Breastfeeding, Polyunsaturated Fatty Acid Levels in Colostrum and Child Intelligence Quotient at Age 5-6 Years

ARTICLE IN PRESS THE JOURNAL OF PEDIATRICS • www.jpeds.com

ORIGINAL ARTICLES

Breastfeeding, Polyunsaturated Fatty Acid Levels in Colostrum and Child Intelligence Quotient at Age 5-6 Years Jonathan Y. Bernard, PhD1,2, Martine Armand, PhD3, Hugo Peyre, MD, PhD4,5, Cyrielle Garcia, PhD3, Anne Forhan, MPH1,2, Maria De Agostini, PhD1,2, Marie-Aline Charles, MD, PhD6, and Barbara Heude, PhD1,2, on behalf of the EDEN Mother-Child Cohort Study Group (Etude des Déterminants pré- et postnatals précoces du développement et de la santé de l'Enfant)* Objective To examine the relationship of polyunsaturated fatty acid (PUFA) in breast milk with children’s IQ. Study design In the French Etude des Déterminants pré- et postnatals précoces du développement et de la santé de l’Enfant (EDEN) mother-child cohort, colostrum samples were collected at the maternity unit. Colostrum omega-6 and omega-3 PUFA were analyzed by gas chromatography. At age 5-6 years, the IQs of 1080 children were assessed using the Wechsler Preschool and Primary Scale of Intelligence-III. The relationships of breastfeeding duration and PUFA levels with children’s IQs were examined by linear regression. Results Full scale IQ of ever breastfed children was 4.5 (95% CI: 2.7, 6.2) higher than never breastfed children in the unadjusted model, but this was not statistically significant in the adjusted model (1.3 points higher [-0.4, 3.0]). Any breastfeeding duration was associated with full scale (0.20 [0.00, 0.41] points/month) and verbal [0.31 [0.09, 0.52]) IQ. Colostrum linoleic acid (LA) levels were negatively associated with Verbal IQ (-0.6 [-1.1, 0.0] points per 1% level increase). Children exposed to colostrum high in LA and low in docosahexaenoic acid (DHA) had lower IQs than those exposed to colostrum high in DHA (3.0 [0.5, 5.5] points) and those exposed to colostrum low in LA and DHA (4.4 [1.6, 7.3] points). Finally, the association between breastfeeding duration and child IQ was stronger when LA levels were high. Conclusions Duration of breastfeeding and colostrum PUFA levels were associated with children’s IQs in the EDEN cohort. These data support breastfeeding and add evidence for the role of early PUFA exposure on childhood cognition. (J Pediatr 2017;■■:■■-■■).

O

bservational studies have shown that breastfed children score higher on cognitive tests, such as IQ, than formula-fed children.1 In our Etude des Déterminants pré- et postnatals précoces du développement et de la santé de l’Enfant (EDEN) cohort, we have highlighted a positive and linear association between breastfeeding duration and cognitive development assessed at ages 2 and 3 years with parent-reported questionnaires.2 This finding since has been replicated by other cohorts with cognitive tests assessed by psychologists.3-5 Whether this relationship is causal remains controversial because other studies have found no association after adjustment for socioeconomic status and maternal IQ.6,7 In addition, if truly causal, the underlying mechanism remains unclear. Human milk provides a nutritional advantage over infant formulas, particularly for lipid contents.8 The 2 series of polyunsaturated fatty acids (PUFAs), the omega-6 (n-6) and the omega-3 (n-3), specifically their long-chain forms (long chain PUFA [LC-PUFA]), are naturally found in human milk and are needed for the developing brain of the fetus and the infant.9 Positive effects on child cognition have been shown in some, but not all randomized controlled trials of formulas enriched in docosahexaenoic acid (DHA, 22:6 n-3) or arachidonic acid (AA, 20:4 n-6) (the most important n-3 and n-6 LC-PUFA, respectively). As the evidence has not been consistent, systematic reviews have been From the 1UMR1153 Epidemiology and Biostatistics inconclusive.10 Sorbonne Paris Cité Centre (CRESS), Developmental Origins of Health and Disease (ORCHAD) Team, Inserm, The published observational studies are limited. In sixty-seven 11-month-old Villejuif, France; 2Paris Descartes University, France; 3 11 Centre National de la Recherche Scientifique, Center for Inuit infants exposed to relatively high levels of breast milk PUFA, Jacobson et al Magnetic Resonance in Biology and Medicine UMR 7339,

AA ALA DHA EDEN FAMEs LA LC-PUFA Omega-3 Omega-6 PUFAs

Arachidonic acid a-linolenic acid Docosahexaenoic acid Etude des Déterminants pré- et postnatals précoces du développement et de la santé de l’Enfant Fatty acid methyl esters Linoleic acid Long-chain PUFA n-3 n-6 Polyunsaturated fatty acids

Aix-Marseille Université, Marseille, France; 4Laboratory of Cognitive Sciences and Psycholinguistics (École Normale Supérieure, École des Hautes Études en Sciences Sociales, Centre National de la Recherche Scientifique), École Normale Supérieure, PSL Research University, Paris, France; 5Child and Adolescent Psychiatry Unit, Assistance Publique - Hôpitaux de Paris, Robert Debré Hospital, Paris, France; and 6Fondation PremUP, Paris, France *List of additional members of the EDEN Mother-Child Cohort Study Group is available at www.jpeds.com (Appendix 1). Funding information is available at www.jpeds.com (Appendix 2). The authors declare no conflicts of interest. 0022-3476/$ - see front matter. © 2016 Elsevier Inc. All rights reserved. http://dx.doi.org10.1016/j.jpeds.2016.12.039

1 FLA 5.4.0 DTD ■ YMPD8897_proof ■ January 9, 2017

THE JOURNAL OF PEDIATRICS • www.jpeds.com found no association between the LC-PUFA levels and cognitive and motor outcomes. In a Spanish cohort, 14-monthold infants who were breastfed for a long duration and consumed colostrum containing a low n-6:n-3 had higher cognitive development than those breastfed for a shorter duration (regardless of colostrum content) and those fed with colostrum containing a higher n-6:n-3 (regardless of breastfeeding duration).12 In their follow-up at 4 years of age, they found no association, suggesting that the effect decreases with age.13 In our EDEN cohort, we previously found a negative association between the levels of linoleic acid (LA, 18:2 n-6), the precursor of the n-6 series, and cognitive development assessed by parents at 2 and 3 years of age.14 In this follow-up evaluation, we examined the relationships of breastfeeding duration and breast milk PUFA levels with children’s IQ at 5-6 years of age.

Methods The EDEN study is a French mother-child cohort which started in 2003 and aimed at examining the role of pre- and postnatal determinants of child development and health.15 Pregnant women (less than 24 weeks gestational age) attending their first antenatal visit in the maternity units of Nancy or Poitiers University Hospitals, France, were invited to participate. Exclusion criteria were twin pregnancies, known diabetes before pregnancy, illiteracy, and intention to move outside the region within the next 3 years. A total of 2002 women were enrolled.15 The study was approved by the ethics research committee (Comité Consultatif de protection des personnes dans la recherche biomédicale) of the Bicêtre Hospital, and by the National Data Protection Authority. From a questionnaire self-administered by the women during pregnancy, we obtained information on the number of siblings, smoking status, alcohol consumption (0, 1-9, ≥10 g/ week), and depression symptoms using the French scale of the Center for Epidemiologic Studies–Depression.16 Prepregnancy body mass index (in kg/m2) was obtained from the selfreported weight before pregnancy. Height was measured during the first pregnancy visit. Parental education level was indicated by the highest diploma declared by both parents. Average household income over the study was calculated from parents’ declarations during pregnancy and every year from 1 to 5 years. Offspring’s sex, gestational age, and birth weight were collected from obstetric and pediatric records. Birth weight z score was calculated according to the French fetal growth reference.17 During the clinical visit at 5-6 years of age, our research assistants administered 21 items from 3 subscales (language stimulation, learning stimulation, and variety in experience) of the French Home Observation for the Measurement of the Environment Inventory-Short Form.18 Feeding modes during the maternity stay and at discharge were obtained from medical records. In questionnaires at 4 months, 8 months, and 1 and 2 years, mothers reported their infants’ feeding modes. Mothers also reported the date of when they stopped breastfeeding. Two variables of breastfeeding duration were calculated according to their intensity level: any

Volume ■■ breastfeeding duration (in exact days and converted into months) and exclusive breastfeeding duration (in exact months). Any breastfeeding was defined as receiving any breast milk (whether exclusively or partially), and exclusive breastfeeding referred to infants receiving neither formula nor milk other than human milk. Additional details are available from prior publications.2,19 About 5 mL of colostrum was collected from lactating new mothers (mean ± SD: 3.9 ± 1.1 days after birth) and stored at -80°C until analysis. Fatty acid methyl esters (FAMEs) were obtained by direct transmethylation of 100 mL of colostrum at 100°C during 1 hour,20 and analyzed by gas chromatography with a fast BPX-70 column (Clarus 600 GC; Perkin Elmer, Waltham, Massachusetts), as previously reported.14 An external standard containing exact amounts of several different FAMEs (GLC 674; Nu-Chek Prep, Waterville, Minnesota) was used to calibrate the peak area of each FAME provided by a flame ionization detector such that it was proportional to its quantity. Twelve PUFAs, expressed as percentages of total milk fatty acids, were identified. These include LA, AA, a-linolenic acid (ALA, 18:3 n-3), eicosapentaenoic acid (20:5 n-3), and DHA. Total levels of n-6 LC-PUFA and of n-3 LC-PUFA (≥20 carbons), as well as several ratios (LA/ALA, AA/DHA, total n-6/ n-3 PUFA) were calculated. Between 5 and 6 years of age, 1 of 2 trained psychologists (1 in each study center) administered the French version of the Wechsler Preschool and Primary Scale of Intelligence-Third Edition.21 They were blinded to the children’s infant feeding history. The core subtests of the battery were assessed (information, vocabulary, word reasoning, block design, matrix reasoning, picture concepts, and coding) to obtain age-adjusted composite scores for verbal, performance, and full scale IQ. Statistical Analyses The characteristics of mother-child dyads with available IQ data at 5-6 years were described as percentages and means (SD), and compared with the population not analyzed. Linear regression was used to examine the associations between breastfeeding and children’s IQs in unadjusted and adjusted models. Adjusted models included the following covariates: study center, maternal age, prepregnancy body mass index, depression during pregnancy, tobacco and alcohol consumption during pregnancy, parity, child’s sex, birth weight z score, gestational age, parental education, household income, and family stimulation score. IQs were compared according to breastfeeding status (as a binary variable; ever vs never breastfed). Among ever breastfed children, associations between continuous variables of breastfeeding duration and children’s IQs were examined. The same analysis was performed in our population of interest (ie, those with available data on colostrum PUFA. Deviation from linearity was tested by adding the squared term of breastfeeding duration into the models. Model diagnostics were performed to identify potential influential observations (outliers and high leverage). Associations between PUFA levels (and ratios) and children’s IQs were assessed by linear regression. The adjusted models accounted for the delay between birth and colostrum

2

Bernard et al FLA 5.4.0 DTD ■ YMPD8897_proof ■ January 9, 2017

ORIGINAL ARTICLES

2017 collection, because breast milk composition can change over this period. Adding/removing the PUFA variables from the models was examined for modification of the estimates of association between breastfeeding duration and children’s IQs. Relationships between breastfeeding duration and IQ according to PUFA levels were examined by introducing the appropriate interaction terms; interactions between PUFAs were also examined. These analyses were first performed with PUFA variables treated continuously, then as categorical variables: each variable was categorized into 2 groups according to its median level (LA: 9.84%; AA: 0.86%; DHA: 0.64%; n-3 LC-PUFA: 1.23%; the total n-6:n-3: 6.17). IQs in these groups of breastfed children were compared with IQs of never breastfed children. Statistical analyses were performed using SAS 9.3 (SAS Institute Inc, Cary, North Carolina).

Results Figure 1 (available at www.jpeds.com) shows the flow of children through the study. Of the 1087 children born after 33 weeks gestation with IQ testing, 7 were excluded as they were outliers with full scale IQ scores < 60. Of the remaining 1080 children, 799 (74.0%) were ever breastfed. The children were assessed on average at 5.7 ± 0.2 years (Table I). The average full scale, verbal and performance IQs were 103.5 ± 12.8, 107.0 ± 13.6, and 99.7 ± 13.3, respectively. Compared with nonparticipants (lost to follow-up or having missing IQ data), the participants had higher education and socioeconomic status (P < .0001) (Table I). Their mothers were older, less likely to smoke, more likely to drink alcohol during pregnancy, and had lower depression score (all P < .0001). AA and DHA levels in

Table I. Characteristics of the participants with and without IQ data at 5-6 years of age*

Characteristics of children Sex, % boy Gestational age at birth, wk Birth weight, z score Breastfeeding Ever breastfed, % Any breastfeeding duration†, mo Exclusive breastfeeding duration†, mo WPPSI-III at 5-6 y Age at assessment, y Full scale IQ Verbal IQ Performance IQ Characteristics of mothers Study center, % Poitiers Age at inclusion, y Primiparous, % Pre-pregnancy body mass index, kg/m2 Tobacco smoking during pregnancy, % Alcohol drinking during pregnancy, % 0 g/wk <10 g/wk ≥10 g/wk Score of depression during pregnancy Characteristics of families Parental education, y Household income class Family stimulation score Colostrum PUFA levels, % of total fatty acids Subsample size, n n-6 PUFA 18:2 n-6 (LA) 20:4 n-6 (AA) Total n-6 LC-PUFA n-3 PUFA 18:3 n-3 (ALA) 20:5 n-3 (EPA) 22:6 n-3 (DHA) Total n-3 LC-PUFA n-6/n-3 PUFA ratios LA/ALA AA/DHA Total n-6/n-3 PUFA

Participants with IQ data included in the analysis (n = 1080)

Participants without IQ data or excluded from the analysis (n = 823)

P

53.0 39.4 (1.5) −0.02 (0.96)

52.0 39.1 (2.0) −0.05 (0.95)

.68 .0002 .58

74.0 4.5 (4.1) 2.5 (2.7)

72.5 4.5 (4.6) 2.5 (2.8)

5.66 103.5 107.0 99.7

(0.15) (12.8) (13.6) (13.3)

.48 .71 .65

— – – – <.0001 <.0001 .22 .05 <.0001 <.0001

58.1 29.7 (4.7) 45.7 23.4 (4.6) 22.1

37.9 28.2 (5.0) 42.9 23.0 (4.6) 32.8

52.3 39.6 8.1 10.9 (7.6)

63.8 30.3 6.0 12.7 (8.6)

13.5 (2.3) 5.0 (1.2) 17.2 (2.2)

13.0 (2.5) 4.7 (1.5) —

539

386

9.80 (1.83) 0.87 (0.16) 2.12 (0.42)

9.91 (1.89) 0.85 (0.16) 2.13 (0.42)

.46 .28 .74

0.64 0.06 0.65 1.24

0.66 0.06 0.63 1.20

(0.22) (0.03) (0.20) (0.31)

.31 .32 .07 .03

16.40 (5.25) 1.46 (0.41) 6.23 (1.56)

.39 .08 .37

(0.22) (0.03) (0.19) (0.32)

16.70 (5.63) 1.41 (0.36) 6.14 (1.51)

<.0001 <.0001 <.0001

EPA, eicosapentaenoic acid; WPPSI-III, Wechsler Preschool and Primary Scale of Intelligence-Third Edition. *Values are % or mean (SD). P values are based on Χ2 and Student t tests, respectively, for categorical and continuous variables. †In ever breastfed children (n = 799 and 587, respectively).

Breastfeeding, Polyunsaturated Fatty Acid Levels in Colostrum and Child Intelligence Quotient at Age 5-6 Years FLA 5.4.0 DTD ■ YMPD8897_proof ■ January 9, 2017

3

THE JOURNAL OF PEDIATRICS • www.jpeds.com

Volume ■■

Table II. Children’s IQs according to breastfeeding status and duration in the EDEN cohort Full scale IQ Unadjusted

Verbal IQ

Adjusted*

Breastfeeding status† Ever breastfed (n = 799) 104.7 (103.8, 105.5) 103.8 (103.0, 104.7) Never breastfed (n = 281) 100.2 (98.7, 101.7) 102.5 (101.1, 103.9) P <.0001 .12 Breastfeeding duration, mo (n = 799)‡ Any breastfeeding 0.34 (0.13, 0.56) 0.20 (0.00, 0.41) P .001 .05 Exclusive breastfeeding 0.47 (0.15, 0.79) 0.25 (−0.07, 0.56) P .005 .12 Breastfeeding duration, mo (n = 539)‡,§ Any breastfeeding 0.33 (0.07, 0.57) 0.22 (−0.03, 0.46) P .01 .08 Exclusive breastfeeding 0.47 (0.10, 0.84) 0.29 (−0.08, 0.66) P .01 .12

Unadjusted

Performance IQ Adjusted*

Unadjusted

Adjusted*

108.3 (107.3, 109.2) 107.4 (106.5, 108.3 100.7 (99.8, 101.6) 100.1 (99.2, 100.9) 103.5 (101.9, 105.0) 105.9 (104.4, 107.4) 96.8 (95.2, 98.3) 98.6 (97.1, 100.2) <.0001 .09 <.0001 .13 0.46 (0.23, 0.69) <.0001 0.60 (0.26, 0.95) .0007

0.31 (0.09, 0.52) .006 0.38 (0.05, 0.72) .03

0.22 (0.00, 0.44) .05 0.24 (−0.09, 0.58) .16

0.10 (−0.12, 0.32 .37 0.06 (−0.28, 0.40) .73

0.49 (0.23, 0.76) .0003 0.66 (0.26, 1.06) .001

0.37 (0.11, 0.63) .005 0.48 (0.10, 0.87) .01

0.13 (−0.12, 0.39) .29 0.17 (−0.21, 0.55) .39

0.04 (−0.22, 0.30) .77 0.01 (−0.38, 0.40) .95

*Models were adjusted for study center, maternal age, prepregnancy body mass index, depression during pregnancy, tobacco and alcohol consumption during pregnancy, parity, child's sex, birth weight z score, gestational age at birth, parental education, household income, and family stimulation score. †Values are means (95% CI) and adjusted means (95% CI). ‡Values are regression coefficients (95% CI). §Models restricted to participants with colostrum PUFA data.

colostrum were 0.87 ± 0.16% and 0.65 ± 0.19%, respectively. There were no differences in PUFA levels between participants and nonparticipants, except total n-3 LC-PUFA levels, which were 0.05 ± 0.02% higher in participants (Table I). Breastfeeding and IQ Ever breastfed children had higher IQs than never breastfed children (b [95% CI]: 4.5 [2.7, 6.2] points for full scale IQ) (Table II). After adjustment, this difference was reduced to 1.3 (−0.4, 3.0) points and no longer significant. Comparable effect sizes were observed for verbal and performance IQs. In ever breastfed children, both any and exclusive breastfeeding durations were positively associated with full scale and verbal IQs, but not with performance IQ (Table II). After adjustment, the effect sizes were reduced by one-third and remained significant for verbal IQ only: a 1-month increase in any breastfeeding and exclusive breastfeeding duration was related to 0.31 (0.09, 0.52) and 0.38 (0.11, 0.63) points increase, respectively. All tests of deviation from linearity were rejected (results not shown). Similar findings were observed when restricting the analysis to the subsample of participants with colostrum data (Table II). Colostrum PUFA Levels and IQ Among the breastfed participants assessed at 5-6 years of age (n = 799), those with PUFA data (n = 539) were breastfed for a longer duration (0.8 [0.2, 1.4] month longer), had higher full scale IQ (2.9 [1.1, 4.8] points), and were less likely to be followed in the Poitiers center (Table III; available at www.jpeds.com). However, there were no differences in parental education and household income. Overall, PUFA levels as continuous variables were not associated with children’s IQs (Table IV; available at www.jpeds.com). The total PUFA n-6:n-3 was negatively associated with verbal IQs (−0.9 [−1.6, −0.1]). A negative

association was observed between LA levels and verbal IQ (−0.6 [−1.1, 0.0] points per 1% level increase). In the models explaining verbal IQ, exclusive breastfeeding duration remained significantly associated with the outcome, with marginal change resulting from the inclusion/removal of PUFA levels. Figure 2 displays adjusted full scale IQs of 3 groups of children (ie, the never breastfed and 2 groups of breastfed children, categorized according to the median level of AA (panel A), total n-3 LC-PUFA (panel B), and total n-6/n-3 PUFA ratio (panel C). The group of children breastfed with colostrum high in AA (“highest AA”) and n-3 LC-PUFA (“highest n-3 LCPUFA”) scored 2.6 (0.5, 4.8) and 2.6 (0.5, 4.7) points higher than the never breastfed children, respectively. The children breastfed with colostrum low in AA and n-3 LC-PUFA had intermediate IQs. A similar pattern was observed with the total n-6:n-3. The children breastfed with colostrum containing a lower ratio scored higher than the never breastfed children. We also observed interactions between LA and DHA levels with full scale IQ as outcome (P value for interaction: .10 and .02 as continuous and 2-category variables, respectively) (Figure 2, D). In other words, DHA was an effect-modifier of the negative association between LA levels and IQs. The children who were breastfed with colostrum high in LA and low in DHA had lower full scale IQs than those exposed to colostrum which was high in DHA (3.0 (0.5, 5.5) points (whatever the level in LA) or low in LA and DHA (4.4 (1.6, 7.3) points). Similar patterns were observed with verbal and performance IQs (results not shown). Finally, the relationship of breastfeeding duration with full scale and verbal IQs differed according to LA levels (Table V): breastfeeding duration was positively associated with IQ among children exposed to high LA levels but not among those exposed to low LA levels.

4

Bernard et al FLA 5.4.0 DTD ■ YMPD8897_proof ■ January 9, 2017

ORIGINAL ARTICLES

2017

Figure 2. Adjusted means (95% CI) of full scale IQ in never breastfed children, and ever breastfed children as 2 groups according to PUFA levels (separated by the median level) in colostrum: AA levels (median: 0.86%; panel A), n-3 LC-PUFA levels (median: 1.23%; panel B), n-6:n-3 (median: 6.17; panel C), LA and DHA levels (median: 9.84% and 0.64%, respectively; panel D). Groups sharing the same superscript letters do not differ from each other.

Discussion In 1080 children from the EDEN cohort, we observed a crude difference of 4.5 IQ points between ever and never breastfed children, which was reduced to a nonclinically significant difference of 1.3 points after accounting for confounders, in particular, parental education and family stimulation. In ever breastfed children, the duration of breastfeeding was positively associated with verbal IQ, but not with full scale and performance IQs. Children’s verbal IQs were negatively associated with LA levels and total n-6:n-3. No association was observed with the other PUFA levels when used as continuous variables; however, when used as 2-category variables, children exposed to higher AA and n-3 LC-PUFA levels had higher IQs than never breastfed children. An interaction between LA and DHA levels was observed: DHA was positively associated with IQs among children exposed to colostrum high in LA. Similarly, breastfeeding duration was positively associated with IQs among children exposed to high LA levels. Overall, these results are consistent with our previous analyses on children’s language and cognition at 2 and 3 years of

age in this same cohort. With a somewhat larger sample size, at 2 and 3 years of age, children’s developmental scores were positively associated with breastfeeding duration and negatively with colostrum LA levels.2,14 These findings were based on parent report of development via questionnaires and may have had declaration bias. In the present study, this limitation was eliminated as IQ at 5-6 years of age was assessed by trained psychologists. The main strength of our study is the availability of breast milk PUFA levels in over 500 mothers, which is, to our knowledge, unique in the literature. Other study strengths include the prospective design, detailed data on infant feeding, and a wide range of confounders accounted for, including parental education and family stimulation. Our study has some limitations. First, as an observational study, residual confounding may still remain. In particular, we were unable to adjust for maternal IQ as this was not available in our cohort. Studies have suggested that maternal IQ may be a confounder of the association between breastfeeding and children’s cognition.6,7 However, this remains controversial as significant associations between breastfeeding and children’s cognition have been reported even after adjusting for maternal IQ or using a cluster-randomized design.4,22 A second limitation is the collection of breast milk, which was

Breastfeeding, Polyunsaturated Fatty Acid Levels in Colostrum and Child Intelligence Quotient at Age 5-6 Years FLA 5.4.0 DTD ■ YMPD8897_proof ■ January 9, 2017

5

THE JOURNAL OF PEDIATRICS • www.jpeds.com

Volume ■■

Table V. Adjusted associations between breastfeeding duration and IQs according to colostrum PUFA levels (n = 539)* Any breastfeeding duration, mo

18:2 n-6 (LA) Lowest† Highest P interaction‡ 20:4 n-6 (AA) Lowest Highest P interaction Total n-6 LC-PUFA Lowest Highest P interaction 18:3 n-3 (ALA) Lowest Highest P interaction 20:5 n-3 (EPA) Lowest Highest P interaction 22:6 n-3 (DHA) Lowest Highest P interaction Total n-3 LC-PUFA Lowest Highest P interaction

Exclusive breastfeeding duration, mo

Full scale IQ

Verbal IQ

Performance IQ

Full scale IQ

Verbal IQ

Performance IQ

b (95% CI)

b (95% CI)

b (95% CI)

b (95% CI)

b (95% CI)

b (95% CI)

−0.11 (−0.46, 0.25) 0.46 (0.11, 0.81) 0.02

0.02 (−0.37, 0.40) 0.62 (0.27, 0.97) 0.03

−0.19 (−0.57, 0.19) 0.22 (−0.14, 0.58) 0.12

0.00 (−0.52, 0.52) 0.53 (−0.02, 1.07) 0.15

0.20 (−0.35, 0.76) 0.74 (0.19, 1.29) 0.13

−0.12 (−0.67, 0.43) 0.09 (−0.48, 0.66) 0.68

0.08 (−0.27, 0.42) 0.35 (−0.01, 0.70) 0.41

0.29 (−0.09, 0.67) 0.42 (0.08, 0.76) 0.68

−0.07 (−0.42, 0.28) 0.14 (−0.25, 0.53) 0.75

−0.08 (−0.60, 0.44) 0.60 (0.07, 1.13) 0.16

0.27 (−0.31, 0.85) 0.62 (0.11, 1.13) 0.49

−0.38 (−0.91, 0.15) 0.35 (−0.23, 0.94) 0.21

0.26 (−0.06, 0.59) 0.18 (−0.19, 0.55) 0.46

0.31 (−0.06, 0.67) 0.47 (0.10, 0.84) 0.73

0.22 (−0.11, 0.56) −0.14 (−0.54, 0.26) 0.06

0.23 (−0.28, 0.74) 0.34 (−0.20, 0.88) 0.66

0.45 (−0.11, 1.01) 0.54 (0.00, 1.09) 0.80

−0.03 (−0.56, 0.50) 0.05 (−0.54, 0.63) 0.54

0.16 (−0.20, 0.52) 0.22 (−0.12, 0.56) 0.93

0.49 (0.11, 0.87) 0.28 (−0.07, 0.64) 0.59

−0.25 (−0.65, 0.15) 0.20 (−0.14, 0.55) 0.16

0.18 (−0.35, 0.72) 0.26 (−0.26, 0.78) 0.99

0.50 (−0.07, 1.07) 0.43 (−0.13, 0.98) 0.96

−0.27 (−0.86, 0.33) 0.12 (−0.41, 0.65) 0.50

0.14 (−0.24, 0.53) 0.20 (−0.13, 0.53) 0.73

0.40 (0.00, 0.80) 0.28 (−0.07, 0.62) 0.32

−0.10 (−0.51, 0.32) 0.10 (−0.24, 0.44) 0.87

0.02 (−0.55, 0.59) 0.39 (−0.12, 0.90) 0.70

0.29 (−0.31, 0.88) 0.57 (0.03, 1.10) 0.93

−0.20 (−0.82, 0.41) 0.08 (−0.45, 0.61) 0.83

0.02 (−0.33, 0.37) 0.42 (0.07, 0.77) 0.28

0.36 (0.00, 0.73) 0.36 (0.00, 0.73) 0.66

−0.24 (−0.60, 0.12) 0.37 (−0.01, 0.75) 0.08

−0.03 (−0.54, 0.49) 0.58 (0.04, 1.13) 0.34

0.33 (−0.21, 0.87) 0.68 (0.12, 1.25) 0.85

−0.29 (−0.82, 0.25) 0.28 (−0.31, 0.87) 0.34

0.09 (−0.24, 0.42) 0.30 (−0.07, 0.67) 0.66

0.26 (−0.10, 0.62) 0.47 (0.09, 0.84) 0.71

−0.05 (−0.39, 0.29) 0.10 (−0.31, 0.51) 0.88

0.01 (−0.50, 0.52) 0.47 (−0.08, 1.02) 0.55

0.26 (−0.30, 0.81) 0.69 (0.14, 1.24) 0.68

−0.17 (−0.7, 0.35) 0.13 (−0.48, 0.73) 0.75

*Models were adjusted for study center, maternal age, prepregnancy body mass index, depression during pregnancy, tobacco and alcohol consumption during pregnancy, parity, child's sex, birth weight Z score, gestational age at birth, parental education, household income, and family stimulation score. †Values are regression coefficients (95% CI) expressed in IQ point per additional month of breastfeeding. ‡Values are P for interaction between breastfeeding duration (continuous) and PUFA levels (2-category).

conducted at a single time point during the first week after birth. Breast milk composition changes over the lactation course,23 and the longitudinal variation in PUFA levels may not have been captured by our measure. Yet, LC-PUFA levels in colostrum have been reported to correlate fairly well with their levels in mature milk.24 Hence, we speculate that measuring LC-PUFA levels in mature milk would not have changed our results drastically. A second limitation comes from the cohort attrition. The sample of participants that were followed up at 5-6 years of age differed from the sample enrolled initially. The characteristics of participants with or without breast milk data were also somewhat different. This selective follow-up limits our ability to generalize our findings. In the adjusted models, we found a nonsignificant difference of 1.3 IQ points between the ever and never breastfed children, although we did find a positive association with verbal IQ when considering breastfeeding duration. The short duration of breastfeeding in our cohort may limit our ability to contrast ever with never breastfed children. Although not clinically significant, this effect-size remains comparable with other studies.1,4,25,26 Larger differences in verbal skills (compared with nonverbal skills) have also been reported, including in our cohort using a wider variety of neuropsychological tests.4,27,28 We found a negative association between children’s IQ and colostrum LA levels. This finding is in line with our previous

study conducted at earlier time points,14 another study in premature infants,29 and an ecological study.30 We also found that children breastfed with higher AA and n-3 LC-PUFA levels have higher IQs than never breastfed children; this is consistent with a large body of literature showing the importance of AA and n-3 LC-PUFA on the development of the infant’s brain.31,32 From midpregnancy to 2 years of age, LC-PUFAs rapidly accumulate in brain cells where they have key roles in membrane structure and signaling.33 Unexpectedly, we found no association between children’s IQ and DHA levels, but we did observe an interaction between DHA and LA levels that is novel to the current literature. When colostrum LA levels were high, lower DHA levels were associated with lower IQs; the group of children exposed to colostrum high in LA and low in DHA had similar IQs to those who were never breastfed. High LA dietary intake reduces the biosynthesis of DHA from ALA in the body by competing with the enzymes involved in PUFA metabolism.34 High dietary intake of LA may also impair the transport of DHA into the brain and inhibit the growth of neurons.35 The expected beneficial effects of DHA could be altered when there is a high intake of LA from breast milk or formula. This hypothesis could explain why DHA supplementation shows limited effects in most randomized controlled trials.36 Finally, we report an interaction between breastfeeding duration and LA levels: breastfeeding duration was positively

6

Bernard et al FLA 5.4.0 DTD ■ YMPD8897_proof ■ January 9, 2017

ORIGINAL ARTICLES

2017 associated with IQ in children exposed to high LA levels but not in children exposed to low levels. This differs from the findings from Guxens et al12 who found an interaction with the n-6:n-3. This might be due to differences in average PUFA levels and breastfeeding duration across the 2 studies (2.5 months on average in our study vs 4-6 months for one-half of their population). An alternative hypothesis is that PUFA levels in colostrum have a limited effect on children’s IQ, and that instead, the negative association of LA with IQ is a reflection of the prenatal exposure to LA through maternal diet. In a previous work, we have reported that maternal LA dietary intake during pregnancy was negatively associated with neurodevelopment at 2 and 3 years of age in never breastfed children.37 In a study of Inuit infants, Jacobson et al11 found a positive association of cord plasma DHA levels with cognitive outcomes, but no association when considering DHA levels in breast milk collected at 1 month postpartum, suggesting a prenatal rather than a postnatal effect. PUFA composition in colostrum and mature milk is highly related to maternal diet, making it hard to disentangle prenatal and early postnatal effects.23,38 In our study, the interaction between LA levels and breastfeeding duration suggests that breastfeeding could compensate postnatally for an inadequate PUFA exposure in utero. Maternal nutrition during pregnancy and lactation could also reflect offspring’s nutrition later in childhood because they share the same dietary environment. Genetic determinants are also shared by the mother and the child, especially the FADS genes, which encode for enzymes involved in the PUFA metabolism.39 FADS variants are associated with breast milk PUFA composition40 and with children’s IQ.41-43 Further studies on gene-environment interactions may improve our understanding of the relationship between breastfeeding, PUFA exposure, and children’s cognition. Lastly, the fact that other human milk bioactives may influence the infant’s neurobiological development through the gut-brain axis is not excluded.44 In conclusion, our study conducted in a French populationbased cohort shows mild associations of children’s IQs at 5-6 years of age with breastfeeding duration and colostrum PUFA levels. We report no evidence of interactions of breastfeeding duration with LC-PUFA levels, but we did report an interaction with LA levels. The interaction between LA and DHA is novel and needs replication in animal and human studies. Overall, our findings support the existing policies to promote breastfeeding. It also provides some evidence on the beneficial roles of AA and n-3 LC-PUFA on child cognition, and suggests that excess LA exposure in early life is unfavorable. ■ We are grateful to the participating families, the midwife research assistants (Lorraine Douhaud, Sophie Bedel, Brigitte Lortholary, Sophie Gabriel, Muriel Rogeon, and Monique Malinbaum) for data collection, the psychologists (Marie-Claire Cona and Marielle Paquinet), and the data entry operators (Patricia Lavoine, Josiane Sahuquillo, and Ginette Debotte). We also thank Wei Wei Pang (employed and supported by the Department of Obstetrics and Gynecology, Yong Loo Lin School of Medicine, National University of Singapore, Singapore) for reviewing the language.

Submitted for publication Jun 7, 2016; last revision received Oct 19, 2016; accepted Dec 13, 2016 Reprint requests: Jonathan Y. Bernard, PhD, Equipe ORCHAD, Centre de Recherche en Épidémiologie et Biostatistiques Sorbonne Paris Cité, 16 Ave Paul Vaillant-Couturier, Villejuif CEDEX 94807, France. E-mail: [email protected]

References 1. Anderson JW, Johnstone BM, Remley DT. Breast-feeding and cognitive development: a meta-analysis. Am J Clin Nutr 1999;70:525-35. 2. Bernard JY, De Agostini M, Forhan A, Alfaiate T, Bonet M, Champion V, et al. Breastfeeding duration and cognitive development at 2 and 3 years of age in the EDEN mother-child cohort. J Pediatr 2013;163:36-42, e1. 3. Quigley MA, Hockley C, Carson C, Kelly Y, Renfrew MJ, Sacker A. Breastfeeding is associated with improved child cognitive development: a population-based cohort study. J Pediatr 2012;160:25-32. 4. Belfort MB, Rifas-Shiman SL, Kleinman KP, Guthrie LB, Bellinger DC, Taveras EM, et al. Infant feeding and childhood cognition at ages 3 and 7 years: effects of breastfeeding duration and exclusivity. JAMA Pediatr 2013;167:836-44. 5. Victora CG, Horta BL, de Mola CL, Quevedo L, Pinheiro RT, Gigante DP, et al. Association between breastfeeding and intelligence, educational attainment, and income at 30 years of age: a prospective birth cohort study from Brazil. Lancet Glob Health 2015;3:e199-205. 6. Jacobson SW, Chiodo LM, Jacobson JL. Breastfeeding effects on intelligence quotient in 4- and 11-year-old children. Pediatrics 1999;103:e71. 7. Der G, Batty GD, Deary IJ. Effect of breast feeding on intelligence in children: prospective study, sibling pairs analysis, and meta-analysis. BMJ 2006;333:945. 8. Andreas NJ, Kampmann B, Mehring Le-Doare K. Human breast milk: a review on its composition and bioactivity. Early Hum Dev 2015;91:62935. 9. Koletzko B, Lien E, Agostoni C, Bohles H, Campoy C, Cetin I, et al. The roles of long- chain polyunsaturated fatty acids in pregnancy, lactation and infancy: review of current knowledge and consensus recommendations. J Perinat Med 2008;36:5-14. 10. Simmer K, Patole SK, Rao SC. Long-chain polyunsaturated fatty acid supplementation in infants born at term. Cochrane Database Syst Rev 2011;4:CD000376. 11. Jacobson JL, Jacobson SW, Muckle G, Kaplan-Estrin M, Ayotte P, Dewailly E. Beneficial effects of a polyunsaturated fatty acid on infant development: evidence from the Inuit of Arctic Quebec. J Pediatr 2008;152:35664. 12. Guxens M, Mendez MA, Molto-Puigmarti C, Julvez J, Garcia-Esteban R, Forns J, et al. Breastfeeding, long-chain polyunsaturated fatty acids in colostrum, and infant mental development. Pediatrics 2011;128:e880-9. 13. Julvez J, Guxens M, Carsin AE, Forns J, Mendez M, Turner MC, et al. A cohort study on full breastfeeding and child neuropsychological development: the role of maternal social, psychological, and nutritional factors. Dev Med Child Neurol 2014;56:148-56. 14. Bernard JY, Armand M, Garcia C, Forhan A, De Agostini M, Charles MA, et al. The association between linoleic acid levels in colostrum and child cognition at 2 and 3 y in the EDEN cohort. Pediatr Res 2015;77:82935. 15. Heude B, Forhan A, Slama R, Douhaud L, Bedel S, Saurel-Cubizolles MJ, et al. Cohort Profile: the EDEN mother-child cohort on the prenatal and early postnatal determinants of child health and development. Int J Epidemiol 2016;45:353-63. 16. Fuhrer R, Rouillon F. La version française de l’échelle CES-D (Center for Epidemiologic Studies-Depression Scale). Description et traduction de l’échelle d’autoévaluation. / The French version of the CES-D (Center for Epidemiologic Studies-Depression Scale). Eur Psychiatr 1989;4:163-6. 17. Mamelle N, Munoz F, Grandjean H. [Fetal growth from the AUDIPOG study. I. Establishment of reference curves]. J Gynecol Obstet Biol Reprod (Paris) 1995;25:61-70. 18. Caldwell B, Bradley R. Home Observation for Measurement of the Environment. 1984.

Breastfeeding, Polyunsaturated Fatty Acid Levels in Colostrum and Child Intelligence Quotient at Age 5-6 Years FLA 5.4.0 DTD ■ YMPD8897_proof ■ January 9, 2017

7

THE JOURNAL OF PEDIATRICS • www.jpeds.com 19. Bonet M, Marchand L, Kaminski M, Fohran A, Betoko A, Charles MA, et al. Breastfeeding duration, social and occupational characteristics of mothers in the French ‘EDEN mother-child’ cohort. Matern Child Health J 2013;17:714-22. 20. Masood A, Stark KD, Salem N Jr. A simplified and efficient method for the analysis of fatty acid methyl esters suitable for large clinical studies. J Lipid Res 2005;46:2299-305. 21. Wechsler D. Wechsler Preschool and Primary Scale of Intelligence (WPPSIIII). 3rd ed. San Antonio (TX): Psychological Corporation; 2002. 22. Kramer MS, Aboud F, Mironova E, Vanilovich I, Platt RW, Matush L, et al. Breastfeeding and child cognitive development: new evidence from a large randomized trial. Arch Gen Psychiatry 2008;65:578-84. 23. Dunstan JA, Mitoulas LR, Dixon G, Doherty DA, Hartmann PE, Simmer K, et al. The effects of fish oil supplementation in pregnancy on breast milk fatty acid composition over the course of lactation: a randomized controlled trial. Pediatr Res 2007;62:689-94. 24. Molto-Puigmarti C, Castellote AI, Carbonell-Estrany X, Lopez-Sabater MC. Differences in fat content and fatty acid proportions among colostrum, transitional, and mature milk from women delivering very preterm, preterm, and term infants. Clin Nutr 2011;30:116-23. 25. Brion MJ, Lawlor DA, Matijasevich A, Horta B, Anselmi L, Araujo CL, et al. What are the causal effects of breastfeeding on IQ, obesity and blood pressure? Evidence from comparing high-income with middle-income cohorts. Int J Epidemiol 2011;40:670-80. 26. Clark KM, Castillo M, Calatroni A, Walter T, Cayazzo M, Pino P, et al. Breast-feeding and mental and motor development at 51/2 years. Ambul Pediatr 2006;6:65-71. 27. Kramer MS, Fombonne E, Igumnov S, Vanilovich I, Matush L, Mironova E, et al. Effects of prolonged and exclusive breastfeeding on child behavior and maternal adjustment: evidence from a large, randomized trial. Pediatrics 2008;121:e435-40. 28. Peyre H, Bernard JY, Hoertel N, Forhan A, Charles M-A, De Agostini M, et al. Differential effects of factors influencing cognitive development at the age of 5-to-6 years. Cogn Dev 2016;40:152-62. 29. Sabel KG, Strandvik B, Petzold M, Lundqvist-Persson C. Motor, mental and behavioral developments in infancy are associated with fatty acid pattern in breast milk and plasma of premature infants. Prostaglandins Leukot Essent Fatty Acids 2012;86:183-8. 30. Lassek WD, Gaulin SJ. Linoleic and docosahexaenoic acids in human milk have opposite relationships with cognitive test performance in a sample of 28 countries. Prostaglandins Leukot Essent Fatty Acids 2014;91:195201. 31. Heird WC, Lapillonne A. The role of essential fatty acids in development. Annu Rev Nutr 2005;25:549-71.

Volume ■■ 32. Brenna JT. Arachidonic acid needed in infant formula when docosahexaenoic acid is present. Nutr Rev 2016;74:329-36. 33. Janssen CI, Kiliaan AJ. Long-chain polyunsaturated fatty acids (LCPUFA) from genesis to senescence: the influence of LCPUFA on neural development, aging, and neurodegeneration. Prog Lipid Res 2014;53:1-17. 34. Blank C, Neumann MA, Makrides M, Gibson RA. Optimizing DHA levels in piglets by lowering the linoleic acid to alpha-linolenic acid ratio. J Lipid Res 2002;43:1537-43. 35. Novak EM, Dyer RA, Innis SM. High dietary omega-6 fatty acids contribute to reduced docosahexaenoic acid in the developing brain and inhibit secondary neurite growth. Brain Res 2008;1237:136-45. 36. Gould JF, Smithers LG, Makrides M. The effect of maternal omega-3 (n3) LCPUFA supplementation during pregnancy on early childhood cognitive and visual development: a systematic review and meta-analysis of randomized controlled trials. Am J Clin Nutr 2013;97:531-44. 37. Bernard JY, Armand M, Forhan A, De Agostini M, Charles M-A, Heude B. Early life exposure to polyunsaturated fatty acids and psychomotor development in children from the EDEN mother-child cohort. OCL 2016;23:D106. 38. Innis SM. Impact of maternal diet on human milk composition and neurological development of infants. Am J Clin Nutr 2014;99:734S-41S. 39. Tosi F, Sartori F, Guarini P, Olivieri O, Martinelli N. Delta-5 and delta-6 desaturases: crucial enzymes in polyunsaturated fatty acid-related pathways with pleiotropic influences in health and disease. Adv Exp Med Biol 2014;824:61-81. 40. Lattka E, Rzehak P, Szabo E, Jakobik V, Weck M, Weyermann M, et al. Genetic variants in the FADS gene cluster are associated with arachidonic acid concentrations of human breast milk at 1.5 and 6 mo postpartum and influence the course of milk dodecanoic, tetracosenoic, and trans-9-octadecenoic acid concentrations over the duration of lactation. Am J Clin Nutr 2011;93:382-91. 41. Caspi A, Williams B, Kim-Cohen J, Craig IW, Milne BJ, Poulton R, et al. Moderation of breastfeeding effects on the IQ by genetic variation in fatty acid metabolism. Proc Natl Acad Sci USA 2007;104:18860-5. 42. Steer CD, Davey Smith G, Emmett PM, Hibbeln JR, Golding J. FADS2 polymorphisms modify the effect of breastfeeding on child IQ. PLoS ONE 2010;5:e11570. 43. Steer CD, Lattka E, Koletzko B, Golding J, Hibbeln JR. Maternal fatty acids in pregnancy, FADS polymorphisms, and child intelligence quotient at 8 y of age. Am J Clin Nutr 2013;98:1575-82. 44. Garcia C, Duan RD, Brevaut-Malaty V, Gire C, Millet V, Simeoni U, et al. Bioactive compounds in human milk and intestinal health and maturity in preterm newborn: an overview. Cell Mol Biol (Noisy-Le-Grand) 2013;59:108-31.

8

Bernard et al FLA 5.4.0 DTD ■ YMPD8897_proof ■ January 9, 2017

ORIGINAL ARTICLES

2017

Appendix 1

Appendix 2

Additional members of the EDEN Mother-Child Cohort Study Group include: Isabella Annesi-Maesano, PhD (Inserm, Paris, France), Jérémie Botton, PhD (University of Paris-Sud, Villejuif, France), Patricia Dargent, PhD (Inserm, Villejuif, France), Blandine de Lauzon-Guillain, PhD (Inserm, Villejuif, France), Pierre Ducimetière, PhD (Inserm, Villejuif, France), Bernard Foliguet, MD (Nancy University Hospital, Nancy, France), Xavier Fritel, MD (Poitiers University Hospital, Poitiers, France), Alice Germa, DMD, PhD (Paris Descartes University, Paris, France), Valérie Goua, MD (Poitiers University Hospital, Poitiers, France), Régis Hankard, MD, PhD (Poitiers University Hospital, Poitiers, France), Monique Kaminski, PhD (Inserm, Paris, France), Béatrice Larroque†, MD (Inserm, Paris, France), Nathalie Lelong, MSc (Inserm, Paris, France), Johanna Lepeule, PhD (Inserm, Grenoble, France), Guillaume Magnin, MD (Poitiers University Hospital, Poitiers, France), Laetitia Marchand, MSc (Inserm, Paris, France), Cathy Nabet, MD (Paul Sabatier University, Toulouse, France), Fabrice Pierre, MD (Poitiers University Hospital, Poitiers, France), Rémy Slama, PhD (Inserm, Grenoble, France), Marie-Josèphe Saurel-Cubizolles, PhD (Inserm, Paris, France), Michel Schweitzer, MD (Nancy University Hospital, Nancy, France), Olivier Thiebaugeorges, MD, PhD (Nancy University Hospital, Nancy, France).

Supported by: Foundation for medical research, National Agency for Research, National Institute for Research in Public health (TGIR cohorte santé 2008 program), French Ministry of Health, French Ministry of Research, INSERM Bone and Joint Diseases National Research and Human Nutrition National Research Programs, Paris–Sud University, Nestlé, French National Institute for Population Health Surveillance, French National Institute for Health Education, the European Union (FP7/2007-2013, HELIX, ESCAPE, ENRIECO, Medall projects), Diabetes National Research Program (through a collaboration with the French Association of Diabetic Patients), French Agency for Environmental Health Safety (now ANSES), Mutuelle Générale de l’Education Nationale, a complementary health insurance, French national agency for food security, French speaking association for the study of diabetes and metabolism (ALFEDIAM). Biological analyses of colostrum samples were funded by the PremUp foundation (foundation for scientific cooperation in connection with pregnancy and prematurity), the Groupe Lipides Nutrition, the ANSSD Program from Inra. Study sponsors were not involved in the study design, collection, analysis and interpretation of the data, the drafting of the manuscript and the decision to submit the manuscript for publication. J.B. postdoctoral fellowship was funded by the French National Research Agency (ANR-12DSSA-0005-01)). C.G.’s PhD was funded by the Conseil Régional Provence-Alpes-Côtes d’Azur and Applications Santé des Lipides.

Breastfeeding, Polyunsaturated Fatty Acid Levels in Colostrum and Child Intelligence Quotient at Age 5-6 Years FLA 5.4.0 DTD ■ YMPD8897_proof ■ January 9, 2017

8.e1

THE JOURNAL OF PEDIATRICS • www.jpeds.com

Volume ■■

Figure 1. Flow chart of the cohort participants included in the present analysis.

Table III. Characteristics of the ever breastfed participants with IQ data, and with or without PUFA data*

Characteristics of children Sex, % boy Gestational age at birth, wk Born preterm (<37 wk), % Birth weight, z score Breastfeeding Any breastfeeding duration, mo Exclusive breastfeeding duration, mo WPPSI-III at 5-6 y of age Age at assessment, y Full scale IQ Verbal IQ Performance IQ Characteristics of mothers Study center, % Poitiers Age at inclusion, y Primiparity, % Prepregnancy body mass index, kg/m2 Tobacco smoking during pregnancy, % Alcohol drinking during pregnancy, % 0 g/wk 1-9 g/wk ≥10 g/wk Score of depression during pregnancy Characteristics of families Parental education, y Household income class HOME-SF score

With PUFA data (n = 539)

Without PUFA data (n = 260)

49.7 39.4 ± 1.4 4.1 −0.05 ± 0.97

55.8 39.4 ± 1.7 7.3 0.01 ± 1.00

4.7 ± 4.2 2.8 ± 2.8

3.9 ± 3.7 2.1 ± 2.3

.004 .0002

5.65 ± 0.17 105.6 ± 12.5 108.9 ± 13.4 101.7 ± 12.7

5.67 ± 0.12 102.7 ± 12.4 107.0 ± 13.4 98.7 ± 13.5

.15 .002 .07 .002

41.4 29.6 ± 4.6 46.8 23.0 ± 4.5 19.5

71.2 30.0 ± 4.9 48.1 23.7 ± 4.6 20.0

<.0001 .26 .73 .05 .86 .16

56.8 36.7 6.5 10.4 ± 7.3

50.4 40.4 9.2 11.0 ± 7.6

.35

13.9 ± 2.3 5.1 ± 1.2 17.1 ± 2.2

13.6 ± 2.2 5.0 ± 1.3 17.5 ± 2.1

.10 .25 .02

P .11 .75 0.05 .44

HOME-SF, Home Observation Measurement of the Environment-Short Form. *Values are % or mean ± SD. P values are based on c2 and Student t tests, respectively, for categorical and continuous variables.

8.e2

Bernard et al FLA 5.4.0 DTD ■ YMPD8897_proof ■ January 9, 2017

ORIGINAL ARTICLES

2017

Table IV. Adjusted associations between colostrum PUFA levels and IQs among breastfed children (n = 539)* Fatty acids n-6 PUFA 18:2 n-6 (LA) 20:4 n-6 (AA) Total n-6 LC-PUFA n-3 PUFA 18:3 n-3 (ALA) 20:5 n-3 (EPA) 22:6 n-3 (DHA) Total n-3 LC-PUFA n-6/n-3 PUFA ratios LA/ALA AA/DHA Total n-6/n-3 PUFA

Full scale IQ

Verbal IQ

Performance IQ

b (95% CI)

b (95% CI)

b (95% CI)

−0.4 (−0.9, 0.2) 6.0 (−0.8, 12.8) 2.3 (−0.3, 4.9)

−0.6 (−1.1, 0.0) 5.2 (−1.9, 12.3) 1.4 (−1.4, 4.1)

−0.1 (−0.7, 0.5) 5.7 (−1.5, 12.9) 2.3 (−0.4, 5.1)

−2.3 4.8 1.2 1.1

0.1 19.5 3.1 1.8

(−7.2, 2.6) (−28.2, 37.8) (−4.3, 6.7) (−2.2, 4.4)

0.0 (−0.2, 0.2) 0.4 (−2.6, 3.4) −0.4 (−1.1, 0.3)

(−5.1, 5.2) (−15, 53.9) (−2.7, 8.8) (−1.7, 5.2)

−4.1 −20.3 −3.2 −1.2

−0.1 (−0.3, 0.1) −0.5 (−3.6, 2.6) −0.9 (−1.6, −0.1)

(−9.3, 1.1) (−55.1, 14.6) (−9.1, 2.6) (−4.7, 2.3)

0.2 (0.0, 0.4) 2.5 (−0.7, 5.6) 0.3 (−0.4, 1.0)

*Models were adjusted for study center, maternal age, prepregnancy body mass index, depression during pregnancy, tobacco and alcohol consumption during pregnancy, parity, child's sex, birth weight z score, gestational age at birth, parental education, household income, family stimulation score, exclusive breastfeeding duration ,and delay between birth and colostrum collection.

Breastfeeding, Polyunsaturated Fatty Acid Levels in Colostrum and Child Intelligence Quotient at Age 5-6 Years FLA 5.4.0 DTD ■ YMPD8897_proof ■ January 9, 2017

8.e3