Dietary intake is a major determinant of both the nutritional status and the general health and well-being of an individual. An optimal diet will supply adequate – but not excessive – amounts of all essential nutrients, while maximizing foods and dietary substances that promote long-term health and avoiding dietary constituents related to ill health(Reference Uauy and Solomons1). Both greater dietary variety (number of different foods and beverages consumed) and dietary diversity (selection from an array of food groups) are associated with more nutritious and more healthful intake patterns(Reference Azadbakht, Mirmiran and Azizi2). What is an inherent reality is that foods are consumed in various meals and meal settings over the course of a day. Moreover, factors of household economics, cultural and culinary conventions and personal convenience will dictate the frequency, size and composition of the meals consumed throughout the day.
A few investigators have analysed individual meal contributions to the day’s intake of macro- or micronutrients. These pioneering studies identified dietary patterns that deviate strongly from recommended population nutrient goals in children(Reference Chitra and Reddy3–Reference Matthys, De Henauw, Devos and De Backer5), adolescents(Reference De Henauw, Wilms, Mertens, Standaert and De Backer6, Reference Sjoberg, Hallberg, Hoglund and Hulthen7) and adults(Reference Ovaskainen, Reinivuo, Tapanainen, Hannila, Korhonen and Pakkala8). The studies on children and adolescents are mostly European studies. The findings emphasize the difference in nutritional value of meals and the association between breakfast consumption and better dietary quality. Unfortunately, no similar results are available for children or adolescents in Guatemala, or even in the broader region. However, studies show a prevailing trend of snack-dominated meal patterns, associated with higher intakes of foods with lower nutrient density (i.e. high in fats and sugars, but low in micronutrients). Further concerns include irregular meal patterns, such as meal skipping, and high fat content in lunch and snacks, patterns to be explored in the present study.
Consumption patterns that are too poor or too rich in macro- or micronutrients, excessive in harmful foods and noxious ingredients, and insufficient in health-protective elements might need to be corrected through education and health interventions. Therefore, knowing the pattern of meals and the nutrient density of these meals might help provide the necessary leverage for corrective change. In Guatemala, micronutrient-enriched products such as micronutrient-fortified cereals as well as sugar fortified with retinyl palmitate are potential major sources of micronutrients for children. However, little is known about how these products are consumed in terms of timing (meal pattern), frequency and quantity.
During a 3-month period in the summer of 2005, 24 h registries of all food and beverage intake were obtained by interactive, pictorial self-recording by children in the third and fourth grades of public and private schools of the city of Quetzaltenango in the western highlands of Guatemala. Our objective was to assess and describe the intakes of energy and main macronutrients across the different eating opportunities in the children’s day, and to relate them the intake of selected micronutrients from the different meals. Below we describe in detail the procedures used to achieve our aim and our findings of meal patterns’ nutrient contributions across social class and gender in this survey of urban Guatemalan schoolchildren. Micronutrient density is particularly important in the diet of children, who require nutrient-dense foods for healthy physical and mental development. Identifying the contribution of foods according to meal type is helpful for identifying the relative contribution of micro- and macronutrients to the diets of children.
Subjects and methods
Subject selection
A total of five public schools, stratified as lower socio-economic status (LSES), and eleven private schools, stratified as higher socio-economic status (HSES), were invited to participate in the study. Within these schools, only children attending third and fourth grades were recruited.
The nature of the study was explained to the teachers and students in the classroom during usual school hours. The children’s legal guardians were informed about the study in writing. They were told that the main objective of the study was to assess usual fruit and vegetable consumption in schoolchildren. Incentives to participate included a free snack (either a bread-based meal or cereal) and the children were allowed to keep the crayons provided for the study.
The study was approved by the Human Subjects Committee of the Center for Studies of Sensory Impairment, Aging and Metabolism (CeSSIAM), Guatemala City and the local education authorities. Only children with a signed consent form from their legal guardians were included in the study.
Questionnaire and sample collection
A single pictorial 24 h prospective diary was collected from April to June 2005. The data collection instrument is a 5-page booklet designed to assess dietary intake with a 24 h time frame. The first page contains written instructions on how to complete the questionnaire. The subject is instructed to draw all the food and drink items consumed within a 24 h time frame and to include all items consumed between meals (snacks) both at school and at home. Children were asked to record all drink and food items consumed since the last meal and for 24 h thereafter. The children were asked to take the booklets home and draw all food and drinks consumed (including candies) and to specify brands and amounts consumed. On the following day, the subjects were interviewed by a research nutritionist, who checked for completeness and estimated the portions of items listed. This interview helped to clarify the described articles and to quantify the food intake by using food models and common household measures. Children who forgot to bring the workbook were given another booklet for that day and the interview was done the next day. All data were recorded on school days and all meals consumed at school were brought from home or bought at the school snack shop. No school lunch was provided, as a usual school day ends at 13.00 hours.
Although the data collection tool was not validated, the instrument and methodology used for data collection were previously used in studies conducted by CeSSIAM in urban schoolchildren from Guatemala City. The estimated median energy intake for the population, of 7829 kJ/d (1870 kcal/d; data not shown), was in close agreement with the WHO recommendations for this age group(Reference Torún9). Our estimate median protein intake of 65 g/d was twice the Recommended Dietary Allowance for this age group(10).
Data analysis
Each food or beverage item reported was coded and entered into a database according to mealtime (breakfast, lunch, dinner or snack). All food and beverages items were codified separately according to preparation or presentation. Different commercial brand names of fruit juices, ready-to-eat cereals, etc. were coded as separate food items. A total of 218 items were listed and their nutrient values were derived primarily from the Central American food composition table (FCT)(11) and secondly from the US Department of Agriculture (USDA) FCT obtained online from the USDA National Nutrient Database for Standard Reference version 16·1 (www.nal.usda.gov/fnic/foodcomp/data). Nutritional content of some products, such as ready-to-eat cereals, were updated. Nutrient values for food items not listed in the FCT were taken from the food labels or other manufacturer data. Typical dishes and composite foods were obtained from everyday Guatemalan recipe books(12–Reference Sierra Franco14) and their nutrient data were added to the FCT.
In order to provide a context to the food culture of the samples of schoolchildren, an illustration of the selection of the principal food and beverage contributors was assembled. For each meal and as disaggregated by gender and socio-economic status (SES) sub-samples, we calculated the cumulative contribution of energy for each food and beverage item reported within that subgroup. We constructed a 4 × 4 panel table by meal and by subgroup, with each panel including the top ten leading contributors to that meal.
The daily distribution of energy, macronutrients and selected micronutrients intakes (vitamin A, vitamin C, vitamin D, thiamin, riboflavin, folate, Ca, Fe and Zn) throughout the day (i.e. breakfast, lunch, dinner and combined snacks) was examined. The mean absolute nutrient intakes and their standard deviations are provided for reference in the Appendix. The focus of analysis, in accordance with the aim of the study, was to assess the distribution of energy, macro- and micronutrients intakes across meals. The focus here was testing differences between the meals (breakfast, lunch, dinner and combined snacks) within each SES or gender group. It was not our intention to generate comparisons within meals across LSES and HSES and girls and boys.
In keeping with the goal of our study, the analysis focuses first on the percentage of the daily intake of a nutrient that a given meal contributes. Thus, we calculated the subject’s consumption of a given nutrient for that meal, divided by the subject’s total daily intake of the nutrient and multiplied by 100. These individual meal-based percentages are presented by SES and gender with comparisons of results for meals classified as breakfast, lunch, dinner and snacks. Next, we compared the percentage contribution for nutrients with the meal’s contribution to energy intake. These proportions were used for a comparative approach to identify the meals where a particular nutrient contribution fell below the overall energy contribution. If nutrient and energy intakes are proportional, the percentage nutrient and total energy contributions should match, one-to-one. In order to assess nutrients that contributed proportionately above or below the day’s nutrient intake, relative to the energy contribution, we arbitrarily defined cut-offs that were 25 % above or below the energy contribution. Thus, a ratio of less than 0·75:1 identifies the meal as contributing relatively less of a nutrient relative to that meal’s energy contribution. An elevated nutrient contribution of greater than 1·25:1 identifies a meal that contributes proportionately more of a nutrient relative to the meal’s energy contribution. For this pattern analysis, we simply tabulated the number of ratios that fell into our pre-established deviant range.
The second analysis, focusing on meal-based nutrient densities, assesses each nutrient relative to the total energy contributed by that meal. Thus, the nutrient consumed at each meal was divided by the total energy. The resulting measure, the meal’s nutrient contribution expressed per 4·187 kJ (1 kcal) consumed, was then multiplied by 1000. As with meal-based percentage intake, energy densities are presented separately for each nutrient, with results stratified by SES and gender, and statistical comparisons are based on differences between the four meals: breakfast, lunch, dinner and snacks.
The final measure compares nutrient densities against estimated nutrient intakes, using a new measure we call ‘critical densities’ modelled after the 1986 Working Group of the Cavendas Foundation for nutrient requirements in Latin America(Reference Bengoa, Torun, Behar and Scrimshaw15). Critical densities were defined as the estimated recommended nutrient intakes, expressed per 4187 kJ (1000 kcal), and representing the amount of the respective nutrients that would achieve the recommended intake when an individual consumed the normative daily energy intake for his or her age and gender group. They were computed for selected micronutrients based on the WHO/FAO Recommended Nutrient Intakes (RNI)(16) and the energy requirements for children proposed by the Institute of Nutrition of Central America and Panama(Reference Torún, Menchú and Elías17). The RNI values were 500 retinol activity equivalents for vitamin A, 35 mg for vitamin C, 5 μg for vitamin D, 0·9 mg for thiamin, 0·9 mg for riboflavin, 300 dietary folate equivalents for folate, 700 mg for Ca, 9 mg for Fe and 5·6 mg for Zn. These nutrient recommendations were normalized to a 4187 kJ (1000 kcal) unit based on the estimated daily energy requirements. Thus, recommended nutrient intakes were divided by energy intake needs estimated to be 7850 kJ (1875 kcal) for boys and 6908 kJ (1650 kcal) for girls. Mean nutrient densities of gender and SES subgroups for each mealtime were then compared with the estimated critical densities, based on group means. Mean estimated intakes below the computed critical densities were considered inadequate.
Statistical methods
Data were analysed with the SPSS statistical software package version 11·0 (SPSS Inc., Chicago, IL, USA). As described above, the analysis focuses on the daily distribution of energy, macronutrients and selected micronutrients intakes (as a proportion of total daily intake) and nutrient densities, per meal. We used repeated-measures ANOVA to examine statistically significant differences in daily nutrient intake distributions and nutrient density distributions between mealtimes (i.e. breakfast, lunch, dinner and combined snacks) within subjects. The focus is testing differences between the meals within each SES and gender subgroup. We considered a probability of 5 % to be significant.
Results
Study population
All five public schools and seven schools invited agreed to participate and were included in the study. The total number of schoolchildren attending third and fourth grades in these twelve schools was 1124 (624 LSES, 500 HSES) and the majority of children were between 8 and 10 years old. A large proportion of children (n 675, 60 %) did not participate in the study for various reasons such as falling outside the age criteria, absence on the day of data collection, ‘forgetting’ the consent form or leaving the data collection booklet at home. The final sample comprised 449 (40 %) children, 219 (49 %) of LSES (113 girls and 106 boys) and 230 (51 %) of HSES (119 girls and 111 boys).
Principal energy sources of one day’s meal-associated intakes
In order to understand the nutritional partition among meals, it is important to know the context of the foods in the meals. Table 1 presents the ten principal food and beverage contributors to the total energy of each class of meal: breakfast, lunch, dinner and snacks. A modal breakfast comprised breakfast cereals and milk with added sugar. Corn tortillas or white bread with fried eggs were also commonly eaten, especially by LSES children. Corn tortillas, a staple food consumed in all three meals, contributed up to 17·5 % of the total energy in LSES boys for lunch. Main energy sources for lunch included chips, white rice and vegetable stew with chicken or beef. Main energy sources for dinner included corn tortilla, sweet bread, coffee with added sugar and fried eggs. Popular snacks included pizza, white bread, crisps and cola drinks.
Table 1 Main sources of daily energy by mealtime, gender and socio-economic status: third- and fourth-grade schoolchildren from Quetzaltenango, Guatemala, 2005
HSES, higher socio-economic status; LSES, lower socio-economic status; RTE, ready-to-eat.
*Food item as a proportion of energy consumed during that meal (%).
†Fluid whole milk with sugar: specifically coded as the milk added to cereals.
‡Sugar: granulated sugar is fortified with vitamin A by Guatemalan law at 10 retinol activity equivalents/g.
§Wheat sweet bread (pan de manteca): a staple containing wheat flour, sugar and shortening.
∥Nutritive beverage (incaparina): a protein-rich popular gruel based on corn and soya with added micronutrients.
¶Artificial drink: any of the sweetened fruit-flavoured non-carbonated beverages often fortified with vitamin C.
**Corn-based tamale (tamalito): lime-treated corn dough baked in corn leaves.
††Maize gruel (atole de masa): lime-treated corn dough prepared as beverage.
Estimated proportion of one day’s energy and macronutrient contribution by mealtime
Means of estimated 1 d intakes of energy and macronutrients by mealtime (i.e. breakfast, lunch, dinner and snacks) are presented as proportions of total daily intake in Table 2. We used repeated-measures ANOVA to examine differences in energy, macronutrient and selected micronutrient distributions between mealtimes (breakfast, lunch, dinner and combined snacks) within subjects. Meal compositions were compared within each of the four subgroups, i.e. HSES boys, LSES boys, HSES girls and LSES girls. Analyses were run separately, testing meal pattern contributions for energy, protein and fat. Of these twelve computations, a significant difference was found in the meals for all macronutrients (P < 0·001) except fat in LSES girls (P = 0·194). Lunch led in terms of energy (P < 0·001), protein (P < 0·001) and carbohydrates (P < 0·001), in all gender and SES subgroups. Lunch led in terms of fat in HSES boys (P < 0·001), whereas breakfast in LSES boys (P = 0·038) and dinner and snacks in HSES girls (P = 0·022) were more important sources of fat. Mean energy intake from lunch ranged from 2319 kJ (554 kcal; 32 % of energy) in LSES girls to 2700 kJ (645 kcal; 29 % of energy) in HSES boys. Dinner and snacks were the lowest sources of energy. Mean energy intake from snacks was between 1562 kJ (373 kcal; 19 % of energy) in LSES girls up to 2244 kJ (536 kcal; 27 % of energy) in HSES boys (Appendix). With respect to the macronutrient:energy ratios, there were few, if any, ratios outside the established boundary limits for the meal-wise fat:energy and carbohydrate:energy percentage ratios. For protein, however, the ratio approached, but did not exceed 1·25, for lunch across the subgroups, and fell clearly below the 0·75 limits for snacks (data not shown).
Table 2 Distribution (as a proportion of total daily intake) of estimated 1 d intakes of energy, macronutrients and selected micronutrients by mealtime, gender and socio-economic status: third- and fourth-grade schoolchildren from Quetzaltenango, Guatemala, 2005
HSES, higher socio-economic status; LSES, lower socio-economic status.
*P value from repeated-measures ANOVA.
Estimated proportion of one day’s micronutrient contribution by mealtime
Means of estimated 1 d intakes of selected micronutrients by mealtime (i.e. breakfast, lunch, dinner and snacks) are presented as proportions of total daily intake in Table 2. We performed within-column repeated-measures ANOVA among the percentage contributions of the four meals for the thirty-six quartets of data involving micronutrients (nine micronutrients by gender and SES). Of these thirty-six computations, a significant difference within the foursome was found for all micronutrients examined (P < 0·001). In general, although lunch led numerically in terms of energy and most macronutrients, it was not the most micronutrient-dense meal. Lunch was the leading source only for vitamin C and Zn in all gender and SES subgroups and folate for LSES girls only. Breakfast led in terms of all other micronutrients examined (i.e. vitamin A, vitamin D, thiamin, riboflavin, folate, Ca and Fe) for all gender and SES subgroups. In general, snacks were the poorest source of all micronutrients.
Turning to a less formally statistical pattern analysis for micronutrients, the 144 ratio values for the percentage contribution of the three macronutrients and nine micronutrients to their corresponding mealtime percentage energy contribution for each specific meal were examined with respect to the boundary criteria. A total of seventy-nine (55 %) were acceptably close to the nutrient:energy contribution concordance ratio of 1:1. An additional thirty-seven (26 %) fell below 0·75:1 and twenty-eight (19 %) were above 1·25:1 (data not shown). Protein:energy ratio in snacks was an example of a consistently below-criterion ratio, as were the corresponding ratios for vitamin D:energy at lunch and in snacks. In general, snacks had the lowest energy contribution ratios for micronutrients, and this was consistent across all gender and SES groups for vitamins A and D, thiamin and Zn. Breakfast was often a meal in which the nutrient:energy contribution ratios greatly exceeded the boundary criterion, a pattern that was consistent across all gender and SES groups for vitamins A and D, Ca and Fe.
Nutrient density by mealtime
Table 3 illustrates mean values and standard deviations for the selected nutrient densities in each mealtime by gender and SES group. We used repeated-measures ANOVA to examine differences in density distributions between mealtimes (breakfast, lunch, dinner and combined snacks) within subjects. Without class distinction, lunch had higher density of protein (P < 0·001) and snacks had higher density of carbohydrates (P < 0·001 in HSES boys and girls, P = 0·034 in LSES boys, P = 0·003 in LSES girls). Mean density of fat was not significantly different between the mealtimes (P = 0·198 in HSES boys, P = 0·206 in LSES boys, P = 0·275 LSES girls), except in HSES girls for which dinner had higher density of fat (P < 0·001). Significant differences were observed for all micronutrients examined, except folate in LSES girls (P = 0·073). Most micronutrients had higher density at breakfast for most gender and SES subgroups. The exceptions were vitamin C (P < 0·001) for which snacks were a major source, riboflavin in LSES girls (P < 0·001) for which dinner was a major source, folate in LSES girls (P = 0·073) for which no differences were observed between meals and Zn (P < 0·001) for which lunch was a major source in boys and LSES girls. Breakfast was a remarkably superior source for vitamin D, Ca and Fe (P < 0·001).
Table 3 Nutrient densities of estimated 1 d intakes of macronutrients and selected micronutrients by mealtime, gender and socio-economic status: third- and fourth-grade schoolchildren from Quetzaltenango, Guatemala, 2005
HSES, higher socio-economic status; LSES, lower socio-economic status; RAE, retinol activity equivalents; DFE, dietary folate equivalents.
*P value from repeated-measures ANOVA.
Critical nutrient density by mealtime
Meals with a nutrient density below the critical density computed according to RNI values and energy requirements are presented in Table 4. Snacks had a nutrient density below the critical density for most micronutrients examined, with some differences between genders and social class. Breakfast had a nutrient density below the critical density for vitamin D in all gender and SES subgroups. In LSES girls only, the critical density for breakfast was also below the standard for vitamin C and folate. Snacks had critical densities for nearly all micronutrients examined with the sole exception of vitamin C.
Table 4 Gender-specific critical densities for micronutrients and meals with average content below critical densities: third- and fourth-grade schoolchildren from Quetzaltenango, Guatemala, 2005
HSES, higher socio-economic status; LSES, lower socio-economic status; RAE, retinol activity equivalents; DFE, dietary folate equivalents.
*Critical density was based on the WHO/FAO vitamin and mineral requirements (Recommended Nutrient Intakes)(16) and a recommended daily energy intake of 7850 kJ (1875 kcal) for boys and 6908 kJ (1650 kcal) for girlsReference Torún, Menchú and Elías(17).
Discussion
Guatemala has traditionally been renowned in the nutritional literature for the description and exploration of nutrient deficiencies(Reference Boisvert, Castaneda, Mendoza, Langeloh, Solomons, Gershoff and Russell18–Reference Majia, Hodges, Arroyave, Viteri and Torun21). At the same time, certain aspects of its traditional Guatemalan cuisine have been associated with good health related to blood pressure(Reference Belizan and Villar22), intestinal function(Reference Kretsch, Crawford and Calloway23) and cardiovascular health(Reference McGill, Stella-Arias and Carbonell24). Increasingly in Latin America, a pattern described as ‘nutrition transition’ has been documented(Reference Albala, Vio, Kain and Uauy25–Reference Barria and Amigo28). The nutrition transition experience is related to demographic and socio-economic changes, dietary changes, increased obesity rates and sedentary lifestyles. It is characterized by dietary changes such as an increase in dietary fat (mostly saturated fat) and the increased availability and preference for high-fat/high-carbohydrate energy-dense foods. In Latin America, these changes have been occurring quickly and unevenly across socio-economic groups. As a consequence a shift from infectious diseases to chronic diseases has been observed. Companion studies in our population have confirmed the emergence of overweight and obesity in the middle-class of Quetzaltenango(Reference Groeneveld, Solomons and Doak29, Reference Groeneveld, Solomons and Doak30) and a lower than recommended consumption of fruits and vegetables (G Montenegro-Bethancourt, unpublished results). The opportunity to look more deeply into the dietary pattern, specifically of how nutrients selectively associate with different meals across the day, has been examined here among 449 schoolchildren of both sexes, attending either public or private schools in the most important metropolitan area of the western highlands of Guatemala.
Certain limitations in the design and methodology are recognized. They derive in part from resource limitations and from limited time of access to each school site and to the subjects within each setting. First, the non-response rate was high which might have resulted in selection bias. The low participation rate is largely caused by the limitations of the data collection time frame combined with the necessary informed consent procedures. There were multiple opportunities for a child to be missed during the five consecutive days of recruitment and data collection. ‘Forgetting’ informed consent forms and leaving data collection booklets at home were common occurrences. While efforts were made to include children with missing data, the time restraints of the data collection period did not permit researchers to return to the same schools. Non-response rate was higher in children of LSES (46 %) compared with children of HSES (35 %). It is, for example, possible that those children with a poor diet may have opted not to participate, or simply that children were disinterested in the extra-curricular activities.
Second, the present study is based on a single day’s register of foods and beverages with the disadvantage of not being representative of the habitual nutrient intake of any individual within the group. As a consequence of this limitation, analysis in the study was conducted at the group rather than individual level, as a single 24 h recall better represents the distribution of the group (and subgroups) intake within the season of the year. With only one day, however, we could not adjust the group averages for variance(Reference Murphy31) and thus the reported distributions are wider than would be conventionally reported with the opportunity for variance adjustments. However, requesting a second day’s registry, even in a sub-sample of our survey population, represents an inconvenience that might have interfered with institutional collaboration or lowered the response rate.
Third, our data rely on self-reporting by children. Paediatric diet researchers have been generally optimistic about the validity of 24 h reporting by children(Reference Baxter, Smith, Litaker, Guinn, Shaffer, Baglio and Frye32–Reference Lytle, Nichaman, Obarzanek, Glovsky, Montgomery, Nicklas, Zive and Feldman34). Lytle et al.(Reference Lytle, Nichaman, Obarzanek, Glovsky, Montgomery, Nicklas, Zive and Feldman34) validated 24 h recalls assisted by food records in third-grade children; they judged prospective pictorial representation as facilitating and this method valid for assessing the dietary intake of children as young as 8 years old for the purpose of group comparison.
Furthermore, the nutritional content of the recipes was determined on the basis of raw ingredients, without considering the losses due to heating treatment during cooking or frying. Thus we could be overestimating the nutritional contribution for the labile vitamins. In addition, there are limitations to the nutrient data obtained from the FCT. Finally, we generally selected our analyses to focus within the daily consumption of nutrients by meals within the various sub-samples of the study, rather than making any systematic effort to identify differences in total consumption or intake adequacy across subgroups.
What was noteworthy in our study was the relative parity for the energy contribution from the various mealtimes across the day of registry. When pooled across social groups (data not shown), the energy contributions from breakfast, dinner and snacks were within a few percentage points of one another (∼23 %), whereas lunch was marginally greater, providing 30 (sd 10) % of daily energy. A parallel lunchtime bulge in relative consumption was seen for protein and carbohydrate, with fat contribution remaining more evenly distributed among the four meals. In contrast, Matthys et al.(Reference Matthys, De Henauw, Devos and De Backer5) found a lower contribution of snacks to energy distribution among meals in Flemish adolescents. In their sample, breakfast and lunch accounted for 32 % and 31 %, respectively, of the day’s energy, whereas snacks contributed only 16 %, with the remaining 21 % coming from the evening meal(Reference Matthys, De Henauw, Devos and De Backer5). Another study in Belgian adolescents found a lower contribution of breakfast to energy distribution (15.7 % for boys and 14.9 % for girls) among meals(Reference Matthys, De Henauw, Bellemans, De Maeyer and De Backer35). Inequality of energy contribution among meals was also the rule in a sample of Swedish adolescents, aged 15 to 16 years, in which the percentage of energy from meals was 20 % and 21 % from breakfast, 16 % and 17 % from lunch, 26 % and 28 % from dinner and 37 % and 35 % from in-between meals in boys and girls, respectively(Reference Sjoberg, Hallberg, Hoglund and Hulthen7). These are European studies in slightly older children, but in the absence of analogous approaches applied to Central American or Latin American children, they represent the only basis for meal-pattern comparison for the juvenile situation.
Several studies have focused on breakfast skipping and breakfast quality. Good breakfast quality has been shown to relate to a better overall dietary pattern(Reference Chitra and Reddy3, Reference Matthys, De Henauw, Bellemans, De Maeyer and De Backer35). Irregular breakfast eating is related to negative lifestyle factors such as smoking, a higher percentage of energy from snack foods and lower intake of micronutrients(Reference Sjoberg, Hallberg, Hoglund and Hulthen7), and also mental distress and lower academic performance(Reference Herrero Lozano and Fillat Ballesteros36–Reference Torres, Carmona, Campillo, Perez and Campillo38). In our study, children rarely skipped breakfast (<1 %) and breakfast was the largest source of essential micronutrients.
With respect to micronutrient contributions in relation to the meal pattern, an additional contrast is seen between our findings and those of the Flemish series(Reference Matthys, De Henauw, Devos and De Backer5). In these Guatemalan third- and fourth-grade schoolchildren, snacks contribute less to the day’s intake of vitamin A, vitamin D, riboflavin, Ca, Fe and Zn than to daily energy. This is similar to the role of snacks’ micronutrient contribution in Finnish adults as reported by Ovaskainen et al.(Reference Ovaskainen, Reinivuo, Tapanainen, Hannila, Korhonen and Pakkala8). By contrast, in the Belgian adolescents, micronutrient intake generally bore a constant relationship to energy intake; there, micronutrient densities were apparently uniform across meals(Reference Matthys, De Henauw, Devos and De Backer5).
It is not sufficient, however, simply to know whether there is insufficiency, adequacy or excess of macro- or micronutrients intake from a diet. The meal-based context of nutrients can only be appreciated when dietary intake focuses on a meal-by-meal assessment of macro- and micronutrients as done here and in companion studies. On the practical side, moreover, knowledge of the nutrient distribution can be used by nutritional professionals as a fulcrum to plan interventions to redress either an excess or a deficiency of a nutrient, using a meal-based perspective in addressing any unhealthful aspect of dietary consumption. In this way, the pattern described would guide the strategy of public health interventions to redress any problems of insufficient or excessive intake of nutrients or dietary constituents. For reducing intake of food components that are associated with poor health, one must know when they are most likely to be eaten. Similarly, to redress deficiencies, one must know which meals are generally rich, or poor, in these nutrients.
The nutrients are consumed in the context of foods and beverages. The selection of foods in Table 1 reflects the preferences of children as well as the cultural norms of the caregivers and the availability, affordability and accessibility of the items in the marketplace. The ten leading items constitute between 68·0 % and 69·5 % of breakfast’s energy, between 48·7 % and 59·3 % for lunch, between 47·3 % and 59·8 % for dinner and between 36·3 % and 52·9 % for combined snacking. In general, the ten main sources accounted for slightly more meal energy for the LSES children, reflecting a lesser variety. Notable across genders and social class is the consumption of corn tortilla. It ranks high in energy contribution to both the midday and the evening meals. The Mayan cuisine, of course, is based on maize, as exemplified in the novel Hombres de Maíz (Men of Corn) by the Guatemalan Nobel Laureate, Miguel Angel Asturias(Reference Asturias39). The Mayan creation myth proclaims that the gods created Man from corn dough. This finding of a corn-rich diet in lower social classes has also been documented in Mexican adolescent girls(Reference Lozada, Flores, Rodriguez and Barquera40). Ready-to-eat breakfast cereals were predominant as a breakfast item, ranking consistently higher in the HSES sample than in its less affluent counterpart; this also confirms the finding for Mexican adolescents(Reference Lozada, Flores, Rodriguez and Barquera40). Several studies have mentioned the importance of ready-to-eat breakfast cereals in terms of nutritional benefits(Reference Chitra and Reddy3, Reference Galvin, Kiely and Flynn41, Reference Gibson42).
The unbalanced distribution of nutrient intake across mealtimes is subject to rapid change. For instance, the Central American and Dominican Republic Free Trade Agreement came into affect between Guatemala and the USA on 1 June 2006. If schoolchildren’s dietary habits evolve under the influence of a broader selection of foods in the marketplace, it could produce major changes from what is currently being eaten at the various meals. To the extent that the nutrient compositions of the new foods are likely to be different, wholesale redistribution of nutrient intake patterns could result. Micronutrients that are currently abundant in the breakfast fare, for instance, may become scarcer.
Latin American public policy has been informed in the past by the concept of critical nutrient density. In a region-wide consensus meeting held by the Cavendas Foundation in Caracas, Venezuela in 1986, an alternative approach to nutrient recommendations, based on nutrient density, was advanced(Reference Bengoa, Torun, Behar and Scrimshaw15). It proposed that a Latin American family eats as a unit; if all nutrients were adequate for every meal, then all members would simultaneously achieve their specific needs. The nutrient density focus for dietary analysis has grown in interest in recent years(Reference Darmon, Darmon, Maillot and Drewnowski43–Reference Backstrand45). The present study informs us is that micronutrient density varies by meal, such that changing the selection patterns for one meal, as with a school meal intervention, could differentially influence the whole day’s supply.
Conclusions
The children of both low and high social class of this urban centre in the Guatemalan highlands had remarkably equivalent and balanced distributions of energy across the four daily meal settings. Protein, carbohydrate and the various vitamins and minerals were generally concentrated into one or two of the meals. This produced unique nutrient densities among the meals. To the degree that certain problems of deficient intake, e.g. vitamin D, remain to be redressed, an understanding of how foods and food groups are combined – within meals and across the day – could be useful in designing the appropriate education and inducement for remedy. The present findings, therefore, place a mathematical face on the complexities of juggling a confluent series of public health aims. We agree with the comments of Perez et al.(Reference Perez, Hoelscher, Brown and Kelder46) that evaluating ‘differences in dietary intake and meal patterns by grade can provide readily accessible information to develop a needs assessment or intervention materials for children’. The meal-based approach may provide guidance to strategies to improve dietary balance in an era of coexisting energy overnutrition and micronutrient inadequacy.
Acknowledgements
The study was funded by grants from the American Institute of Cancer Research (AICR), the Sight and Life Organization, and the Department of Health Sciences, Vrije Universiteit, The Netherlands. We are most grateful to the Quetzaltenango Health and Education Authorities and to the students of the Universidad Rafael Landivar, Quetzaltenango for their help with data collection. Above all we are grateful to the staff of the twelve schools, and to the children and their parents or guardians, who participated so cheerfully. We are also grateful to Professor Jaap Seidell from the Vrije Universiteit of Amsterdam, for his collaborative partnership with CeSSIAM. The authors’ responsibilities were as follows: M.V. participated in data analysis, interpretation of results, writing and editing of the manuscript. G.M.-B. conducted the research as part of the Master programme for International Public Health and wrote the protocol, collected the data and participated in data analysis. L.D.J.K. provided statistical advice. C.M.D. and N.W.S. contributed in the design of the study, supervision, interpretation of results and writing of the manuscript. There were no conflicts of interest.
Appendix
Estimated 1 d intakes of energy, macronutrients and selected micronutrients by mealtime, gender and socio-economic status: third- and fourth-grade schoolchildren from Quetzaltenango, Guatemala, 2005
HSES, higher socio-economic status; LSES, lower socio-economic status; RAE, retinol activity equivalents; DFE, dietary folate equivalents.
