Cognitive and Neurologic Development of the Premature, Small for

Gestational Age Infant Through Age 6: Comparison by Birth Weight and

Gestational Age

Cecelia M. McCarton, MD*; ma F. Wallace, PhD*IJ1J; Michael Divon, MD#; and

Herbert C. Vaughan, Jr, MD*,**

ABSTRACT. Objective. To compare the neurologic ABBREVIATIONS. SGA, small for gestational age; AGA, appro-
and cognitive outcomes of 129 premature small for ges- priate for gestational age; AGA-BW, appropriate for gestational
tational age (SGA) infants with 300 premature appropri- age infants matched for birth weight; IVH, intraventricular hem-
ate for gestational age (AGA) infants through 6 years of orrhage; BPD, bronchopulmonary dysplasia; NHI, neonatal health
age. index; MDI, mental development index; VIQ, verbal intelligence
Design. Infants born at 37 weeks gestational age quotient; PIQ, performance intelligence quotient; ANOVA, anal-
ysis of variance.
and 2500 g with birth weight 2 standard deviations or
more below the mean birth weight for gestational age
were categorized as SGA. Cognitive and neurologic out-
Full-term and premature newborns whose birth
comes of SGA and AGA prematures at 1, 2, 3, and 5
weight is significantly less than the expected mean
and/or 6 years of age were compared when the infants
were stratified by gestational age in 2-week intervals or for their gestational age are considered small for
by birth weight in 500-g intervals. The association be- gestational age (SGA). Premature SGA infants are
tween SGA/AGA and neurologic status on cognitive out- felt to be at double biologic jeopardy because of a
comes at each age was also examined. shortened gestational period compounded by intra-
Results. SGA infants had significantly poorer cogni- uterine growth retardation. However, the direct con-
tive scores at each age when compared with AGA infants sequences of intrauterine growth retardation in the
of similar gestational ages. Normal neurologic status was premature infant are unclear. Studies of the early
more likely at all assessments for the AGA than for SGA
outcome of premature SGA infants have shown con-
infants of comparable gestational age. There were no
flicting results1 with rates of handicap ranging from
differences between SGA and AGA children in cognitive
or neurologic outcomes at any age when grouped by
low2’3 to high.46 School-age outcome studies also
birth weight. Cognitive impairment was closely associ- have differed in the relative incidence of academic
ated with neurologic abnormality in both SGA and AGA and intellectual deficits in preterm appropriate for
groups. There was, nevertheless, a significant effect of gestational age (AGA) as compared with SGA
SGA on cognitive outcome independent of neurologic infants.7’8
status at all ages except 3 years. These outcome studies have varied considerably in
Conclusions. Irrespective of degree of prematurity, constitution of the samples and other aspects of
SGA infants are at greater risk for neurodevelopmental
methodology that are likely to account for much of
impairment than are equally premature AGA infants.
the disparity in findings on the impact of SGA on
The cognitive impairment can be largely, but not en-
neurologic and cognitive development. First of all,
tirely, attributed to a higher incidence of neurologic ab-
normalities in the SGA infants at each gestational age. investigators have used different criteria for classify-
Pediatrics 1996;98:1167-1178; preterm, small for gesta- ing infants as SGA. Commonly used criteria are
tional age, cognitive outcome, neurologic outcome. based on Lubchenco’s growth curves,9 which define
an SGA infant as less than the 10th percentile birth
weight for gestational age, or Usher’s standards,1#{176}
From the *Rose F. Kennedy Center for Research in Mental Retardation and
which use 2 standard deviations below the mean
Human Development, Albert Einstein College of Medicine, Bronx, New birth weight (ie, the 3rd percentile) for gestational
York; the Department of Pediatrics, Albert Einstein College of Medicine, age as the cutoff. Moreover, the birth weight distri-
Bronx, New York; the §Department of Otolaryngology, Albert Einstein
butions are different in the two reference groups on
College of Medicine, Bronx, New York; the IlCenter for Research in Educa-
tion, Research Triangle Institute, Research Triangle Park, North Carolina;
which the criteria are based so that it is difficult to
the #{182}Department of Pediatrics, School of Medicine, University of North compare results of studies using different classifica-
Carolina at Chapel Hill, Chapel Hill, North Carolina; the #Department of tion criteria.
Obstetrics and Gynecology. Albert Einstein College of Medicine, Bronx, Studies of cohorts born in the 1960s primarily fo-
New York; and the **Departments of Neurology and Neuroscience, Albert
cused on the consequences of low birth weight in
Einstein College of Medicine, Bronx, New York.
Received for publication Mar 21, 1995; accepted Dec 27, 1995. which the analysis of the developmental outcome of
Reprint requests to (C.M.C.) Department of Pediatrics, Albert Einstein SGA infants was of secondary interest. Research dur-
College of Medicine, Rose F. Kennedy Center for Research in Mental Re- ing this period contrasted heterogeneous groups of
tardation and Human Development, 1300 Morris Park Aye, Bronx, NY
SGA and AGA infants who were not equivalent in
10461.
PEDIATRICS (ISSN 0031 4005). Copyright © 1996 by the American Acad- birth weight or gestational age.1113 Other studies had
emy of Pediatrics. no comparison group for the SGA infants.5 More

PEDIATRICS
Downloaded from www.aappublications.org/news at Indonesia:AAP Vol.September
Sponsored on 98 No. 16,
6 December
2019 1996 1167
recently, some investigators compared preterm SGA neonatal intensive care units at the teaching hospitals of the Albert
infants with a group of AGA infants of similar birth Einstein College of Medicine. The 429 children represent 93% of
the 461 children recruited whose parents agreed to have them
weight.2’4”4”5 Because the AGA groups were less ma-
followed by the LIFE Program who met criteria for entrance into
ture than the SGA infants, they were at greater risk the study and had data at one or more of the data points. There
for the complications of prematurity that may have were 32 children who were lost between discharge and follow-up
led to poorer developmental outcomes. Thus, these at I year of age and who are not part of the present report.
The 429 children included in the present investigation were
studies biased the results in favor of the SGA infants.
born between 1975 and 1987, weighed 2500 g at birth, and had a
Recent studieslilS have examined the impact of gestational age 37 weeks using the Dubowitz criteria,2#{176} which is
both birth weight and gestational age on the short- a postnatal assessment of gestational age. The criterion for SGA
and long-term developmental outcomes of prema- was a birth weight less than the 3rd percentile for gestational age
using the normative data provided by Usher and McClean.’#{176}
ture SGA infants. In two of the three studies, differ-
Based on this criterion, 129 (30%) of the entire sample were clas-
ences were found between SGA infants and the AGA
sified as SGA. The remaining 300 were classified as AGA. None of
infants matched for gestational age, either in a higher the children enrolled in the study had congenital malformations,
rate of neurologic abnormalities at I year (Pena et chromosomal anomalies, or known exposure to illicit drugs in
al’6, or lower cognitive scores at 1, 2, and 3 years of utero. All families were able to communicate in English. Each
family was fully informed about the research protocol at enroll-
age (Sung et al’7). In both studies the SGA infants
ment and gave written consent. The procedures were approved by
generally had similar cognitive scores to the AGA the Committee on Clinical Investigation of the Albert Einstein
infants matched for birth weight (AGA-BW). How- College of Medicine.
ever, in the groups followed by Sung and colleagues, As is shown in Table 1, the two groups are similar in demo-

the AGA-BW group had a higher rate of suspicious! graphic characteristics. There were no significant differences be-
tween SGA and AGA groups in gender, maternal age, maternal
abnormal neurologic findings through 3 years of age
education, ethnicity, or socioeconomic status using the Hollings-
than both the SGA and AGA-GA groups. In contrast, head index2’ as measured or gender. More than half of the chil-
Robertson and colleagues18 found no differences in dren in this study were from the lowest two social class categories,
academic performance or cognitive abilities at 8 and over 85% in each group were of African-American or His-
panic ethnicity.
years between their sample of SGA infants in com-
The perinatal status variables and medical complications for
parison with either of their groups of AGA prema- the SCA and AGA groups are also presented in Table I . Data were
tures. Although these studies used a 10th percentile not available for every infant, particularly information regarding
cutoff for defining SGA, their small sample sizes may intraventricular hemorrhage (IVH) and bronchopulmonary dys-
plasia (BPD) for the infants born before 1980. As expected, there
account for the discrepancies in findings.
was a significant difference in birth weight between the SGA and
Previously, we’ reported on the results of a study AGA infants, with the average birth weight of the SGA infants
in which we compared the outcome of 83 SGA pre- being about 450 g less than the AGA infants (P < .001). There was
term infants (classified using the Lubchenco stan- also a difference between the SGA and AGA children in the
dards) to 240 AGA preterms while statistically con- average head circumference at birth (P < .001 ) and in the percent-
age whose head circumference was subnormal for their gesta-
trolling for both birth weight and gestational age.
tional age (< 3rd percentile). On the other hand, with the excep-
With both birth weight and gestational age covaried, tion of days spent in the hospital, there were no significant
no differences were found in cognitive status differences between SGA and AGA infants in any other perinatal
through age 3. We found a significantly greater mci- status variable or medical complications. Although the SGA in-

dence of neurologic abnormalities at I year in the fants spent a longer time in the hospital as newborns (P < .001),
they were not significantly sicker than the AGA group once their
SGA group as compared with the AGA group. When
birth weight was taken into account. This is reflected in their
the children were stratified by gestational age, dif- Neonatal Health Index scores (NHI),2’ which is a summary mea-
ferences were found only between SGA and AGA sure of the degree of their illness, calculated based on length of
children whose gestational age was 30 to 31 weeks. stay in the neonatal nursery, adjusted for birth weight, and stan-
dardized to have a mean of 100 (higher scores represent better
In the present study we examined neurologic and
neonatal health).
cognitive outcome in a larger group of 429 premature
infants with birth weights s2500 g, the majority of
whom were followed from birth through 6 years of Procedures
age. We employed the more stringent 2 standard All children were scheduled for developmental evaluations at
1 , 2, and 3 years of age postterm (ie, corrected for prematurity).
deviation criterion of Usher and McClea&#{176} to define
Although the majority of children were also scheduled to he seen
SGA. We performed separate comparisons of the at 5 and/or 6 years of age uncorrected for prematurity, children
cognitive performance and neurologic status of SGA born after July 1986 were not routinely scheduled for assessments
versus AGA infants when stratified by either birth after 3 years of age. When children were seen at both 5 and 6 years,
weight or gestational age across the range of prema- the 6-year data were used.
Table 2 presents the number of children seen at each age and
turity. We further examined the association between Table 3 provides a comparison of demographic measures, perina-
neurologic impairment and cognitive performance in tal status variables, and medical complications in children seen
an effort to determine the extent to which cognitive and not seen at 6 years. Despite the unavailability of data at 5 or
impairment might be secondary to the brain dys- 6 years in approximately one third of the original sample, the rates
of missing data are comparable in SGA and AGA groups. With the
function associated with abnormal neurologic status.
exception of maternal age in the SGA group, there are no differ-
ences in background or perinatal characteristics between the chil-
METHODS dren with and without missing data at 5 and/or 6 years within the
AGA or SGA groups.
Subjects At I and 2 years of age, the Bayley Scales of Infant Develop-
The 429 premature infants in this study were followed by the ment were administered. The Bayley Mental Development Index
Longitudinal Infant Follow-up and Evaluation (LIFE) Program at (MDI) has a mean of 100 and a standard deviation of 16. At 3 years
the Rose F. Kennedy Center of the Albert Einstein College of of age, the majority of children received the Stanford-Binet Intel-
Medicine. All children were in-born and treated in one of the three ligence Scale, Form L-MP The 94 infants born between 1984

1168 DEVELOPMENTDownloaded
OF from
PREMATURE SMALL FOR GESTATIONAL
www.aappublications.org/news AGE INFANTS
at Indonesia:AAP Sponsored on September 16, 2019
TABLE 1. Medical and Demographic Background Characteristics

Group P

AGA SGA
(n = 300)* (n = 129)

Gender-Males 130 (43%) 57 (44%) .954

Gestational age at birth M (SD) 32.5 (2.4) 32.5 (2.6) .780
Birth weight, g M (SD) 1645.7 (425.4) 1196.4 (321.5) <.001
Birth head circumference, cm M (SD) 29.3 (4.1) 26.7 (2.6) <.001
Subnormal birth head circumference 25 (9%) 82 (67%) <.001
Apgar, 5 mm M (SD) 7.9 (2.0) 7.7 (2.0) .302
Asphyxia (Apgar <4) 16 (6%) 8 (7%) .886
Days hospitalized M (SD) 38.9 (29.4) 65.4 (33.4) <.001
Neonatal health index M (SD) 91.2 (16.2) 89.2 (19.1) .329
Respiratory distress syndrome 135 (55%) 60 (52%) .743
Bronchopulmonary dysplasia 16 (6%) 10 (9%) .435
Intraventricular hemorrhage 28 (28%) 9(35%) .676
Maternal age at birth M (SD) 25.1 (5.8) 24.9 (6.3) .698
Maternal education, years M (SD) 11.9 (2.3) 11.5 (2.3) .096
SES (Hollingshead) M (SD) 28.8 (12.7) 27.7 (11.6) .396
Lowest two social classes 152 (56%) 75 (60%) .560
Ethnicity .266
African-American 142 (47%) 74 (57%)
Hispanic 116 (39%) 42 (33%)
White/other 42 (14%) 13 (10%)

Abbreviations: AGA, appropriate for gestational age; SGA, small for gestational age; M, mean; SD, standard deviation; IVH, intraven-
tricular hemorrhage; RDS, respiratory distress syndrome; SES, socioeconomic status; BPD, bronchopulmonary dysplasia; NHI, neonatal
‘ilth index.
* Sample size for AGA was less than 300 for the following variables: 100 for IVH; 247 for RDS; 270 for SES and lowest two social classes;
276 for BPD; 285 for days hospitalized and NHI; 289 for head circumference and subnormal head circumference; 290 for Apgar and
asphyxia; 295 for maternal education; 297 for maternal age.
:1:Sample size was less than 129 for the following variables: 26 for IVH; 115 for RDS; 117 for BPD; 123 for days hospitalized, NHI, head
circumference and subnormal head circumference; 124 for Apgar and asphyxia; 125 for SES and lowest two social classes; 125 for maternal
education 128 for maternal age.

TABLE 2. umber of Children W ho Missed Visits* who ranged in age from 7 months to 3 years. Agreement in
classification was .90.
Group

AGA (n = 300) SGA (n = 129) Data Analysis

Number Number % Number Number % We compared the cognitive and neurologic outcomes of SGA
Seen Missed Seen Missed and AGA prematures grouped for either similar birth weight or
gestational age. Separate analyses were performed: one in which
Year I 299 1 .3 128 1 .7 SGA and AGA prematures were stratified by gestational age in
Year 2 277 23 8 120 9 7 2-week intervals (ie, 29, 30-31, 32-33, 34-35, and 36-37 weeks at
Year 3 275 25 8 118 Il 8 birth) and one in which SGA and AGA prematures were stratified
Year 5 and/or 6 206 66 24 85 32 27 by birth weight in 500-g intervals (ie, 1000, 1001-1500, 1501-
Abbreviations: AGA, appropriate for gestational age; SGA, small 2000, and 2001-2500 grams). Analysis of variance (ANOVA) were
used to examine the cognitive outcomes and and Mantel-Haen-
for gestational age.
* These numbers reflect number who missed both neurological szel procedures27 were used to examine neurological outcomes. A
and cognitive evaluations. significance level of .05 was adopted for all tests.
:1:There were 29 AGA and 12 SGA children not scheduled for 5/6
year appointments; they are not included in these figures. RESULTS

Cognitive Outcome
through 1987 received the Stanford-Binet Intelligence Scale, 4th
Table 4 presents the cognitive scores of the AGA
Edition,24 a revision of the previous L-M edition. Both versions
provide IQ scores with a mean of 100 and a standard deviation of and SGA premature infants from I to 5 and!or 6
16. At 5 years, the Wechsler Preschool and Primary Scale of years of age. At each age, cognitive performance was
Intelligence25 was used; and at 6 years, the Wechsler Intelligence slightly but consistently better in the AGA than the
Scale for Children-Revised25 was used. Both scales provide full- SGA infants: by 9.1 points at 1 year, 6.3 points at 2
scale, verbal IQ scores (VIQ), and performance IQ (PIQ) scores
years, 4.6 points at 3 years and 6 points at 5 and!or
with a mean of 100 and a standard deviation of 15. Assessors were
unaware of the children’s SGA status. 6 years. Cognitive outcomes were statistically exam-
A neurologic examination was also administered at 1, 2 and 3 med using separate 2-way ANOVA for each group-
years corrected age and at 5 or 6 years of age uncorrected for ing method: (SGA!AGA status X gestational age
prematurity. Children were classified as normal, suspicious, or
strata, and SGA!AGA status X birth weight strata).
abnormal. A rating of normal was given for a completely normal
neurologic examination. A rating of suspicious was given for an
Because each main effect is statistically adjusted for
examination with some atypical neurologic findings in tone, re- the other, the significance test for SGA provided a
flexes, gait, or movement, for which no specific diagnosis was test of the independent effect of gestational age (birth
available. A rating of abnormal was given when the individual weight) strata on cognitive test scores.
displayed a definite abnormality for which a diagnosis could be
Figure 1 presents the means of each of the cogni-
given, such as cerebral palsy, severe hypotonicity, microcephaly,
blindness, or sensorineural hearing loss. Interscorer reliability for tive measures (ie, MDI and IQ scores) for each
neurologic diagnosis was examined on a sample of 100 children 2-week gestational age stratum. As seen in the Figure,

Downloaded from www.aappublications.org/news at Indonesia:AAP Sponsored on September 16, 2019

ARTICLES 1169
TABLE 3. Background Characteristics in Children Seen and Not Seen at 6 Years

Group

AGA SGA

Seen Not Seen P Seen Not Seen P

Gender-Males 87 (43%) 32 (48’) ) .560 40 (487 ) 13 (39%) .550

Gestational age at birth M (SD) 32.8 (2.4) 32.6 (2.2) .640 32.6 (2.6) 32.7 (2.5) .787
Birth weight, M (SD) 1694.6 (423) 1663.0 (406) .593 1212.3 (335) 1222 (2961 .883
Birth head circumference, M (SD) 29.6 (4.7) 29.2 (2.0) .315 26.7 (2.6) 27.0 (2.5) .618
Subnormal birth head circumference 19 (9%) 5 (8V ) .830 55 (667 ) 20 (61V ) .779
Apgar, M (SD) 8.1 (1.8) 7.8 (2.0) .324 7.7 (2.1) 7.9 (1.9) .547
Asphyxia 11 (6%) 5 (8V) .730 8 (107) 3 (10%) 1.00
Days hospitalized M (SD) 36.9 (30.7) 39.9 (26.7) .508 67.9 (35.8) 58.6 (27.7) .190
Neonatal health index M (SD) 91.6 (17.0) 87.8 (13.7) .113 86.8 (20.3) 92.6 (15.2) .151
Respiratory distress syndrome 87 (52%) 30 (54%) .971 40 (54’7) 12 (39%) .222
Bronchopulmonary dysplasia 9 (5%) 1 (2V ) .459 6 (8V ) I (3V ) .671
Intraventricular hemorrhage 14 (26%) 6 (29%) 1.00 4 (27V) 0 (0%) 1.00
Maternal age at birth M (SD) 25.6 (5.9) 24.5 (5.6) .186 25.9 (6.4) 22.7 (5.8) 0.14
Maternal education, M (SD) 12.0 (2.4) 11.7 (2.0) .334 11.7 (1.9) 11.0 (3.1) .227
SES (Hollingshead) M (SD) 29.5 (13.8) 26.9 (10.3) .119 28.2 (11.1) 26.6 (13.3) .501
Lowest two social classes 99 (54%) 40 (66’7) .156 50 (60’7c) 21 (66’7) .752
Ethnicity .176 .899
African-American 93 (46%) 33 (46V ) 48 (57V ) 18 (55%)
Hispanic 74 (36% ) 29 (43% ) 27 (32% ) 1 1 (33%)
White/other 37 (18%) 5 (8%) 9 (11%) 4 (12%)

Abbreviations: AGA, appropriate for gestational age; SGA, small for gestational age; SES, socioeconomic status.

TABLE 4. Cognitive Test Scores* greater birth weight was seen regardless of whether
Group the infant was AGA or SGA.

AGA SGA

M SD n M SD n Neurological Outcome
Neurologic outcomes at each age are presented in
Year 1 (MDI) 99.3 19.2 296 90.2 20.6 126
Year 2 (MDI) 86.8 18.2 269 80.5 18.0 111 Table 5. The percentage of children classified as nor-
Year 3 (IQ) 84.0 17.4 263 79.4 16.5 110 mal increases at each assessment age, mainly due to
Year 5 and/or 6 (IQ) 91.4 16.5 206 85.4 14.5 85 the decrease in the percentage of children initially
Year 5 and/or 6 (VIQ) 90.1 17.1 206 85.2 15.0 85
classified as suspicious, most of whom were classi-
Year 5 and/or 6 (PIQ) 94.3 16.0 205 88.1 14.4 85
fied as normal at later ages. Neurologic outcomes
Abbreviations: MDI, mental development index; IQ, intelligence
were examined as follows. Due to the ambiguity of
quotient; VIQ, verbal intelligence quotient; PIQ, performance in-
telligence quotient.
the suspicious category, especially at the younger
* Testing was performed at corrected ages through 3 years and ages, we performed two separate analyses, one com-
was uncorrected at 5/6 years. paring the likelihood of a normal versus not normal
neurologic status (ie, suspicious and abnormal cate-
gories combined) in the AGA and SGA groups, and
higher MDI!IQ scores were generally associated with
another comparing the risk of an abnormal outcome
increasing gestational ages. The superior test scores of
to the normal and suspicious outcomes combined.
the AGA children in all but one of the gestational age
When the overall association was significant as mdi-
strata at age 3 are also evident. Results of the first
cated by the statistic, further tests were performed
ANOVA indicated that gestational age was signifi-
controlling for gestational age and for birth weight
cantly associated with cognitive scores at age I (P <
.001), age 2 (P = .004)], age 3 (P .027), and age 5 =
using the Mantel-Haenszel procedure,27 which pro-
vides a test of association and an estimate of the
and/’or 6 (P < .001), with lower scores as a function of
immaturity. The SGA children had lower cognitive test degree of association or odds ratio (O’nii;).
scores at 1 year (P < .001), 2 years (P = .002), 3 years In the initial analyses, the suspicious and abnormal
(P = .025), and 5 and!or 6 years (P = .005). At 5 and!or groups were collapsed into a deviant category. When
6 years, both VIQ and PIQ scores were also lower in the the AGA and SGA groups were compared, the per-
SGA group: VIQ (P .033); PIQ (P = .003). In addition,
=
centage of AGA children classified as completely
the difference between SGA and AGA children was normal was significantly greater at every age (1 year:
similar across gestational age groups as indicated by P < .001; 2 years: P = .003; 3 years: P = .004; 5 and!or
the nonsignificant interaction effects at every age. 6 years: P = .001 ). Inasmuch as the effect of SGA was
When cognitive scores were grouped according to significant at all ages of assessment, the Mantel-
birth weight, they were also found to differ across Haenszel procedure was used to examine the SGA
birth weight groups at 1, 2, and 5 and!or 6 years of effect independent of gestational age and of birth
age (1 year: P < .001; 2 years: P = .004; 5 and!or 6 weight.
years: P < .001). However, there were no differences The proportion of neurologically abnormal and
as a function of SGA!AGA status at any age. Com- suspect infants for each gestational age stratum and
parable improvement in cognitive performance with age of assessment are depicted in Fig 2. An excess of

1170 Downloaded
DEVELOPMENT OF from www.aappublications.org/news
PREMATURE SMALL at Indonesia:AAP Sponsored
FOR GESTATIONAL on September 16, 2019
AGE INFANTS
 1 YEAR MDI

34.35 36-37

2 YEAR MDI

Fig Cognitive test scores as a function of gesta- 0

tional
1.
age for AGA and SGA preterms.
3YEARIO U SGA

26.29 30.31 32.33 34.35 36-37

5/6 YEAR 0

26-29 30-31 32-33 34.35 36-37

GESTATIONAL AGE STRATA (weeks)

abnormal and suspicious examinations for the SGA

TABLE 5. Neurolog ic Outcome
infants at each assessment across most gestational
age groupings is evident. There was, moreover, a Group
striking increase in neurologic abnormalities in both AGAn(%) SGAn(%)
SGA and AGA infants below 32 weeks estimated
Year 1
gestational age which was especially pronounced for
Normal 194 (65) 50 (39)
the SGA group. When we compared the percentage Suspicious 82 (27) 58 (45)
of children classified as normal in the SGA and AGA Abnormal 28 (8) 20 (16)
groups stratified by gestational age, we found that Year 2
AGA infants had a 2.3 to 3.3 greater chance of being Normal 219 (80) 78 (65)
Suspicious 27 (10) 22 (18)
classified as normal at each age of examination than Abnormal 29 (10) 20 (17)
did SGA infants of similar gestational age (1 year Year 3
#{176}mh 3.28, P < .001; 2 year #{176}‘mli 2.3, P < .005; 3 Normal 228 (84) 82 (70)
year #{176}mh 2.34, P < .005; 5 and!or 6 year #{176}‘mh Suspicious 14 (5) 11 (9)
Abnormal 31 (12) 24 (20)
2.85, P < .005).
Year 5 and/or 6
When the children were stratified by birth weight, Normal 179 (88) 60 (71)
no association was found between normal neuro- Suspicious 5 (2) 9 (11)
logic status and SGA!AGA status at any age. The Abnormal 20 (10) 15 (18)
odds of being classified as neurologically normal Abbreviations: AGA, appropriate for gestational age; SGA, small
were no better in the AGA group than in the SGA for gestational age.

Downloaded from www.aappublications.org/news at Indonesia:AAP Sponsored on September 16, 2019

ARTICLES 1171
 1 YEAR

2 YEAR

AGA abnormal
26-29 30-31 32-33 34-35 36.37
. SGA abnormal
Fig 2. Neurological deviances as a function of
gestational age in AGA and SGA preterms.
0 AGA suspicious
U SGA suspicious

26-29 30.31 32.33 3435 36.37

5/6 YEAR
I 00

80

60

40

20

26-29 30-31 32-33 34.35 36-37

GESTATIONAL AGE STRATA ‘weeks

group when equated by birth weight (1 year O’fl;I, TABLE 6. Classification of Neurological Conditions at 3 Years
1.53, not significant; 2 years #{176}‘mh 1.04, not signif- Group
icant; 3 years #{176}‘mh 1.26, not significant; 5 and,/or 6
AGA SGA
years #{176}mh .67, not significant).
Looking at the other side of the coin, we compared Abnormal (n = 31) (n = 24)
those children classified as neurologically abnormal Cerebral palsy 25 14
Hypotonia 0 1
with the combined group of those who were normal Blind 2 2
and suspicious. Significant overall associations were Deaf I I
present at I year (P = .02); and 3 years (P = .026). Microcephaly 2 3
Estimates of the odds ratios with gestational age Chronic seizures 0 1
Others I 2
stratified indicate that the odds of being classified as
Suspicious (n = 14) (n = 11)
abnormal were about 2 times as great in the SGA Increased tone and reflexes in 9 8
group at both I and 3 years of age (1 year #{176}‘n;h lower extremities
2.39, P < .025; 3 years #{176}‘mh 2.13, P < .025). In Slight hypotonia I 0
Awkward gait 3 2
contrast, with birth weight stratified, the odds of
Intention tremors 1 1
being classified as neurologically abnormal were
equal in the SGA and AGA groups (1 year #{176}‘nh Abbreviations: AGA, appropriate for gestational age; SGA, small
for gestational age.
.84, not significant; 3 years #{176}‘mh 1.06, not signifi-
cant).
Table 6 lists the neurologic diagnoses in the SGA some confidence, and the number of suspicious cases
and AGA groups at 3 years of age. At this age it was diminished to one fifth those so designated at I year
possible to establish a neurologic diagnosis with of age. Cerebral palsy was the most common abnor-

Downloaded from www.aappublications.org/news at Indonesia:AAP Sponsored on September 16, 2019

1172 DEVELOPMENT OF PREMATURE SMALL FOR GESTATIONAL AGE INFANTS
mality in both SGA and AGA groups, accounting for The mean cognitive scores for each neurologic sta-
58% and 81 % of the diagnoses, respectively. In the tus grouping at each age of assessment are presented
suspicious category, findings of mild hypertonia, hy- in Fig 3. With the exception of 3 years, children
potonia, or motor incoordination were of similar in- classified as normal had the highest cognitive test
cidence in each group. Thus, there was no difference scores. The suspicious group had slightly lower
in the pattern of neurologic diagnoses or deviant scores, whereas the abnormal children had consider-
findings between the SGA and AGA groups. ably lower scores. Furthermore, within each neuro-
logic category the AGA children had superior scores
Association Between Neurologic Outcome and at every age.
Cognitive Performance Three-way ANOVAs (neurologic status X AGA!
Because cognitive performance was poorer and SGA x gestational strata) were performed to assess
neurologic deviance was more prevalent in the SGA the effects of each independent variable on cognitive
group, we sought to evaluate a possible association outcomes at 1, 2, 3, and 5 and!or 6 years of age.
between the presence of neurologic impairment and Gestational age was only associated with cognitive
cognitive performance. In the analyses at each age test scores at ages 1 year (P = .012) and 3 years (P =
we used the classifications based on the 3-year neu- .009). On the other hand, neurologic status was in-
rologic examination. Comparisons were made only dependently associated with cognitive status at each
for the gestational age strata because SGA status was age (1 year: P < .001; 2 years: P < .001; 3 years: P <
unrelated to either cognitive or neurologic outcome .001; 5 and,/or 6 years: P < .001). However, except at
when the groups were stratified by birth weight. 3 years of age, SGA status was also independently

I YEAR MDI

2 YEAR MDI

Fig 3. Cognitive test scores as a function of neuro-

logical status at 3 years in AGA and SGA preterms. 0
USGA

516 YEAR 0

Downloaded from www.aappublications.org/news at Indonesia:AAP Sponsored on September 16, 2019

ARTICLES 1173
related to cognitive test scores, (1 year: P < .001; 2 and 12.2 points at 5 and!or 6 years. The difference
years: P = .027; 5 and!or 6 years: P = .02), with SGA between normal and suspicious was significant at
children for the most part doing less well than AGA only I year: 7 points.
children within gestational age strata. It is worth To further assess the impact of neurologic abnor-
noting that cognitive test scores were more variable mality on individual cognitive functioning, we com-
within the neurologically suspect and abnormal pared the percentage of children in the SGA and
groups. AGA groups with cognitive scores more than 2 stan-
After controlling for gestational age and neuro- dard deviations below the standardization mean at
logic status the significant differences in test scores each age of testing (ie, < 68 ages 1, 2, 3, years and <
between AGA and SGA groups at each year were as 70 at 5 and 6 years) as a function of their neurologic
follows: 6.7 points at 1 year, 4.3 points at 2 years, and status. This analysis produced a striking result,
4.6 points at 5 and!or 6 years. After adjusting for which is depicted in Fig 4.
gestational age and SGA status, the difference be- Whereas less than 10% of the individuals with
tween children classified as neurologically abnormal normal neurologic status scored at a level indicative
and normal was 27.9 points at I year, 20.5 points at 2 of mental retardation, between 28% and 69% of the
years, 16.6 points at 3 years, and 13.1 points at 5 individuals in the neurologically abnormal group
and/’or 6 years. After adjusting for gestational age were substantially impaired in cognitive function.
and SGA status, the difference between children clas- Although there was a consistently larger proportion
sified as suspicious and abnormal was 20.9 points at of children with mental retardation in the SGA in-
I year, 19.5 points at 2 years, 19.6 points at 3 years, fants at each age of assessment, these differences

1 YEAR

80

60

40

20

0
Normal Suspicious Abnormal

2 YEAR

80

60

40

20

0
Normal Suspicious Abnormal
Fig 4. Percentage of AGA and SGA preterms
with cognitive retardation as a function of neu-
3 YEAR
rological status.
0 AGA
U SGA

5/6 YEAR

80

60

40

20

0
I I

Normal
I
Suspicious
I Abnormal

Downloaded from www.aappublications.org/news at Indonesia:AAP Sponsored on September 16, 2019

1174 DEVELOPMENT OF PREMATURE SMALL FOR GESTATIONAL AGE INFANTS
were not statistically significant. Within the neuro- also differed in whether the SGA!AGA comparison
logically normal group, the odds of being classified was based on gestational age or birth weight, and in
as mentally retarded at I year were greater in the the duration of follow-up.
SGA children than the AGA children (O’mh 373, The definition of SGA presents formidable diffi-
P < .05). It is evident, however, that mental retarda- culties, some of which have been discussed by Gold-
tion is principally associated with neurologic abnor- enberg et al.#{176} These problems include errors in de-
malities in both the SGA and AGA group. termination of gestational age, adequacy of the
normative sample of birth weight distributions, and
DISCUSSION the cutoff criterion for definition of SGA. Estimation
The premature SGA infant has often been consid- of gestational age, whether based on dates or phys-
ered to be at risk from the adverse effects of both ical criteria at birth, is subject to inherent errors that
prematurity and intrauterine undernutrition. Al- affect both the accuracy of normative standards and
though it is generally accepted that prematurity in- gestational age estimates within study samples. Even
creases the risk of cognitive and neurologic impair- when these errors are not systematic they can reduce
ment, and that lower birth weight increases the the reliability of findings in studies using small sam-
incidence of developmental disability,28’29 and incre- ples. In the present study, the consistent pattern of
mental adverse neurodevelopmental effect of intra- findings across gestational age strata suggests that
uterine growth retardation in the SGA premature has possible errors in estimation of gestational age did
proven difficult to demonstrate. The findings of this not appreciably reduce the reliability of our results.
study clarify some of the discrepancies in reported Small sample size is pervasive in the literature on
effects of intrauterine growth retardation on the cog- effects of SGA, and most conclusions are based on a
nitive and neurologic development of premature in- small number of cases. For example, three recent
fants. In a relatively large sample of infants, we have studies1618 that contrasted SGA samples with com-
found significantly greater cognitive and neurologic parison groups equated for either gestational age or
morbidity in SGA infants from I year through 5 to 6 birth weight had samples ranging from 16 to 36
years of age in comparison to AGA premature in- infants in each group. These small samples can pro-
fants of similar gestational age. When AGA and SGA vide results prone to sample bias and cannot be used
infants were equated for birth weight there were no to systematically examine differences in outcome as a
differences between the groups in cognitive perfor- function of gestational age or other variables.
mance or neurologic status. Study samples have differed widely in the range of
Except for birth weight and head circumference at birth weight and!or gestational age. Our sample
birth, the AGA and SGA groups were similar in comprised premature infants 2500 g. Some other
demographic characteristics, perinatal status, and studies have confined their population to infants
neonatal morbidity. These similarities provide some <1500 gm at birth.2’4’14’1617’31’32 Systematic effects re-
evidence that the developmental differences were lated to birth weight and!or gestational age can be
linked to intrauterine brain growth rather than other obscured in studies with small sample sizes and
factors. Moreover, cognitive impairment was closely restricted range of gestational age. Inasmuch as neu-
associated with the presence of neurologic abnormal- rologic abnormalities and cognitive impairment both
ity in both groups. The higher incidence of neuro- increase with decreasing gestational age, it is to be
logic deficit in SGA infants may be responsible for anticipated that the range of gestational age and
the relatively greater cognitive impairment in the birth weight of study samples will influence the
SGA infants. Although cognitive and neurologic sta- overall incidence of neurobehavioral abnormalities.
tus have often been assessed in previous studies, to The present study found the expected systematic
our knowledge a close association between these effect of gestational age on neurologic and cognitive
indices of brain function has not been described pre- status, which emphasizes the need for larger samples
viously. In the following discussion we first examine representing a broad range of gestational ages and
some of the methodological issues in studies of the birth weights for reliable estimates of neurobehav-
effects of prematurity and intrauterine growth retar- ioral morbidity in premature infants.
dation on neurodevelopmental status, and then con- A distinction should be made between factors,
sider the implications of our findings for the etiology such as gestational age estimation, that can influence
of cognitive impairment in the SGA premature findings within a study and those that principally
infant. affect comparisons across studies. The problem of
comparability of results across studies cannot readily
Methodological Issues be resolved, due to differences in the normative stan-
A number of methodological differences character- dards and in the criterion used for SGA. Because the
ize studies that have addressed the outcome of in- statistical criteria for SGA depend on the specific
fants considered to be SGA. Variations in sample birth weight distribution within a normative sample,
characteristics and other methodological factors it is virtually impossible to ensure that any two sam-
seem likely to account for differences in reported ples can be compared against a standard population
outcomes. Sample-related issues include the defini- mean unless they use the same distribution. For ex-
tion of SGA, size of sample, range of gestational ages ample, the study by Robertson et al,18 which used the
and birth weights in the sample; inclusion or exclu- Lubchenco growth curves with a 10th percentile cut-
sion of specific perinatal medical conditions, and off, noted that this cutoff corresponded to the 5th
socioeconomic and ethnic composition. Studies have percentile of the Province of Alberta birth weight

Downloaded from www.aappublications.org/news at Indonesia:AAP Sponsored on September 16, 2019

ARTICLES 1175
statistics. Thus, a given nominal cutoff criterion can logic and cognitive functioning of AGA and SGA
designate differing proportions of infants as SGA infants by separately stratifying the samples with
depending on the birth weight distribution of the respect to each variable. In the present study, the
normative sample. Furthermore, variations in birth effects of differing characteristics of separate AGA
weight distributions across different ethnic groups comparison samples have been avoided by using the
from which normative samples are drawn pose an same group of AGA infants stratified either as to
apparently insuperable obstacle to meaningful com- gestational age or birth weight for comparison with
parisons across studies. At present, the best that can the similarly stratified SGA infants. This approach
be done is to characterize the demographic and pen- excludes the possibility that adventitious differences
natal characteristics of both the normative and study in AGA comparison samples might lead to spurious
samples as fully as possible, and to use a SGA stan- conclusions. In our analyses significant differences in
dard based on clearly defined cutoff weights at each cognitive outcome were only found when comparing
gestational age. SGA premature infants to AGA premature infants of
In the present study we used a stringent criterion, similar gestational ages.
namely 2 standard deviations or the 3rd percentile Few studies have reported follow-up beyond 2
below mean birth weight at each gestational age, years of age, and results of follow-ups with school
using the normative data of Usher and McClean.1#{176} age children have been inconsistent. As with other
Others have used the 10th percentile of various nor- comparisons, some of the inconsistencies in findings
mative samples2’6’14’16’17’31 as the SGA criterion. De- are related to methodological differences. These dif-
spite the problems associated with the SGA criterion, ferences in methodology include matching AGA and
the results of the present study appear sufficiently SGA groups on birth weight,14 including children
consistent to permit conclusions that do not rely on with chromosomal abnormalities or syndromes,7’8 or
comparisons with other studies in which different relying on small samples.18 In contrast, our sample of
normative samples and cutoff criteria have been 94 premature SGA infants and 229 AGA premature
used. infants at 5 andi’or 6 years of age provides the largest
In addition, the exclusion or inclusion of infants sample size to date for a follow-up study of this
with specific medical conditions such as chromo- duration. With this relatively large sample, it was
somal abnormalities and intrauterine viral infections possible to demonstrate significant differences in
will have an obvious impact on the incidence of cognitive outcome and in neunologic status between
severe neurodevelopmental abnormalities in a given SGA and AGA infants when compared by gesta-
sample. In our study we excluded known causes of tional age, but not by birth weight.
SGA that could spuriously increase the apparent im-
pact of intrauterine growth retardation. However, Implications for the Etiology of Neurobehavioral
other subtle medical differences in sample constitu- Dysfunction
tion may have an impact on outcome extraneous to The findings of this study support the presump-
the effects of SGA. For example, the inclusion of tion that premature SGA infants are subject to a
infants transferred to neonatal intensive care units higher risk of developmental neurobehavioral im-
can lead to biased results. When only more stable or pairment not only from their prematurity but from
healthy infants are transferred, the chances of a bet- intrauterine undernutrition as well. The present
ten developmental outcome may be increased.14 Con- study makes it clear that there is a consistently
versely, institutions that receive only the sickest and greaten impairment in both cognitive function and
more medically complicated infants are likely to neurologic status in SGA infants when compared
have a higher incidence of developmental abnormal- with the complementary group of non-SGA infants
ities. Because our study infants were all in-bonn, of similar gestational age. However, a key finding of
there was a reduced possibility of a medically biased this study is the strong association between neuro-
sample attributable to variations in transport prac- logic abnormality and cognitive impairment. This
tices. association is found for both the SGA and AGA
In the past, a major methodological issue has been groups, with the lower cognitive performance of the
the appropriate basis for comparing SGA and AGA SGA group being related to their higher incidence of
infants. Studies that have matched the groups on the neurologic dysfunction.
basis of birth weight have compared the SGA infants The potential neurologic differences between SGA
with AGA infants who are less mature and thus, more and AGA infants appear evident from birth. In com-
susceptible to complications of prematurity, especially panison to AGA infants, the SGA infants in our study
in the very low birth weight category. This may ac- were noted to have a head circumference that was
count for the findings of a higher incidence of neuro- 2.5-cm smaller and more likely to be subnormal.
logic abnormalities or lower cognitive test scores in Whereas head circumference is an indirect measure-
AGA infants when contrasted with SGA infants on the ment of brain growth, there is evidence that head
basis of birth weight.3”17 Pena et aP6 called attention to circumference correlates with brain volume and
the desirabffity of comparing SCA and AGA infants of brain weight.33 The association between reduced
the same gestational age so that the effect of intrauter- head size and developmental morbidity is also well
me growth retardation on infants of comparable matu- documented, as previous investigators have shown
nity can be directly examined. that neonatal head size is associated with both neu-
Our project also assessed the relative importance rological and cognitive outcomes. Although known
of birth weight versus gestational age in the neuro- medical complications of prematurity (eg, IVH, RDS)

Downloaded from www.aappublications.org/news at Indonesia:AAP Sponsored on September 16, 2019

1176 DEVELOPMENT OF PREMATURE SMALL FOR GESTATIONAL AGE INFANTS
did not differentially affect our SGA infants, it is 4. Fitzhardinge PM, Kalman E, Ashby S. Pape KE. Present status of the
infant of very low birth weight treated in a referral neonatal intensive
possible that the retardation of brain growth might
care unit in 1974. In: Elliot K, O’Connor M, eds. Major Me;,tal Handicap:
have enhanced the adverse affect of perinatal com- Met/teds a;,iI Costs of Preve;,tion. Ciba Foundation Symposium 59. Am-
plications on neurologic status. sterdam, The Netherlands: Elsevier; 1978:139-144
Like Dnillien,35 we found that cerebral palsy was 5. Commey JO, Fitzhardinge PM. Handicap in the pre-term small-for-
the most common neurologic diagnosis in both SGA gestational age infant. J Pediatr. 1979;94:779-786
6. Kitchen W, Ford G, Orgill A, et al. Outcome in infants with birth weight
and AGA children. Inasmuch as the neurologic ab- 500 to 999 gm: a regional study of 1979 and 1980 births. I Pediatr.
normalities of both groups of infants are similar and 1984;1 04:921-927
not attributable to known genetic and infectious 7. Calame A, Fawer CL, Claeys V. Arrazola L, Ducret S, Juanin L. Neuro-
causes of gross brain disorders, the overriding factor developmental outcome and school performance of very-low-birth-
weight infants at 8 years of age. Eur J Pediatr. 1986;145:461-466
in determining the type of neurologic problem ap-
8. l-Iadders-Algra M, l-Iuisjes HJ, Towen BCL. Preterm or small-for-
pears to be prematurity and not whether a child is gestational age infants: neurological and behavioral development at the
SGA on AGA. Cerebral palsy is often considered to age of 6 years. Eur J Pediatr. 1988;147:460-467
be the result of hypoxic-ischemic encephalopathy, 9. Lubchenco L, Hansmann C, Boyd E. Intrauterine growth in weight,

which has its greatest incidence in the smallest pre- length and head circumference as estimated from live births at gesta-
tional ages from 26 to 42 weeks. Pediatrics. 1966;37:403-408
mature infants.36 It is possible that the excess of neu-
10. Usher R, McClean F. Intrauterine growth of live-born Caucasian infants
rologic damage, especially cerebral palsy, in the SGA at sea level: standards obtained from measurements in 7 dimensions of
infants may be attributable to an enhanced vulnera- infants born between 25 and 44 weeks gestation. I Pediatr. 1969;74:
bility of the SGA infant to hypoxia and its deleterious 901-910
11 . Eaves LC, Nutall JC, Klonoff H, et al. Developmental and psychological
effects on the brain.
test scores in children of low birth weight. Pediatrics. 1970;45:9-20
There was, nevertheless, some cognitive impair-
12. Drillien CM. The small for date infant: etiology and prognosis. Pediatr
ment even in the SGA infants who were considered Cli;, Norti, A;;,. 1970;1 7:9-21
to be neurologically normal. Evidence for direct cen- 13. Francis-Williams J, Davies PA. Very low birth weight and later intelli-
tral nervous system effects of intrauterine undernu- gence. Dez’ Med Cl,ild Neurol. 1974;16:709-728
14. Vohr BR, Oh W. Growth and development in preterm infants small for
tnition is primarily based on animal studies. Labora-
gestational age. I Pediatr. 1983;103:941-945
tory studies in newborn rats and guinea pigs that 15. Hack M, Fanaroff AA. The outcome of growth failure associated with
have experienced intrauterine growth retardation preterm birth. Cli;, Obstet Gynecol. 1984;27:647-663
have shown decreased neonatal body weight, brain 16. Pena IC, Teberg AJ, Finello KM. The premature small-for-gestational-

weight, and reduced amounts of brain DNA, protein, age infant during the first year of life: comparison by birth weight and
gestational age. I Pediatr. 1988;113:1066-1073
and myelin lipids.3741 Chase and colleagues37 de-
17. Sung I-K, Vohr B, Oh W. Growth and neurodevelopmental outcome of
scnibed similar biochemical alterations in human fe- very low birth weight infants with intrauterine growth retardation:
tal brains following intrauterine growth retardation. comparison with control subjects matched by birth weight and gesta-
It is not known which cellular elements are impacted tional age. I Pediatr. 1993;123:618-624
18. Robertson CMT, Etches PC, Kyle JM. Eight-year school performance
by intrauterine undernutrition, although the impair-
and growth of preterm, small for gestational age infants: a comparative
ment of myelination would suggest glial dysfunc- study with subjects matched for birth weight or for gestational age.
tion. Thus, there is continuing uncertainty concern- I Pediatr. 1990;116:19-26
ing the effects of intrauterine undernutrition on the 19. McCarton C, Wallace IF. The cognitive and neurologic development of
developing brain. Nevertheless, our finding of a rel- the premature small-for-gestational age infant In: Divon MY, ed. Al;-
;or;;ial Fetal Grout!,: IUGR a;,l Macrosomia. New York, NY: Elsevier
ative cognitive impairment in the neurologically nor-
Science Publishing Co; 1991:319-337
ma! SGA infants is the first indication that intrauter- 20. Dubowitz LMS, Dubowitz V, Goldberg C. Clinical assessment of ges-
me growth retardation may have an adverse impact tational age in the newborn infant. J Pediatr. 1970;77:1-10
on cognitive function, independent of the neunobe- 21. Scott DT, Bauer CR, Kraemer HC, Tyson J. A neonatal health index for
preterm infants. Pediatr Res. 1989;25:263A
havioral impairment associated with the neurologic
22. Bayley N. Baile,I ScaIc’s t;f Infa;,t Develop;;ze;:t. New York, NY: Psycho-
complications of prematurity.
logical Corp; 1969
23. Terman LM, Merrill MA. Sta;iford-Bi;,et l;,telli’e;,ce Scale, Fore, L-M.
Boston, MA: Houghton Mifflin; 1973
ACKNOWLEDGMENTS
24. Thorndike RL, Hagen EP, Sattler, JM. Standord-Binet l;,te’lli’e;,ct’ Scale.
This study was supported in part by National Institute of 4th ed. Chicago, IL: Riverside Publishing; 1986
Health grants HD-01799, HD-20435, and NS 19748; March of 25. Wechsler D. Weclsler Presclool a;d Pri;;,ar,,, Scale of l;,telli’’e;,ce. New
Dimes Birth Defects Foundation grant No. 12-7; the Robert Wood York, NY: Psychological Corp; 1967
Johnson Foundation (ID #9532); Maternal and Child Health Bu- 26. Wechsler D. Wecl,sler l;tellige;ce Scale for Cl,ildre;,-Revised. New York,
reau (MCJ-360593-OI-02-0); and the Pew Charitable Trust (No. NY: Psychological Corp; 1974
88-01305). 27. Fleiss J. Statistical Methods for Rates a;,d Proportio;s. 2nd ed. New York,
We thank the staff of the LIFE Program for recruiting and NY: Wiley; 1981
maintaining the study cohorts, Anna Brattson for preparing the 28. Hoy EA, Bill JM, Sykes DEl. Very low birth weight: a long-term devel-
figures, and Anita Levine and Lori Hill for their assistance in opmental impairment? I;,t J Bel,av Dcv. 1988;11:37-67
preparing the manuscript. 29. McCormick MC. The contribution of low birth weight to infant mortal-
ity and childhood morbidity. N E;igl I Med. 1985;312:82-90
30. Goldenberg RL, Cutter GR, Hoffman HJ, Foster JM, Nelson KG, Hauth
REFERENCES
JC. Intrauterine growth retardation: standards for diagnosis. Am I Obstet
1. Teberg AJ, Wather FJ, Pena IC. Mortality, morbidity and outcome of the Gy;iecol. 1989;161 :271-277
small-for-gestational age preterm infant. Sc;;,i;, Peri;,atol. 1988;12:84-89 31. Koops BL. Neurologic sequelae in infants with intrauterine growth
2. Vohr BR, Oh W, Rosenfield AG, Cowett RM, Berstein J. The premature retardation. / Reproductive Med. 1978;21 :343-351
small for gestational age infant: a 2-year follow-up study. A;;, J Obstet 32. Kumar SP, Anday EK, Saks LM, et al. Follow-up studies of very low
Gtj;,t’col. 1979;133:425-431 birth weight infants (1,250 grams or less) born and treated within a
3. Calame A, Ducret 5, Juarin L, Plancherel B. High risk appropriate for perinatal center. Pediatrics. 1980;66:438-444
gestational age (AGA) and small for gestational age (SGA) preterm 33. Winick M, Ross P. The effect of severe early malnutrition on cellular
infants. Hel,’ Paediatr Act;;. 1 983;38:39-50 growth of human brain. Pediatr Res. 1969;3:181-184

Downloaded from www.aappublications.org/news at Indonesia:AAP Sponsored on September 16, 2019

ARTICLES 1177
34. Gross SJ, Kosmetatos N, Grimes CT, et al. Newborn head size and 38. Chow BF, Lee CJ. Effect of dietary restriction of pregnant rats on body
neurological status: predictors of growth and development in low birth weight gain of the offspring. I N,,tr. 1964;82:10-18
weight infants. A;;, I Dis C/,ild. 1978;132:753-756 39. Venkatachalam PS, Ramanathan KS. Severe protein deficiency during
35. Drillien CM. Atiology and outcome in low birth weight infants. Dci’ Med gestation in rats on birth weight and growth of offspring. I;,dia;, J Mcd
Cl,ild Nc’,irol. 1 972;14:563-574 Rt’s. 1966;54:402
36. Stanley FJ, English DR. Prevalence of and risk factors for cerebral palsy 40. Wigglesworth JS. Foetal growth retardation. Br Mcd B;;!!. 1966;22:13-15
in a total population cohort. Dci’ Med Cl,ild Ncurol. 1986;28:559-568 41. Zamenhof 5, van Marthens E, Margolis FL. DNA (cell number) and
37. Chase HP, Dabiere CS, Welch NN, O’Brien D. Intrauterine undernutri- protein in neonatal brain: alteration by maternal dietary protein restric-
tion and brain development. Pediatrics. 1971;47:491-500 tion. Scie;,ct’. 1968;1 60:322-323

PLACEBO VS NONTREATMENT

We often and wrongly equate the response seen in the placebo arm of a clinical
trial with the placebo effect. In order to obtain the true placebo effect, other
nonspecific effects can be identified by including an untreated control group in
clinical trials.

Ernst E, Resch KL. Concept of true and perceived placebo effects. Br Med 1. 1995;311 :551-553.

Submitted by Student

Downloaded from www.aappublications.org/news at Indonesia:AAP Sponsored on September 16, 2019

1178 DEVELOPMENT OF PREMATURE SMALL FOR GESTATIONAL AGE INFANTS
Cognitive and Neurologic Development of the Premature, Small for Gestational Age
Infant Through Age 6: Comparison by Birth Weight and Gestational Age
Cecelia M. McCarton, Ina F. Wallace, Michael Divon and Herbert G. Vaughan, Jr
Pediatrics 1996;98;1167

Updated Information & including high resolution figures, can be found at:
Services http://pediatrics.aappublications.org/content/98/6/1167
Permissions & Licensing Information about reproducing this article in parts (figures, tables) or in its
entirety can be found online at:
http://www.aappublications.org/site/misc/Permissions.xhtml
Reprints Information about ordering reprints can be found online:
http://www.aappublications.org/site/misc/reprints.xhtml

Downloaded from www.aappublications.org/news at Indonesia:AAP Sponsored on September 16, 2019

 Cognitive and Neurologic Development of the Premature, Small for Gestational Age
Infant Through Age 6: Comparison by Birth Weight and Gestational Age
Cecelia M. McCarton, Ina F. Wallace, Michael Divon and Herbert G. Vaughan, Jr
Pediatrics 1996;98;1167

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/98/6/1167

Pediatrics is the official journal of the American Academy of Pediatrics. A monthly publication, it has
been published continuously since 1948. Pediatrics is owned, published, and trademarked by the
American Academy of Pediatrics, 141 Northwest Point Boulevard, Elk Grove Village, Illinois, 60007.
Copyright © 1996 by the American Academy of Pediatrics. All rights reserved. Print ISSN: 1073-0397.

Downloaded from www.aappublications.org/news at Indonesia:AAP Sponsored on September 16, 2019