| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Original Research |
Kellogg-USA, Battle Creek, Michigan (S.C., C.A.C.), Berkeley, California
School of Public Health, University of California, Berkeley, California (M.D., C.J.P.B., G.B.)
Address reprint requests to: Gladys Block, PhD, University of California, Berkeley, School of Public Health, 140 Warren Hall, Berkeley, CA 94720-7360. E-mail: gblock{at}uclink.berkeley.edu
 |
ABSTRACT |
|---|
 TOPFOOTNOTES ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSCONCLUSIONACKNOWLEDGMENTSREFERENCES |
|---|
Methods: Data from the Third National Health and Nutrition Examination Survey (NHANES III), a large, population-based study conducted in the United States from 1988 to 1994, were analyzed for breakfast type, total daily energy intake, and BMI. The analyzed breakfast categories were "Skippers," "Meat/eggs," "Ready-to-eat cereal (RTEC)," "Cooked cereal," "Breads," "Quick Breads," "Fruits/vegetables," "Dairy," "Fats/sweets," and "Beverages." Analysis of covariance was used to estimate adjusted mean body mass index (BMI) and energy intake (kcal) as dependent variables. Covariates included age, gender, race, smoking, alcohol intake, physical activity and poverty index ratio.
Results: Subjects who ate RTEC, Cooked cereal, or Quick Breads for breakfast had significantly lower BMI compared to Skippers and Meat and Egg eaters (p 
0.01). Breakfast skippers and fruit/vegetable eaters had the lowest daily energy intake. The Meat and Eggs eaters had the highest daily energy intake and one of the highest BMIs.
Conclusions: This analysis provides evidence that skipping breakfast is not an effective way to manage weight. Eating cereal (ready-to-eat or cooked cereal) or quick breads for breakfast is associated with significantly lower body mass index compared to skipping breakfast or eating meats and/or eggs for breakfast.
Key words: breakfast type, body mass index, energy intake, feeding patterns
|
INTRODUCTION |
|---|
|
TOPFOOTNOTESABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSCONCLUSIONACKNOWLEDGMENTSREFERENCES |
|---|
30kg/m2) [6] increased almost twofold among U.S. adults 20 to 74 years of age, from 12.8 percent to 22.5 percent [3]. During the same time period, more Americans began skipping breakfast despite cultural beliefs regarding the importance of breakfast eating and scientific knowledge about the meal’s contribution to sufficient micronutrient intakes and a healthful macronutrient profile [10,11]. Between 1965 and 1991, the proportion of adults skipping breakfast increased from 14 to 25 percent [12]. Reasons individuals give for skipping breakfast commonly include lack of time for the preparation and consumption of food and concerns about excess body weight [10,11,13].
An inverse relationship has been found between BMI and breakfast consumption [14,15]. Breakfast eaters tend to have lower body mass index than breakfast skippers [16], and obese individuals are more likely to skip breakfast or consume less energy at breakfast [15,17–20]. Studies have shown that the energy intake of normal and underweight persons is more evenly distributed throughout the day than that of the obese [18]. Night Eating Syndrome, characterized in part by habitual breakfast skipping, increases in prevalence with increasing severity of obesity and was associated with less weight loss in an outpatient program [21]. Moderately obese women lose more weight when they consume 70% of their daily energy intake before noon instead of in the afternoon and evening [22]. Additionally, lean individuals lose weight when consuming a 2000-calorie meal at breakfast but tend to gain weight if the meal is eaten at dinner [23].
Several hypotheses have been proposed to explain why breakfast skipping is associated with a decreased ability to lose weight. Breakfast skipping can lead to overeating later in the day—for example, having one big meal in the evening [24,25]. In contrast, eating breakfast is associated with increased eating frequency. Increased eating frequency may in turn promote less efficient energy utilization by increasing dietary induced thermogenesis, leading to lower BMI [26].
In the study presented here, we hypothesized that not only is breakfast consumption itself associated with BMI, but also that the type of food eaten at breakfast affects BMI. Therefore, we investigated the relationship between breakfast type, body mass index and energy intake using data from the Third National Health and Nutrition Examination Survey (NHANES III), a large, population-based study conducted in the United States from 1988 to 1994. Because of the large body of evidence showing that both BMI and breakfast habits are highly associated with age, gender, race, socioeconomic status (SES), smoking and other lifestyle factors [11,12,14,27–30], we investigated the effects of breakfast type independently of these factors.
|
MATERIALS AND METHODS |
|---|
|
TOPFOOTNOTESABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSCONCLUSIONACKNOWLEDGMENTSREFERENCES |
|---|
Physical examinations, which included measurements of standing height and weight used to calculate body mass index, were conducted year-round in mobile examination centers. Food intake information was obtained via a 24-hour dietary recall. Interviews were conducted in English (86%), Spanish or English-Spanish (13%) or another language (1%). Details of the procedure utilized for interviews have been published previously [33,34].
Study Sample
Subjects who had unreliable dietary recall records, as identified by NHANES III, were excluded from our analysis. Pregnant women were excluded on the basis that a BMI over 25 kg/m2 for a pregnant woman might not indicate overweight or obesity. Individuals younger than 18 years of age were also excluded because BMIs of adolescents are not comparable to BMIs of adults due to rapid changes in height during adolescence. Additionally, subjects whose self-reported ethnicity was not white, African American or Hispanic were excluded, because the number of persons of other races in NHANES III was too small to permit valid inferences about this subgroup. After these exclusions, the remaining sample size was 16,452 individuals.
Design of Breakfast Categories
Breakfast, in NHANES III, was defined as any food or beverage consumed in a meal occasion named by the respondent as breakfast (English) or desayuno (Spanish). Subjects that took no food or beverage at breakfast, excluding water, were categorized as breakfast skippers.
Foods consumed at breakfast were initially categorized based on the USDA food-coding scheme in order to categorize all possible foods consumed. Some further modifications were then applied to define commonly-understood breakfast types. We divided the grain products into several groups of foods: Ready-to-eat cereal, Cooked cereal, and two breads groups. We subdivided the breads into "Breads" and "Quick breads" in order to differentiate between a breakfast consisting mainly of toast from one consisting mainly of pastries, for example, which might have different implications for BMI. The ten resulting breakfast categories were "Skippers," "Dairy," "Meat and Eggs," "Fruits and Vegetables," "Ready-to-eat cereal," "Cooked cereal," "Breads," "Quick Breads," "Fats and Sweets" and "Beverages."
An individual was assigned to a breakfast category if either 1) all food and beverages consumed at breakfast were from that breakfast category or 2) food and beverages from that category contributed more calories to the meal than those of any other breakfast category. For example, a person who had a breakfast of orange juice (providing 200 kcal) and toast (100 kcal) would be assigned to the "Fruits and Vegetables" category. Additives to a food in a breakfast category were included in the calculation of calories from that category. For example, milk used on cereal was counted as part of the cereal; cream or sugar added to coffee was counted as part of the coffee. As a result, persons assigned to the "Dairy" category usually had obtained most of their breakfast calories from milk alone (or cheese). A person who had coffee (3 kcal) with cream (100 kcal) (and nothing else) would be assigned to the "Beverages" group, not the "Dairy" group. Thus, this analysis reflects the effect of breakfast type, not necessarily of food group.
Statistical Analysis
Analysis of covariance was used to estimate adjusted mean body mass index (BMI) and energy intake (kcal) as dependent variables. Body mass index is calculated as weight in kilograms divided by height in meters squared (kg/m2). Dietary data from the NHANES III 24-hour recall were used to estimate average daily intake of energy for each breakfast group. The energy variable represents kilocalorie intake for the entire day, not just from the breakfast meal, in order to investigate the impact of breakfast type on daily calorie intake.
Means were adjusted for gender, race, age, poverty level, smoking habits, alcohol intake and physical activity. These variables were chosen as covariates because they could confound the relationship between breakfast type and BMI or energy intake. Age and gender are well-established major determinants of BMI and energy intake [11,14]. The other covariates are included as control for the possibility that the type of breakfast subjects consumed could simply be an indicator for a healthy lifestyle pattern or for cultural or economic patterns, rather than having a more direct relationship with BMI and energy intake.
The physical activity variable was specified as the sum of intensity rating multiplied by times per month for each of the activities listed in the NHANES III dataset. The alcohol variable was specified as the number of times beer, wine and hard liquor (combined) were consumed per month. Smoking was a yes/no variable referring to current smoking habits. To estimate poverty, the Poverty Income Ratio (PIR) was used, with values of 0.0 to 11.889. If the PIR was less than 1.3 (less than 1.3 times the poverty level), then subjects were grouped as below the poverty level, otherwise they were grouped as above the poverty level. The level of 1.3 corresponds to government guidelines for the eligibility cut-off for economic assistance. The statistical model for BMI was additionally adjusted for age squared to improve the fit of the model and improve control of confounding by age. The covariate age squared was not significant in the energy model, thus was not included as a covariate in that model.
The adjusted mean BMI and daily energy intake of the different breakfast groups were compared with each other using analysis of covariance. Because of the large sample size, differences were considered statistically significant only if p 0.01. All analyses included an adjustment for over-sampling and non-response and were carried out using SUDAAN for IBM PC [35], a data analysis package for multi-stage sample designs. This method accounts for different sample weights and the effects of the complex sample design on variance estimation.
|
RESULTS |
|---|
|
TOPFOOTNOTESABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSCONCLUSIONACKNOWLEDGMENTSREFERENCES |
|---|
|
|
|
0.01 was considered statistically significant.
|
|
In the analysis for energy intake by breakfast category, breakfast Skippers and Fruit and vegetable eaters had the lowest energy intake throughout the day (Table 5). These two groups were significantly different from all other categories except the Fats and sweets and Beverages groups, which also had a low total daily energy intake. The Meat and eggs eaters had the highest daily energy intake, which was significantly different from energy intake of all groups except the Dairy, Quick breads, and the Cooked cereal consumers. The energy intakes of the Ready-to-eat cereal, Cooked cereal, Bread and Quick bread groups, all containing grain products, were close to the energy intake mean of all groups (2,199 ± 18.91 kcal), but were higher than those of the Skippers, the Fats and sweets and the Fruit and vegetable eaters, and were lower than the Meat and eggs group energy intake, with various of these pair-wise comparisons reaching significance at p 0.01 (see superscripts in Tables 4 and 5).
|
DISCUSSION |
|---|
Twenty percent of the subjects in this NHANES III cohort skipped breakfast, making it the largest group of all ten investigated categories. This percentage is similar to the number reported by Haines et al. [12], who found that in 1991 25% of US adults skipped breakfast. Also, an earlier published report by Zabik [36] found that 24% of young women regularly skipped breakfast.
The breakfast skippers were one of the three groups with the highest BMIs (breakfast Skipper, Meat and eggs, Dairy). Our finding of an association between breakfast skipping and high BMI is consistent with data in the literature showing that breakfast skipping may lead to imbalanced eating later in the day (e.g., one big meal in the evening) [24], a dietary pattern associated with obesity [5,25,37,38].
Studies by Zabik et al. [36] and by Resnicow et al. [39] found that daily energy intakes appeared not to be affected when breakfast is skipped [11]. However, we found that the skippers have the lowest total daily energy intake of all groups, despite having high BMIs. An explanation for this could be that subjects who skip breakfast are already overweight, are trying to lose weight and are also limiting their daily energy intake from other food sources throughout the day. From these data it cannot be determined whether the association is causal (skipping breakfast leading to poor dietary patterns leading to obesity) or correlational (persons who skip breakfast are already overweight and are trying to control it by skipping breakfast). In either case, it is clear from these and other data that skipping breakfast does not lead to attaining or maintaining a healthy weight.
Of the five breakfast categories that represented, at least 10% of the population, subjects in the Meat and eggs group had the highest BMI and subjects in the Ready-to-eat cereal group had the lowest BMI, and these values were significantly different from one another at p 0.01. The high energy intake of the Meat and eggs group throughout the day is consistent with their high BMI. It is likely that this group obtained a substantial amount of their total daily calorie intake from the fat of the breakfast itself. In contrast, the energy intake of the Ready-to-eat cereal group (2,241 ± 25.3 kcal) is actually slightly higher than the overall mean energy intake (2,199 ± 18.91 kcal), a fact that appears to be inconsistent with their low BMI. It is possible that the composition of the diet, particularly in terms of fat and fiber intake, may be a factor in maintaining a healthy body weight.
Among the less common breakfast types, Fats and sweets (4.2% of all breakfast types) and Dairy (4.6%) consumers averaged higher BMIs and similar or somewhat higher daily energy intakes compared to the other groups. Individuals who took Dairy products at breakfast averaged the highest BMI of all breakfast categories (27.11 ± 0.4 kg/m2); however, this BMI only differed significantly from that of Cooked cereal consumers, but not from those of RTEC and Quick bread eaters, at p 0.01. These groups appear to represent people "on the run," eating only candy or soda, or grabbing a glass of milk or a piece of cheese. Their higher BMI would appear to support the notion that "dysregulated" eating patterns are associated with obesity, instead of or in addition to total energy intake per se.
Another less common breakfast type was the Fruit and vegetable eaters (4.5% of all breakfast types). Their BMI was similar to the overall mean, while their kilocalorie intake was lower than most other categories. The lower daily energy intake may reflect subjects who tend to eat meals with low energy content such as avoiding foods high in fat, factors that may have been incompletely adjusted in our control for several factors associated with a healthy lifestyle. On the other hand, it may be that the presence of other factors, such as fiber or micronutrients, has an influence on obesity.
Subjects who ate any types of breakfast that contained or consisted of grains (cereals and breads) had energy intakes that were close to the overall mean energy intake of all groups combined. At the same time, these groups also had lower BMIs than the Skippers and the Meat and eggs eaters. These products—cereals and certain types of breads—not only provide energy, but they can also contain a substantial amount of insoluble fiber [40,41]. This could explain why the two breakfast types with the lowest BMI were Ready-to-eat cereal and Cooked cereal consumers.
The Breads category appears to be an anomaly in this analysis. Subjects in this group consumed carbohydrate and presumably some fiber, but nevertheless had a BMI almost identical to the overall average (26.49 kg/m2). This was in contrast to subjects who consumed Ready-to-eat cereals, Cooked cereals and Quick breads, who had lower BMIs than those of other groups. We interpret this to again represent the impact of a "dysregulated" dietary pattern: persons who were categorized in this breakfast type did not have a breakfast that contributed substantial energy from dairy or meat groups, or from cooked or ready-to-eat cereals, or from fruit. Rather, they seem to have simply grabbed a piece of toast as their "breakfast."
It is possible that fiber is an important factor in this equation. Food with moderate and high fiber content, such as cooked cereal, some ready-to-eat cereals, fruit and vegetables, stays longer in the gastrointestinal tract and therefore contributes to a longer lasting satiety feeling [42,43]. Eating foods rich in fiber therefore could possibly prevent overeating of foods with high fat contents and thus high in energy. This could prevent overweight and obesity when combined with other weight-loss efforts such as increased physical activity.
These results could also support a hypothesis that the fat content at breakfast influences BMI. Some research has shown that eating a diet high in carbohydrate and low in fat is helpful in successful weight-loss maintenance [44], consistent with the higher BMI of subjects eating Meat and eggs, and with the lower BMI of subjects eating cereals, cooked or ready-to-eat. A study by Stanton and Keast [45] found that individuals whose breakfasts included a ready-to-eat cereal had a significantly lower total daily intake of fat, thus received significantly lower percentages of calories from fat. This is consistent with a study by Huang et al. concluding that eating breakfast, especially cereals and breads, reduced fat intake and helped increase fiber intake of young adults [46]. At the same time, these results are consistent with an interpretation that high carbohydrate, or high fiber, at breakfast influences BMI, since cereals and breads tend to be simultaneously low in fat and higher in carbohydrate and fiber. The data we present cannot differentiate between these two hypotheses.
Our data do strongly suggest that skipping breakfast, or eating a high-fat, low-fiber breakfast (as in the Meat and eggs category), is associated with a higher BMI. Among the other breakfast types, however, our observations are consistent with a hypothesis that either 1) a low-fat breakfast is associated with lower BMI or 2) that it is the organized meal style typified in cereal or quick bread breakfasts, rather than the energy, carbohydrate, fiber or fat contents of the various breakfast types, that ultimately influences BMI. The most consistent pattern in these data is that it is the eating of a fairly "organized" breakfast that is associated with appropriate body mass regulation. Ready-to-eat cereals, cooked cereals and quick breads such as pancakes are typically eaten at home and represent an attempt to have a "conventional" breakfast. In contrast, a quick soda, glass of milk, piece of toast or cup of coffee represent an "on-the-run" behavior that we consider to be conducive to poor dietary habits and obesity.
|
CONCLUSION |
|---|
|
TOPFOOTNOTESABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSCONCLUSIONACKNOWLEDGMENTSREFERENCES |
|---|
|
ACKNOWLEDGMENTS |
|---|
|
TOPFOOTNOTESABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSCONCLUSIONACKNOWLEDGMENTSREFERENCES |
|---|
|
FOOTNOTES |
|---|
|
TOPFOOTNOTESABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSCONCLUSIONACKNOWLEDGMENTSREFERENCES |
|---|
Received December 22, 2000. Accepted January 3, 2003.
|
REFERENCES |
|---|
|
TOPFOOTNOTESABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSCONCLUSIONACKNOWLEDGMENTSREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  A. K Kant, M. B Andon, T. J Angelopoulos, and J. M Rippe Association of breakfast energy density with diet quality and body mass index in American adults: National Health and Nutrition Examination Surveys, 1999-2004 Am. J. Clinical Nutrition, November 1, 2008; 88(5): 1396 - 1404. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 E.-K. Groholt, H. Stigum, and R. Nordhagen Overweight and obesity among adolescents in Norway: cultural and socio-economic differences J. Public Health Med., September 1, 2008; 30(3): 258 - 265. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 C. E. O'Neil and T. A. Nicklas A Review of the Relationship Between 100% Fruit Juice Consumption and Weight in Children and Adolescents American Journal of Lifestyle Medicine, July 1, 2008; 2(4): 315 - 354. [Abstract] [PDF]
|
|
|
|
 |
|
 H. A. Raynor, R. W. Jeffery, A. M. Ruggiero, J. M. Clark, L. M. Delahanty, and for the Look AHEAD (Action for Health in Diabetes) Weight Loss Strategies Associated With BMI in Overweight Adults With Type 2 Diabetes at Entry Into the Look AHEAD (Action for Health in Diabetes) Trial Diabetes Care, July 1, 2008; 31(7): 1299 - 1304. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. L. J. Greenwood and J. B. Stanford Preventing or Improving Obesity by Addressing Specific Eating Patterns J Am Board Fam Med, March 1, 2008; 21(2): 135 - 140. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 L. R. Purslow, M. S. Sandhu, N. Forouhi, E. H. Young, R. N. Luben, A. A. Welch, K.-T. Khaw, S. A. Bingham, and N. J. Wareham Energy Intake at Breakfast and Weight Change: Prospective Study of 6,764 Middle-aged Men and Women Am. J. Epidemiol., January 15, 2008; 167(2): 188 - 192. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 N. F. Krebs, J. H. Himes, D. Jacobson, T. A. Nicklas, P. Guilday, and D. Styne Assessment of Child and Adolescent Overweight and Obesity Pediatrics, December 1, 2007; 120(Supplement_4): S193 - S228. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. K Kant and B. I Graubard Secular trends in patterns of self-reported food consumption of adult Americans: NHANES 1971-1975 to NHANES 1999-2002. Am. J. Clinical Nutrition, November 1, 2006; 84(5): 1215 - 1223. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
T. Wilsgaard, B. K. Jacobsen, and E. Arnesen Determining Lifestyle Correlates of Body Mass Index using Multilevel Analyses: The Tromso Study, 1979-2001 Am. J. Epidemiol., December 15, 2005; 162(12): 1179 - 1188. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
H. R Farshchi, M. A Taylor, and I. A Macdonald Deleterious effects of omitting breakfast on insulin sensitivity and fasting lipid profiles in healthy lean women Am. J. Clinical Nutrition, February 1, 2005; 81(2): 388 - 396. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
