 |
|
 |
|
 |
 |
|
 |
 |
 |
 |
 |
 |
 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
PEDIATRICS Vol. 104 No. 3 September 1999, pp. 468-475
,, and
From the * Department of Paediatrics and Child Health, Flinders
University of South Australia; and the Department of Paediatrics and
Child Health, Flinders Medical Centre, Bedford Park (Adelaide),
Australia.
 |
ABSTRACT |
|---|
Top
Abstract
Methods
Results
Discussion
References
|
|---|
Objective. To determine if dietary long-chain polyunsaturated fatty acids (LCPUFA) affect the growth of formula-fed infants relative to breastfed infants.
Methods. A randomized, double-blind trial of three formula-fed groups and a parallel reference group of breastfed infants was conducted. The three treatments were a placebo (no LCPUFA), docosahexaenoic acid (DHA) supplemented, and DHA plus arachidonic acid supplemented formulas fed for 12 months. Infant weight, length, head circumference, and fatty acid status were assessed at 6, 16, 34 weeks, and 1 year of age. Anthropometrics were repeated at 2 years of age.
Results. There were no differences between the randomized formula groups for weight, length, or head circumference even after adjustment for gender, gestational age, and the actual age at assessment. Post hoc regressions demonstrated a small negative association between DHA status at 16 weeks of age and weight at 1 and 2 years.
Conclusions. Dietary LCPUFA do not influence growth of healthy term infants to a clinically significant degree. Key words: docosahexaenoic acid, arachidonic acid, growth, infants, breast milk, infant formula, long-chain polyunsaturated fatty acids.
The dietary long-chain polyunsaturated fatty acid (LCPUFA)
to receive most attention during infancy has been docosahexaenoic acid
(DHA, 22:6n-3). DHA occurs in breast milk at levels that range from
0.1% to 1.0% of total fatty acids depending on the level in the
maternal diet.1 Most formulas for term infants are devoid
of LCPUFA such as DHA. Studies demonstrating higher levels of DHA in
the cerebral cortex of breastfed infants compared with formula-fed
infants2,3 have sparked many randomized trials of formula
supplemented with DHA to investigate neural outcomes during
infancy.4-7 However, the interpretation of these trials
has been difficult because of the concurrent addition of other LCPUFA,
such as arachidonic acid (AA, 20:4n-6) and eicosapentaenoic acid (EPA,
20:5n-3), that may also have clinically relevant functions. These n-6
and n-3 homologues are the indispensable precursors for the synthesis of eicosanoids that operate as regulators of cell and tissue functions throughout the body. Both AA and EPA are found to be exchangeable in
the membranes of most cells and EPA tends to produce eicosanoids of
attenuated activity compared with those produced from AA.
Our first randomized trial of formula supplemented with LCPUFA used an
oil that contained more EPA (0.58% total fatty acids) than DHA (0.36%
total fatty acids).4 Supplemented infants had an 8-fold
increase in erythrocyte EPA, a 2.5-fold increase in DHA, and a
reduction in AA status compared with infants fed the unsupplemented
formula. These fatty acid perturbations did not affect infant growth,
although the study had limited power.4 In contrast, a
trial in preterm infants using a similar fatty acid supplement has
reported decreased growth in preterm infants fed supplemented formula
relative to those fed placebo formula.8 Furthermore, it
was postulated that the reduction in infant AA status, as a consequence
of n-3 LCPUFA supplementation, was the factor responsible for poorer
growth.9 To assess the effect of the reduction in infant
AA status associated with n-3 LCPUFA supplementation, we undertook the
current trial in which two supplemental regimens were compared with
unsupplemented formula. One test formula contained DHA at a level of
0.35% total fatty acids and had minimal EPA; the other test formula
contained DHA and AA at a level of 0.34% total fatty acids.
There have been two other n-3 polyunsaturated fatty acid
intervention trials in term infants; one demonstrated no effect of supplementation on infant growth,6 whereas the other suggested decreased weight gain in supplemented infants compared with
those fed placebo formula.10 The current trial was
specifically designed and was adequately powered to address whether
dietary LCPUFA affect infant growth.
Study Participants
Healthy, white, term infants born at Flinders Medical Centre
between December 1993 and November 1994 were eligible for entry into
the trial. Infants were excluded if they were small-for-gestational age, there was evidence of congenital disease, if mothers had insulin-dependent diabetes, or had a history of drug or alcohol abuse.
To avoid any possibility of influencing the mothers' feeding choice,
only those who chose to formula-feed their infants and never attempted
breastfeeding were approached to enter the randomized trial of formula
feeding. Each mother and her family had the opportunity to discuss the
study with a research nurse. Written informed consent was obtained by
day 7 postpartum and before discharge from hospital. After consent,
infant-mother pairs were randomly assigned to one of three formulas.
The random number list was computer-generated and formula assignment
was according to sequentially numbered, opaque, and sealed envelope.
One of the investigators (M.A.N.) not involved in the day-to-day
clinical management of the trial prepared the randomization and
assignment schedule. Investigators and families were blinded to the
randomization.
A reference group of breastfed infants was also recruited for the
trial. Written informed consent was obtained in the same manner as with
formula-fed infants. Full breastfeeding was defined as no more than 200 mL of formula per week in the first 16 weeks of life and no more than
200 mL of formula per day between the ages of 16 weeks and 1 year.
Study Formulas
Nestec Ltd (Konolfingen, Switzerland) supplied the
formulas in powder form in 1-kg cans. Reconstitution instructions were identical, as was the packaging. All formulas had similar nutrient composition. The protein, fat, and carbohydrate content were 1.5, 3.4, and 7.6 g/100 mL, respectively. The fatty acid composition of each
study formula is shown in Table 1.
Formulas contained either no LCPUFA (placebo), or 0.35% DHA as total
fatty acids from tuna oil, or 0.34% DHA and 0.34% AA as total fatty
acids from an egg phospholipid fraction (Table 1).
TABLE 1
METHODS
Top
Abstract
Methods
Results
Discussion
References
Fatty Acid Composition (% Total Fatty Acids) of the Test Formulas and
Human Milk Fed to the Reference Breastfed Infants
Protocol
Between day 5 and 7 of life, infant and parental demographic information was collected4,11 and a 200-µL blood sample taken from infants by heel-prick for baseline analysis of plasma and erythrocyte phospholipid fatty acids. The blood sample was timed to coincide with the routine Guthrie screening test. Formula-fed infants commenced the study formula as soon as this blood sample was taken. Mothers of formula-fed infants were instructed to feed their assigned formula as the only milk until their infant's first birthday. Mothers of breastfed infants were encouraged to continue breastfeeding for as long as possible. The introduction of solid foods for both breastfed and formula-fed infants was discouraged before 4 months of age. Research nurses, dietitians, and lactation consultants were available to attend to feeding difficulties and nutrition enquiries.
Infants were assessed at 6, 16, and 34 weeks of age as well as 1 and 2 years of age. At each assessment infant growth was measured by weight, length, and head circumference. Information regarding infant well being and tolerance was collected at 6 and 16 weeks of age because these were the only assessment points in which formula was the sole source of nutrition for infants in the randomized trial. Blood samples were also collected by heel-prick at 6, 16, 34 weeks, and 1 year of age for the assessment of plasma and erythrocyte phospholipid fatty acids.
The Committee on Clinical Investigations (Ethics) at Flinders Medical Centre approved the trial protocol.
Methods
Growth Infants were weighed undressed on a calibrated Seca Baby Balance (Model 727, Seca, Hamburg, Germany). Length was determined in the supine position to the nearest 0.5 cm using an infant measuring mat. A nurse took the length measurement with the help of an assistant who held the infant in position against the head-board. Head circumference was measured at the largest occipitofrontal circumference to the nearest 0.1 cm with a nonstretch tape.
Fatty Acid Analyses Plasma and erythrocytes were separated by centrifugation and the erythrocytes were washed three times with isotonic saline. Plasma and erythrocyte lipids were extracted with chloroform:methanol12 and chloroform:propanol,13 respectively. The phospholipid fractions of both plasma and erythrocyte lipid extracts were separated by thin-layer chromatography, methylated, and quantified by capillary gas chromatography.14
Sample Size and Statistics
We hypothesized that a difference in growth because of LCPUFA treatment would need to be similar in magnitude to the weight differences often observed between breastfed and formula-fed infants to be clinically relevant. On this premise, it was calculated that a sample size of 25 infants per formula-feeding group would be sufficient to detect a mean difference in weight of ~700 g at 8 months and 1 year of age with a 0.05 level of significance and 80% power.
At each assessment age, comparisons of weight, length, and head circumference between randomized formula groups were made by analysis of covariance. Gender, gestational age, size at birth, birth order, postnatal age at assessment, and maternal smoking were considered as covariates. Because there were no differences between the growth parameters of the formula-fed infants at any age, all data of formula-fed infants were combined and compared with the growth of breastfed infants also using analysis of covariance. Only infants fully breastfed to 34 weeks of age were included in the analyses. Gender, gestational age, size at birth, birth order, postnatal age at assessment, and maternal smoking were considered as covariates in these analyses.
The association between measures of growth (weight, length, and head circumference) and infant DHA, AA, and EPA status in formula-fed infants was assessed using linear multiple regression. DHA, AA, and EPA were considered as possible independent variables because these were the only LCPUFA altered in the diets of formula-fed infants. Two models were constructed at each assessment age, one using the LCPUFA at the matching assessment age and the other using the 16-week fatty acid data as an indicator of maximal treatment-induced change. Gender, birth size, gestational age, birth order, actual age at each assessment, parental smoking, education, and social score were also considered as independent variables for the regression models. The independent variables were chosen for the regression analysis if they formed correlations with growth measures (P < .2), and were not collinear. All analyses were performed using SPSS for Windows 6.0 (SPSS Inc, Chicago, IL).
| |
RESULTS |
|---|
|
Top
Abstract
Methods
Results
Discussion
References
|
|---|
Study Sample
A total of 83 infants were randomly allocated to three formula
groups
28 infants to the placebo formula, 27 to the DHA-alone formula,
and 28 to the DHA + AA formula. A parallel, nonrandomized group of
breastfed infants (n = 63) was also recruited. All
infants were of similar gestational age at birth, birth order, and had mean Apgar scores of 9 at 5 minutes (Table
2). Gender was evenly balanced in all
groups except in the group fed the DHA + AA formula in which 64% of
infants were boys (Table 2). The rates of maternal smoking, the
mother's level of education, and social score were similar among the
randomized formula groups. Mothers of infants in the breastfed
reference group had lower rates of smoking (P < .05),
a more prestigious social score (P < .001), and
attained a higher level of education (P < .05)
compared with mothers of formula-fed infants (Table 2).
|
By 34 weeks of age, 15 out of 83 (18%) of formula-fed infants were withdrawn from the trial. Eleven infants ceased the study formula and their mothers' withdrew consent (5 from the placebo group, 3 from the DHA-alone group, and 3 from the DHA + AA group), 2 families moved interstate (both from the placebo group), 1 infant was withdrawn for an unrelated medical problem (DHA-alone group) and another for an unknown reason (DHA + AA formula group). Forty-six breastfed infants (73%) completed the trial to 34 weeks of age. Seventeen breastfed infants were withdrawn; 5 mothers reported they were too busy, 4 changed their minds about participation in the trial, 2 families moved interstate, 2 ceased breastfeeding and withdrew their consent, 1 infant was diagnosed with cataracts, and 3 withdrew for unknown reasons. At 6, 16, and 34 weeks of age, 45, 33, and 23 infants, respectively, were fully breastfed.
Among the randomized groups, the infants who completed the trial were similar in gestational age, Apgar scores, had mothers with similar levels of education and social scores as those who originally entered the trial (Table 2). There was a tendency, however, for more girls (12 out of 41) than boys (3 out of 42) and proportionately more smokers (8 out of 30) that nonsmokers (7 out of 53) to be withdrawn from the trial. In the nonrandomized, breastfed reference group, the 46 infants who completed the trial to 34 weeks of age had similar characteristics to the 63 infants initially enrolled. However, the subgroup of 23 infants who were fully breastfed for at least 34 weeks had mothers with lower smoking rates and higher levels of education (Table 2).
Between 34 weeks and 2 years of age, 7 were lost to follow-up. Five infants were from the DHA + AA formula group, 1 from the DHA-alone formula group, and 1 was from the placebo group. No infants from the breastfed group were lost to follow-up between 34 weeks and 2 years.
Infant Fatty Acid Profiles
The plasma phospholipid DHA levels of infants fed formula with DHA
(
0.35% as total fatty acids) were higher than breastfed infants
that received 0.20% DHA (Table 3).
Placebo formula-fed infants had the lowest level of plasma phospholipid
DHA that was 42% of the value of breastfed infants and 34% of
the DHA-treated formula groups. Plasma DHA of breastfed infants
remained similar to baseline (day 5) levels until 34 weeks of age
whereas DHA-treated formula groups displayed increases in plasma DHA
greater than baseline and placebo formula-fed infants demonstrated a
dramatic decrease from baseline (Fig 1).
Plasma phospholipid EPA levels were similar in infants fed the DHA
formula and breast milk but was lower in placebo and DHA + AA
formula-fed infants (Table 3). Among the formula groups, plasma 22:5n-3
was lowest in the group with highest dietary LCPUFA and highest in the
group with no dietary LCPUFA (Table 5). In contrast, the 22:5n-3 level
of breastfed infants (in which 22:5n-3 contributes 0.19% of total
dietary fatty acids) exceeded that observed in all formula-fed groups.
|
|
|
Within the randomized formula groups, infants fed the DHA + AA formula attained the highest plasma phospholipid AA levels, the placebo formula group had intermediate AA values, whereas infants fed the DHA formula had the lowest levels of plasma AA (Table 3, Fig 1). Although the plasma phospholipid AA of all infants decreased between day 5 and week 6 of age, the differences in plasma AA between groups were evident throughout infancy (Fig 1). The fall in AA (7.7% to 5.9% total fatty acids) was greater than the increase in EPA (0.2% to 0.5% total fatty acids) indicating that DHA as well as EPA can replace AA in plasma phospholipids. All fatty acid trends observed in infant plasma phospholipids were mirrored in infant erythrocyte phospholipids (data not shown).
Growth of Formula-fed Infants
There was no difference in weight, length, or head circumference
between the formula groups at any age, even after adjusting for gender,
gestational age, and postnatal age at assessment (Table
4). Regressions examining the independent
factors related to growth of formula-fed infants are shown in Table
5. Not surprisingly, the most consistent
factor to emerge at all time points for each of the growth parameters
was gender. Maternal smoking had a consistently negative effect on
infant length. Importantly, neither AA nor EPA was associated with any
of the outcome measures at any time points. When the models were
constructed using the fatty acid data matching each assessment time, no
LCPUFA emerged in any model. When 16-week fatty acid data were used,
DHA emerged as the only fatty acid predicting weight at 1 and 2 years
of age. The magnitude of this effect was small; 
118 g (95%
confidence interval [CI] 233, 3) at 1 year and 205 g (95% CI
41 to 368) at 2 years. At both ages, 16-week DHA status predicted
<2% of infant weight.
|
Growth of Breastfed and Formula-fed Infants
Figure 2 displays the weight length
and head circumference of fully breastfed and formula-fed boys and
girls. The reported differences in weight and length at 34 weeks, and 1 and 2 years of age were all within the limits of expected, standard growth15 and more consistent among girls than boys. The
95% CIs for differences in weight, of breastfed and formula-fed infants were as follows: 996 g, 151 g (P < .01, n = 91) at 34 weeks, 1267 g, 313 g
(P < .005, n = 88) at 1 year, and
1284 g, +52 g (P < .08, n = 84) at 2 years of age. Breastfed infants were also shorter than formula-fed
infants at these ages (2.4 cm, 0.2 cm [P < .05, n = 91] at 34 weeks; 2.5 cm, 0.2 cm
[P < .05, n = 88] at 1 year; 2.8
cm, +0.04 cm [P < .06] at 2 years of age).
|
Tolerance
The frequency of parents reporting infant restlessness, rash, vomiting, diarrhea, and constipation was similar in all formula-fed groups at both 6 and 16 weeks of age (data not shown).
| |
DISCUSSION |
|---|
|
Top
Abstract
Methods
Results
Discussion
References
|
|---|
The aim of our trial was to determine if LCPUFA treatment of formula-fed infants influenced growth. Our study induced large changes in infant levels of AA and DHA and these changes were maintained for at least 12 months. Despite this, we observed no differences in weight, length, or head circumference between randomization groups after adjustment for gender and other variables known to influence infant growth. Our trial was sufficiently powered to detect clinically relevant changes in weight and length, and obviously detected the well-known differential between breastfed and formula-fed infants and between boys and girls. Although our regression analyses suggested an association between 16-week infant DHA status and weight at 1 and 2 years, these regression models accounted for <2% of the infant weight at these ages. Furthermore, the CI for this association represents 20% of the effect one would expect as a result of breastfeeding. The relevance of this observation is unclear given the small effect size, the lack of association between DHA status and weight at other time points, and the fact that an association does not prove causality. However, it may be argued that a negative association between DHA and weight in late infancy and early childhood may be beneficial because it represents a shift toward the weight pattern of healthy, breastfed children.
Two other trials in term formula-fed infants using similar LCPUFA treatments also showed no differences in the growth of formula-fed groups but neither could they detect differences between breastfed and formula-fed infants.6,7 Despite the relatively large sample size of these trials, there are clearly issues of sensitivity in relation to growth measures. In contrast, Dewey et al16 with a similar sample size at 12 months of age reported a weight differential of ~700 g between breastfed and formula-fed infants.
Only one study has reported a relationship between growth and LCPUFA status in term infants.10 At 17 weeks of age, Jensen et al10 found an association between a range of n-6 LCPUFA (AA, 22:4n-6, 22:5n-6) and weight after correcting for birth weight. However, this association was not found at any other time points and these investigators did not report any influence of gender in the model used, which was a consistent predictor of growth in our data. In addition, the data of Jensen et al10 might not be comparable to our own because they varied the ratio of linoleic acid to alpha-linolenic acid and did not include LCPUFA in their infant diets.
There is controversy regarding the most appropriate growth for healthy infants. Studies have pointed out that the weight of breastfed infants during late infancy tracks below the 50th percentile of common reference (National Center for Health Statistics) growth charts and some differences in length between breastfed and formula-fed infants are also evident.16 Indeed, there have been attempts to construct reference growth charts for breastfed infants and a recommendation to use these growth charts as a reference for all infants.17 The growth differential between breastfed and formula-fed infants cannot be accounted for by LCPUFA. In other words, mimicking the DHA and AA status of breastfed infants does not result in a comparable growth pattern.
It is interesting to note that although reports from early studies in preterm infants suggested that dietary n-3 LCPUFA and perturbations in infant LCPUFA status inhibit growth relative to placebo formula-fed infants, more recent reports have not confirmed these finding. The studies reported by Carlson and coworkers8,9 used EPA-rich fish oils resulting in large reductions in AA status which was cited as a causative factor in growth inhibition. The failure of the later preterm studies to detect an effect of LCPUFA on growth has been suggested to be because of the balance of AA and DHA (together with low EPA) in these diets.18-20 It may also be that the growth of infants in these early studies was affected by factors other than the dietary fats. The large range of AA values recorded in the formula-fed infants in our study were similar to levels reported in preterm infants and yet we found no association between infant AA status and growth. Taken together with our data we can find little evidence for concerns about growth because of dietary fatty acids in term infants.
| |
ACKNOWLEDGMENTS |
|---|
This trial was funded by Nestec Ltd, Switzerland, and the Australian National Health and Medical Research Council (Public Health).
We thank Jenny Osmond, Mary Wooden, Cordula Blank, Jane Armstrong, Lyn Pullen, Ela Zielinski, Dani Dixon, and Dr Mark Hoby for their administrative and technical support.
| |
FOOTNOTES |
|---|
Received for publication May 26, 1998; accepted Jan 12, 1999.
Address correspondence to Robert A Gibson, PhD, Director, Child Nutrition Research Centre, Child Health Research Institute, Women's & Children's Hospital, North Adelaide, SA 5006, Australia. E-mail: rgibson{at}flinders.edu.au
| |
ABBREVIATIONS |
|---|
LCPUFA, long-chain polyunsaturated fatty acids; DHA, docosahexaenoic acid; AA, arachidonic acid; EPA, eicosapentaenoic acid; CI, confidence interval.
| |
REFERENCES |
|---|
|
Top
Abstract
Methods
Results
Discussion
References
|
|---|
This article has been cited by other articles:
 |
|
 |
|
  J T. Brenna, B. Varamini, R. G Jensen, D. A Diersen-Schade, J. A Boettcher, and L. M Arterburn Docosahexaenoic and arachidonic acid concentrations in human breast milk worldwide Am. J. Clinical Nutrition, June 1, 2007; 85(6): 1457 - 1464. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
L. M Arterburn, E. B. Hall, and H. Oken Distribution, interconversion, and dose response of n-3 fatty acids in humans Am. J. Clinical Nutrition, June 1, 2006; 83(6): S1467 - 1476S. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. Makrides, R. A Gibson, T. Udell, K. Ried, and the International LCPUFA Investigators Supplementation of infant formula with long-chain polyunsaturated fatty acids does not influence the growth of term infants Am. J. Clinical Nutrition, May 1, 2005; 81(5): 1094 - 1101. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 N. Auestad, D. T. Scott, J. S. Janowsky, C. Jacobsen, R. E. Carroll, M. B. Montalto, R. Halter, W. Qiu, J. R. Jacobs, W. E. Connor, et al. Visual, Cognitive, and Language Assessments at 39 Months: A Follow-up Study of Children Fed Formulas Containing Long-Chain Polyunsaturated Fatty Acids to 1 Year of Age Pediatrics, September 1, 2003; 112(3): e177 - 183. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
I. B. Helland, O. D. Saugstad, L. Smith, K. Saarem, K. Solvoll, T. Ganes, and C. A. Drevon Similar Effects on Infants of n-3 and n-6 Fatty Acids Supplementation to Pregnant and Lactating Women Pediatrics, November 1, 2001; 108(5): e82 - 82. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
P. Rump, R. P Mensink, A. D. Kester, and G. Hornstra Essential fatty acid composition of plasma phospholipids and birth weight: a study in term neonates Am. J. Clinical Nutrition, April 1, 2001; 73(4): 797 - 806. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. M. PhD, M. A. Neumann, K. Simmer, and R. A. Gibson A Critical Appraisal of the Role of Dietary Long-Chain Polyunsaturated Fatty Acids on Neural Indices of Term Infants: A Randomized, Controlled Trial Pediatrics, January 1, 2000; 105(1): 32 - 38. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
 |
||
|
||
Home | My Pediatrics | Journal Information | Current Issue | Past Issues | Subscriptions & Services | Contact Us Click here for faster international access © 1999 American Academy of Pediatrics. All rights reserved. |
||
|
||
) represent the median value of fully breastfed infants and
the open squares (
) the median value of formula-fed infants. Error
bars represent ±1 SD. The reference growth percentiles are taken from
National Center for Health Statistics data.