ARTICLE

Beneficial Effects of Breast Milk in the Neonatal

Intensive Care Unit on the Developmental Outcome
of Extremely Low Birth Weight Infants at 18 Months
of Age
Betty R. Vohr, MDa, Brenda B. Poindexter, MD, MSb, Anna M. Dusick, MDb, Leslie T. McKinley, MS, RDa, Linda L. Wright, MDc, John C. Langer,
MScd, W. Kenneth Poole, PhDd, for the NICHD Neonatal Research Networka

aDepartment of Pediatrics, Brown Medical School, Providence, Rhode Island; bDepartment of Pediatrics, Indiana University School of Medicine, Indianapolis, Indiana;
cCenter for Research for Mothers and Children, National Institute of Child Health and Human Development, Rockville, Maryland; dRTI International, Research Triangle
Park, North Carolina

The authors have indicated they have no financial relationships relevant to this article to disclose.

ABSTRACT
OBJECTIVE. Beneficial effects of breast milk on cognitive skills and behavior ratings
have been demonstrated previously in term and very low birth weight infants.
www.pediatrics.org/cgi/doi/10.1542/
Extremely low birth weight infants are known to be at increased risk for devel- peds.2005-2382
opmental and behavior morbidities. The benefits of breast milk that is ingested in doi:10.1542/peds.2005-2382
the NICU by extremely low birth weight infants on development and behavior
Key Words
have not been evaluated previously. human milk, extremely low birth weight,
Bayley, outcomes
METHODS. Nutrition data including enteral and parenteral feeds were collected pro-
Abbreviations
spectively, and follow-up assessments of 1035 extremely low birth weight infants LC-PUFA—long-chain polyunsaturated
at 18 months’ corrected age were completed at 15 sites that were participants in fatty acids
ELBW— extremely low birth weight
the National Institute of Child Health and Human Development Neonatal Research AAP—American Academy of Pediatrics
Network Glutamine Trial between October 14, 1999, and June 25, 2001. Total CA— corrected age
NICHD—National Institute of Child Health
volume of breast milk feeds (mL/kg per day) during hospitalization was calculated. and Human Development
Neonatal characteristics and morbidities, interim history, and neurodevelopmental BSID II—Bayley Scales of Infant
Development II
and growth outcomes at 18 to 22 months’ corrected age were assessed.
BRS—Behavior Rating Scale
MDI—Mental Development Index
RESULTS. There were 775 (74.9%) infants in the breast milk and 260 (25.1%) infants
PSI—Psychomotor Development Index
in the no breast milk group. Infants in the breast milk group were similar to those VLBW—very low birth weight
in the no breast milk group in every neonatal characteristic and morbidity, Accepted for publication Jan 31, 2006
including number of days of hospitalization. Mean age of first day of breast milk for Address correspondence to Betty R. Vohr, MD,
Women and Infants Hospital, 101 Dudley St,
the breast milk infants was 9.3 ⫾ 9 days. Infants in the breast milk group began to Providence, RI 02905. E-mail: Betty_Vohr@
ingest non– breast milk formula later (22.8 vs 7.3 days) compared with the non– brown.edu

breast milk group. Age at achieving full enteral feeds was similar between the PEDIATRICS (ISSN Numbers: Print, 0031-4005;
Online, 1098-4275). Copyright © 2006 by the
breast milk and non– breast milk groups (29.0 ⫾ 18 vs 27.4 ⫾ 15). Energy intakes American Academy of Pediatrics
of 107.5 kg/day and 105.9 kg/day during the hospitalization did not differ between
the breast milk and non– breast milk groups, respectively. At discharge, 30.6% of

PEDIATRICS Volume 118, Number 1, July 2006 e115

Downloaded from http://pediatrics.aappublications.org/ by guest on October 18, 2017
infants in the breast milk group still were receiving Mental Development Index would be 10 ⫻ 0.53, or 5.3
breast milk. Mothers in the breast milk group were points.
significantly more likely to be white (42% vs 27%), be
CONCLUSIONS. An increase of 5 points potentially would
married (50% vs 30%), have a college degree (22% vs
6%), and have private health insurance (34% vs 18%) optimize outcomes and decrease costs by decreasing the
compared with the no breast milk group. Mothers who number of very low birth weight children who require
were black, had a low household income (ⱕ$20 000), or special education services. The societal implications of a
had higher parity were less likely to provide breast milk 5-point potential difference (one third of an SD) in IQ
feeds. The analysis of outcomes between the any human are substantial. The potential long-term benefit of re-
milk and no human milk groups were adjusted for ma- ceiving breast milk in the NICU for extremely low birth
ternal age, maternal education, marital status, race/eth- weight infants may be to optimize cognitive potential
nicity, and the other standard covariates. Children in the and reduce the need for early intervention and special
breast milk group were more likely to have a Bayley education services.
Mental Development Index ⱖ85, higher mean Bayley
Psychomotor Development Index, and higher Bayley
Behavior Rating Scale percentile scores for orientation/
engagement, motor regulation, and total score. There N UMEROUS BENEFICIAL EFFECTS of breast milk have
been demonstrated for term and near-term infants,
including improved cognitive skills,1–10 improved behav-
were no differences in the rates of moderate to severe
cerebral palsy or blindness or hearing impairment be- ior ratings,11–13 and decreased rates of infection.14–19 Im-
tween the 2 study groups. There were no differences in proved neurodevelopment has been related to the pres-
the mean weight (10.4 kg vs 10.4 kg), length (80.5 cm vs ence of long-chain polyunsaturated fatty acids (LC-
80.5 cm), or head circumference (46.8 cm vs 46.6 cm) PUFA; arachidonic and docosahexaenoic), which are
for the breast milk and no breast milk groups, respec- found in human milk but not bovine milk.20–29 Prema-
tively, at 18 months. Multivariate analyses, adjusting for ture infants are immunologically immature at birth and
confounders, confirmed a significant independent asso- may have deficiencies of LC-PUFA because accretion
ciation of breast milk on all 4 primary outcomes: the occurs in the third trimester. Decreased rates of infection
mean Bayley (Mental Development Index, Psychomotor have been associated with the enhanced immunologic
Development Index, Behavior Rating Scale, and inci- properties of breast milk, including lactoferrin, ly-
dence of rehospitalization). For every 10-mL/kg per day sozyme, and secretory IgA.30,31 Extremely low birth
increase in breast milk ingestion, the Mental Develop- weight (ELBW) infants are at increased risk for neuro-
ment Index increased by 0.53 points, the Psychomotor developmental disability and for rehospitalization after
Development Index increased by 0.63 points, the Behav- discharge from the NICU.32 In the past, this vulnerable
ior Rating Scale percentile score increased by 0.82 population of high-risk neonates has had limited expo-
points, and the likelihood of rehospitalization decreased sure to breast milk in the NICU. However, in 1997 and
by 6%. In an effort to identify a threshold effect of breast 2005, the American Academy of Pediatrics (AAP)33,34
milk on Bayley Mental Development Index and Psy- published position statements recommending breast
chomotor Development Index scores and Behavior Rat- milk for premature and other high-risk infants by breast-
ing Scale percentile scores, the mean volume of breast feeding and/or using the mother’s own expressed milk.
milk per kilogram per day during the hospitalization was The purpose of this study was to identify the rate of
calculated, and infants in the breast milk group were breast milk ingestion in the NICU in a cohort of ELBW
divided into quintiles of breast milk ingestion adjusted infants who were born after the publication of the 1997
for confounders. Overall, the differences across the feed- AAP statement and to assess the relationship of breast
ing quintiles of Mental Development Index and Psy- milk ingestion with developmental and behavior test
chomotor Development Index were significant. There scores and rehospitalization after discharge from the
was a 14.0% difference in Behavior Rating Scale scores NICU. It was hypothesized that ELBW infants who re-
between the lowest and highest quintiles. For the out- ceived breast milk in the NICU would have (1) higher
comes (Mental Development Index, Psychomotor De- Bayley35 mental, psychomotor, and behavior scores at 18
velopment Index, Behavior Rating Scale, and Rehospi- to 22 months’ corrected age (CA) and (2) fewer rehos-
talization ⬍1 year), only the values for the ⬎80th pitalizations between hospital discharge and 18 to 22
percentile quintile of breast milk feeding were signifi- months’ CA.
cantly different from the no breast milk values. In our
adjusted regression analyses, every 10 mL/kg per day METHODS
breast milk contributed 0.53 points to the Bayley Mental The study cohort was derived from the 1433 infants who
Development Index; therefore, the impact of breast milk were enrolled prospectively in the Glutamine Trial36 at
ingestion during the hospitalization for infants in the 15 sites of the National Institute of Child Health and
highest quintile (110 mL/kg per day) on the Bayley Human Development (NICHD) Neonatal Research Net-

e116 VOHR et al
Downloaded from http://pediatrics.aappublications.org/ by guest on October 18, 2017
work between October 1999 and June 2001. A total of months’ CA.32 BSID II scores of 100 ⫾ 15 represent the
251 infants died before discharge, and 23 infants died mean ⫾ 1 SD. An abnormal score (⬍70) is ⬎2 SD below
before the 18-month visit. A total of 124 infants were the mean. BRS percentile scores range from 1 to 99;
lost to follow-up. The final sample consisted of 1035 scores of 11 to 25 are considered questionable, and
(89.4%) of 1159 ELBW infant survivors on whom fol- scores of 1 to 10 are nonoptimal.
low-up data were available. Mothers of infants who The primary caregiver or adult who brought the child
were seen in follow-up were more likely to have re- for the visit stayed with the child during the BSID II
ceived prenatal care than mothers of infants who were examination, which was administered early in the clinic
not seen (93% vs 86%; P ⬍ .01, respectively). There visit before the medical assessment and interviews. Ex-
were no differences in infant characteristics or morbidi- aminers were not able to administer parts or all of the
ties between those who were and were not seen in BSID II successfully to 95 children who were seen. The
follow-up. following reasons were given: acute illness (n ⫽ 10),
The Glutamine Trial encouraged early parenteral nu- language barrier (n ⫽ 2), behavior problem (n ⫽ 24),
trition but did not prescribe an enteral feeding protocol. developmental delay (n ⫽ 18), and other (n ⫽ 41). The
Nutrition data were collected daily until the infants were other category included children who had sensory loss
on full enteral feeds (ⱖ462 kJ/kg per day). Thereafter, (blindness or deafness) and could not be administered
data were collected on Monday, Wednesday, and Friday the BSID II items. These data therefore were not in-
of each week until discharge or 120 days; values were cluded in the analysis. Although every effort was made
interpolated for days of the week on which the data to test children within the window of 18 to 22 months’
were not collected: Tuesday data were the weighted CA, 22 infants were evaluated outside the window be-
average of Monday and Wednesday data. The total vol- cause of illness or tracking issues. These data were in-
ume of breast milk feeds per kilogram per day for the cluded because the BSID II is age adjusted.
duration of hospitalization was calculated. A total of 775 Social and economic status information, including
(74.9%) of the infants received some breast milk during maternal and paternal education and occupation, mari-
their NICU hospitalization. None of the centers used tal status, insurance status, and income level, and a
banked breast milk during the trial. The proportion of detailed interim medical history, including data on hear-
infants who received breast milk feeds at centers ranged ing and vision status, were obtained. Hearing status in-
from a low of 39.1% to a high of 97.3%. Information formation was obtained from the parent and follow-up
was not collected on breast milk ingestion after dis- audiologic test results when available. Hearing impair-
charge. Neonatal characteristics and morbidities were ment was defined as use of hearing aids. A history of
collected, including gestational age by best obstetric es- postdischarge eye examinations and procedures was ob-
timate, birth weight, and neonatal course. Institutional tained from the parent. In addition, a standard eye ex-
Review Board approval and informed consent were ob- amination was completed to evaluate tracking, esotro-
tained as previously described.36 pia, nystagmus, or roving eye movements. Blind was
Infants were evaluated at 18 to 22 months’ CA. The defined as functional corrected vision ⬍20/200.
assessment included standardized interim medical his- Statistical analyses were completed by RTI Interna-
tory, a developmental evaluation, neurologic assess- tional. Bivariate analyses for group differences (any
ment, and physical examination that included growth breast milk versus no breast milk) consisted of t tests, ␹2,
parameters.32 A neurologic examination that was based or Fisher’s exact tests. We evaluated the relationship
on the Amiel-Tison37 assessment was performed by cer- between breast milk feeding and developmental out-
tified examiners. The neurologic assessment included an comes in 3 ways: (1) whether any breast milk feeding
evaluation of tone, strength, reflexes, angles, and pos- was beneficial (breast milk yes versus breast milk no),
ture. Infants were scored as normal when no abnormal- (2) whether the amount of breast milk feeding was
ities were observed in the neurologic examination. Ce- associated with better developmental outcomes (breast
rebral palsy was defined as a nonprogressive central milk as a continuous variable), and (3) whether there
nervous system disorder characterized by abnormal was a threshold beyond which breast milk feeding pro-
muscle tone in at least 1 extremity and abnormal control vided benefits. Infants who received breast milk were
of movement and posture. divided into quintiles of mean intake of breast milk per
The Bayley Scales of Infant Development II (BSID kilogram per day during the entire hospitalization, in-
II),35 including the Mental Scale, Motor Scale, and the cluding 0 for days with no intake, for analysis. Mean
Behavior Rating Scale (BRS), were administered by volume of human milk (mL/kg per day) for the entire
testers who were trained to reliability by 1 of 4 study hospitalization was entered as a continuous variable.
Gold Standard examiners. Examiner certification at sites Multivariate analyses to evaluate the effects of breast
was obtained by the successful completion of 2 video- milk on outcomes consisted of multiple linear-regression
taped demonstrations of accurate performance and scor- and logistic-regression analyses. Adjustments were made
ing of the BSID II on preterm children at 18 to 22 for the following confounders: mother’s age, education,

PEDIATRICS Volume 118, Number 1, July 2006 e117

Downloaded from http://pediatrics.aappublications.org/ by guest on October 18, 2017
marital status, and race/ethnicity and infant’s gestational TABLE 2 Maternal Characteristics by Breast Milk Feeding
age, gender, culture-positive sepsis, intraventricular Breast Milk No Breast Milk P
hemorrhage 3 to 4,38 periventricular leukomalacia, oxy- N (%) 775 (74.9) 260 (25.1) —
gen at 36 weeks’ CA, necrotizing enterocolitis, and Maternal age, y
weight ⬍10th percentile at 18 months. ⬍20 113 (15) 45 (17)
20–29 363 (47) 128 (49) .2793
ⱖ30 299 (39) 87 (33)
RESULTS
Married 387 (50) 78 (30) .001
There were 775 (74.9%) infants in the breast milk group Education
and 260 (25.1%) infants in the no breast milk group. Less than high school graduate 206 (28) 83 (35)
Table 1 shows the infant characteristics and morbidities High school graduate 197 (27) 94 (40) .0001
of the breast milk and no breast milk groups. Infants in Some college 162 (22) 35 (15)
College degree or above 163 (22) 15 (6)
the breast milk group were similar to those in the no
Private insurance or HMO 253 (34) 43 (18) .001
breast milk group in every neonatal characteristic and Race/ethnicity
morbidity, including number of days of hospitalization. Black 303 (39) 167 (64)
Mean age of first day of breast milk for the breast milk White 325 (42) 71 (27) .0001
infants was 9.3 ⫾ 9 days. Infants in the breast milk group Hispanic 133 (17) 18 (7)
Other 14 (2) 4 (2)
began to ingest non– breast milk formula later (22.8 vs
Income last year ⬍$20 000 280 (42) 127 (59) .001
7.3 days) compared with the non– breast milk group. Parity, mean ⫾ SD 2.1 ⫾ 1.3 2.7 ⫾ 1.8 .001
Age at achieving full enteral feeds was similar between Data are n (%) except where indicated. HMO indicates health maintenance organization.
the breast milk and non– breast milk groups (29.0 ⫾ 18
vs 27.4 ⫾ 15). Energy intakes of 107.5 kg/day and 105.9
kg/day during hospitalization did not differ between the
breast milk and non– breast milk group, respectively. comes between the any breast milk and no breast milk
There were center differences in the rate of use of hu- groups adjusted for maternal age, maternal education,
man milk fortifiers; however, 74.7% of the breast milk marital status, race/ethnicity, and the other standard
feeds without parenteral nutrition were fortified (61.7% covariates. Children in the breast milk group were more
with breast milk fortifiers; 13% with other nutritional likely to have a Bayley Mental Development Index
additives). At discharge, 30.6% of infants in the breast (MDI) ⱖ85, higher mean Bayley Psychomotor Develop-
milk group still were receiving breast milk. ment Index (PDI), and higher Bayley BRS percentile
Table 2 shows the maternal characteristics of the 2 scores for orientation/engagement, motor regulation,
study groups. Mothers in the breast milk group were and total score. There were no differences in the rates of
significantly more likely to be white (42% vs 27%), be moderate to severe cerebral palsy or blindness or hearing
married (50% vs 30%), have a college degree (22% vs impairment between the 2 study groups
6%), and have private health insurance (34% vs 18%) There were no differences in the mean weight (10.4
compared with the no breast milk group. Mothers who kg vs 10.4 kg), length (80.5 cm vs 80.5 cm), or head
were black, had a low household income (ⱕ$20 000), or circumference (46.8 cm vs 46.6 cm) for the breast milk
had higher parity were less likely to provide breast milk and no breast milk groups at 18 months’ CA. Rehospi-
feeds. talization between birth and the first birthday was lower
Table 3 shows the results of the analyses of the out- in the breast milk group (23.3% vs 30.1%; P ⬍ .028)
compared with the no breast milk group. There were no
differences in rehospitalization rates between the first
TABLE 1 Infant Characteristics and Morbidities birthday and the visit at 18 to 22 months’ CA (data not
Breast Milk No Breast Milk P
shown).
Multiple-regression analyses, adjusting for confound-
N (%) 775 (74.9) 260 (25.1%) —
Female gender 425 (55) 140 (54) .781 ers, confirmed a significant independent association of
Birth weight, g 785 ⫾ 129 794 ⫾ 133 .363 breast milk on all 4 primary outcomes: the mean Bayley
Gestational age, wk 26.5 ⫾ 2 26.7 ⫾ 2 .331 (MDI, PDI, BRS, and incidence of rehospitalization) as
IVH grade 3–4 84 (11) 25 (10) .586 shown in Table 4. For every 10-mL/kg per day increase
PVL 29 (4) 10 (4) .931
in breast milk ingestion, the MDI increased by 0.53
Proven necrotizing enterocolitis 54 (7) 26 (10) .113
Early-onset sepsis ⱕ72 h 6 (1) 3 (1) .568 points, the PDI increase by 0.63 points, the BRS percen-
Late-onset sepsis ⬎72 h 311 (40) 91 (35) .142 tile score increased by 0.82 points, and the likelihood of
Bronchopulmonary dysplasiaa 349 (46) 120 (47) .714 rehospitalization decreased by 6%. We repeated the
Days in hospital 99 ⫾ 42 96 ⫾ 41 .301 multivariate analysis excluding the 237 (30.6%) chil-
Data are mean ⫾ SD or n (%). IVH indicates, intraventricular hemorrhage; PVL, periventricular dren who were discharged on breast milk to determine
leukomalacia.
a Bronchopulmonary dysplasia was defined as oxygen administration at 36 weeks’ postmen- whether breast milk effects remained for infants who
strual age. received breast milk only in the NICU. Effects of breast

e118 VOHR et al
Downloaded from http://pediatrics.aappublications.org/ by guest on October 18, 2017
 TABLE 3 Outcomes by Breast Milk Feeding
Outcomes Breast Milk, n (%) No Breast Milk, n (%) Adjusted P
or Mean ⫾ SD or Mean ⫾ SD
MDI 79.9 ⫾ 18 75.8 ⫾ 16 .0709
MDI ⬍85 421 (58.1) 168 (70.9) .0355
PDI 84.6 ⫾ 19 81.3 ⫾ 17 .0227
PDI ⬍85 307 (42.9) 116 (49.2) .0939
BRS
Orientation/engagement 53.8 ⫾ 30 48.5 ⫾ 29 .0351
Emotional regulation 53.0 ⫾ 29 47.1 ⫾ 29 .0825
Motor quality 50.1 ⫾ 35 44.1 ⫾ 33 .0296
Total 52.0 ⫾ 31 45.6 ⫾ 30 .0280
Blind in both eyes 4 (0.5) 0 (0.0) .5773a
Hearing aids 9 (1.2) 4 (1.6) .7472a
Cerebral palsy (moderate/severe) 43 (5.6) 16 (6.2) .7327a
Rehospitalizations before first birthday 180 (23.3) 78 (30.1) .028
Adjustments were made for the following confounders: mother’s age, education, marital status, and race and infant’s gestation, gender, sepsis,
IVH 3 to 4, PVL, O2 at 36 weeks, necrotizing enterocolitis, weight ⬍10th percentile at 18 months, and volume of breast milk (mL/kg per day) for
every day of hospitalization (a continuous variable).
a Unadjusted P.

TABLE 4 Effect of Breast Milk Feeding (mL/kg per day) on no breast milk group. Breast milk intake by quintile
Developmental Outcomes and Rehospitalization: Multiple ranged from 1.0 mL/kg per day to 110.6 mL/kg per day.
Regression Results In addition, total breast milk on days when only breast
Outcomes Parameter P milk was ingested was calculated and ranged from 22.1
Estimate mL/kg per day to 124.0 mL/kg per d. The proportion of
MDI 0.53 .0002 infants who were discharged on breast milk ranged from
PDI 0.63 ⬍.0001 ⬃0.7% for the lowest quintile to 85.1% for the highest
BRS 0.82 .0025
quintile. Mean Bayley MDI, PDI, and BRS were calcu-
Logistic regression OR 95% CI lated for the quintiles, and the relationship between
Rehospitalization for infection or respiratory 0.94 0.90–0.98 BSID II scores and quintiles of breast milk are shown.
morbidity
For MDI, there was a 13.1-point difference between
Estimates are adjusted for standard covariates listed previously. OR indicates odds ratio; CI,
confidence interval.
ⱕ20th quintile and ⬎80th quintile, and for PDI, there
was an 8.8-point difference between ⱕ20th quintile
and ⬎80th quintile. Overall, the differences across the
milk on all of our study outcomes remained significant. feeding quintiles of MDI and PDI were significant (mod-
(data not shown). el-adjusted P ⬍ .0044 and P ⬍ .0027, respectively).
In an effort to identify a threshold effect of breast milk There was a 14.0% difference in BRS scores between the
on Bayley MDI and PDI scores and BRS percentile lowest and highest quintiles (P ⬍ .028). For the out-
scores, we calculated the mean volume of breast milk comes presented in Table 4 (MDI, PDI, BRS, and Rehos-
per kilogram per day during the hospitalization, and pitalization ⬍1 year), only the values for the ⬎80th
infants in the breast milk group were divided into quin- percentile of breast milk feeding are significantly differ-
tiles of breast milk ingestion adjusted for confounders as ent from the no breast milk values at P ⬍ .01 within the
shown in Table 5. Quintiles were analyzed relative to the multiple-regression analyses. The values for MDI, PDI,

TABLE 5 Breast Milk Feeding Measures, Developmental Outcomes, and Rehospitalization by Quintiles of Breast Milk Intake During
Hospitalization
No Breast Milk Percentile Adjusted P
ⱕ20th 20th–40th 40th–60th 60th–80th ⬎80th
Total breast milk (mL/kg per day) for every day of hospitalization 0.0 1.0 7.3 24.0 63.8 110.6 —
Total breast milk (mL/kg per day) on days breast milk received 0.0 22.1 45.0 66.8 95.3 124.0 —
% Discharged on breast milk 0.0 0.7 2.1 8.6 40.2 85.1 —
Mean MDI score 75.8 74.2 76.9 78.3 80.4 87.3a ⬍.0044
Mean PDI score 81.3 80.6 82.7 84.2 84.4 89.4a .0027
Mean total BRS percentile score 45.6 44.8 52.1 50.1 51.8 58.8a .0281
Rehospitalized ⬍1 y 30.2 25.2 32.2 26.0 23.2 12.7a .0460
a Value significantly different from no breast milk value at P ⬍ .01.

PEDIATRICS Volume 118, Number 1, July 2006 e119

Downloaded from http://pediatrics.aappublications.org/ by guest on October 18, 2017
BRS, and rehospitalization at the 40th to 60th percen- ELBW infants during their NICU stay was associated
tiles and the 60th to 80th percentiles trended toward with higher Bayley MDI and PDI developmental scores,
significant. higher BRS scores, and fewer rehospitalizations during
the first year. These effects of breast milk on Bayley MDI
DISCUSSION scores remained significant after adjustment for impor-
ELBW infants are known to be at increased risk for tant environmental confounders, including marital sta-
developmental delays, neurosensory impairments, poor tus, education, health insurance, income, race/ethnicity,
growth, and rehospitalization after discharge. The pri- and parity. Although there was a 6% reduction in re-
mary objective of perinatal and neonatal studies of hospitalization, other factors are associated with in-
ELBW infants has been to identify interventions that creased likelihood of rehospitalization, including smok-
improve these outcomes. Breast milk has been shown ing in the household and child care attendance, which
consistently to have beneficial effects for term infants. were not available.
Studies in premature infants have shown both positive There were no differences between the any breast
effects1–6,8–10,39 and no effects40 on cognitive outcome and milk and no breast milk study groups in growth param-
positive effects on behavior ratings.11–13 eters at 18 to 22 months’ CA, which we attribute to
Our study to investigate the effects of breast milk on current NICU and postdischarge management of provid-
BSID II scores and rehospitalization would be difficult to ing enhanced nutritional intake for ELBW infants. En-
assess in a breast milk/no breast milk randomized trial teral feeds consist of either high-nutrient preterm for-
because method of feeding is determined by the mother. mulas or breast milk supplemented with breast milk
In addition, the AAP recommends breast milk for pre- fortifier and preterm formulas.
mature infants; therefore, there is no true equipoise. In There also were no differences in neurosensory out-
the studies of Lucas et al,1–6 randomization to study comes, including the number of children with cerebral
group was made after the mothers made the choice to palsy or blindness or hearing impairment. The differ-
bottle or breastfeeding. These studies compared breast ences in rehospitalization in the first year of life seem to
milk in various combinations to term formula, preterm be associated with fewer serious infections and respira-
formula, and banked breast milk. The prospective de- tory illnesses reported by the parents after discharge and
tailed daily neonatal enteral, including breast milk, and are suggestive of enhanced immune protection. This
parenteral nutritional intake data that were collected effect was not observed in the second year of life. Al-
longitudinally as part of the NICHD Network Glutamine though we do not have postdischarge data on breast
Trial36 on a large population of ELBW infants who were milk ingestion, we do know that only 30.6% of mothers
assessed at 18 to 22 months’ CA allowed us to examine were still providing breast milk at discharge. It is well
the relationship between breast milk and outcome. The documented that the percentage of mothers of ELBW
nutritional intake data revealed that 75% of the infants infants who still provide breast milk at discharge de-
received some breast milk during their NICU stay and creases rapidly after discharge.45–47
30.6% of infants were still receiving breast milk at the On the Bayley BRS, which is administered by a Bay-
time of discharge. ley examiner who is unaware of the neonatal feeding
Consistent with previous reports,41,42 mothers who history, infants who were fed breast milk had signifi-
provided breast milk were more likely to be married, be cantly higher (more optimal) scores for orientation/en-
better educated, have private health insurance, have gagement, emotional regulation, motor quality, and to-
higher income, and have lower parity. Black mothers tal behavior scores. Higher ratings for orientation,
were less likely to provide breast milk. It is of interest motor, and state regulation have been reported for
that there were no differences in any of the neonatal breastfed infants using the Brazelton Neonatal Behavior
characteristics or morbidities compared in bivariate anal- Assessment Scale.11 Enhanced arousability also has been
ysis. This is similar to an NICHD Network study43 that demonstrated in breastfed infants.12 Our data suggest
evaluated the relationship between late-onset sepsis and that effects of breast milk on the behavior of ELBW
breast milk. The bivariate relationship between any infants continue to be observed at 18 to 22 months’ CA.
breast milk feeding and sepsis versus no sepsis was not We do not, however, have information allowing us to
significant. However, on multivariate logistic regression, differentiate effects of breastfeeding from breast milk
it was identified that for every 10% increase in breast provided by tube or bottle feeding.
milk, there was a 5% decrease in the rate of sepsis. There are several mechanisms by which breast milk
Furman et al44 also reported a dose relationship between ingestion by infants might contribute to more optimal
increasing breast milk ingestion and decreasing rate of cognitive/behavioral outcomes. It has been postulated
sepsis. that the act of feeding at the breast, as well as the
Despite the limited amount of breast milk ingested interaction between mother and child, account for im-
during the NICU hospitalization, our results support our proved cognitive and behavioral outcome.48 However,
hypothesis that breast milk ingestion in the NICU by few ELBW infants in the NICU are feeding at the breast,

e120 VOHR et al
Downloaded from http://pediatrics.aappublications.org/ by guest on October 18, 2017
and none are breastfeeding exclusively. Differences in optimize outcomes and decrease costs by decreasing the
parenting style and maternal intelligence, which were number of VLBW children who required special educa-
not a part of this study protocol, also have been proposed tion services. Lucas et al5 reported that increased volume
to be important. These factors are difficult to measure of breast milk was significantly associated with higher
and cannot be ruled out. Our results, however, were test scores in VLBW infants and that effects remained
adjusted for maternal education. Another factor that was significant at 7.5 to 8 years of age. Although critics of
not evaluated is that there is greater bioavailability of breast milk studies have suggested that higher family
protein in breast milk compared with formula. The socioeconomic status is the major contributor to higher
fourth proposed mechanism is the role of specific com- developmental scores, both in our own cohort and in
ponents of breast milk, including LC-PUFA (arachidonic that of Lucas et al,5 breast milk effects remained signifi-
acid and docosahexanoic acid), choline, glycoproteins, cant after adjustment of measures of socioeconomic sta-
phospholipids, growth factors, and hormones, on cogni- tus confounders. A limitation of our study is that we did
tive function.20,36,49–51 These mechanisms may comple- not assess maternal IQ. The potential long-term benefit
ment each other. of receiving breast milk in the NICU for the ELBW infant
Our findings suggest that the effects of breast milk on and family may be to optimize cognitive potential and to
the ELBW population are similar to that previously reduce the needs for early intervention and special ed-
found in very low birth weight (VLBW) infants.1–6 The ucation services. Long-term follow-up of ELBW infants
mean gestation of infants in our study was lower at 26.6 is needed to determine whether these effects persist at
weeks. We speculate that the 12 to 14 weeks before term school age. A weakness of this study is that we do not
may be an important window of opportunity for this have data on use of breast milk after discharge, and we
vulnerable population of infants. Active brain develop- do not have specific information on the home environ-
ment, neurogenesis, migration, and synaptogenesis oc- ment.
cur during this time, and brain development may be
particularly responsive to “maternal nutrition.” Previous CONCLUSION
studies have shown beneficial effects of arachidonic acid The provision of breast milk to ELBW infants during the
and docosahexanoic acid in breast milk on neurodevel- neonatal period is an easy-to-implement, cost-effective
opment.20–24,26–29 During the study period, preterm for- intervention with a potential payoff that includes better
mulas with these nutrients were not available. Addi- developmental outcomes, more optimal behavior, and
tional prospective studies need to be initiated to examine fewer rehospitalizations.
the relationships between breast milk ingestion and
brain development before term and after discharge. ACKNOWLEDGMENTS
Strengths of the study include the large study popu- Members of the NICHD Neonatal Research Network:
lation, the detailed nutritional intake data collected, and Alan Jobe, MD, PhD, Chairman*; University of Cincin-
the finding of breast milk effects after controlling for nati, University of California at San Diego (U10
known environmental and biological confounders. The HD40461): Neil N. Finer, MD*, Yvonne Vaucher, MD‡,
13.1-point difference in Bayley MDI scores (74.2– 87.3) Chris Henderson, RCP, Martha Fuller, RN, MSN; Case
between the lowest and highest quintiles adjusted for Western Reserve University (U10 HD21364), Avroy A.
environmental confounders at 18 months is remarkable Fanaroff, MB, BCh*, Dee Wilson, MD‡, Nancy Newman,
and provides a measure of the gap between those with RN, Bonnie Siner, RN; University of Cincinnati (U10
the highest and lowest amount of breast milk feeding. In HD27853, M01 RR 08084): Edward F. Donovan, MD*,
our regression analyses adjusted for important maternal Jean Steichen, MD‡, Marcia Mersmann, RN, Teresa
social/environmental and infant biological confounders, Gratton, RN; Emory University (U10 HD27851): Barbara
every 10 mL/kg per day of breast milk contributed 0.53 J. Stoll, MD*, Ira Adams-Chapman, MD‡, Ellen Hale,
points to the Bayley MDI. Therefore the impact on the RN; Indiana University (U10 HD27856, M01 RR 00750):
Bayley MDI of breast milk for infants who were in the James A. Lemons, MD*, Anna Dusick, MD‡, Lucy
highest quintile and ingested 110 mL/kg per day would Miller, RN, Leslie Richard, MD; University of Miami
be 10 ⫻ 0.53, or 5.3 points. The societal implications of (U10 HD21397): Shahnaz Duara, MD*, Charles R.
a 5-point potential difference (one third of an SD) in IQ Bauer, MD‡, Ruth Everett, RN; National Institute of
are substantial.52 Annually, ⬃56 000 (1.4%) of infants in Child Health and Human Development: Linda L. Wright,
the United States are born VLBW.53 Reports54–57 indicate MD, Rosemary Higgins, MD, James Hanson, MD; Uni-
that 50% of VLBW infants (⬃28 000 children) require versity of New Mexico (U10 HD27881, M01 RR00997):
remedial or special education services at school age. Lu-Ann Papile, MD*, Conra Backstrom, RN; Research
Hack56 reported a history of lower academic achieve- Triangle Institute (U01 HD36790): W. Kenneth Poole,
ment in adult VLBW survivors for whom the IQ was 5 PhD, Abhik Das, PhD, Betty Hastings, Carolyn Petrie,
points lower than that in term control subjects (87 vs 92, MS; Stanford University (U10 HD27880, M01 RR
respectively). An increase of 5 points potentially would 00070): David K. Stevenson, MD*, Susan R. Hintz, MD‡,

PEDIATRICS Volume 118, Number 1, July 2006 e121

Downloaded from http://pediatrics.aappublications.org/ by guest on October 18, 2017
M. Bethany Ball, BS; University of Tennessee at Mem- 15. Savilahti E, Jarvenpaa AL, Raiha NC. Serum immunoglobulins
phis (U10 HD21415): Sheldon B. Korones, MD*, Henri- in preterm infants: comparison of human milk and formula
feeding. Pediatrics. 1983;72:312–316
etta Bada, MD‡, Tina Hudson, RN; University of Texas
16. Howie PW, Forsyth JS, Ogston SA, Clark A, Florey CD. Pro-
Health Science Center at Houston (U10 HD21373): Jon tective effect of breast feeding against infection. BMJ. 1990;
E. Tyson, MD, MPH*, Brenda Morris, MD‡, Georgia 300:11–16
McDavid, RN; University of Texas Southwestern Medical 17. Frank AL, Taber LH, Glezen WP, Kasel GL, Wells CR, Paredes
Center (U10 HD40689): Abbot R. Laptook, MD*, Sue A. Breast-feeding and respiratory virus infection. Pediatrics.
Broyles, MD‡, Susie Madison, RN, Jackie Hickman, RN; 1982;70:239 –245
18. Covert RF, Barman N, Domanico RS, Singh JK. Prior enteral
Wayne State University (U10 HD21385): Seetha Shan- nutrition with human milk protects against intestinal perforation
karan, MD*, Yvette Johnson, MD‡, Geraldine Muran, in infants who develop necrotizing enterocolitis [abstract]. Pediatr
RN, Debbie Kennedy, RN; Women and Infants Hospital Res. 1995;37:305A
(U10 HD27904): William Oh, MD*, Betty Vohr, MD‡, 19. Bier J, Oliver TL, Ferguson A, Vohr B. Human milk reduces
Angelita Hensman, RN, Lucy Noel, RN; Yale University outpatient infections in very low birth weight infants [ab-
stract]. Pediatr Res. 1999;45:120A
(U10 HD27871, M01 RR 06022): Richard A. Ehren-
20. Uauy R, Hoffman DR. Essential fatty acid requirements for
kranz, MD*‡; Patricia Gettner, RN, Elaine Romano, RN; normal eye and brain development. Semin Perinatol. 1991;15:
University of Alabama at Birmingham (U10 HD34216): 449 – 455
Waldemar A. Carlo, MD*, Myriam Peralta, MD‡, Monica 21. Carlson SE, Werkman SH, Rhodes PG, Tolley EA. Visual-acuity
V. Collins, RN, Vivien Phillips, RN. *Principal investiga- development in healthy preterm infants: effect of marine-oil
tor; ‡follow-up principal investigator. supplementation. Am J Clin Nutr. 1993;58:35– 42
22. Bourre JM, Francois M, Youyou A, et al. The effects of dietary
alpha-linolenic acid on the composition of nerve membranes,
enzymatic activity, amplitude of electrophysiological parame-
REFERENCES ters, resistance to poisons and performance of learning tasks in
1. Lucas A, Morley R, Cole TJ, Gore SM. A randomised multicen- rats. J Nutr. 1989;119:1880 –1892
tre study of human milk versus formula and later development 23. Agostoni C, Trojan S, Bellu R, Riva E, Giovannini M, et al.
in preterm infants. Arch Dis Child Fetal Neonatal Ed. 1994;70: Neurodevelopment quotient of healthy term infants at 4
F141–F146 months and feeding practice: the role of long-chain polyunsat-
2. Lucas A, Morley R, Cole TJ. Randomised trial of early diet in
urated fatty acids. Pediatr Res. 1995;38:262–266
preterm babies and later intelligence quotient. BMJ. 1998;317:
24. Agostoni C, Trojan S, Bellu R, Riva E, Bruzzese MG, Giovannini
1481–1487
M. Developmental quotient at 24 months and fatty acid compo-
3. Lucas A, Fewtrell MS, Morley R, et al. Randomized outcome
sition of diet in early infancy. Arch Dis Child. 1997;76:421– 424
trial of human milk fortification and developmental outcome
25. Carlson SE, Werkman SH, Peeples JM, Cooke RJ, Tolley EA.
in preterm infants. Am J Clin Nutr. 1996;64:142–151
Arachidonic acid status correlates with first year growth in
4. Lucas A, Gore SM, Cole TJ, et al. Multicentre trial on feeding
preterm infants. Proc Natl Acad Sci U S A. 1993;90:1073–1077
low birthweight infants: effects of diet on early growth. Arch Dis
26. Birch E, Birch D, Hoffman D, Hale L, Everett M, Uauy R.
Child. 1984;59:722–730
Breast-feeding and optimal visual development. J Pediatr Oph-
5. Lucas A, Morley R, Cole TJ, Lister G, Leeson-Payne C. Breast
thalmol Strabismus. 1993;30:33–38
milk and subsequent intelligence quotient in children born
27. Birch EE, Birch DG, Hoffman DR, Uauy R. Dietary essential
preterm. Lancet. 1992;339:261–264
fatty acid supply and visual acuity development. Invest Ophthal-
6. Lucas A, Morley R, Cole T, Gore S. A randomized multicenter
study of human milk versus formula and later development in mol Vis Sci. 1992;33:3242–3253
preterm infants. Arch Dis Child. 1994;71:288 –290 28. Makrides M, Simmer K, Goggin M, Gibson R. Erythrocytes
7. Horwood LJ, Darlow BA, Mogridge N. Breast milk feeding and docosahexaenoic acid correlates with the visual response of
cognitive ability at 7– 8 years. Arch Dis Child Fetal Neonatal Ed. healthy term infants. Pediatr Res. 1993;33:425– 427
2001;84:F23–F27 29. Innis SM, Nelson CM, Rioux MF, King DJ. Development of
8. Gale CR, Martyn CN. Breastfeeding, dummy use, and adult visual acuity in relation to plasma and erythrocyte omega-6
intelligence. Lancet. 1996;347:1072–1075 and omega-3 fatty acids in healthy term gestation infants. Am J
9. Pollock JI. Mother’s choice to provide breast milk and devel- Clin Nutr. 1994;60:347–352
opmental outcome. Arch Dis Child. 1989;64:763–764 30. Goldblum RM, Schanler RJ, Garza C, Goldman AS. Human
10. Jacobson SW, Jacobson JL, Dobbing J, Beijers RJW. Breast- milk feeding enhances the urinary excretion of immunologic
feeding and intelligence. Lancet. 1992;339:926 factors in low birth weight infants. Pediatr Res. 1989;25:
11. Hart S, Boylan LM, Carroll S, Musick YA, Lampe RM. Brief 184 –188
report: breast-fed one-week-olds demonstrate superior neu- 31. Hutchens TW, Henry JF, Yip T-T, et al. Origin of intact lacto-
robehavioral organization. J Pediatr Psychol. 2003;28:529 –534 ferrin and its DNA-binding fragments found in the urine of
12. Horne RS, Parslow PM, Ferens D, Watts AM, Adamson TM. human milk-fed preterm infants. Evaluation by stable isotope
Comparison of evoked arousability in breast and formula fed enrichment. Pediatr Res. 1991;29:243–260
infants. Arch Dis Child. 2004;89:22–25 32. Vohr BR, Wright LL, Dusick AM, et al. Neurodevelopmental
13. Ooylan LM, Hart S, Porter KB, Driskell JA. Vitamin B-6 con- and functional outcomes of extremely low birth weight infants
tent of breast milk and neonatal behavioral functioning. J Am in the National Institute of Child Health and Human Develop-
Diet Assoc. 2002;102:1433–1438 ment Neonatal Research Network, 1993–1994. Pediatrics. 2000;
14. Hylander MA, Strobino DM, Dhanireddy R. Human milk feed- 105:1216 –1226
ings and infection among very low birth weight infants. Pedi- 33. American Academy of Pediatrics, Work Group on Breastfeed-
atrics. 1998;102(3). Available at: www.pediatrics.org/cgi/ ing. Breastfeeding and the use of human milk. Pediatrics. 1997;
content/full/102/3/e38 100:1035–1039

e122 VOHR et al
Downloaded from http://pediatrics.aappublications.org/ by guest on October 18, 2017
34. American Academy of Pediatrics, Section on Breastfeeding. benefits and challenges, incidence and duration, and barriers to
Breastfeeding and the use of human milk. 2005;115:496 –506 breastfeeding in preterm infants. Adv Neonatal Care. 2005;5:
35. Bayley N. Bayley Scales of Infants Development-II. San Antonio, 72– 88; quiz 89 –92
TX: Psychological Corporation; 1993 47. Killersreiter B, Grimmer I, Buhrer C, Dudenhausen JW,
36. Poindexter BB, Ehrenkranz RA, Stoll BJ, et al. Parenteral glu- Obladen M. Early cessation of breast milk feeding in very low
tamine supplementation does not reduce the risk of mortality birthweight infants. Early Hum Dev. 2001;60:193–205
or late-onset sepsis in extremely low birth weight infants. 48. Doyle LW, Rickards AL, Kelly EA, Ford GW, Callanan C.
Pediatrics. 2004;113:1209 –1215 Breastfeeding and intelligence. Lancet. 1992;339:744 –745
37. Amiel-Tison C. Neuromotos status. In: Taeusch HW, Yogman 49. Fewtrell MS, Morley R, Abbott RA, et al. Double-blind, ran-
MW, eds. Follow-up Management of the High-Risk Infant. Boston, domized trial of long-chain polyunsaturated fatty acid supple-
MA: Little, Brown & Company; 1987:115–126 mentation in formula fed to preterm infants. Pediatrics. 2002;
38. Papile LA, Burstein J, Burstein R, Koffler H. Incidence and 110(pt 1):73– 82
evolution of subependymal and intraventricular hemorrhage: 50. O’Connor DL, Hall R, Adamkin D, et al. Growth and develop-
a study of infants with birth weights less than 1,500 gm. ment in preterm infants fed long-chain polyunsaturated fatty
J Pediatr. 1978;92:529 –534 acids: a prospective, randomized controlled trial. Pediatrics.
39. Horwood LJ, Fergusson DM. Breastfeeding and later cognitive
2001;108:359 –371
and academic outcomes. Pediatrics. 1998;101(1). Available at:
51. Amin SB, Merle KS, Orlando MS, Dalzell LE, Guillet R. Brain-
www/pediatrics.org/cgi/content/full/101/1/e9
stem maturation in premature infants as a function of enteral
40. Furman L, Wilson-Costello D, Friedman H, Taylor HG, Minich
feeding type. Pediatrics. 2000;106(pt 1):318 –322
N, Hack M. The effect of neonatal maternal milk feeding on the
52. Weiss B. Pesticides as a source of developmental disabilities.
neurodevelopmental outcome of very low birth weight infants.
Ment Retard Dev Disabil Res Rev. 1997;3:246 –256
J Dev Behav Pediatr. 2004;25:247–253
53. Martin JA, Kochanek KD, Strobino DM, Guyer B, MacDorman
41. Nolan L, Goel V. Sociodemographic factors related to breast-
feeding in Ontario: results from the Ontario Health Survey. Can MF. Annual summary of vital statistics: 2003. Pediatrics. 2005;
J Public Health. 1995;86:309 –312 115:619 – 634
42. Lucas A, Cole TJ, Morley R, et al. Factors associated with 54. Taylor HG, Minich N, Bangert B, Filipek PA, Hack M. Long-
maternal choice to provide breast milk for low birthweight term neuropsychological outcomes of very low birth weight:
infants. Arch Dis Child. 1988;63:48 –52 associations with early risks for periventricular brain insults.
43. Meinzen-Derr J, Poindexter BB, Donovan EF, et al. The role of J Int Neuropsychol Soc. 2004;10:987–1004
human milk feedings in risk of late-onset sepsis [abstract]. 55. Litt J, Taylor HG, Klein N, Hack M. Learning disabilities in
Pediatr Res. 2004;55:393A children with very low birthweight: prevalence, neuropsycho-
44. Furman L, Taylor G, Minich N, Hack M. The effect of maternal logical correlates, and educational interventions. J Learn Dis-
milk on neonatal morbidity of very low-birth-weight infants. abil. 2005;38:130 –141
Arch Pediatr Adolesc Med. 2003;157:66 –71 56. Hack M, Flannery DJ, Schluchter M, Cartar L, Borawski E,
45. Callen J, Pinelli J, Atkinson S, Saigal S. Qualitative analysis of Klein N. Outcomes in young adulthood for very-low-birth-
barriers to breastfeeding in very-low-birthweight infants in the weight infants. N Engl J Med. 2002;346:149 –157
hospital and postdischarge. Adv Neonatal Care. 2005;5:93–103 57. Saigal S. Follow-up of very low birthweight babies to adoles-
46. Callen J, Pinelli J. A review of the literature examining the cence. Semin Neonatol. 2000;5:107–118

PEDIATRICS Volume 118, Number 1, July 2006 e123

Downloaded from http://pediatrics.aappublications.org/ by guest on October 18, 2017
 Beneficial Effects of Breast Milk in the Neonatal Intensive Care Unit on the
Developmental Outcome of Extremely Low Birth Weight Infants at 18 Months of
Age
Betty R. Vohr, Brenda B. Poindexter, Anna M. Dusick, Leslie T. McKinley, Linda L.
Wright, John C. Langer and W. Kenneth Poole
Pediatrics 2006;118;e115
DOI: 10.1542/peds.2005-2382

Updated Information & including high resolution figures, can be found at:
Services http://pediatrics.aappublications.org/content/118/1/e115
References This article cites 52 articles, 23 of which you can access for free at:
http://pediatrics.aappublications.org/content/118/1/e115.full#ref-list-
1
Subspecialty Collections This article, along with others on similar topics, appears in the
following collection(s):
Fetus/Newborn Infant
http://classic.pediatrics.aappublications.org/cgi/collection/fetus:newb
orn_infant_sub
Permissions & Licensing Information about reproducing this article in parts (figures, tables) or
in its entirety can be found online at:
https://shop.aap.org/licensing-permissions/
Reprints Information about ordering reprints can be found online:
http://classic.pediatrics.aappublications.org/content/reprints

Pediatrics is the official journal of the American Academy of Pediatrics. A monthly publication, it
has been published continuously since . Pediatrics is owned, published, and trademarked by the
American Academy of Pediatrics, 141 Northwest Point Boulevard, Elk Grove Village, Illinois,
60007. Copyright © 2006 by the American Academy of Pediatrics. All rights reserved. Print ISSN:
.

Downloaded from http://pediatrics.aappublications.org/ by guest on October 18, 2017

 Beneficial Effects of Breast Milk in the Neonatal Intensive Care Unit on the
Developmental Outcome of Extremely Low Birth Weight Infants at 18 Months of
Age
Betty R. Vohr, Brenda B. Poindexter, Anna M. Dusick, Leslie T. McKinley, Linda L.
Wright, John C. Langer and W. Kenneth Poole
Pediatrics 2006;118;e115
DOI: 10.1542/peds.2005-2382

The online version of this article, along with updated information and services, is
located on the World Wide Web at:
http://pediatrics.aappublications.org/content/118/1/e115

Pediatrics is the official journal of the American Academy of Pediatrics. A monthly publication, it
has been published continuously since . Pediatrics is owned, published, and trademarked by the
American Academy of Pediatrics, 141 Northwest Point Boulevard, Elk Grove Village, Illinois,
60007. Copyright © 2006 by the American Academy of Pediatrics. All rights reserved. Print ISSN:
.

Downloaded from http://pediatrics.aappublications.org/ by guest on October 18, 2017