Objective. To determine whether dietary long-chain polyunsaturated fatty acids (LCPUFA), such as docosahexaenoic acid (DHA) and arachidonic acid, affect visual evoked potential (VEP) acuity of formula-fed infants, relative to breastfed infants. A secondary objective was to assess the effect of LCPUFA on Bayley's mental developmental index (MDI) and psychomotor developmental index (PDI).
Methods. Formula-fed infants were randomly allocated, in a double-blind manner, to either a placebo (no LCPUFA;n = 21), DHA supplemented (n = 23), or DHA+arachidonic acid supplemented formula (n = 24). Infants were fed their assigned formula from the first week of life until 1 year of age. A parallel reference group of breastfed infants was recruited and followed (n = 46). Infant VEP acuity was assessed at 16 and 34 weeks, and Bayley's MDI and PDI were assessed at 1 and 2 years of age.
Results. There were no differences among the randomized formula groups for VEP acuity at either 16 or 34 weeks of age. Breastfed infants had better VEP acuity at 34 weeks of age, but not at 16 weeks, compared with all formula-fed infants. Bayley's MDI and PDI were similar in the 3 formula-fed groups at 1 and 2 years. Breastfed infants had higher MDI scores than formula-fed infants at 2 years of age even after adjusting for environmental variables.
Conclusions. LCPUFA supplementation did not influence VEP acuity development in these well-nourished, formula-fed infants.
- docosahexaenoic acid
- arachidonic acid
- visual evoked potential
- breast milk
- infant formula
- long chain polyunsaturated fatty acids
- LCPUFA =
- long-chain polyunsaturated fatty acids •
- DHA =
- docosahexaenoic acid •
- EPO =
- evening primrose oil •
- EPA =
- eicosapentaenoic acid •
- GLA =
- γ-linolenic acid •
- PDI =
- psychomotor developmental index •
- AA =
- arachidonic acid •
- VEP =
- visual evoked potential •
- logMAR =
- log of the minimum angle of resolution •
- MDI =
- mental developmental index •
- HSQ =
- home screening questionnaire •
- ANCOVA =
- analysis of covariance •
- CI =
- confidence interval •
- ALA =
- α-linolenic acid
The long-chain polyunsaturated fatty acid (LCPUFA) docosahexaenoic acid (DHA, 22:6n-3) has been implicated in the development of the brain and neural networks. The frontal1and parietal2 cortices of infants fed human breast milk have higher concentrations of DHA than do the cortices of those fed infant formula. This observation has been attributed to the fact that human breast milk is a rich source of DHA, whereas most infant formulas contain no DHA. This has resulted in speculation that adding DHA to infant formulas will result in improved neural function. This hypothesis has been central to several randomized clinical trials of formula feeding in attempts to mimic the beneficial effects of breastfeeding on early childhood development. Our initial trial was the first to demonstrate that healthy term infants randomly allocated to a formula with DHA have improved visual acuity development, compared with infants fed placebo, unsupplemented formula.3 Some subsequent trials have supported these findings,4 ,5whereas others have demonstrated no effect of DHA supplementation.6 ,7
In our previous trial,3 we compared a formula containing .36% DHA as total fatty acids with an unsupplemented formula. The DHA in the test formula was provided by microencapsulated oil that was an equal mixture of fish oil and evening primrose oil (EPO). Therefore, this supplement also provided .58% total fatty acids as eicosapentaenoic acid (EPA, 20:5n-3) and .27% as γ-linolenic acid (GLA, 18:3n-6). In light of the reports that related high-EPA fish oils with reduced growth in preterm infants8 and a lower Bayley's Psychomotor Developmental Index (PDI)9 compared with unsupplemented formula-fed preterm infants, we undertook the current trial. The purpose of our study was to compare growth and developmental indices of infants fed formula with .35% DHA and .10% EPA or .34% DHA and .34% arachidonic acid (AA, 20:4n-6) with a placebo, unsupplemented formula, and a breastfed reference group. In this article, we report the outcomes of the visual evoked potential (VEP) assessments at 16 and 34 weeks of age and the Bayley's Developmental Indices at 1 and 2 years. The detailed growth and infant fatty acid data are reported elsewhere.10
Infants were eligible for the trial if they were white, born at term (>37 weeks' gestation) and were appropriate weight for gestational age. Infants were excluded if there was evidence of congenital disease and complications during pregnancy.10Mothers who consented and made a decision to formula feed their infant were randomly allocated to 1 of 3 formulas. Infants were allocated to their formula group according to a sequentially numbered, sealed envelope that contained the formula assignment.10 Mothers who consented and intended to breastfeed their infant were included in the reference group. Full breastfeeding was defined as breastfeeding with <200 mL of formula per week in the first 16 weeks of life and ≤200 mL of formula per day between the ages of 16 weeks and 1 year.10 Any formula used by mother–infant pairs in the breastfed reference group was the mother's own choice. The rationale for not providing formulas to breastfeeding mothers was to avoid any possibility of influencing a mother's decision to persist with breastfeeding and change to infant formula. At the time the trial was conducted there were no commercially available formulas for term infants that contained LCPUFA.
Nestec Ltd (Konolfingen, Switzerland) supplied the formulas in powder form in 1-kg cans. All formulas had similar nutrient compositions as well as identical packaging and reconstitution instructions. The protein, fat, and carbohydrate content of all formulas were 1.5, 3.4, and 7.6 g per 100 mL, respectively. Only the fatty acid composition of each study formula varied. Formulas contained either no LCPUFA (placebo) or .35% DHA as total fatty acids from tuna oil or .34% DHA and .34% AA as total fatty acids from an egg phospholipid fraction (Table 1).
Infants had a VEP test to determine visual acuity at 16 and 34 weeks of age by 1 of 3 trained operators. An ophthalmic examination by a pediatric ophthalmologist was conducted before infants were 16 weeks of age. The pupils were dilated, the fundus examined, and cycloplegic refraction conducted. Infants with refraction outside the range of −3 to +5 diopters or with severe astigmatism (≥1.75 diopters) were excluded from the VEP aspect of the study. Global indices of development were assessed at 1 and 2 years using the Bayley's Scales of Development. At 1 year, 1 of 2 testers trained in developmental assessment conducted that Bayley's test, and at 2 years a single tester (1 of the initial 2) conducted all the Bayley's assessments. Infants also had a blood sample taken by heel prick at 16 and 34 weeks and 1 year of age. Plasma and erythrocyte fatty acid profiles were quantified by capillary gas chromatography.10
VEP Acuity Determination
VEPs were recorded under transient conditions using the Enfant 4010 system (Neuroscientific Corp, Farmingdale, NY). Infants were seated, with a parent, 1 m away from a 19-in monitor presenting high contrast black and white checkerboard pattern reversal (2 Hz) stimuli. The active electrode was placed 3 cm above the inion (Oz), the reference electrode at Fz and the inactive electrode was on the forehead. Two recordings were performed at each checkerboard pattern and these were subsequently averaged. Seven different checkerboard patterns were tested at each age (7, 10, 14, 20, 28, 42, and 55 minutes arc at 16 weeks and 3, 7, 10, 14, 20, 28, and 42 minutes arc at 34 weeks). The P100 latency of the averaged VEP at each checkerboard pattern was recorded. The peak to peak amplitude of the VEP (N1-P100) response was measured and plotted against log of the angle subtended by each check size. The linear portion of the plot was extrapolated to 0 μV to give the theoretical value that would just elicit a response (log of the minimum angle of resolution [logMAR]). Hence, lower logMAR values represent better visual acuity, because these infants are capable of eliciting better responses to smaller checkerboard patterns. Points were excluded from the regression if they were not on the linear portion of the stimulus–response function or represented amplitudes of ≤2 μV. VEP acuity extrapolations were accepted as valid only if there were at least 3 points and the regression line was significant (P < .05). The process of conducting the VEP test was similar to our earlier published work,3 although the hardware and software were updated to allow a greater range of checkerboard patterns to be presented to the infant, and thus more points were available for a more reliable extrapolation.
Bayley's Scales of Childhood Development
The Bayley's I Scales of Childhood Development were used to assess all children. Mental developmental index (MDI) and PDI were standardized for age, based on established reference norms.11 If the child refused to cooperate from the beginning of the appointment, the Bayley's test was not commenced and another appointment was scheduled. If the initial test proceeded and the infant then became uncooperative, a repeat test was scheduled in 3 months time. This was done to avoid biasing the test result because the child had learned aspects of the test. All mothers also completed the Home Screening Questionnaire (HSQ), which assesses the quality of the home environment.12
Sample Size and Statistical Analysis
We estimated that a sample size of 15 infants per formula feeding group would be sufficient to detect a mean difference of .2 log units in VEP acuity at both 16 and 34 weeks of age with a .05 level of significance and at least 90% power. To allow for loss to follow-up and unsuccessful VEP acuity extrapolation from the testing procedure, we aimed to enroll 25 infants per feeding group.
At 16 and 34 weeks of age, comparisons of VEP acuity between randomized formula groups were made by analysis of covariance (ANCOVA). The following variables were considered as possible covariates in these analyses: gestational age, gender, birth size, birth order, 5-minute Apgar score, parental smoking, education and social scores, age at testing, and the assessor. Because there were no differences between VEP acuity of formula-fed infants at any age, all data of formula-fed infants were combined and compared with the VEP acuity of breastfed infants also using ANCOVA. Both fully and partially breastfed infants were included in these comparisons. In a parallel set of analyses, ANCOVA models also were used to compare Bayley's MDI and PDI scores among infants randomly allocated to the 3 formula groups, at both 1 and 2 years of age. Environmental variables that were associated significantly with Bayley's indices were adjusted for in these analyses. Comparisons between breastfed and formula-fed infants in MDI and PDI also were explored using ANCOVA. All analyses were completed on an intention-to-treat basis.
Multiple linear regression models were constructed to further investigate the impact of environmental and nutritional variables on indices of development. All measured environmental and nutritional variables that may have an impact on development were considered as possible independent variables for these regression models. These variables included gender, gestational age, birth size, birth order, parental smoking, education and social scores, HSQ score, number of siblings, infant plasma and erythrocyte DHA, EPA, and AA, age at each test, size at each test, and the assessor. Models were constructed using the independent variables that were associated (P< .2) with the dependent variable. Independent variables were removed from a given model if the independent variable's presence or absence did not influence the model. No independent variables in the final regression models were found to be colinear. All analyses were performed using SPSS for Windows Version 6.0 (SPSS Inc, Chicago, IL).
A total of 83 infants were randomly allocated to 1 of 3 formula groups: 28 to the placebo formula, 27 to the DHA formula, and 28 to the DHA+AA formula. By 34 weeks of age, 15 (18%) of 83 infants were withdrawn from the trial. The reasons for withdrawal were similar among groups.10 Only 1 infant was withdrawn because of cataracts detected as part of the ophthalmic examination. A parallel, nonrandomized group of breastfed infants (n = 63) also was recruited. Of the 63 infants, 46 (73%) completed the trial to 34 weeks of age.10 At 16 and 34 weeks of age, 33 and 23 infants, respectively, were fully breastfed.
Among the randomized groups, the infants who completed the trial were similar in birth characteristics and were from families of similar social background (Table 2). There was a tendency, however, for proportionately more boys to be enrolled in the DHA+AA group, compared with the other 2 formula groups. In the nonrandomized, breastfed reference group, the 46 infants who completed the trial to 34 weeks of age had similar characteristics to the 23 fully breastfed infants (Table 2). In general, the breastfed infants had parents who were less likely to smoke, had attained a higher level of education, and had more prestigious social scores compared with formula fed infants.
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, and 1 was from the placebo group. No infants from the breastfed group were lost to follow-up between 34 weeks and 2 years of age.
There were no differences in VEP acuity between any of the randomized formula groups at either 16 or 34 weeks of age, before and after adjusting for the significant covariates of gender, postconceptional age, birth weight, and maternal smoking (Table 3). Further, the proportion of infants with a measurable VEP in response to the smallest checkerboard pattern was not different among formula groups at both 16 and 34 weeks of age. There was a trend for proportionately fewer infants in the DHA formula group to respond to patterns subtending 3′ of arc at 34 weeks of age compared with the other formula groups (P = .2). The VEP acuity of all formula-fed infants improved with age regardless of dietary grouping (Table 3). There was no effect of VEP operator on infant VEP acuity.
Because there was no difference in VEP acuity between formula groups, these were combined and compared with infants who were breastfed. At 16 weeks of age, the VEP acuity of breastfed and formula-fed infants was similar. However, at 34 weeks, infants who were fully breastfed were able to resolve smaller checkerboard patterns than infants who were formula-fed (logMAR breastfed n = 14 vs formula-fedn = 44; .28 ± .15 vs .41 ± .18;P < .01). Similarly, a greater proportion of fully breastfed infants had detectable response to 3′ arc checkerboard patterns at 34 weeks of age, compared with infants who were fed formula (87% vs 66%; P < .06). These trends were also apparent when the VEP acuity of all breastfed (partially and fully) infants were compared with those fed formula, but did not reach statistical significance.
To determine whether any dietary or environmental factors independently influenced VEP acuity, separate regression models were constructed at 16 and 34 weeks of age. At 16 weeks of age the significant predictors of VEP acuity were birth weight, gender, and maternal smoking (Table 4). Male gender and maternal smoking were associated with poorer VEP acuity. No dietary or environmental variables significantly predicted VEP acuity at 34 weeks of age (Table 4).
Bayley's MDI and PDI
At 1 year of age, no repeat Bayley's tests were necessary, whereas at 2 years, 7 repeat tests were conducted: 4 from the placebo formula group, 2 from the DHA formula group, and 1 breastfed infant. At repeat testing, 4 infants still did not complete the entire test and PDIs were not recorded in these infants (1 placebo formula-fed infant, 2 DHA formula-fed infants, and 1 breastfed infant). The PDI results of 1 breastfed infant at 1 and 2 years of age were excluded because this infant was diagnosed with hypotonia by her pediatrician (independent of the study).
Bayley's MDI and PDI values were not different in the 3 formula-fed groups at 1 and 2 years of age (Table 5). Bayley's MDI and PDI values also were compared between breastfed and formula-fed infants. At 1 year of age, MDI scores of breastfed and formula-fed infants were not different, but at 2 years of age, MDI scores of breastfed infants were higher than those of formula-fed infants even after adjusting for the significant covariates of gender and number of siblings (95% confidence interval [CI]: 4.4,21.7 for comparison with fully breastfed infants; 95% CI: 3.0,16.8 for comparison with all breastfed infants). PDI scores of breast and formula-fed infants were similar at 1 and 2 years of age.
There was a significant decrease in MDI scores of formula-fed infants between 1 and 2 years of age (111 ± 14; n = 61 at 1 year vs 105 ± 17; n = 61 at 2 years) that was independent of diet. This temporal change is difficult to explain and not apparent in the PDI scores of formula-fed infants or either of the development scores for breastfed infants.
Regression analyses were conducted to determine the influence of independent nutritional and environmental variables on Bayley's MDI and PDI scores at both 1 and 2 years of age (Table 6). No fatty acid variables significantly predicted MDI scores at either 1 or 2 years. Feeding mode was the only nutritional variable to predict MDI with formula feeding resulting in lower MDI scores (95% CI: −6.8,−3.0). Although environmental variables such as parental education, occupational prestige, and HSQ scores were associated with Bayley's MDI at 1 and 2 years of age, only weight (at 1 year) and birth order, feeding mode, and gender (at 2 years) significantly predicted MDI.
The only fatty acid variable to significantly influence 1-year PDI was plasma AA at 1 year of age (95% CI:−4.6,−.2). Plasma EPA (P = .19) and erythrocyte EPA (P = .08) at 1 year were positively associated with infant PDI but did not enter the regression model. At 2 years of age, 16-week erythrocyte EPA and DHA and plasma EPA were associated with PDI, but none of these variables significantly predicted PDI in the regression model. The significant environmental variables to influence PDI scores at 1 year were the assessor, maternal education, number of siblings, and the infant's age at testing, whereas at 2 years of age head circumference (at 2 years), number of siblings, and maternal smoking predicted PDI (Table 6).
MDI at 1 year was associated with MDI at 2 years (r = .44; P < .001; n = 107) and similarly PDI at 1 year was correlated with PDI at 2 years (r = .50; P < .001; n = 102).
A major finding of this study was that we could not detect an effect of dietary LCPUFA on visual acuity in this group of healthy term infants who had received formula from birth. Not only were there no group effects but no LCPUFA emerged as an influencing factor in the regression model describing the relationship between diet and VEP acuity. There are 2 other full reports in the literature that show a similar lack of an effect of dietary LCPUFA in formula-fed term infants.6 ,7 In contrast, our previous study3and reports from other laboratories4 ,5 have shown in randomized clinical trials differences in visual acuity that seem to be attributable to dietary LCPUFA.
The lack of effect of dietary LCPUFA on VEP acuity in this rigorously conducted trial was surprising for a number of reasons. First, we had previously demonstrated an effect of dietary LCPUFA3 and the VEP apparatus was known to be more sensitive than the equipment used in an earlier study.13 Further, at the 16-week assessment, the numbers in each group allowed for sufficient power to detect changes had they been present. For these reasons, we think that it is important to reevaluate the evidence relating to dietary LCPUFA and visual function.
In the current study, we used either low EPA tuna oil or egg phospholipid containing a balance of AA and DHA but little EPA. This was different from our earlier study, in which high EPA fish oil was used.3 Other studies also have used a variety of LCPUFA sources.4–7 One hypothesis is that in those studies in which an effect has been reported,3 the diets have generally contained high levels of total n-3 LCPUFA including EPA as well as DHA. This is not to suggest that the active ingredient in these early studies was necessarily EPA but rather that the presence of this compound may have had a sparing effect on the oxidation of DHA and thus more DHA may have been available for incorporation into neural tissues. Certainly one of the largest and most consistent effects reported in preterm infants conducted by Birch and coworkers14 ,15supplemented the basic formula with nearly 1% total n-3 LCPUFA. Other studies in preterm and term infants reporting either weak effects or no effects have supplemented diets with relatively modest levels of n-3 LCPUFA (.1%-.35% DHA or total n-3 LCPUFA). However, this theory is not supported by a randomized trial of maternal DHA supplementation among breastfed infants in which increasing DHA in breast milk from ∼.2% to 1.1% total fatty acids resulted in no measurable effect on infant VEP acuity at either 12 or 16 weeks of age.13Alternatively, Heird et al16 proposed that visual effects of dietary supplementation with n-3 LCPUFA were seen only in studies in which the placebo diet contained <2% α-linolenic acid (ALA). However, this second hypothesis is not consistent with the current literature, because there are now 2 trials with <2% ALA in the placebo diet that demonstrate no effect of LCPUFA treatment7 (and the current trial) and 4 trials with >2% ALA in the placebo formula that demonstrate beneficial effects of n-3 LCPUFA supplementation on the measures of visual function.4 ,5 ,17 ,18 Therefore, it would seem that neither the amount of n-3 LCPUFA added or the PUFA content of the background diet (or placebo formula) is capable of explaining the disparate outcomes of LCPUFA trials conducted during infancy. We also believe that it is unlikely that the form of dietary LCPUFA (triglycerides, phospholipids) is relevant to the recipient infant provided that LCPUFA are present, because the diet-induced changes in infant LCPUFA status are similar regardless of the dietary source.
Our study was designed with sufficient power (80%) to detect a .15-.2 log unit difference in VEP acuity, allowing us to detect differences of about this magnitude between breastfed and formula-fed infants. Thus at 16 weeks, it seems unlikely that the lack of difference between the formula treatments was attributable to a type II error. However, because of the higher proportion of unsuccessful VEPs at 34 weeks, the power dropped to 72%, reducing the level of confidence in the result at this age. It is important to recognize that even in studies that report an influence of dietary LCPUFA on visual acuity, the effect is not generally large. For example, Birch et al5 recently reported a small effect (.1–.15 log units) of LCPUFA supplementation in VEP acuity in term infants but no effect in visual acuity measured using cards. Carlson et al4 reported a similarly small effect but in a measure of acuity by Teller cards. The biggest predictor of VEP acuity in our study was birth weight such that the larger the child the better the acuity score (negative association of weight with logMAR). Certainly Jorgensen and coworkers7have reported that birth weight was a major predictor of sweep VEP acuity. Because birth weight may be associated with greater maturity at birth, it is important that this and other potential confounding variables are adequately controlled in studies of this type. The medium- and long-term clinical significance of small differences in visual acuity during infancy remain to be established. Further, it may well be that other more sensitive and clinically relevant outcome measures may be required to establish if there is functional essentiality of dietary LCPUFA.
The secondary aim of our trial was to assess the effect of LCPUFA supplementation on Bayley's MDI and PDI at 1 and 2 years of age. The MDI and PDI scores of all randomly allocated infants were similar. This result was not surprising given the relatively small number of infants per group and the similar parental and environmental background of these groups. It is interesting to note that no study in term infants has demonstrated an effect of LCPUFA supplementation using the Bayley tests19 and only one preterm study has claimed an effect of fish oil supplementation in preterm infants.20 It is, however, important to note that none of these trials have been adequately powered to measure an effect of dietary LCPUFA on the Bayley's MDI and PDI and all these studies are subject to type II error. Thus, there seems to be little evidence to prove or disprove a role for LCPUFA in global mental development as assessed by the Bayley test.
In our post hoc regression analyses including both the randomly allocated formula-fed infants and nonrandomized breastfed infants, formula feeding was negatively associated with Bayley's MDI at 2 years of age. Other studies have demonstrated that children who were breastfed as infants have better developmental scores than children who were formula-fed after adjustment for some, but not all, environmental factors that may confound the relationship between breastfeeding and developmental outcome.21–24 In our study, birth order and gender also significantly predicted MDI at 2 years of age. We have reported similar effects of gender on Bayley's MDI scores in a separate group of infants who were all initially breastfed.25
The issue that has been hotly debated is whether the differences reported between breastfed and formula-fed infants are attributable to procedural/methodological differences or factors present in breast milk that are not contained in formula. An alternate hypothesis that has been suggested is that there may be factors in formula that have a negative effect on neurodevelopment.26 The complexity of interpretation of interventions of early nutrition and behavior in both animal and human studies has been addressed,27 and it is clear that few firm conclusions can be made regarding the functional essentiality of any single dietary factor in the developing central nervous system.
Based solely on the results of this trial, we can find little evidence in support of an effect of dietary DHA on development of visual acuity in healthy term infants. Given that this trial had the capacity to detect differences in VEP acuity of .15 to .20 log units, it may be that any effect of dietary DHA is small and below this limit. Whether effects smaller than this can translate into developmental differences later in life or are reflected in other developmental capacities is unknown.
Our observation is consistent with a number of trials;6 ,7other trials have demonstrated small positive effects of dietary DHA.4 ,5 The questions that surround all such results are the degree to which the reported results can be accounted for by possible bias or confounding and the degree that the reported best effect is clinically relevant in the short and long term.
The trial was supported by Nestec Ltd, Switzerland, the MS McLeod Research Trust, and the Australian National Health and Medical Research Council.
We thank Jenny Osmond, Mary Wooden, Bronwen Paine, Cordula Blank, Dr John Pater, Lyn Pullen, Jane Armstrong, Brett Jeffrey, Ela Zielinski, Dani Bryan, and Dr Mark Hoby for their administrative and technical support.
- Received March 23, 1999.
- Accepted June 24, 1999.
- Address correspondence to Robert A. Gibson, PhD, Child Nutrition Research Centre, Child Health Research Institute, North Adelaide, SA 5006, Australia. E-mail:
Dr Makrides is currently at the University of Adelaide Department of Obstetrics and Gynaecology, Women's and Children's Hospital, North Adelaide, SA 5006, Australia.
- Makrides M,
- Neumann MA,
- Byard RW,
- Simmer K,
- Gibson RA
- ↵Carlson SE. Lipid requirements of very-low-birth-weight infants for optimal growth and development. In: Dobbing J, ed. Lipids, Learning, and the Brain: Fats in Infant Formulas. Columbus, OH: Ross Laboratories; 1993:188–214
- Makrides M,
- Neumann MA,
- Simmer K,
- Gibson RA
- ↵Bayley N. Manual for the Bayley Scales of Infant Development. Bekeley, CA: The Psychological Corporation; 1969
- ↵Coons C, Gay E, Fandal A, Frankenburg W. The Home Screening Questionnaire Reference Manual. Denver, CO: JF Kennedy Child Development Center, University of Colorado Health Science Center; 1981
- Birch EE,
- Birch DG,
- Hoffman DR,
- Uauy R
- Birch DG,
- Birch EE,
- Hoffman DR,
- Uauy RD
- Carlson SE,
- Werkman SH,
- Rhodes PG,
- Tolley EA
- Carlson SE,
- Werkman SH,
- Tolley EA
- ↵Scott DT, Janowsky JS, Carroll RE, Taylor JA, Auestad N, Montalto MB. Formula supplementation with long-chain polyunsaturated fatty acids: are there developmental benefits? Pediatrics. 1998;102(5). URL: http://www.pediatrics.org/cgi/content/full/5/e59
- Paine BJ,
- Makrides M,
- Gibson RA
- ↵Borstein MH. Nutrition and development: observations and implications. In: Dobbing J, ed. Developing Brain and Behaviour: The Role of Lipids in Infant Formula. San Diego, CA: Academic Press; 1997:475–510
- ↵Wainwright PE, Ward GR. Early nutrition and behavior: a conceptual framework for critical analysis of research. In: Dobbing J, ed. Developing Brain and Behaviour: The Role of Lipids in Infant Formula. San Diego, CA: Academic Press; 1997:387–426
- Daniel A. Power, Privilege and Prestige: Occupations in Australia. Melbourne, Australia: Longman-Cheshire; 1983
- Copyright © 2000 American Academy of Pediatrics