 
|
|
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| 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.
MethodsSurvey—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.
ResultsAll 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%).
DiscussionFood 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: erush@aut.ac.nz
References:
|
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| | Current
issue | Search journal |
Archived issues | Classifieds
| Hotline (free ads) Subscribe | Contribute | Advertise | Contact Us | Copyright | Other Journals |
|