Energy and macronutrient intake and dietary pattern among school children in Bahrain: a cross-sectional study

This is a cross-sectional descriptive study involving qualitative and quantitative variables. The research was conducted from January/1999 to May/2001. The target population was Bahraini boys and girls of primary, intermediate, and secondary levels studying in public schools of the 11 populated regions of Bahrain.

The sample size was determined as per standard method recommended by the WHO Expert Committee (1995) for studies on nutritional status [21]. By applying the standard formula, the estimated sample size obtained was 2443. The sample size finally chosen was 2594; however 32 students declined to participate in the study giving a response rate of 98.7%. A multi-stage sampling design was chosen that combined multi-cluster and simple random sampling methods. Cluster sampling was used in two successive stages: first, for selection of the number of Bahraini students in proportion to the total number of students enrolled in each region; second for allocation of the schools in each region in proportion to the population size of the region. The allocated number of students from each level was chosen by randomly selecting few students from each class. There were 175 government schools, 88 for girls and 87 for boys enrolling a total of 104,189 Bahraini students (52,885 girls and 51,304 boys) from ages 6 to 18 years. Out of 54,782 primary level, 25,779 intermediate level and 23,628 secondary level students, 1386 (2.5%), 596 (2.3%) and 612 (2.6%) students respectively, were chosen for the study sample. The distribution of primary schools in the different regions ranged from 1 to 11, intermediate and secondary schools from 0 to 4. Selection of schools and number of students from each school is given in appendix 1/additional file 1 (decision by the editorial board)

Food frequency data was obtained for the entire sample along with other nutritional status parameters that were objectives of a larger study. For the energy and macronutrient component of the study, a sub-sample of 500 students (~20% of the total sample), was selected by following a similar multi-stage sampling design as mentioned above. Four students out of these refused to participate in the interview questionnaire on the 24-hour dietary recall, thus giving a response rate of 99% for this aspect of the study.

Permission to conduct the study was obtained from the Ethical committees of the Ministry of Health and Ministry of Education. Inclusion of a student in the study was determined by parental consent. All permission letters were sent to parents 3-4 weeks prior to the study allowing them adequate time to take a decision. Confidentiality of information was assured to the parents and the children. Parents were informed if a student was found to have a nutritional health problem as determined by the different assessment methods that were part of the larger study.

The study instrument was a questionnaire which included information on the age and gender of the child, a 24-hour dietary recall and food frequency. Age was recorded in completed years and verified by the researcher (NG) from school records. While the food-frequency questionnaire was self-administered by all the selected students in the classroom, information on the 24-hour dietary recall was obtained for a sub-sample by interviews that were conducted in a special room in the school premises. Interviews were conducted by a trained nurse, a nutritionist or one of the investigators (NG). Care was taken to conduct the interviews in a non-judgmental manner. For children ≤ 10 years of age, the interview for the 24-hour dietary recall was conducted in the presence of their mothers who were invited to attend the school on the day of the study. Most of the mothers in the girls' section attended the session but fewer mothers came in the boys' section. Mothers, who failed to attend the session, were contacted on telephone or home-visited by one of the interviewers to get data on the 24-hour dietary recall of their children. Effort was made to ensure that the child was present with his/her mother at home during the interview session. Data for the 24-hour recall was obtained for all food eaten on the previous day from wake-up to bed time. Information collected represented food consumed on a week-day. Week-end food intake may be variable and hence may not be representative of the usual dietary pattern for most days. Mothers and children were not informed about the method of data collection (24-hour dietary recall) prior to the study to avoid any bias which may have occurred due to change in the diet of children on the day before the interview.

During the interview, samples of local household dishes and utensils (different sizes of bowls, plates, cups, glasses, and spoons) were displayed to the child/mother. They were then shown pictures of common foods eaten in these dishes/utensils to indicate portion sizes consumed. One of the investigators (NG) had weighed different portions of various food items, placed them in the dishes/utensils and photographed them prior to the survey. After a student/mother as child proxy had indicated the portion size of food consumed, the corresponding weight of that food was recorded in the 24-hour dietary questionnaire. The latter had a list of 156 different food/drink items including: milk and dairy products, meat group, fast food types (local & international), bread and cereal group, mixed dishes eaten locally, vegetables and fruits, desserts and snacks, beverages (coffee & tea, juice, and fruit drinks), and other miscellaneous items. Some of these items were further classified, e.g.: type of milk (whole, low fat, or skim), type of meat eaten (boiled, fried, or grilled). There were open-ended questions for certain food items, e.g. type of fruit, vegetable and snack.

Mean energy and macronutrient intakes of the children were compared with the Dietary Reference Values (DRV) of the United Kingdom by age groups and gender [22]. Energy intake was compared to the Estimated Average Requirements (EARs) and macronutrient intake to the Reference Nutrient Intake (RNI) which is a value two notional standard deviations above EAR.

In addition, Energy percent (E%) from protein, carbohydrate and fat from total calories consumed, were assessed and compared to the DRVs recommended by the Committee on Medical aspects of Food Policy (COMA) [22].

To describe the potential effects of diet on serum lipid levels, dietary cholesterol levels and the Polyunsaturated to Saturated Fat Ratio (P: S ratio) were estimated. Dietary cholesterol intake of ≥ 300 mg and a P: S ratio of less than 1 was considered unacceptable [23].

The food items were grouped into 7 categories based on similarity of nutrient profiles [24]. These categories were: (1) meats & alternatives, including fish, chicken, red meat, eggs, and legumes; (2) milk & dairy products; (3) cereals, including rice, bread, cornflakes, and pasta; (4) fruits & vegetables including leafy and other vegetables; (5) sweets & snacks, including, cakes, chocolates, biscuits, candies, and crisps; (6) sauces & spreads, including ketchup, hot sauce, jams, and butter; (7) soft drinks, including all fizzy drinks (regular & diet), fruit juices, tea and coffee. Response categories were never, 1-2 times a week, 3-4 times a week, and 5-7 times a week (almost daily).

Dietary analyses were performed using Nutritionist 5 (First Data Bank, version 1.6, 1998) [25]. It is an American food data base and nutrient analysis software that includes over 17,000 food items with approximately 80 nutrients and nutrient factors. It was also possible through this software to enter the ingredients of Bahraini food recipes and obtain their nutrient content. Averages of energy and nutrient data by age groups and gender were then exported into Microsoft Excel 2000 (version 9.0). Statistical analyses were performed using Statistical Package for the Social Sciences (version 11.5, SPSS, 2002, SPSS Inc., Chicago, IL) for frequencies, means, standard deviations, and other descriptive and inferential statistics [26]. The chi-square, Kruskal-Wallis and Mann-Whitney tests were used where appropriate [27] A P-value less than 0.05 was considered to indicate statistical significance and a value of less than 0.01 as strongly significant.

Height and weight of these children was measured and BMI calculated as an objective of the larger study. Details of methods for the anthropometric measurements are described elsewhere [28]. We adopted the criteria of the World Health Organization for overweight status [29]. A comparison between the energy intake and the BMI status of students showed that the mean energy intake of the non-overweight (1775 Kcal/day) was significantly lower than those at-risk of overweight or overweight/obese (2182 Kcal/day) (p < 0.05). With these findings one can make some general assumptions on the validity of the students' reported energy intake.

Quality controls to improve reliability and validity included 1) a protocol that specified exact techniques for interviewing, recording and calculating results; 2) standardizing food portion sizes in commonly used dishes/utensils for quantification of foods and beverages consumed; 3) use of the American food data base, Nutritionist 5 [25]; 4) rigorous training of interviewers prior to the study and 5) involvement of mothers in the interview for children ≤10 years old.

Overall, the mean energy intake of Bahraini students ranged between 82.5% and 103% of the recommended EARs for different age groups and gender. The percentage of mean energy intake when compared to the EAR values decreased with increasing age of boys and girls. Overall the mean energy intake in boys was significantly higher than that in girls (p < 0.01) and this pattern was consistent at all ages. (Table 1).

Table 1 shows that the mean intake of protein exceeded the RNI at all ages and in both genders. Children 7-10 years old consumed about 2.5 times more protein relative to RNI values whereas adolescent boys and girls consumed 1.5 times the RNI values. The Protein E% from was 15.6% in boys and 15.2% in girls and these findings were close to the recommendations set by COMA [22] on energy intake from Protein (15%). The Protein E% did not change significantly with increasing age in boys and girls.

The mean daily intake of carbohydrate was significantly higher in boys (276.8 g ± 73.3) than in girls (232.4 g ± 60.7) (p < 0.01) (Table 1). This trend remained consistent for all age groups. In boys, carbohydrate intake increased significantly with age up to 18 years (p < 0.01) but in girls, intake increased up to 14 years and decreased thereafter. The carbohydrate E% was 52.9% in boys and 52.1% in girls which is close to the recommendations of COMA [22] for energy intake from carbohydrate (50%). While in girls the E% from carbohydrate did not change with age, in boys, it significantly increased with age (p < 0.05).

The mean daily intake of total sugars (intrinsic and extrinsic sugars) of 101.3 g ± 37.3 for boys and 89.1 g ± 36.1 for girls was high according to the maximum recommended intake of 60 g/day by the DRV of UK [22]. Sugar consumption increased while lactose intake decreased with age. The mean daily intake of dietary fiber was significantly higher in boys (13.5 g ± 5.8) than in girls (12.4 g ± 5.9) (p < 0.05) (Table 1). For adolescents these values apparently fell short of the current recommendations for fiber intake [30‐32].

The mean daily total fat intake for boys (75.7 g ± 22.5) was higher than that for girls (67 g ± 19.5) (p < 0.01) and this trend remained consistent in all age groups. Mean fat intake decreased in late adolescent girls. Overall, boys had significantly higher mean daily intakes of monounsaturated and polyunsaturated fats than girls (p < 0.01; p < 0.01) (Table 1). The mean P: S ratio was 0.6 ± 0.2 for both boys and girls. The mean daily intake of cholesterol was 287.2 mg ± 131.6 for boys and 245.1 mg ± 121.7 for girls (p < 0.01). The Total fat E% was 32.4% for boys and 33.6% for girls and this was slightly lower than the recommendations set by COMA [22] for energy intake from fat (35%). The Total fat E% did not change with age in girls but significantly decreased with age in boys (p < 0.01) (Table 1). The E% from saturated fat were slightly higher in girls (10.8%) than in boys (9.9%) (p < 0.01). The E% from saturated and polyunsaturated fats significantly decreased with age in boys (p < 0.01; p < 0.05)

According to the recommendations of the DRV standards of COMA for children above 5 years, saturated fat and total fat should not provide more than 11% and 35% of total dietary energy respectively [22]. We found that 50.2% and 37.8% of girls and boys respectively, exceeded these limits for saturated fat and 46.9% and 39.2%, respectively for total fats. Diets of 35.8% girls and 50.2% boys contained more than 300 mg of cholesterol per day which is the suggested upper limit [33, 34].

Analysis was done to determine the association of mean energy intake, E% of macronutrients, dietary cholesterol, sugar and dietary fiber with diets having Total fat E% at 3 levels (< 30%, between 30-37%, and ≥ 38%). Table 2 shows that while Carbohydrate E%, sugar and dietary fiber were significantly higher in the less than more dense-fat diets (p < 0.01; p < 0.01; p < 0.01), a reverse trend was observed for fatty acids (saturated, monounsaturated, and polyunsaturated) in the diet of girls and boys. Though not statistically significant, intake of dietary cholesterol (mg/1000 Kcal) was also found to increase with increasing density of total fat in diets. The P: S ratio remained below 1 at all 3 levels of fat-dense diets.

Table 3 shows that bread and cereal, meat group and sweets and snacks were the most commonly consumed daily food items among students. Only half of the children were consuming milk and its products and one-fourth of them were taking vegetables and fruits on a daily basis. However, regular soda drinks were consumed daily by approx. 50% of the children. Daily consumption of sweets and snacks was significantly more common in girls (64.2%) than in boys (47.5%) (p < 0.01). On the other hand, meat, milk and its products and tea/coffee were more popular with boys than girls (p < 0.01; p < 0.05; p < 0.01). Figure 1 shows that daily consumption of most food items in girls and milk and its products in boys decreased with age but intake of sweets and snacks increased with age in both genders.