sensitivity analysis including BCF stability revealed that the majority of participants in both arms were in the Stable category (Table S2). Regression analyses demonstrated that accounting for BCF stability did not change our main findings that an increase in average BCF was associated with more weight loss in ILI and that this effect was attenuated after adjusting for physical activity and diet (Table S3).
We evaluated the use of MR and found that nearly all participants undergoing ILI received MR, with 1986 out of 1999 participants receiving MR during 0–6 months and 1980 out of 1999 participants receiving MR during 7–12 months. Higher BCF was associated with higher MR usage at both time points (Table S4). Bivariate linear regression showed that MR usage in ILI was associated with both greater BCF and weight loss (data not shown). Adjusting for MR usage modestly attenuated the association between BCF and weight loss, indicating that MR is likely a partial mediator of the effect of BCF on weight loss (data not shown), which is an expected finding given that the MRs were primarily used to replace breakfast and lunch.
Discussion
We found that breakfast consumption was associated with greater weight loss in older adults with overweight/obesity and T2DM who were randomized to an intensive lifestyle intervention, independent of baseline sociodemographic, anthropometric, and diabetes-related factors. Specifically, each one-day higher average weekly breakfast consumption was associated with an additional 0.5% weight loss. This effect appeared to be mediated in part by increased physical activity. In contrast, we found no association between breakfast consumption and weight outcomes or physical activity levels in the DSE group. These findings suggest that breakfast consumption may enhance the weight loss efficacy of an intensive lifestyle program.
The results of our study are consistent with prior observational analyses, which have reported an inverse relationship between habitual breakfast consumption and weight over time.6–8,19,20 Additionally, two observational studies evaluated breakfast consumption and weight in the context of weight loss maintenance.21,22 The National Weight Control Registry Study and the MedWeight Study, which was another registry of weight loss maintainers and regainers, both found that daily breakfast consumption was associated with better weight loss maintenance,21 but only in men in the MedWeight Study.22 In contrast, interventional trials in adults with overweight/obesity have not found a positive effect of breakfast consumption on weight loss.9–11,23 Notably, these studies recruited young and middle-aged adults with mean ages in the 30s to 40s, who had overweight/obesity but were otherwise healthy. Additionally, existing interventional breakfast trials were limited by the lack of controlled feeding9,10,23 with either no reporting of caloric intake10,11 or non-isocaloric intake between treatment arms,23 small sample size, 9–11,23 lack of reporting9–11 or controlling for physical activity,23 and heterogeneity in the timing, composition, and size of breakfast.9–11,23 Given these limitations, the effect of breakfast consumption on weight in the absence of other dietary or physical activity changes remains unclear.
The role of breakfast consumption as a component of a multi-faceted behavioral weight loss intervention on weight outcomes and the impact of breakfast consumption in specific populations such as older adults and those with T2DM are clinically relevant knowledge gaps in the literature. We identified only one prior study that examined the effect of breakfast consumption as part of a behavioral weight loss intervention, on body weight. In this secondary analysis of a randomized clinical trial that compared a statewide wellness program to a combination of the wellness program with a Diabetes Prevention Program-based lifestyle intervention in 211 adults, increased breakfast eating was associated with greater weight loss, regardless of the randomized treatment.24 Our study builds on this evidence with a large sample size of nearly 4000 participants and the unique study cohort of middle-aged and older adults with overweight/obesity and T2DM. Due to the extensive characterization of our study population, we were able to assess the effect of breakfast consumption while adjusting for key anthropometric, sociodemographic, and diabetes-related variables such as insulin use. We were also able to further explore the potential mechanisms related to the effect of BCF on weight by adjusting for diet and physical activity levels, which were 2 key components of the Look AHEAD intensive lifestyle intervention.
While breakfast consumption alone may not be adequate to achieve successful weight loss, it may contribute to healthier eating and physical activity patterns which are critical for weight reduction. Our findings suggest that the beneficial effect of breakfast consumption on weight loss may be mediated by increased physical activity. While the ILI program employed both caloric restriction and increased physical activity, self-reported physical activity levels had a more pronounced attenuating effect on the relationship between BCF and weight outcomes in our models. Additionally, physical activity levels differed significantly among the 3 different BCF groups in the ILI group, while the estimated daily caloric intake levels did not. To further support the relationship between BCF and physical activity, our cross-sectional analysis of baseline data revealed an association between higher BCF and increased physical activity at baseline. Similar to our findings, several cross-sectional and prospective observational studies in pediatric populations have demonstrated a positive relationship between breakfast frequency and physical activity.25–27 In adult populations, the National Weight Control Registry study found that those who routinely ate breakfast reported more physical activity than non-breakfast eaters.21 Furthermore, a 6-week intervention of breakfast vs extended overnight fast in adults with obesity and in lean adults resulted in greater physical activity during the morning in the breakfast group.23,28
The exact mechanisms underlying the potential benefit of breakfast consumption on weight loss are not known. One potential explanation involves the effect of breakfast on metabolic and hormonal responses to foods consumed later in the day.29,30 A study in men with normal weight showed that consuming breakfast compared to skipping the same meal increased postprandial satiety hormones and lowered glucose and insulin levels in response to a subsequent meal.29 In contrast, another study did not find any beneficial adaptations related to energy expenditure or appetite-regulating hormones with breakfast consumption in adults with obesity.31 Another possible explanation is that the redistribution of calories towards the morning with consequently fewer calories consumed in the evening helps to synchronize food intake with circadian regulation of metabolism, as this allows the majority of food intake to occur during our biologically active hours, as dictated by our endogenous circadian rhythm. A study in 93 women with overweight/obesity and metabolic syndrome who were randomized to either high-calorie breakfast or high-calorie dinner with the same total daily energy intake and identical macronutrient composition for 12 weeks found that a high-calorie breakfast induced a 2.5-fold greater weight loss than high-calorie dinner.32 Additionally, the high-calorie breakfast group exhibited greater ghrelin suppression and satiety scores throughout the day compared to the high-calorie dinner group. Similarly, a controlled feeding study demonstrated that eating on a daytime schedule (8am-7pm) compared to a delayed schedule (12pm-11pm) lowered body weight in healthy non-obese volunteers over 8 weeks.33 In contrast, Hutchison et al found that 7 days of early (8am-5pm) vs delayed (12pm-9pm) time-restricted feeding in men at risk for T2DM produced no difference on weight.34 Hoddy et al found that in the context of an alternate day fasting intervention, consuming the fast day meal at lunch vs dinner led to similar body weight reductions.35
Our study has several limitations. First, causality regarding breakfast consumption and weight loss cannot be inferred in this observational analysis due to the possibility of unmeasured, confounding health behaviors. Notably, we did not have data on sleep timing and duration, which can impact breakfast intake36 and physical activity levels.37 Second, participants provided BCF by a questionnaire which did not specify a standardized definition of breakfast and participants’ interpretation of breakfast may vary considerably. Additionally, the questionnaire did not specify the macronutrient or caloric contents of breakfast. As previous NHANES data have reported that breakfast types have differential effects on body weight,5,38 this may be an unmeasured confounder. Future studies may consider more in-depth characterization of breakfast composition and to compare the effects of breakfast with different compositions within the same BCF group. However, we did have dietary assessment on a subset of participants that showed lower fat intake and increased carbohydrate intake for those who ate breakfast every day in both treatment arms. Third, self-reported questionnaires, which we used to assess physical activity and diet, contain known limitations including the tendency for over-estimating physical activity39,40 and under-estimating dietary intake.41 Fourth, physical activity and dietary assessment were collected in a subset of the Look AHEAD sample so that our models including these variables had a relatively small sample size; while this could bias our results toward finding false associations, physical activity and diet were measured in random samples of participants, rather than in a self-selected sample, which should lower the risk of type 1 error. Lastly, the effect of BCF on weight loss was only seen in participants in ILI so generalizability of our findings to free-living individuals, including those attempting weight loss in the absence of a structured intervention, requires further study.
Conclusion
We found that in the Look AHEAD trial, one of the largest behavioral weight loss trials in older adults with overweight/obesity and T2DM, eating breakfast appeared to facilitate greater weight loss during an intensive lifestyle intervention by inducing increased physical activity. To optimize weight loss interventions, the relationship between breakfast consumption and other weight loss behaviors should be further explored. Long-term, large-scale randomized clinical trials with controlled feeding and objective assessments of physical activity are needed to confirm our findings and the relationship between breakfast consumption and physical activity in weight loss effectiveness in adults with obesity. Future studies should also explore the role of breakfast consumption frequency and sustained weight loss after the conclusion of the Look AHEAD intervention period. Our study provides additional insight into behaviors that may enhance weight loss efficacy, which have important practical implications given the challenges of achieving and sustaining meaningful weight loss in individuals with overweight and obesity.
Supplementary Material
Study Importance Questions:
1. What is already known about this subject?
Observational studies have found that breakfast skipping is associated with obesity and cardiometabolic diseases.
Limited interventional studies have not found a positive impact of breakfast consumption on weight loss.
2. What are the new findings in your manuscript?
In one of the largest behavioral weight loss trials in adults with overweight/obesity and type 2 diabetes, breakfast consumption was associated with greater weight loss in the intensive lifestyle intervention group only, independent of demographic and clinical factors.
The effect of breakfast consumption on weight loss may be mediated by increased physical activity.
3. How might your results change the direction of research or the focus of clinical practice?
Breakfast consumption may be an important factor for improving the efficacy of behavioral weight loss interventions, and its relationship with physical activity should be further studied.
For clinicians who treat individuals with overweight/obesity, greater emphasis could be placed on behavioral factors that are related to dietary patterns such as breakfast consumption.
Acknowledgement
This study used the Look AHEAD public access dataset. Our use of the dataset was in accordance with the terms of the Look AHEAD Study Data Use Agreement.
DISCLOSURE:
JMC and NMM report grant support from the National Institutes of Health under award number 5U01DK057149. NMM reports grant support from the American Heart Association under grant number 17SFRN33590069.
FUNDING:
National Institutes of Health 5T32HL110952 (D. Duan)
EOS Foundation gift (D. Duan)
Johns Hopkins KL2 Clinical Research Scholars Program KL2TR003099 (S. Pilla)
National Institutes of Health 5U01DK057178 (B. Laferrère)
American Heart Association 17SFRN33590069 (N. Maruthur)
National Institutes of Health 5U01DK057149 (J. Clark, N. Maruthur)
Footnotes
References
- 1.Hales CM, Carroll MD, Fryar CD, Ogden CL. Prevalence of Obesity and Severe Obesity Among Adults: United States, 2017–2018.; 2020. Accessed July 4, 2020. https://www.cdc.gov/nchs/products/databriefs/db360.htm. [PubMed]
- 2.Mekary RA, Giovannucci E, Willett WC, van Dam RM, Hu FB. Eating patterns and type 2 diabetes risk in men: Breakfast omission, eating frequency, and snacking. Am J Clin Nutr. 2012;95(5):1182–1189. doi: 10.3945/ajcn.111.028209 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Shafiee G, Kelishadi R, Qorbani M, et al. Association of breakfast intake with cardiometabolic risk factors. J Pediatr. 2013;89(6):575–582. doi: 10.1016/j.jped.2013.03.020 [DOI] [PubMed] [Google Scholar]
- 4.Rong S, Snetselaar LG, Xu G, et al. Association of Skipping Breakfast With Cardiovascular and All-Cause Mortality. J Am Coll Cardiol. 2019;73(16):2025–2032. doi: 10.1016/j.jacc.2019.01.065 [DOI] [PubMed] [Google Scholar]
- 5.Cho S, Dietrich M, Brown CJP, Clark CA, Block G. The Effect of Breakfast Type on Total Daily Energy Intake and Body Mass Index: Results from the Third National Health and Nutrition Examination Survey (NHANES III). J Am Coll Nutr. 2003;22(4):296–302. doi: 10.1080/07315724.2003.10719307 [DOI] [PubMed] [Google Scholar]
- 6.Odegaard AO, Jacobs DR, Steffen LM, van Horn L, Ludwig DS, Pereira MA. Breakfast frequency and development of metabolic risk. Diabetes Care. 2013;36(10):3100–3106. doi: 10.2337/dc13-0316 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Purslow LR, Sandhu MS, Forouhi N, et al. Energy Intake at Breakfast and Weight Change: Prospective Study of 6,764 Middle-Aged Men and Women. Am J Epidemiol. 2008;167(2):188–192. [DOI] [PubMed] [Google Scholar]
- 8.van der Heijden A, Hu FB, Rimm EB, van Dam RM. A Prospective Study of Breakfast Consumption and Weight Gain among U.S. Men. Obesity. 2007;15(10):2463–2469. doi: 10.1038/oby.2007.292 [DOI] [PubMed] [Google Scholar]
- 9.Schlundt DG, Hill JO, Sbrocco T, Pope-Cordle J, Sharp T. The Role of Breakfast in the Treatment of Obesity: A Randomized Clinical Trial. Am J Clin Nutr. 1992;55:645–651. [DOI] [PubMed] [Google Scholar]
- 10.Dhurandhar EJ, Dawson J, Alcorn A, et al. The effectiveness of Breakfast recommendations on weight loss: A randomized controlled trial. Am J Clin Nutr. 2014;100(2):507–513. doi: 10.3945/ajcn.114.089573 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Geliebter A, Astbury NM, Aviram-Friedman R, Yahav E, Hashim S. Skipping breakfast leads to weight loss but also elevated cholesterol compared with consuming daily breakfasts of oat porridge or frosted cornflakes in overweight individuals: a randomised controlled trial. J Nutr Sci. 2014;3:1–7. doi: 10.1017/jns.2014.51 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.LeCheminant GM, LeCheminant JD, Tucker LA, Bailey BW. A randomized controlled trial to study the effects of breakfast on energy intake, physical activity, and body fat in women who are nonhabitual breakfast eaters. Appetite. 2017;112:44–51. doi: 10.1016/j.appet.2016.12.041 [DOI] [PubMed] [Google Scholar]
- 13.Bonnet JP, Cardel MI, Cellini J, Hu FB, Guasch-Ferré M. Breakfast Skipping, Body Composition, and Cardiometabolic Risk: A Systematic Review and Meta-Analysis of Randomized Trials. Obesity. 2020;28(6):1098–1109. doi: 10.1002/oby.22791 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Sievert K, Hussain SM, Page MJ, et al. Effect of breakfast on weight and energy intake: Systematic review and meta-analysis of randomised controlled trials. BMJ. 2019;364. doi: 10.1136/bmj.l42 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Ryan D, Espeland M, Foster G, et al. Look AHEAD (Action for Health in Diabetes): Design and methods for a clinical trial of weight loss for the prevention of cardiovascular disease in type 2 diabetes. Control Clin Trials. 2003;24(5):610–628. doi: 10.1016/S0197-2456(03)00064-3 [DOI] [PubMed] [Google Scholar]
- 16.Look AHEAD Research Group, Wadden T, West D, et al. The Look AHEAD Study: A Description of the Lifestyle Intervention and the Evidence Supporting It. Obesity (Silver Spring). 2006;14(5):737–752. doi: 10.1038/oby.2006.84 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Wesche-Thobaben JA. The development and description of the comparison group in the Look AHEAD trial. Clinical Trials. 2011;8(3):320–329. doi: 10.1177/1740774511405858 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Wadden TA, West DS, Neiberg RH, et al. One-year weight losses in the look AHEAD study: Factors associated with success. Obesity. 2009;17(4):713–722. doi: 10.1038/oby.2008.637 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Otaki N, Obayashi K, Saeki K, Kitagawa M, Tone N, Kurumatani N. Relationship between breakfast skipping and obesity among elderly: cross-sectional analysis of the Heijo-Kyo Study. J Nutr Health Aging. 2017;21(5):501–504. [DOI] [PubMed] [Google Scholar]
- 20.Monzani A, Ricotti R, Caputo M, et al. A systematic review of the association of skipping breakfast with weight and cardiometabolic risk factors in children and adolescents. What should we better investigate in the future? Nutrients. 2019;11(2). doi: 10.3390/nu11020387 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Wyatt HR, Grunwald GK, Mosca CL, Klem ML, Wing RR, Hill JO. Long-Term Weight Loss and Breakfast in Subjects in the National Weight Control Registry. Obesity Research. 2002;10:78–82. doi: 10.1038/oby.2002.13 [DOI] [PubMed] [Google Scholar]
- 22.Brikou D, Zannidi D, Karfopoulou E, Anastasiou CA, Yannakoulia M. Breakfast consumption and weight-loss maintenance: results from the MedWeight study. British Journal of Nutrition. 2016;115:2246–2251. [DOI] [PubMed] [Google Scholar]
- 23.Chowdhury EA, Richardson JD, Holman GD, Tsintzas K, Thompson D, Betts JA. The causal role of breakfast in energy balance and health: a randomized controlled trial in obese adults. The American Journal of Clinical Nutrition. 2016;103(3):747–756. doi: 10.3945/ajcn.115.122044 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Megson M, Wing R, Leahey TM. Effects of breakfast eating and eating frequency on body mass index and weight loss outcomes in adults enrolled in an obesity treatment program. Journal of Behavioral Medicine. 2017;40(4):595–601. doi: 10.1007/s10865-017-9828-0 [DOI] [PubMed] [Google Scholar]
- 25.Timlin MT, Pereira MA, Story M, Neumark-Sztainer D. Breakfast eating and weight change in a 5-year prospective analysis of adolescents: Project EAT (eating among teens). Pediatrics. 2008;121(3). doi: 10.1542/peds.2007-1035 [DOI] [PubMed] [Google Scholar]
- 26.Schembre SM, Kun Wen C, Davis JN, et al. Eating breakfast more frequently is cross-sectionally associated with greater physical activity and lower levels of adiposity in overweight Latina and African American girls. Am J Clin Nutr. 2013;98:275–281. doi: 10.3945/ajcn.112.045849 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Zakrzewski-Fruer JK, Gillison FB, Katzmarzyk PT, et al. Association between breakfast frequency and physical activity and sedentary time: A cross-sectional study in children from 12 countries. BMC Public Health. 2019;19(1):1–11. doi: 10.1186/s12889-019-6542-6 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Betts JA, Richardson JD, Chowdhury EA, Holman GD, Tsintzas K, Thompson D. The causal role of breakfast in energy balance and health: A randomized controlled trial in lean adults. American Journal of Clinical Nutrition. 2014;100(2):539–547. doi: 10.3945/ajcn.114.083402 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Astbury NM, Taylor MA, Macdonald IA. Breakfast Consumption Affects Appetite, Energy Intake, and the Metabolic and Endocrine Responses to Foods Consumed Later in the Day in Male Habitual Breakfast Eaters. The Journal of Nutrition. 2011;141(7):1381–1389. doi: 10.3945/jn.110.128645 [DOI] [PubMed] [Google Scholar]
- 30.Leidy HJ, Racki EM. The addition of a protein-rich breakfast and its effects on acute appetite control and food intake in breakfast-skipping adolescents. International Journal of Obesity. 2010;34(7):1125–1133. doi: 10.1038/ijo.2010.3 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Chowdhury EA, Richardson JD, Gonzalez JT, Tsintzas K, Thompson D, Betts JA. Six Weeks of Morning Fasting Causes Little Adaptation of Metabolic or Appetite Responses to Feeding in Adults with Obesity. Obesity. 2019;27(5):813–821. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32.Jakubowicz D, Barnea M, Wainstein J, Froy O. High caloric intake at breakfast vs. dinner differentially influences weight loss of overweight and obese women. Obesity. 2013;21(12):2504–2512. [DOI] [PubMed] [Google Scholar]
- 33.Allison KC, Hopkins CM, Ruggieri M, Spaeth AM, Ahima RS, Zhang Z, Taylor DM, Goel N. Prolonged, Controlled Daytime versus Delayed Eating Impacts Weight and Metabolism. Curr Biol. 2021;31(3):650–657.e3. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34.Hutchison AT, Regmi P, Manoogian ENC, Fleischer JG, Wittert GA, Panda S, Heilbronn LK. Time-Restricted Feeding Improves Glucose Tolerance in Men at Risk for Type 2 Diabetes: A Randomized Crossover Trial. Obesity (Silver Spring). 2019;27(5):724–732. [DOI] [PubMed] [Google Scholar]
- 35.Hoddy KK, Kroeger CM, Trepanowski JF, Barnosky A, Bhutani S, Varady KA. Meal timing during alternate day fasting: Impact on body weight and cardiovascular disease risk in obese adults. Obesity (Silver Spring). 2014;22(12):2524–31. [DOI] [PubMed] [Google Scholar]
- 36.Al-Hazzaa HM, Alhussain MH, Alhowikan AM, Obeid OA. Insufficient sleep duration and its association with breakfast intake, overweight/obesity, socio-demographics and selected lifestyle behaviors among Saudi school children. Nature and Science of Sleep. 2019;11:253–263. doi: 10.2147/NSS.S225883 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37.Thomas EA, Zaman A, Cornier MA, Catenacci VA, Tussey EJ, Grau L, Arbet J, Broussard JL, Rynders CA. Later Meal and Sleep Timing Predicts Higher Percent Body Fat. Nutrients. 2020;13(1):73. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 38.Deshmukh-Taskar P, Nicklas TA, Radcliffe JD, O’neil CE, Liu Y. The 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): Public Health Nutrition. 2012;16(11):2073–2082. doi: 10.1017/S1368980012004296 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 39.Tucker JM, Welk GJ, Beyler NK. Physical activity in U.S. adults: Compliance with the physical activity guidelines for Americans. American Journal of Preventive Medicine. 2011;40(4):454–461. doi: 10.1016/j.amepre.2010.12.016 [DOI] [PubMed] [Google Scholar]
- 40.Dyrstad SM, Hansen BH, Holme IM, Anderssen SA. Comparison of self-reported versus accelerometer-measured physical activity. Medicine and Science in Sports and Exercise. 2014;46(1):99–106. doi: 10.1249/MSS.0b013e3182a0595f [DOI] [PubMed] [Google Scholar]
- 41.Schoeller DA. Limitations in the assessment of dietary energy intake by self-report. Metabolism. 1995;44(SUPPL. 2):18–22. doi: 10.1016/0026-0495(95)90204-X [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.