Journal of the New Zealand Medical Association, 05-September-2008, Vol 121 No 1281
Food frequency information—relationships to body composition and apparent growth in 4-year-old children in the Pacific Island Family Study
Elaine Rush, Janis Paterson, Vladimir Obolonkin
Worldwide patterns of diet and activity affect body composition and future health throughout the lifecycle. Obesity rates in New Zealand have risen sharply over the past decade. In 2002/2003 in a national survey of health1 using ethnic-specific body mass index (BMI) cut-off points 56% of New Zealand adults aged ≥20 years were classified as overweight (35%) or obese (21%) while more than 80% of Pacific adults were classified as overweight or obese.
In the 2002 National Children’s Nutrition Survey,2 more than 60% of Pacific children were classified as overweight or obese using the Cole cut-offs3 to define overweight and obesity; furthermore, the Pacific population in New Zealand are over-represented in adverse social4 and health statistics.1,5
These statistics have immediate implications for child health and wellbeing and longer term developmental consequences. Obese children are more likely to be obese in adulthood6 and obesity in early life is associated with cardiometabolic syndrome in adulthood.7
Energy intake exceeding energy expenditure (over-nutrition) will result in an increase in body fatness but children have an added requirement for specific nutrients to enable optimal growth.8 Growth, particularly at critical periods,9 is determined by nutrient intake and energy expenditure matching.10
Dietary recommendations throughout the life course include foods from each of the protein, cereal, fruit and vegetable, and dairy groups. Food patterns, related to the variety, frequency, and portion sizes of foods from each group determine nutrient quality and balance. Food frequency questionnaires may be used to determine key foods, dietary balance and patterns from a whole food perspective.
The food frequency questionnaire developed for New Zealand children in 2000 shows good short term repeatability11 and identified dietary patterns in the 2002 child nutrition survey.2 Pacific children were identified as eating more breakfast cereal, drinking less milk but more sugary drinks than European children.
The Pacific Island Family (PIF) Study was designed to increase knowledge of the health, psychosocial, and behavioural characteristics of Pacific peoples with recruitment of mothers (n=1376) of children born at Middlemore Hospital in South Auckland in 2000.12
Initial data showed that 6 weeks after the birth of the baby 40% of the Pacific Island families reported that sometimes food ran out due to lack of money.13 We have tracked the growth of these children to 4 years and show that compared with the World Health Organization (WHO) growth standard14 Pacific babies were born heavy, and over 4 years increased weight faster and between 2 and 4 years increased height faster than the reference breastfed child independent of pre and post natal factors. Maternal smoking decreased the rate of weight gain and children who were not breast fed gained weight faster.14
The aim of this investigation was to record the most frequently eaten foods, analyse the dietary pattern and to identify associations of food choices with body composition and growth characteristics at the 4-year measurement point of the PIF Study.
Survey—From the 1376 children recruited at birth, 1048 (76.2%) were retained at age 4 years.
The qualitative food frequency questionnaire used in the New Zealand Child Nutrition Survey (2002)2 was administered to the parent of each of 739 children as part of the 4-year assessment in the Pacific Island Family Study.
“How often over the last 4 weeks” was asked for 111 foods with options for other foods to be included in each section. Six foods were considered specifically Pacific: cooked green banana, cassava, taro, coconut cream, boiled corned beef, and canned corn beef. Frequency of consumption was then multiplied by a fraction per day (Table 1) to weight the responses so that they could be summated as frequency/day.
Table 1. Frequency of consumption of food and the weighting factor applied to standardise to a daily rate
Each food was classed as either a source of carbohydrate, fat, protein, dairy or fruit and vegetables (Table 2). For example pies, burgers, and sausage meats were categorised as protein, legumes/baked beans were also protein, caloric drinks were carbohydrate and spreads and sauces, convenience meals and snacks were classified by their major ingredient. Foods were then subclassified as nutrient high or low. The British nutrient profiling score15was used to discriminate foods according to their nutrient (and energy) density.
Table 2. Classification of the 111 foods in the children’s nutrition survey2 food frequency questionnaire by major nutrient or as dairy or fruit and vegetable and by nutrient and energy density15
Birth weight was obtained from the birth records. Body weight and height were measured at 2 and 4 years by using standardised equipment and procedures. BMI was calculated as weight in kilograms divided by squared height in meters.
At 4 years, direct measurements of resistance (R), reactance (X), impedance (Z), and phase (θ) to a 50 kHz signal using a bioimpedance analyser (Model Imp 4, Impedimed, Queensland, Australia) with a tetrapolar arrangement of self-adhesive electrodes (Red Dot 2330, 3M Healthcare, St Paul, MN, USA) were made.
These measurements were carried out with the child lying supine, the arms near but not touching the body and the legs abducted. The skin of the right hand and foot was swabbed with alcohol before the electrodes were placed. Source electrodes were placed on the dorsal surface of the foot over the distal portion of the second metatarsal, and on the hand on the distal portion of the second metacarpal.
Sensing electrodes were placed at the anterior ankle between the tibial and fibular malleoli and at the posterior wrist between the styloid processes of the radius and ulna. The child was lying still for at least 5 min before the measurements were made. The average of repeated measurements of R and X agreeing to within 1 ohm of each other was used in subsequent analyses. Resistance measurements were used to derive body fat percentage using the prediction equation of Rush et al.16
Statistics—Results are presented as means ± SD. Patterns of association between body composition and frequency per day of food groups and subcategories of nutrient density were investigated using bivariate Pearson correlation. Data were analysed using SPSS (version 13) software (SPSS Inc, Chicago, IL). P values <0.05 were considered significant.
All measurements were made and a food frequency questionnaire for the child completed by the parent for 355 girls and 384 boys. Table 3 shows the physical characteristics of the study population. On average, the z-scores for BMI, weight, and height are significantly above the ideal child.17 Male and female were similar and for purposes of this descriptive analysis have not been separated.
Table 3. Characteristics of study cohort
From the food frequency questionnaire bread and milk were the reported foods most frequently eaten (Table 4). Three choices were given for the “main type of bread you usually eat? tick one box”. White bread only was ticked by 77% and a further 11% had white and grain (wholemeal or mixed grain) breads ticked as well as white bread.
Only 11% ticked that only grain breads were consumed. Milk was consumed by 7% never or less than once a month. For the “kind of milk” there also were also multiple responses possible—85% of those who drank milk ticked standard milk, 11.5% light blue, and 3% each the trim and extra calcium, 1% consumed soy.
The mean total number of times that specific foods were reported as consumed by a child each day was 30. The first most frequently consumed 20 foods accounted for 37% of the frequency of eating and the next 20 for 21%—i.e. 58% of the total frequency was included in 40 items in the questionnaire.
Nutrient poor food including powdered drinks, noodles, tomato sauce, and potato crisps and corn snacks were in the top 20. Total milk including flavoured and food drinks was consumed at a similar rate to bread 1.63 vs 1.32 times in a day. More traditional Pacific food consumption was also investigated and on average was 5% of the dietary pattern. Taro was eaten on average 0.29 times a day, cassava 0.12, canned corned beef 0.22, boiled corned beef 0.13, fish 0.30, and cooked green banana 0.26. Foods high in white flour or sugar made up more than 20% of the diet.
Table 4. Forty most frequently eaten foods ranked by mean frequency of consumption each day
FV=Fruit and vegetable; P=Protein; F=Fat; C=Carbohydrate; D=Dairy; +Higher nutrient, lower energy; – Lower nutrient, higher energy.
Associations (Pearson r, p<0.05) between growth and body size characteristics with frequency/day of food categories were systematically examined and key aspects of the pattern of association detected are summarised in Table 5.
Birth weight was positively associated with the proportion of nutrient dense food (p=0.003), fruit and vegetable intake (p=0.0005) and negatively with the nutrient poor protein foods (p=0.02). The more times carbohydrate foods were consumed in a day the lower the birth weight (p=0.04).
Four-year weight, BMI, body fatness, and weight gain over the 4 years showed a consistent pattern not seen for the 2 -year old body composition variables. Higher frequency of consumption of fruit and vegetables and of fat foods was associated with less weight gain (p=0.05, 0.009) and a smaller BMI (p=0.04, 0.008). Total protein and dairy consumption were also positively associated with a larger body size (p<0.05).
The most influential foods on 4-year BMI were the relative frequency of consumption of all fat foods (-ve, p=0.008) and good protein (+ve, p=0.01) foods.
Table 5. Pattern of association between growth and body size characteristics and frequency of consumption of different food groupings (relative to total food frequency)
+ve means that there was a significant positive association and –ve a significant negative association. Blank cells indicate no association seen.
The recommended frequency of consumption fruit (2+ a day) and vegetables (3+ a day) for 2–12 y old children in NZ18 was compared to the frequency reported. For fruit 442 (60%) consumed fruit at least twice a day and the four top fruit were apples or pears, oranges or mandarins at ~0.8 serve day; banana consumption was similar.
Vegetables were eaten at least three times a day by 255 (35%) the most popular being tomato sauce, mixed vegetables, and carrots. Noodles were eaten at least once a week by 659 (89%) and 375 (50%) consumed breakfast cereal at least once a day. Rice was eaten at least once a day by 150 (20%).
Food patterns, growth and their associations have been examined in this cohort of Pacific Island children in South Auckland. Foods frequently eaten by these Pacific Island children have been identified and bread and milk are the foods most frequently consumed. Associations of the quality of the food pattern with body size and growth trajectory have also been demonstrated.
A higher proportion of fruit and vegetables in the diet is associated with a higher birth weight but lower BMI and weight gain over 4 years. Conversely protein and dairy foods, both good sources of protein, are associated with increased weight gain and BMI. A higher proportion of fat foods are associated with less weight gain. Each finding will be discussed in turn.
Bread and milk are staples of the NZ diet for children and adults alike.2,19 White bread and standard milk were consumed by the majority of children frequently—10% of the diet. These common food sources point to areas where nutrient density could be improved very easily—addition of whole grain flours to the bread and removal of saturated fat from the milk—small changes in composition with very little potential for change in taste and palatability would make a big difference to nutrient density and intake.
Price and availability would also need to be addressed to make this a viable intervention. Consumption of whole grains and less saturated fat, affordable price and access are all known to be associated with improved health.20
We have reported previously that at birth this cohort were not food secure13. We have also reported that at birth babies of non smokers were larger than those of smokers14,21 and that breast feeding was associated with a smaller body size at age 4.14
Higher fruit and vegetable intake is a marker of a more healthy (and maybe wealthy) lifestyle.22 In this study a higher birth weight was associated with a higher fruit and vegetable intake reported at age four. A lower BMI at age 4 was also associated with a higher fruit and vegetable intake.
This study is limited in a number of aspects including many confounders and unknowns. The food frequency questionnaire is not quantitative as portion size is not asked but it does quantify the number of times in a given period that foods are consumed.
At age 4 the reported foods have been ranked and a 4-week pattern analysed. The data was collected over a year—there is no correction for seasonal availability. Similarly there is no correction for socioeconomic status, body size for amount of food consumed or physical activity to name a few key confounders. But it is unlikely that there were large changes in the family food environment over the last four years23 and the associations with birth weight support this.
The foods most frequently eaten agree with findings in the national children’s nutrition survey and endorse the validity of the questionnaire. Specific ethnic foods such as taro were included in the analysis but were only 5% of the foods chosen. Conversely high sugar and white flour foods comprise 20% of the foods.
Food classification and consequent analysis or scoring food patterns is difficult given the complexity of identifying “good” and “bad” foods. Consumption of foods high in whole grain breads and cereals, fruits and fruit juices, and raw vegetables, and low in processed meat, butter, cheese, margarine, and meat in an adult European population is predictive of low prospective weight change.24
Responses to food frequency questionnaires rank a population well25 but are fraught with inconsistency. Although the median and the mean of the reported number of food items per day look reasonable (26 and 30 respectively) the maximum and minimum were 7 to 150 food items per day which reflects the difficulty of getting accurate information from questionnaires related to food consumption.
We did not exclude any data to avoid adding more bias. But broad food group and nutrient density classification methods such as that used here15 are useful to identify areas where dietary patterns and frequency of consumption are associated with increased body fatness or other measures of health, growth, and development.
We have provided a list of foods most commonly consumed by Pacific children at 4 years and identified some potential for change in the food supply. We demonstrate that measures of nutrient quality of foods and within food groups of diet over the previous 4 weeks are associated with growth over the past 4 years and present body size. The issue of protein and dairy foods being associated with larger body size merits further investigation.
Competing interests: None known.
Author information: Elaine Rush, Professor of Nutrition, Centre for Physical Activity and Nutrition Research, Faculty of Health and Environmental Sciences, Auckland University of Technology, Auckland; Janis Paterson, Professor, Director Pacific Island Family Study, School of Public Health and Psychosocial Studies, Faculty of Health and Environmental Sciences, Auckland University of Technology, Auckland; Vladimir Obolonkin, Research Scientist, LIC, Hamilton
Acknowledgements: The Pacific Islands Families (PIF) Study is funded by the Foundation for Research, Science & Technology, the Health Research Council of New Zealand, and the Maurice & Phyllis Paykel Trust. The authors gratefully acknowledge the families who participate in the study as well as other members of the research team. In addition we thank the PIF Advisory Board for their guidance and support.
Correspondence: Professor Elaine Rush, Centre for Physical Activity and Nutrition Research, Faculty of Health and Environmental Sciences, Auckland University of Technology, Private Bag Box 92006 Auckland, New Zealand 1142. Fax +64 (0)9 9219746; email: email@example.com
issue | Search journal |
Archived issues | Classifieds
| Hotline (free ads)
Subscribe | Contribute | Advertise | Contact Us | Copyright | Other Journals