Breakfast: Shaping Guidelines from Food and Nutrient Patterns
Breakfast: Shaping Guidelines from Food and Nutrient Patterns
Michael Gibney and Irina Uzhova
The value of regular breakfast intake for nutrition, health, and wellbeing is widely recognized. Data exist, with varying degrees of confirmation, that a regular breakfast intake may help in body weight management, in cognitive function, and in cardiometabolic health. Regular breakfast consumers also enjoy a more optimal total daily intake of nutrients compared to breakfast skippers. A critically important question that must be addressed by policy makers is the approach to defining an optimal breakfast. Such a definition would involve the enumeration of quantitative nutrient guidelines which should transcend geographic patterns and ensuing food-based dietary guidelines that would reflect local foods and gastronomic customs.
Most governments that have issued standards for optimal breakfasts have done so in the context of food groups such as cereals, breads, fruits, and fruit juices, dairy produce, and pulses. None have based their recommendations on any underlying objective evidence derived from any analysis of existing breakfast patterns. Two countries, the USA and Mexico, have issued statutory guidelines on nutrient and food choices for optimal breakfasts in schools. In both cases, the nutrient recommendations are simply energy-adjusted targets based on prevailing dietary guidelines for adults. Moreover, neither enumerates specific targets for micronutrients and neither do they elaborate on the evidence base for the targets set for food intakes to achieve these nutritional guidelines. In the UK, a voluntary nutrient standard for school meals recommends a universal figure of 20% of daily intake based on the frequently observed finding that breakfast supplies about 20% of daily energy intake. The reality is that the intakes of many desirable nutrients, the minerals and vitamins in particular, contribute far more than 20% to total daily intake.
Cluster and principal component analyses have been used to identify patterns of breakfast food intake. These studies have frequently allowed the intakes of nutrients of different breakfast food clusters to be described in terms of their contribution to both breakfast nutrients and to overall daily nutrients. What is clear from an analysis of some of these data is that intake of nutrients across clusters of breakfast foods is very large compared to the variation that exists for the same clusters as regards total daily intake. In simple terms, good breakfast patterns do not automatically lead to optimal daily nutritional patterns. Equally, poor quality of breakfast nutrients does not determine a poor overall daily nutritional pattern.
One approach which is currently being examined by the International Breakfast Research Initiative is to assign each individual in a national survey with a value for their nutrient-rich food (NRF) index. The NRF index measures dietary quality for both macro- and micronutrients and has been extensively used across the globe to provide a quantitative score of overall dietary quality. For each of 6 national surveys (Canada, Denmark, France, Spain, UK, and USA), tertiles of NRF scores are created, indicating increasing overall energy-adjusted dietary quality. The intake of foods and nutrients for each individual at breakfast is computed for each of the 3 NRF tertiles. This approach is to be explored to establish if optimal breakfast nutrient targets can be defined. As regards food-based dietary guidelines, these will alter across different gastronomic traditions, but some critically important points need to be noted before food advice for breakfasts is issued. The average population intake of a food is a function of 2 elements: the percentage of the population consuming the food and the mean intake of the food among consumers of that food. The average intake of yogurt across NRF tertiles is constant for yogurt consumers. However, as NRF scores improve, more people eat yogurt. This approach is essential for public health nutrition in improving breakfast food choice toward achieving an optimal nutrient intake.
Whilst there is extensive literature on the health benefits of a regular breakfast, there are few guidelines to help policy makers to issue specific targets on optimal nutrient intake at breakfast or the selection of foods to attain these targets. The food and nutritional ad- vice on breakfast offered by most governments is confined to simple advice on food servings. The USA and Mexico typify the few countries that have attempted to issue specific nutrient targets for breakfast. However, these simply reflect general nutrient guidelines for adults, adjusted to suit lower energy needs of toddlers and school children. Little guidance is issued on micronutrient intake, and the advice on food choice does not appear to be linked to patterns of nutrient intake. The application of cluster and principal component analysis, which is used to determine the patterns of daily or breakfast food consumption and also link them to nutrient intake, greatly improved our understanding of optimal breakfast choices. Using 6 national nutrition surveys (Canada, Denmark, France, Spain, the UK, and the USA), the International Breakfast Research Initiative has opted to score each individual with a measure of overall daily nutritional quality (based on the nutrient-rich food index). It is hoped that options for the derivation in breakfast nutrient targets and associated food-based guidelines will arise from an analysis of tertiles of this score. Ultimately, meal-based advice will become the basic building block for digitally based personalized dietary analysis and guidelines.
It is widely accepted that breakfast is the most important meal of the day. Breakfast skipping has been associated with poorer overall nutrient intakes with implications for weight and glycemic control, cognition, and cardiovascular disease. Given the importance of breakfast in determining overall diet quality, it is not surprising that specific food and nutrient targets have been proposed for breakfast meals. This contrasts with other meals (snacks, lunch, and evening meal) where such data are extremely limited and almost exclusively directed to advice on snacks. This is not unexpected since, intuitively, it is held that consumers are far more likely to repeat the same breakfast choice over the week than they are to repeat, for example, the same main meal over the week. This concept was analyzed for the Irish National Adult Nutrition Survey for those subjects who were studied over 3 weekdays and 1 weekend day [unpubl. data]. The remaining subjects varied in the number of weekdays and weekend days. The results are given in Table 1. A quarter of the population always had the same break- fast over the entire week. If weekdays alone are considered, then 42% of subjects had the same breakfast on all 3 days. Many other combinations of similarity of weekday and weekend breakfasts can be considered, but the key point is that there is a strong tendency for the same breakfast to be consumed quite regularly over the week. Equally, it is worth noting that for one quarter of the subjects, the weekend breakfast differed from the weekday breakfast. Thus, given the repeatability of breakfast, at least through the working week, the identification of specific guidelines for target nutrient intake and accompanying food in- take patterns remain an attractive possibility which would be valuable to policy makers and ultimately consumers.
Maximizing Quality Breakfast Uptake
There are 3 nonnutritive attributes of breakfast which need to be safeguarded in attempting to design any optimal breakfast intake. Breakfasts must be affordable, convenient, and appetizing. On an average school day in the US, about 14 million children take part in the School Breakfast Program; nearly 12 million of them were low-income children who received a free or reduced-price school breakfast . Thus, breakfasts, if they are to be prepared and eaten at home, must be affordable by the most economically vulnerable groups. The same US data also show that a major reason for schoolchildren to miss breakfast at home, especially in rural areas, is the long commute to school. Moreover, participation in a school breakfast program is hampered by commuting delays. Thus, convenience in breakfast preparation is essential. Finally, we know that children turn away from foods when flavor is compromised. One classic example was in Oregon where it was decided to abandon the sale of chocolate-flavored milk in favor of plain milk in 11 elementary schools . The net effect was a 12% fall in sales and a 24% rise in milk wastage. Thus, palatability is a third key element in shaping school breakfast programs.
Table 1. Percent of subjects with breakfast meals similar and different for weekdays and weekends from the National Adult Nutrition Survey of Ireland
Present Governmental Guidelines on Food and Nutrition Standards for Breakfast
Many governments and state agencies have issued advice on the foods to possibly include in the development of a nutritious breakfast. These data are largely qualitative and are not based on any evidence arising from analyses of prevailing patterns of food choice at breakfast. Moreover, there is a lack of consistency in the definition of recommended categories: “cereal” is a category recommended in Ireland , while in New Zealand that is extended to “porridge or muesli” or to a specific branded breakfast cereal , and in the UK it is defined as “low fat, low sugar cereals” . Two countries, Mexico and the USA, have issued more detailed guidelines on target nutrient intakes and recommended food selection for break- fast. In the case of Mexico, specific guidelines are issued for nutrient intakes at breakfast (Tables 2, 3). The values listed for preschool children and schoolchildren correspond on an energy-adjusted basis to the national dietary guidelines for Mexico . The official nutritional guidelines for breakfast for Mexican children are accompanied by detailed standards for specific foods (data not shown). In the case of the USA, the focus is on recommending specific foods and then to set out very precise nutrient profiles which all foods have to follow (Table 4) . For most nutrients, there are qualifying issues related to specific foods. Some are expressed as percent energy, some as units per serving, and some as a maximum percentage of the total weight of food. In both cases, little evidence is presented as to the scientific basis for choosing the particular foods and their portion sizes. In addition, both reports offer limited guidance on micronutrient intake of food composition standards. In the US, school breakfasts are expected to provide a minimum of 25% of the recommended daily allowance (RDA), presumably to mirror an expected energy intake of 25% of daily intake. Taken together, all of these governmental recommendations highlight the need for a more scientifically rigorous approach to recommendation as to optimal food and nutrient intakes at breakfast.
Table 2. Average daily total energy and macronutrients for breakfast as recommended in Mexico 
Table 3. Nutrient intake data across 6 clusters of Mexican breakfast (minimum and maximum nutrient intakes and fold increase for intakes at breakfast and the total day)
Table 4. Nutritional standards for foods used in federal school food programs in the USA 
Cluster Analysis and Its Application to Optimal Breakfast Food and Nutrient Choices
Several published studies have sought to exploit the statistical methodology of nearest-neighbor analysis through either principal component or cluster analysis. The application of cluster analysis to describe the pattern of breakfast intake among Mexican children identified 6 meaningful clusters . Due to the variety in dietary patterns identified, ranging from traditional Mexican to Westernized, the nutrient intakes varied considerably across clusters. What might be high in one is low in another and vice versa. This is true when cluster-based nutrient intakes are considered for breakfast only or for the full day, but the level of variation is far higher in the former (Tables 2, 3). The ratio between the cluster with the lowest intake of any nutrient to that of the cluster with the highest intake varies far more for breakfast than for the day as a total. This makes sense since different meals throughout the day have quite different overall nutritional profiles. Other approaches to a combination of principal component and cluster analysis have taken a total overview of nutrient intake. Thus, a French study showed that when this approach was applied to breakfast food, 4 clusters were identified. Data on food intake are presented in the study, but the nutrient profile is given as an overall score for breakfast nutrients . In this study, the mean adequacy ratio (MAR) was used. The MAR expresses the intake of each nutrient as a percentage of the RDA, and the mean of all these intakes is the MAR score. Clear differences in the MAR score were evident for breakfast with a range from 18 to 30. Another study, carried out on a sample of German adults, identified 4 breakfast patterns using principal component analysis and linked them to the a priori-defined breakfast quality index , which in this instance combined both food and nutrient intakes into scores (vegetables, fruits, whole grains, sugar-sweetened beverages, fruit juice, nuts, legumes, red and processed meat, trans fats, long- chain (n-3) fatty acids (eicosapentaenoic acid and docosahexaenoic acid), and total polyunsaturated fatty acids. The breakfast quality index was also directly related to health outcomes covering glycemic control, cardiovascular disease, and weight control. Thus, while cluster and principal component analysis re- mains a useful tool in determining optimal breakfast nutrient and food intake, there are major limitations in setting out gradations in optimal nutrient in-takes for breakfast.
The International Breakfast Research Initiative
Recognizing the need to explore the possibility of a more structured and harmonized approach to defining nutrient intakes for breakfast, a number of scientists from both sides of the Atlantic have come together under the sponsorship of Cereal Partners Worldwide to explore options to make this possible. Data from 6 national nutrition surveys will be used (Canada, USA, Denmark, France, Spain, and the UK) and will provide estimates of nutrient and food intakes in as harmonized a means as possible. The intention is to explore some options for an a priori-defined scoring. Initially, the use of the nutrient-rich food index will be explored with the subjects examined across tertiles of this score based on the daily nutrient intakes . Breakfast nutrient intakes will then be examined, and from that analysis, some harmonized approach to the development of guidelines for optimal nutrient intakes at breakfast will be proposed. Clearly, wherever possible, existing international nutrient guidelines will be used. However, bearing in mind that the proposed guidelines are for just 1 meal during the day, many shortcomings with existing guidelines will have to be overcome. For the micronutrients, a blanket 20% of the RDA is proposed in the USA. However, current data show that intakes at breakfast for many micronutrients well exceed 20% of the RDA, and quite how these higher prevailing intakes should be protected and promoted remains to be seen.
Table 5. Number of days to attain an average portion of target foods in the total population and the population of consumers only with data also presented as percent consumers for the total population
Exploring Strategies to Set Goals to Optimize Food Intakes at Breakfast
The average daily intake of a food is a function of 3 elements:
- The average mean daily intake of the total population, which will include both consumers of the food and nonconsumers (includes zero values).
- The percentage of the population that consumes the food.
- The intake of the food among consumers only (no zero values included).
Most data on food intake simply report mean daily total population intake and, inevitably, that leads to recommendations to seek an increase or decrease in average daily intake. However, when consumers only are considered, it is of- ten observed that such consumers are at the maximum reasonable level of intake of a target food. Thus, the main problem that we face in shaping advice on food choices is that it is the percentage of the population that consumes the target food that needs to be addressed rather than the level of intake among those who regularly consume a given food. Table 5 provides the data from the Irish National Adult Nutrition Survey and compares food intake for the lowest and highest tertiles of the nutrient-rich food index for the total population and for consumers only. When the former is considered, because of the presence of a significant number of zero intakes, the time taken in days for an average serving to be consumed is very high and, within that, higher at the lower tertile. However, when nonconsumers are excluded, there is clear evidence that an average portion is consumed about once every 2 days on average for almost all foods. Creating advice to consumers who do not consume a food requires a different strategy than the challenge of asking regular consumers to eat the food more regularly and at higher portions. Most of the progress in changing population habits will be made by the former, that is getting nonconsumers to begin to become consumers. Simple score sheets might arise from the International Breakfast Research Initiative, which will allow educators to help individuals score their breakfast quality based on whether or not they eat a food or how often they consume it.
Promoting Healthy Breakfasts in the Digital Era
Unless scientifically strong agencies engage in the use of social media and associated digital technologies to promote healthy eating, the gap will be filled by poorly qualified opportunists. Bearing in mind the growing dominance of the digital revolution in the lives of younger people, there is little choice but to engage with them within this technology. One of the most comprehensive tests of digitally delivered dietary analysis and guidance was the Food4Me study, which covered 1,300 subjects across 7 EU centers over a 6- month period . The results clearly show that the digitally led intervention significantly improved target nutrient intakes. Most of these studies are built around the role of individual foods in delivering a nutritional input. The future will have to abandon a food-by-food approach and focus on meals. Breakfast, which tends to have a constancy of consumption patterns (Table 1), would seem a good place to start. Defining meals using data mining techniques such as decision tree analysis or neural network analysis has proved problematic. This has led to an expert-based approach in deciding the typical dishes chosen for a particular meal, and this has proven the most successful route to date. In an ideal world, a consumer would see images of candidate dishes for that meal on the screen of their digital device. Thus, for breakfast, one image might depict a cereal-based breakfast. Another might depict a coffee-and-croissant or a cooked-meat-and-egg breakfast type. The consumer would click on the dish that they most frequently choose and then, with drop-down menus, refine the image to match exactly their own breakfast. This would be repeated for all eating occasions. Those choices would be entered into the analytical end of the personalized nutrition service, emanating in advice on the daily nutrients, the intakes of which might need to be improved. The suggested eating patterns would be entirely meal based and thus would allow the user to create a weekly menu towards an optimal diet. Empowering consumers to make choices which best suit them personally is the future, and bland advice such as food pyramids is the past.
Breakfast is without doubt a very important meal and one which is sufficiently constant to merit its own nutrient guidelines and subsequent food-based advice. For any proposal for the development of such to be successful, a number of conditions must be met. The proposals must be based on science, must take account of fortification, must not conflict with WHO targets, and must be amenable to reality checking. Hopefully, the International Breakfast Research Initiative will set this process in motion.
1 Food Research and Action Center (2017): School Breakfast Scorecard 2016–2017. http://frac.org/wp-content/uploads/school- breakfast-scorecard-sy-2015–2016.pdf (accessed March 7, 2018).
2 Hanks AS, Just DR, Wansink B: Chocolate milk consequences: a pilot study evaluating the consequences of banning chocolate milk in school cafeterias. PLoS One 2014;9:e91022.
3 Food Safety Promotion Board (2018): What is a healthy breakfast? http://www.safefood. eu/Start/Healthy-Living/What-is-a-healthy- breakfast.aspx (accessed March 7, 2018).
4 Office of Children’s Commissioner (2014): Guidelines for school food programmes. http://www.occ.org.nz/assets/Publications/ Guidelines-for-School-Food-Programmes. pdf (accessed March 7, 2018).
5 National Health Service: Healthy breakfasts (for people who hate breakfast). https://www. nhs.uk/livewell/loseweight/pages/healthy- breakfasts.aspx (accessed March 7, 2018).
6 Secretaría de Salud: Acuerdo mediante el cual se establecen los lineamientos generales para el expendio o distribución de alimentos y bebidas preparados y procesados en las escuelas del Sistema Educativo Nacional. [Agreement establishing the general guidelines for the sale or distribution of foods and beverages prepared and processed in the schools of the National Educational System.] Mexico City, Di- ario Oficial de la Federación, 2014.
7 Food and Nutrition Service, United States Department of Agriculture: National School Lunch Program and School Breakfast Pro- gram: nutrition standards for all foods sold in school as required by the Healthy, Hunger- Free Kids Act of 2010. Fed Regist 2016;81: 3–5.
8 Afeiche MC, Taillie LS, Hopkins S, et al: Breakfast dietary patterns among Mexican children are related to total-day diet quality. J Nutr 2017;147:404–412.
9 Lepicard EM, Maillot M, Vieux F, et al: Quantitative and qualitative analysis of breakfast nutritional composition in French schoolchildren aged 9–11 years. J Hum Nutr Diet 2017;30:151–158.
10 Iqbal K, Schwingshackl L, Gottschald M, et al: Breakfast quality and cardiometabolic risk profiles in an upper middle-aged German population. Eur J Clin Nutr 2017;71:1312– 1320.
11 Drewnowski A: Uses of nutrient profiling to address public health needs: from regulation to reformulation. Proc Nutr Soc 2017;76: 220–229.
12 Celis-Morales C, Livingstone KM, Marsaux CF, et al: Effect of personalized nutrition on health-related behaviour change: evidence from the Food4Me European randomized controlled trial. Int J Epidemiol 2017;46:578– 588.