Nutrition during early childhood has an important and long-lasting influence on child growth and cognitive development(Reference Benton1–Reference Northstone and Emmett3). Brain growth and development is most rapid and critical during the first 2 years of life, but development of the frontal lobes continues throughout early childhood(Reference Myers, Thurston and Pearson4, Reference Bryan, Osendarp and Hughes5). Certain nutrients make important contributions to optimum brain and neurological function including Zn, Fe, iodine, folate, Se and long-chain PUFA(Reference Bryan, Osendarp and Hughes5). Many of these nutrients are involved in metabolic processes such as the production of enzymes or essential cofactors, some of which are central to brain growth and development(Reference Bourre6, Reference Benton7). Zn is recognised as important for growth and neurodevelopment, and low intakes have been associated with behavioural problems such as attention-deficit hyperactivity disorder in children(Reference DiGirolamo, Ramirez-Zea and Wang8). Fe and iodine are critical nutrients in early brain development, but deficiencies are commonplace among school-aged children in both developing and developed countries and are a major cause of impaired motor skills and cognitive function(Reference Benton1, Reference Melse-Boonstra and Jaiswal9). Folate and vitamin B12 are necessary for the methylation pathway, an integral process to the synthesis of DNA and neurotransmitters, while Se contributes to thyroid hormone synthesis and may affect mood(Reference Benton1, Reference Bourre6). There is evidence that DHA (n-3 PUFA) and arachidonic acid (n-6 PUFA), long-chain PUFA which are most important during fetal and infant development, can offer postnatal benefits to child performance and behaviour(Reference Ryan, Astwood and Gautier10–Reference McNamara, Able and Jandacek12).
A poor diet during childhood may be associated with low intakes of these nutrients and evidence from longitudinal studies suggests that dietary habits formed in childhood can often continue into later life(Reference Mikkila, Rasanen and Raitakari13, Reference Mikkila, Rasanen and Raitakari14), thus highlighting the importance of assessing childhood diet. Assessment of nutrient adequacy, which has been defined as intake of nutrients which either meets or exceeds the appropriate Dietary Reference Value (DRV) for that population without being excessive(15), is a useful tool in dietary surveys of children.
An additional interesting aspect of the relationship between nutrition and child development described by Strain et al. (Reference Strain, Bonham and Duffy16) is the suggestion that nutrients in fish may ameliorate the potentially neurotoxic effects of methylmercury exposure. This issue is of special concern within high-fish-eating populations such as the Republic of Seychelles. For this reason, the Seychelles Child Development Nutrition Study (SCDNS) is investigating nutritional intakes and status of pregnant women and their children. The aim of the current study was to analyse dietary intakes and investigate adequacy of potentially important nutrients for cognitive development among 5-year-old Seychellois children.
The present study is a follow-up analysis of children born to mothers initially recruited for the SCDNS. In 2001, a total of 300 pregnant women were recruited within Mahé, the main island of the Republic of Seychelles. This cohort was considered representative of the Seychelles population owing to study design and recruitment, which have been described in detail previously(Reference Bonham, Duffy and Robson17). As part of that longitudinal study, offspring of the women were re-evaluated in 2006 when they were approximately 5 years old. There were complete dietary data for a total of 229 children who participated in the follow-up study. The study was reviewed and approved by the appropriate Research Subjects Review Boards of collaborating partners, the Ministry of Health of Seychelles and the University of Rochester.
Dietary intakes were assessed using two weighed 4 d food diaries, of which one was supplied to parents/guardians and the other to teachers, who were requested to keep records of each child's food intake over four consecutive days (two weekdays and two weekend days) when at home and at school. Diet diaries were available in both English and Creole languages. Both parents/guardians and teachers were supplied with Ravenscourt weighing scales to assist with recording portion weights of food and drinks consumed by the children, including information on amount, brand, recipe and preparation method used. When it was not possible to weigh meals received at school, researchers and trained nutritionists in the Seychelles assisted with calculating portion sizes of meals eaten by the children during school hours. Nutritionists in the Seychelles reviewed the diaries within one week of completion, with any errors and omissions being clarified with children and their families. Researchers also visited the family home/school frequently to monitor the weighing and recording of foods by respective parents/guardians and teachers. Food and nutrient intakes were analysed using the nutrient database package WISP version 3·0 (Tinuviel Software, Warrington, UK), which contains data from the UK Nutrient Databank(18) and was previously supplemented with food composition and recipe data for foods typical to the Seychelles diet(Reference Robson, Bonham and Duffy19). In the present study, the WISP database was further augmented with food composition data of foods commonly consumed by Seychellois children. Owing to incomplete PUFA composition data for all fish species, it was not possible to analyse PUFA intakes accurately within the study.
Dietary data were coded into seventeen food groups and expressed as g/d: cereal & cereal products; milk & milk products; eggs; vegetables; fruit; sugars, preserves & snacks; white fish; fatty fish; crustaceans; molluscs; fish products & dishes; red meat; white meat; meat products & dishes; fats and oils; beverages; and other foods (which included the sum of herbs & spices, nuts & seeds, soups, sauces & miscellaneous foods). Fish were classified as being white (lean) or fatty based on knowledge of their total fat and fatty acid content(20). For five children, dietary records were available only for three of the four surveyed days; therefore all calculations for these children were modified according to the three recorded days. Children's heights and weights were measured according to standardised procedures by nutritionists, using calibrated equipment at the Child Development Centre in Seychelles, and BMI was calculated as [weight (kg)]/[height (m)]2.
Assessing validity of energy intake
In order to assess the validity of reported energy intake (EIrep) in dietary surveys of children, it is important to use appropriate cut-offs for both age and sex in order to account for variability in growth and physical activity(Reference Zlotkin21). In the present study, the method described by Black and Cole(Reference Black and Cole22) for evaluating energy intakes was employed. Energy expenditure (EE) was estimated using published sex- and age-specific equations which were developed using doubly labelled water data measured in children of a similar age(23).
A low active physical activity level (PAL = 1·4–1·6) was assumed for all children based on doubly labelled water data(Reference Rennie, Jebb and Wright24), which allowed a constant corresponding physical activity coefficient value of 1·16 to be used in these equations. The estimated energy requirement (EER) was calculated as the EE + 84 kJ/d (20 kcal/d)(23). The validity of EIrep was then assessed against EE by use of the EIrep:EE ratio and by obtaining a 95 % confidence limit for the agreement between these two measurements. In order to calculate the confidence limit, the within-subject CV for EIrep and EE were initially calculated and the mean value for both the CV of EIrep (CVEI) and the CV of EE (CVEE), taken as 8·2 % for the present study(Reference Black and Cole22), were used to calculate a combined CV (CVC):
where d is the number of days of dietary assessment. The acceptable range, or confidence limit, for EIrep:EE was then calculated as:
Those children with a ratio below this limit were then classified as under-reporters, while normal reporters were classified as those children with an EIrep:EE within the acceptable range.
Assessment of nutritional adequacy
Nutritional adequacy was evaluated by applying a statistical equation to adjust the distribution of nutrient intakes reported from the 4 d assessment period, thus correcting for intra-individual variability which creates inaccuracies in dietary assessment(25). Adjusted nutrient intakes were then compared with the Estimated Average Requirement (EAR) according to UK DRV(25, 26) for healthy children of the same sex and age. In the absence of specific nutritional guidelines existing for the Seychelles population, the UK DRV have been used and these DRV are considered to be appropriate guidelines for this population(Reference Bonham, Duffy and Robson17). The EAR is known as the best reference value with which to examine the prevalence of inadequate nutrient intakes within a group(23). This method has been described as the most accurate compared with alternative methods of assessing nutrient adequacy(Reference Tabacchi, Wijnhoven and Branca27). In the case of Na, K, Cu, Se and iodine, the Recommended Nutrient Intake (RNI) has been used for comparison in the absence of an established EAR. The RNI, however, can only be used to estimate nutritional adequacy for intakes above, but not below, this level(25). Furthermore, the EAR cut-point method cannot be used to assess energy intakes, as this approach would break a statistical assumption of the method, namely that the nutrient intake is not related to requirement(Reference Gibson28). Therefore we compared energy intakes against the group mean EER instead. Nutrient inadequacy was evaluated as the percentage of children with adjusted intakes below the EAR.
Statistical analyses were conducted using the SPSS for Windows statistical software package version 18·0 (SPSS Inc., Chicago, IL, USA). Data for all variables were initially tested for normality prior to analysis and descriptive data were expressed as means and standard deviations or medians and 5th, 95th percentiles as appropriate. Nutrient intakes were adjusted using a method described by Beaton(Reference Beaton29), where within- and between-individual variations in nutrient intakes were calculated before applying the equation:
These intakes were then compared with the UK EAR for each nutrient by calculating the percentage of children with intakes below this level. Independent t tests and one-sample t tests were used to determine differences in intakes between boys and girls and between the group including and excluding under-reporters, respectively. In all instances, a two-tailed P value of ≤0·05 was considered statistically significant.
The group consisted of 118 boys and 111 girls. Their mean age was 5·62 (sd 0·31) years and their mean height and weight values (1·15 (sd 0·05) m) and 19·99 (sd 3·79) kg) were above the 75th percentiles for height and weight according to the Seychelles growth curve data for children. The mean BMI of the boys (15·05 (sd 1·97) kg/m2) and girls (15·02 (sd 2·08) kg/m2) was between the 50th and the 75th percentiles for BMI, placing them within the normal height for weight category for their age group. Using specific height and weight data for each child, the mean EE was estimated at 6121 (sd 586) kJ/d (1463 (sd 140) kcal/d). The mean EIrep and daily variance in energy intake were 4841 (sd 1364) kJ/d (1157 (sd 326) kcal/d) and 27·75 %, respectively, which gave a 95 % confidence limit of the agreement between EIrep:EE of 0·63–0·95. According to this cut-off, 21 % of children were classified as under-reporters (n 49) and 79 % as normal reporters (n 180).
Energy and nutrient intakes
Table 1 presents mean dietary intakes of energy and nutrients for the entire cohort (n 229) and separately for the group excluding under-reporters (n 180). When mean nutrient intakes were compared between these groups, intakes of energy, protein, fat, K, Zn and thiamin were significantly higher (P < 0·05) in the group which excluded under-reporters. Under-reporting did not have an effect on intake of any other nutrient and there were no significant differences in nutrient intakes between boys and girls. We therefore carried out the following analyses of nutrient adequacy on the entire cohort (n 229) in order to include as many valid dietary entries as possible and because the detected degree of under-reporting impacted only some nutrient intakes.
%E, percentage of total energy intake.
*P < 0·05 indicates a significant difference in nutrient intakes between all children including under-reporters (n 229) and children excluding under-reporters (n 180), as assessed by a one-sample t test.
Table 2 presents adjusted nutrient intakes for the entire cohort (n 229) and for boys (n 118) and girls (n 111) separately, showing the proportion with inadequate intakes below the UK EAR. Although only adult guideline intakes exist for total fat and carbohydrate (33 % and 47 % of daily energy) within the UK DRV, adjusted mean intakes of all children met these guidelines. Every child met recommended protein requirements, while ≥86 % of children were deemed as having adequate intakes of many nutrients, namely K, Ca, Mg, Fe, Se, thiamin, riboflavin, folate and vitamins B12 and C. Adjusted mean energy intakes of boys (4769 (sd 524) kJ/d (1139·84 (sd 125·15) kcal/d)) and girls (4759 (sd 633) kJ/d (1137·43 (sd 151·41) kcal/d)) fell below the EER by 439 and 423 kJ/d (105 and 101 kcal/d), respectively. Compared with the EAR, ≥92 % of children were found to have inadequate intakes of Zn, vitamin A and niacin and a further 30 % of boys and 23 % of girls, respectively, had vitamin B6 intakes below the EAR. A large proportion of children (≥86 %) were found to have intakes below the RNI for Cu and iodine. The RNI for vitamin D among this age group is set at zero, as it is expected that children's requirements will be met through exposure to sunlight(Reference Poskitt, Cole and Lawson30). Adequacy of vitamin E intakes was also not assessed in this age group given that there exists only a safe intake level for vitamin E for infants in the UK(26). There were no significant differences in the number of children with inadequate intakes between boys and girls; however, adjusted mean intakes of vitamin E were significantly greater in girls than boys (P = 0·05).
EAR, Estimated Average Requirement; %E, percentage of total energy intake.
a,bMean values within a row with unlike superscript letters were significantly different between boys and girls, as assessed by independent-samples t test (P = 0·05).
*For energy, the estimated energy requirement (EER) is given based on the children's sex, age, height, weight and activity level.
†Population guide for adults only.
‡UK Recommended Nutrient Intake in the absence of an EAR.
§As retinol equivalents.
∥As niacin equivalents.
Consumption of food groups and contribution to nutrient intakes
Table 3 shows the children's reported daily intake of food groups and the contribution of food groups to energy, protein and nutrients which are potentially important for cognitive development. Children consumed most food groups, with cereal products, milk products and fruit being the major food groups in the diet for all consumers. These food groups were consumed by children in daily median (5th, 95th percentile) quantities of 56·25 (33·97, 98·74), 60·83 (0, 151·00) and 70·58 (0, 198·67) g/d, respectively. All children consumed vegetables, with a daily median (5th, 95th percentile) amount of 23·19 (9·42, 76·28) g/d. Noticeably, fatty fish and white meat, of which chicken was the major component, were consumed in similar quantities (median (5th, 95th percentile): 26·00 (0, 77·00) and 33·50 (0, 120·75) g/d) by a similar proportion of children (74 % and 78 %) and provided similar amounts of protein (8 % and 9 %), respectively. Cereal products and milk products contributed the largest amounts of energy, protein and nutrients to the diet, and the most commonly consumed cereal and milk products were white rice and cheddar cheese. These two food groups combined provided over 50 % of Fe, Cu, Zn, iodine and folate intakes. Vegetables also provided at least 10 % of Fe, Cu, vitamin B6 and folate intakes. The median daily fatty fish consumption of 26·00 g/d provided 14 % of Se intake and 10 % of vitamin B12 intake.
P5, 5th percentile; P95, 95th percentile.
*Includes soups, sauces & miscellaneous, nuts & seeds, herbs & spices, custom foods.
The present study is the first report of dietary intakes of young children living in the Republic of Seychelles and shows that the diet of 5-year-old children within the SCDNS provides adequate amounts of Ca, Mg, Fe, Se, folate and vitamin B12. These nutrients are recognised as important for growth and development in children(Reference Benton1, Reference Nelson, Naismith and Burley31, Reference Singh32) and suggest that this group of children in Seychelles, unlike those in many other countries(Reference Serra-Majem, Ribas-Barba and Perez-Rodrigo33–Reference Lambert, Agostoni and Elmadfa35), are not at risk of impaired cognitive development caused by Fe-deficiency anaemia. Furthermore, their diet provides sufficient intakes of folate and vitamin B12, which are necessary to maintain neurological integrity(Reference Bryan, Osendarp and Hughes5). An adequate Se intake may also have benefits for child development based on evidence that it can protect against possible neurotoxic effects of methylmercury to which all fish-eating populations are exposed(Reference Ralston and Raymond36). Given the children's frequent consumption of cereal products and milk products, fruits, vegetables and fish, one might expect these children to meet the EAR for this range of micronutrients. Cereal products, including rice and fortified breakfast cereals, were the largest contributors to intakes of Fe, Se and folate in this cohort, as has been reported in other populations(Reference van den Boom, Serra-Majem and Ribas37, 38). Milk products were the second largest contributing food group to nutrients, not surprisingly as milk imported to Seychelles is powdered and fortified with these micronutrients.
Reported energy intakes among ≥86 % of boys and girls in this cohort were approximately 418 kJ/d (100 kcal/d) below the mean EER, suggesting that children were not eating enough to meet their energy requirements. These intakes are lower than energy intakes reported for children of a similar age living in European countries and the developing country of Papua New Guinea(Reference Lambert, Agostoni and Elmadfa35, Reference Ferro-Luzzi, Norgan and Durnin39). Bovet et al. (Reference Bovet, Chiolero and Madeleine40) previously reported that underweight was common among Seychellois children, with approximately 20 % of boys and girls attending public schools being classed as grade 1 thinness according to international standards(Reference Cole, Flegal and Nicholls41). Despite an overall rise in the prevalence of obesity and overweight in Seychelles(Reference Bovet, Chiolero and Madeleine42), it is possible that underweight and associated low energy intakes are still an issue among young children. Subsequent follow-up of these children would be required to determine whether the low energy intakes reported by the present study compromise children's growth.
While many nutrients were found to be consumed in adequate amounts, at least 80 % of children had suboptimal intakes of Zn, vitamin A and niacin, and a further 23 % of girls and 30 % of boys were not meeting recommended intakes of vitamin B6. Deficiency of vitamin A can affect visual development and growth(Reference West43), deficiency of niacin can affect psychological functions(Reference Singh32) and deficiency of Zn is associated with poor attention span and impaired motor development(Reference Hubbs-Tait, Kennedy and Droke44). For a large proportion of children reported Cu and iodine intakes were also below the RNI; however, in the absence of an EAR these low intakes cannot be interpreted as inadequacies(25), but neither should they be disregarded based on the essential role of iodine in neurodevelopment(Reference Melse-Boonstra and Jaiswal9). While it is likely that the low energy intakes detected in this cohort could account for these low nutrient intakes, it is also possible that the UK DRV are not appropriate standards with which to compare dietary intakes of Seychellois children. It was particularly unexpected that Seychellois children would have low intakes of the above nutrients in view of their average fish consumption of 42 g/d; an intake which is notably higher than the UK total fish intake of 20 g/d among 4–10-year-olds(38, Reference Sirot, Dumas and Leblanc45). Fish is rich in many nutrients including iodine, Zn, Cu, vitamin A and n-3 long-chain PUFA(46). However, the intakes of several nutrients in the present study could have been underestimated as a result of incomplete nutrient composition data in the current nutritional database for all fish species consumed in Seychelles(Reference Bonham, Duffy and Robson17). Indeed, for this same reason, it was not possible to investigate long-chain PUFA intakes in the current study. This possible underestimation should also be taken into consideration when interpreting the low intakes of some nutrients found in the present study.
Several steps were taken in our study to reduce the limitations which can be associated with dietary assessment in children(Reference Livingstone and Robson47) and to strengthen the validity of the data presented, including the use of weighing scales to measure portion size and the statistical techniques used to detect dietary under-reporting and to remove intra-individual variability in nutrient intake. However, it was not possible to weigh all school meals consumed and although portion sizes for children were calculated and used in this instance, this approach may have impacted or affected the accuracy of dietary values obtained. The lack of specific DRV for the Seychelles may also be considered a potential limitation of the present study. As this cohort continues to be longitudinally assessed within the SCDNS, it would be of interest for future research to assess if these patterns of nutrient intake persist as children age and whether they are related to cognitive measures in these children.
Overall, the results of the present study indicate that this group of 5-year-old children in the Seychelles has adequate intakes of many nutrients deemed to be important for growth and cognitive development (protein, Ca, Mg, Fe, Se and B vitamins). Intakes of energy, Zn, vitamin A and niacin were found to be below this age group's requirements, but this may possibly reflect inaccuracies in our nutritional database. These findings on nutrient intakes of this group of children within the SCDNS should be valuable from a public health perspective with regard to implementing nutritional guidelines for children in the Seychelles population.
This study was supported by the US National Institute of Environmental Health Sciences, National Institutes of Health (grants 5-R01-ES010219, R01-ES015578 and 2-T32-ES007271); the European Union (contract EU FP6-2004-FOOD-3-A PHIME, Public Health Impact of long-term, low-level Mixed Element Exposure in susceptible population strata); and by the Government of Seychelles. There are no conflicts of interest. A.J.M. prepared the manuscript and analysed the dietary data. M.S.M. and M.P.B contributed to the dietary analysis. M.P.B., E.M.M., C.E., S.R. and J.F. collected and reviewed dietary data in the Seychelles. T.W.C., G.J.M., P.W.D., C.F.S., J.M.W.W. and J.J.S. were responsible for the study design. All authors read the manuscript and contributed to its finalisation. The authors thank Juliette Henderson and nurses at the Ministry of Health for all their work with the participants in the Seychelles, and also acknowledge the participation of all study participants involved.