Nutrient Content of Consumed Elementary School Lunches: A Pilot Study from Sweden

Please note that this study was published before the SY2014-15 implementation of the Smart Snacks Nutrition Standards for Competitive Food in Schools, as required by the Healthy, Hunger-Free Kids Acts of 2010. As such, certain research relating to food in schools may not be relevant today.

Swedish law states that all students are entitled to a free, nutritious lunch, composed according to National Food Agency (NFA) recommendations, every day for all nine years of elementary school (The Swedish Parliament & Department of Education, 2010). The lunch should provide 25–35% of daily nutritional and energy requirements (National Food Agency [NFA], 2011). A guide for the translation of these recommendations into actual food items was published in 2003 (NFA, 2003). There are no further recommendations for schools other than that the food should be nutritious and the menus composed using NFA guidelines.

A recent pilot study on school lunches (Rosander et al., manuscript in preparation) showed that the macronutrient content and distribution of lunches were generally in line with NFA recommendations. Micronutrient levels were also in line with NFA recommendations, with the highest levels found in meals composed of the largest amount of organic food items, and levels of sodium chloride were generally higher than advised in the school lunches. Students could choose freely from a nutritious lunch buffet containing vegetables and salads, hot dishes (including a vegetarian alternative), bread, and milk or water. However, fruits were only offered in small amounts as part of entrees or salads. Pupils with allergies or those whose diets, for example excluded pork were offered alternative dishes with corresponding nutrient content. All students were accompanied by teachers at lunch time who were supposed to inspire them to make healthy food choices, spend enough time at the table and not waste food, which is a problem in Sweden as well as in other countries (NFA, 2013; Smith & Cunningham-Sabo, 2013). The NFA (2011) has stated that “Children who eat a lunch that meets their needs with regard to macro- and micronutrients perform better in school”. Studies also suggest that both breakfast and lunch school meals are important for student learning and cognitive function (Kleinman et al., 2002).

In addition to energy content, the overall macro- and micronutrient composition of school meals is also of great importance. The proportions of different macronutrients have a significant effect on both satiety and mental performance (Wurtman et al., 2003). Nutritional status is now considered a critical factor for brain function, with recent research supporting the argument that a nutritious lunch improves student behavior and concentration during afternoon lessons (Dani, Burrill, & Demmig-Adams, 2005). Deficiencies in essential fatty acids may also be linked to a number of different disorders such as ADHD (attention deficit hyperactivity disorder) (Dani et al., 2005).

At present the nutrient contribution from school lunches for students in Sweden is not well understood, and the students’ actual lunch intake is unclear. In 2003 the NFA undertook a national survey of food and nutrient intake in Swedish children using food frequency questionnaires. The results revealed that although meal energy distribution was adequate, saturated fat intake was above the recommended level of 14% of energy. While there currently is no NFA recommendation for sodium consumption in children, the student intake of 1–3 g (5–7 g salt) was considered twice the “desirable level”.

In Sweden, students are able to eat ad libitum from a free lunch buffet composed according to NFA recommendations. This contrasts with the U.S., where the overall aim is increased amounts of whole grains, fruits, and vegetables. Also in the U.S. lunches are designed to contain a certain number of calories (adapted for different age groups) and meals are not always free. Nevertheless, one overall aim seems to be the same for both countries: to provide a nutritious lunch containing increased amounts of vegetables.

The aim of this study was to investigate the average nutrient content of school lunches consumed by healthy ten-year old students eating a diet without restrictions, in the context of present Swedish recommendations regarding food composition and intake.

Methods

Students 
Three elementary schools (denoted 1–3) from a city in southern Sweden, all with identical lunch menus, were randomly selected to be included in the study. Fourth-grade children at these schools agreed to participate in the study by submitting informed consent forms signed by their parents. A questionnaire addressing students’ opinions of school food such as whether they were hungry during the day and if they wasted food was completed by 136 of the 152 fourth-grade students at the participating schools. Eighty-three percent of the students reported that they were hungry in the morning, and 59% indicated that they liked the food served for school lunches. Sixty-one percent stated that they rarely wasted food, but sometimes did so because they were too full or simply did not like the food.

Students whose meals were analyzed were selected based on several criteria, including the absence of food allergies and consumption of special foods. Based on this information, one girl and one boy in the fourth grade (aged 10 years) were randomly chosen from each school, representing a total of six students. Each participating student was asked to state on how many occasions she or he was physically active during leisure time. Students were supposed to spend at least 15 minutes of their 50 minute lunch break at the lunch table. The time each student spent on each meal were recorded by the accompanying dietitian. Data regarding body mass index (BMI), physical activity level, time spent eating, and amount and type of drink consumed were obtained during the week.

Lunch Collection 
School menus (Figure 1) were planned by a dietitian using NFA food database values (NFA, 2011). Another dietitian who was present during lunch instructed each student to eat as she or he normally would and simultaneously collected an identical portion to that selected by the student (double portion method). At the end of the meal, the dietitian removed from these portions an amount of food identical to that not consumed by the student. The (identical) portions consumed during the week were then collated to form one composite lunch sample for each student (a total of six different student lunches) that was stored frozen until analysis. The double portion method (Keys & Kimura, 1970) can be employed to overcome problems associated with the study of fourth-grade children, such as the low accuracy issues reported when using 24-hour dietary recalls (Baxter, Royer, Hardin, Guinn, & Smith, 2007). The double portion method reduces recall problems since study participants are not required to remember anything. Participants in the present study received no prior information as to what nutritious food is; the analyses thus describe only the actual nutrient content of the lunches consumed.

 

Day	Hot Meal Alternatives
Monday	Pancakes, Root Vegetables, Cranberries Vegetarian Taco Stew with Rice Minced Vegetables, Ketchup
Tuesday	Fish Gratin with Dill, Shrimp and Mashed Potatoes Sausage Casserole with Rice Parsnip Gratin
Wednesday	Meatballs, Sauce, Potatoes Hot Dog, Macaroni, Ketchup Greek Pie
Thursday	Minestrone Soup, Bun, Cheese Hungarian Goulash with Potato Vegetarian Spring Rolls, Bulgur, Soy
Friday	Tacos Chicken Nuggets with Curry Sauce and Rice Potato Pancakes with Lingonberries

 

Figure 1. Menus for lunches served in Schools 1–3, Spring 2011
Note: At every school the children could choose freely from a buffet containing hot vegetables (peas, broccoli, corn) and cold vegetables (grated carrots, cucumber, tomatoes) leaf lettuces, hot meal alternatives, bread, and milk. The same menu was used in all schools, but School 3 prepared lunch for School 3 and lunches delivered to School 1, and School 2 prepared lunches on site.

Nutrient Analyses
All collected meals were frozen immediately after collection and stored at -20 °C (-4 °F). Composite samples were sent to an accredited laboratory in Sweden for analyses, with each composite sample examined in triplicate. Samples were analyzed for protein, total fat, saturated fat, monounsaturated fat, polyunsaturated fats (omega-3 and omega-6), retinol, ß-carotene, vitamin E, vitamin C, folate, thiamin, riboflavin, calcium, magnesium, sodium, iron, zinc, water, and ash content, with results expressed as mean daily intake for the week. Carbohydrate and total energy levels were also calculated using the values for protein, total fat, water, and ash. Levels of retinol (RE) were calculated as ß-carotene (µg/12) since the former were below the limit of detection (<21 µg) in all samples.

Beverages were not included in the composite sample or laboratory analyses since they are not considered part of the meals in terms of which nutrient recommendations are based. However, they were instead weighed and recorded separately with the nutritive value of milk obtained from the NFA database. Student lunch contents were compared to the NFA recommendations for nutrient intake for ten year olds (NFA, 2011).

Results And Discussion

This pilot study describes the actual energy and macronutrient content, as well as their distribution in terms of energy, micronutrient content and density of lunches consumed by six students at three Swedish elementary schools during Spring 2011. Nutrient intake was assessed in terms of total intake, and quality in terms of energy distribution and micronutrient density. This enabled the comparison of meals of different mass (g), while also revealing whether portion size or composition should be changed. The precision of the chemical analyses was satisfactory, with the coefficient of variation (standard deviation/mean*100) ranging between 3 and 5% for protein, fats, water, ash, vitamin E, and riboflavin, between 5 and 10% for ß-carotene, folic acid, calcium, magnesium, and vitamin C, and between 10 and 15% for sodium, iron, and zinc.

Students
In Table 1, student lunches are denoted G1–G3 for girls and B1–B3 for boys. NFA guideline values are included for reference. The upper limit for the ‘normal’ weight of ten year old children is BMI<20; the BMI values of students participating in the present study ranged from 12 to 21. However, the individual with BMI 21 was more muscular than the other students and was therefore considered of ‘normal’ weight despite being 5% above the recommended value. Average eating times varied between 12–17 minutes, with those students spending a shorter time eating also consuming fewer g and kcal. The overall daily mean weight of consumed meals ranged from around 150 to just under 400 g. As a consequence, energy intake differed substantially between students. Whereas four participated two or three times in organized physical activity during their leisure time, the remaining two were not physically active. All students were able to eat freely from the buffet, and yet only two ate the amount expected. These students came from the same school and were both physically active during leisure time.

Energy and Macronutrients
Three of the students, including the two non-active individuals, received less than 50% of recommended energy from their lunches (Table 1). This low energy intake could cause hunger, which in turn may affect learning. Hungry children tend to exhibit a greater number of behavioral, emotional and academic problems, including elevated levels of aggression and anxiety (Kleinman et al., 2002). Severe protein deficiency can also have pronounced effects on cognition and behavior in children (Dye, Lluch, & Blundell, 2000; Fischer, Colombani, & Wenk, 2004; Kleinman et al., 2002), and thus it is essential that children eat a sufficient amount. In the present study, protein consumption was adequate only in the two meals containing recommended amounts of energy, although recommended percent of energy values for protein were observed in all students’ lunches.

Table 1. Student Characteristics and Energy/Macronutrient Content of School Lunches Consumed at Three Swedish Elementary Schools
	NFAa	School 1		School 2		School 3
		G1b	B1b	G2b	B2b	G3b	B3b
BMIc	<20	18.0	12.0	17.4	18.0	18.0	21.1
Eating time (min)		12	12	16	15	17	15
Physical activity/week		2–3	2–3	0	0	2–3	2–3
Weight of portion (g)		156	187	173	158	382	383
Milk consumption (g)	200d	118	0	44	0	140	200
Energy (kcal)	637	210	406	313	282	686	588
Energy (kcal)	637	263	406	335	282	764	687
Protein (g)	23	8.6	18	11	10	24	21
Protein (g) b	23	13	18	13	10	29	28
Fat (g)	25	7.5	19	15	11	22	20
Fat (g) b	25	9.3	19	16	11	24	23
-SFA (g)	7	3.0	7.4	5.2	4.0	8.0	7.2
-SFA (g) b	7	4.2	7.4	5.6	4.0	9.3	9.2
-MUFA (g)	7–11	3.6	9.2	6.7	5.3	10	9.8
-Omega 3 (g)	0.7	0.11	0.46	0.27	0.28	0.57	0.95
-Omega 6 (g)	2.8–6.3	0.56	1.61	1.76	1.58	3.0	2.7
Carbohydrates (g)	80–95	26	38	33	33	93	76

^aNational Food Agency recommendations (33% of daily intake) are included for reference (NFA, 2011).
^bStudents are identified as girls (G1, G2, G3) and boys (B1, B2, B3).
^cBMI groups for 10-year-old children: normal weight BMI <20, overweight BMI 20–24, obese BMI > 24. Definition of BMI for underweight children not included.
^d200g milk of the 500g recommended by NFA, which is also used when planning lunches.

Although meal fat content varied from 7.5 g to 22 g (Table 1), none exceeded the upper limit for a 637 kcal meal of 25 g total fat or the upper limit recommended for percentage of energy from fat. SFA levels were, however, above recommendations in the two meals containing sufficient amounts of energy (G3 and B3). All meals contained recommended levels of MUFA, which for a 637 kcal meal should be between 7 and 11 g, while only one of the six student meals (B3) satisfied guidelines for omega-3 fat intake, which should be approximately 0.7 g in a meal of 637 kcal. The two students undertaking no organized physical activity during their leisure time (G2 and B2) also consumed meals containing the highest percent of energy as fat, whereas those meals containing the recommended (and highest) amount of energy were also characterized by the lowest proportion of fat (G3, B3). Two of the students’ meals (B1 and B3) reached the guideline for percentage of energy from omega-3 fatty acids.

Carbohydrates should, according to NFA aims, provide 50–60% of energy, which in a 637 kcal meal is equivalent to about 84 g. Only one (G3) of the six lunches reached this value; the average proportion of energy derived from carbohydrates varied between 44–57% of energy. While all six meals met guideline levels with respect to the average proportion of energy derived from protein (14–19% of energy; recommended 10–20 % of energy), half were below and half were above the recommended upper limit of 35% of energy for the average proportion of energy derived from fat. Therefore, to meet Swedish nutritional recommendations, some fat should be replaced by carbohydrates in five of the meals. This replacement may be problematic, however, since levels of vitamin E (from e.g. vegetable oil) were already low in the studied lunches.

The observed results regarding macronutrient distribution are in line with those of a previous Swedish study examining children’s energy and nutrient intake (NFA, 2003). In both studies recommended levels of protein were obtained (NFA recommendation: 23g/ 637 kcal). However, the subjects consumed too much saturated fat (NFA recommendation: 7g/ 637 kcal) and sodium (NFA recommendation: 0.5g/ 637 kcal), and an insufficient amount of complex carbohydrates (NFA recommendation: 84g/ 637 kcal). One possible source of fat, sodium and protein is ready-made food such as fish fingers and chicken nuggets. These meals are popular with the students and are served as lunch alternatives (Figure 1). Variations in the ratio of energy obtained from carbohydrates to fat and protein in one meal can affect the body’s glycemic index (GI), ghrelin response and feeling of satiety during the following meal (Lomenick, Melguizo, Mitchell, Summar, & Anderson, 2009).

Micronutrients 
The children’s lunches were characterized by considerable variation in micronutrient content (Table 2). Larger portions contained more of most nutrients, with the exception of vitamin C and iron. Only one (B1) lunch contained the recommended amount of vitamin C. Two lunches were characterized by less than half of recommended levels of seven micronutrients (retinol, vitamin C, riboflavin, calcium, magnesium, iron, and zinc), with four particularly low in iron.

Table 2. Micronutrient Content of School Lunches Consumed by Students at Swedish Elementary Schools as a Percentage of NFA Recommendations
	NFAa	School 1		School 2		School 3
Micronutrient		G1b	B1b	G2b	B2b	G3b	B3b
Retinolc	198 (RE)	20	9	6	7	14	15
Vitamin E	2.45 (a-TE)	40	82	51	58	113	124
Vitamin C	17 (mg)	88	182	28	5	36	60
Folate	66 (µg)	33	33	26	48	315	244
Folate (incl. milk)d	66 (µg)	50	33	30	48	329	262
Thiamin	0.365 (mg)	140	104	79	52	274	416
Riboflavin	0.43 (mg)	20	17	22	17	30	46
Riboflavin (incl. milk)d	0.43 (mg)	57	17	35	17	76	107
Calcium	297 (mg)	23	33	26	23	35	56
Calcium (incl. milk)d	297 (mg)	47	33	40	23	93	134
Magnesium	92 (mg)	25	47	32	33	96	56
Magnesium (incl. milk)d	92 (mg)	49	47	37	33	114	134
Sodium	0.51 (g)	165	184	149	102	308	300
Iron	3.6 (mg)	31	57	27	20	81	86
Zinc	3.1 (mg)	45	100	35	31	110	126
Zinc (incl. milk)d	3.1 (mg)	71	100	42	31	126	155

^aNational Food Agency recommendations (33% of daily intake) are included for reference (NFA, 2011).
^bStudents are identified as girls (G1, G2, G3) and boys (B1, B2, B3).
^cRetinol (RE) content calculated using ß-carotene data (µg/12)
^dNutrients for which 200 grams of milk significantly raised the level of the nutrient have been described as “nutrient incl. milk”.

Two samples (G3 and B3) contained recommended levels of vitamin E, folate and zinc, but not of vitamin C and iron; these lunches also contained recommended amounts of riboflavin, calcium and magnesium, when they were consumed with milk. Dairy products represent one of the main dietary sources of riboflavin, which means that children who are lactose intolerant or allergic to cows’ milk should consume other riboflavin-rich foods. None of the lunches contained recommended levels or densities of iron, while only two (G3 and B3) met guideline amounts for zinc. This is significant since iron deficiency is a worldwide problem (World Health Organization, 2013) and may affect enzymes involved in cognition and behavior (Muñoz & Humeres, 2012).

No lunch was characterized by recommended retinol densities (RE/1000 kcal, calculated from ß-carotene), while only one (B3) contained advised vitamin C and magnesium densities (Table 3). However, all meals met guideline densities for vitamin E and thiamin. Most lunches were closer to recommendations in terms of their nutrient densities than amounts, indicating that four of the children simply needed to consume more of the same food.

Table 3. Micronutrient Densities (weight/1000 kcal) in School Lunches Consumed by Students at Swedish Elementary Schools as a Percentage of NFA Recommendations
	NFAa	School 1		School 2		School 3
Micronutrients	Weight/1000 kcal	G1b %	B1b %	G2b %	B2b %	G3b %	B3b %
Retinolc	335 (RE)	55	13	11	15	12	15
Vitamin E	3.77 (a-TE)	122	131	106	132	108	138
Vitamin C	33 (mg)	213	225	45	9	26	53
Folate	188 (µg)	56	29	29	60	162	147
Folate (incl. milk)d	188 (µg)	82	29	33	60	167	153
Thiamin	0.50 (mg)	475	183	183	133	292	517
Riboflavin	0.59 (mg)	71	36	57	43	43	57
Riboflavin (incl. milk)d	0.59 (mg)	179	36	93	43	136	143
Calcium	419 (mg)	75	58	58	58	36	67
Calcium (incl. milk)d	419 (mg)	147	58	92	58	120	163
Magnesium	147 (mg)	74	71	66	71	89	103
Magnesium (incl. milk)d	147 (mg)	143	71	74	71	103	129
Sodium	0.75 (g)	500	289	311	232	289	332
Iron	6.70 (mg)	79	76	47	39	63	79
Zinc	4.61 (mg)	143	166	78	74	106	144
Zinc (incl. milk)d	4.61 (mg)	224	166	92	74	125	174

^aNational Food Agency recommendations (33% of daily intake) are included for reference (NFA, 2011).
^bStudents are identified as girls (G1, G2, G3) and boys (B1, B2, B3).
^cRetinol (RE) density calculated using ß-carotene data (µg/12)
^dNutrients for which 200 grams of milk significantly raised the level of the nutrient has been described as nutrient incl. milk

Although, no extra sodium (as salt) was available on the school dining tables, all lunches were generally too high in sodium, with densities 2–5 times higher than the recommended 0.75 g/1000 kcal. This may potentially indicate a student preference for ready-made foods such as chicken nuggets and fried fish. A sodium intake twice the recommended amount has been noted for Swedish 10 year olds (NFA, 2013). Since this might be a risk factor for raised blood pressure in later life (Scientific Advisory Committee on Nutrition, 2003), sodium content and consumption should be reduced in school lunches.

Food and Nutrition Quality and Awareness
The school food served was nutritious if consumed in sufficient amounts (Rosander et al., manuscript in preparation). However, none of the actually consumed lunches reached recommended levels of vitamin C, and all were low in ß-carotene and folate. NFA (2011) guidelines promoting increased fruit and vegetable consumption (to meet the recommended 500 g per day for both adults and children aged 10 and above) are highly relevant for the students in the present study. Although fruit and vegetables represent good sources of these nutrients, fruits were rarely served during lunch at the investigated schools, and vegetable consumption was only half of that recommended for 10 year old children (NFA, 2013). However, in a questionnaire conducted prior to the study, approximately 85% of the fourth-grade students reported eating vegetables for lunch at least 2–3 times per week, while the same percentage knew that vegetables form part of a healthy diet. Despite this, the six students whose lunches were analyzed apparently consumed less than the recommended amount. The Swedish National Food Survey (NFA, 2003) also found that children consumed insufficient fruit and vegetables (only half the amount recommended). However, such low intake is not unique to Sweden, but is rather a world-wide trend with average daily consumption by children across the European continent being only 80 g (European Nutrition and Health Report, 2009). Also, data regarding the fruit and vegetable intake of 6,513 children aged 2–18 in the US were recently compared to recommendations outlined in the 1999–2002 National Health and Nutrition Examination Survey and the 2005 US Dietary Guidelines for Americans. According to that study, the leading source of total fruit was 100% fruit juice, while that for total vegetables was French fries. Of children aged 6–11, 26% met recommendations for fruit intake and only 16% met recommendations for vegetable intake (Lorson, Melgar-Quinonez, & Taylor, 2009).

Conclusions And Application

Three of the six students obtained less than half of the energy recommended by the NFA, while all analyzed lunches were low in several nutrients commonly found in fruit and vegetables, such as vitamin E, vitamin C, magnesium, and ß-carotene. The students’ lunches were generally closer in quality (weight/1000 kcal) than quantity (weight) with respect to NFA guidelines, with four of the six needing to consume more food in order to reach recommended amounts. Previous results (Rosander et al., manuscript in preparation.) showed that it was possible for students to consume a nutritious lunch, provided that the portion size was large enough and included sufficient amounts of vegetables, as well as a glass of milk. All schools offered nutritionally equivalent alternatives for students allergic to milk. Although students at these schools could serve themselves as much as they wanted of the various food items on the menu, they may not possess sufficient knowledge or motivation to choose a healthy lunch or have the appetite to eat a full portion. Also their average time spent on eating (12–17 min) may have been too short, or their appetites not sufficient to satisfy nutrient needs. There might be a need to design and serve more nutritionally dense lunches for students eating small portions. For those students not eating sufficient amounts, it is especially important to stimulate appetite. The presence of adults in the lunch room is also important and might help to reduce distractions experienced during lunch time.

Discussions about not providing soft drinks and sweets at schools have led to the removal of these items in several Swedish schools. However, students are allowed to bring snacks such as fruits and sandwiches. Therefore, it is possible that the students not eating enough had eaten other foods prior to school lunch. It would be better if morning snacks generally contained less filling food. Snacks eaten at school should be fruit, vegetables, or if needed, whole grain sandwiches with appropriate toppings.

The consumed lunches varied considerably in terms of their micro- and macronutrient content. Although the proportion of energy derived from protein was sufficient, the protein content was too low in those lunches containing insufficient amounts of energy. The fact that the proportion of SFA should be decreased and that of carbohydrates increased is indicative of a need for increased fruit, vegetable, and starch consumption. This is important since, according to NFA, fruits and vegetables are the foods that are most wasted (NFA, 2013).

All students in the present study consumed too much sodium. A survey examining the food intake of Swedish children (NFA, 2003) found that their sodium intake is twice or more the recommended level. This could be reduced by decreasing the use of ready-made foods such as chicken nuggets and fish fingers in school lunches, and serving fish and chicken dishes prepared at the school kitchens. Since breads contain various amounts of salt, the type of bread should be carefully chosen.

Due to the limited sample size, it was not possible to statistically test differences. Another limitation of the study is that no information was obtained on what other foods and snacks students had consumed prior to eating school lunches. Furthermore, the period of food sampling was short, only one week. Further studies on more students, including information on breakfast and morning snacks, are necessary to validate the results of the present pilot investigation and give more specific information on school lunches in Sweden.

However, this study describing the nutrient content of six school lunches consumed by Swedish elementary students provides preliminary data that may be useful when planning further research. Since recommendations are given for a group and not for individuals, there may be a need to design and serve more nutritionally dense lunches for students eating small portions. Also, knowledge of food and nutrition is very important for good and sustainable eating habits. As part of the curriculum, information on nutrition should be provided to all students and staff, and, if possible, parents.

Acknowledgments

This work was supported by Landskrona municipality, KRAV (Sweden), the AAA Foundation (Gothenburg, Sweden), the Swedish Dairy Association, Lantmännen and Santa Maria AB (Sweden). We thank the children and Landskrona schools participating in the study for their cooperation.