|Home | About | Journals | Submit | Contact Us | Français|
The newborn classified as growth-restricted on birth weight curves, but not on fetal weight curves, is classified prenatally as small for gestational age (SGA), but postnatally as appropriate for gestational age (AGA).
To see (1) to what extent the neurodevelopmental outcomes at 24 months corrected age differed among three groups of infants (those identified as SGA based on birth weight curves (B-SGA), those identified as SGA based on fetal weight curves only (F-SGA), and the referent group of infants considered AGA, (2) if girls and boys were equally affected by growth restriction, and (3) to what extent neurosensory limitations influenced what we found.
Observational cohort of births before the 28 week of gestation. Outcome measures: Mental Development Index (MDI) and Psychomotor Development Index (PDI) of the Bayley Scales of Infant Development II.
B-SGA, but not F-SGA girls were at an increased risk of a PDI < 70 (OR=2.8; 95% CI: 1.5, 5.3) compared to AGA girls. B-SGA and F-SGA boys were not at greater risk of low developmental indices than AGA boys. Neurosensory limitations diminished associations among girls of B-SGA with low MDI, and among boys B-SGA and F-SGA with PDI < 70.
Only girls with the most severe growth restriction were at increased risk of neurodevelopmental impairment at 24 months corrected age in the total sample. Neurosensory limitations appear to interfere with assessing growth restriction effects in both girls and boys born preterm.
Infants are considered growth restricted, or small for gestational age (SGA) if their birth weights are in the lowest decile on a gestational-age specific birth weight curve. Birth weight distributions have lower means, medians and 10th centiles than estimated fetal weight distributions, especially among preterm newborns . Some newborns whose birth weight is considered appropriate by neonatal standards are growth restricted by fetal standards (Figure 1). This phenomenon appears to reflect the preferential preterm birth of growth restricted infants .
We refer to infants who are SGA on birth weight curves as B-SGA, and to those who are SGA on estimated fetal weight curves, but appropriate for gestational age (AGA) on birth weight curves, as “F-SGA” . Fetal weight estimates are considered reliable [3, 4].
Among children who had been born at an early gestational age or a low birth weight, those whose birth weight was in the lowest decile for gestational age tend to be at increased risk of cognitive limitations [5–10], but not always [11–14]. The authors of a recent report raised the possibility that their finding no association between fetal growth restriction and cognitive limitations might have been due to their excluding children who had neurosensory limitations. . Recently, both the sex of the child  and the magnitude of the growth restriction  have been identified as factors that influence the relationship between growth restriction and development in children born at a low gestational age.
In this paper, we compare the development of extremely low gestational age newborns at age two years who were B-SGA (relatively severe growth restriction) and those who were F-SGA (less severe growth restriction) to their AGA peers. In addition, we explore the influence of the infant’s sex and neurosensory impairment on the relationship between growth-restriction and development.
The ELGAN study prospectively enrolled infants born between 23 0/7 and 27 6/7 weeks of gestation, from March 2002 until August 2004, at 14 institutions . Institutional Review Board approval was obtained at each study site and written informed consent was obtained from every one of the 1249 mothers of the 1506 infants enrolled.
Study participants were assigned to a B-SGA, F-SGA or AGA group before data analysis. Even though all B-SGA infants are, by definition F-SGA, we restrict the term F-SGA to those children who were not classified as B-SGA. The three weight groups are mutually exclusive (Figure 1).
Of the three separate fetal-weight distributions [16–18], and the six birth-weight distributions [19–24] we evaluated, we chose the fetal  and the birth  distributions that were sex-specific and provided the most reasonable number of growth-restricted newborns. No newborn had a dysmorphic syndrome or congenital infection associated with growth restriction.
Maternal race was collected by maternal report. Details about factors associated with delivery and antenatal steroids are presented elsewhere . The most reliable measure of gestational age is the timing of fertilization therapy or fetal ultrasound performed before the 14th week. When this information was not available, we accepted less reliable measures such as fetal ultrasound performed after the 14th week, the date of last menstrual period, and the gestational age as recorded by clinicians.
Anthropometric measures, including body weight, length and head circumference, were recorded throughout the hospitalization and at follow-up.
The diagnoses of patent ductus arteriosis, pneumothorax, pulmonary interstitial emphysema, pulmonary hemorrhage and bacteremia were collected from the medical record. Modified Bell staging criteria were used to classify necrotizing enterocolitis .
Mode of ventilation and days of oxygen were noted. An infant was considered to have bronchopulmonary dysplasia if supplemental oxygen was provided at 36 weeks postmenstrual age. Retinal examinations were obtained in keeping with clinical guidelines and retinopathy of prematurity was diagnosed by ELGAN Study ophthalmologists . Growth velocity (grams/kg/day) between days 7 and 28 was calculated using the formula 1000*[(wt28−wt7)/wt7]/21.
Ventriculomegaly and an echolucent lesion were identified by cranial ultrasound scans. The procedures for obtaining scans and interpreting them have been described elsewhere .
Parents were invited to bring their infants for neurodevelopmental follow-up testing at approximately 24 months corrected age. Of the 1200 children who survived to 2 years, 1103 (92%) had a developmental assessment, with 75% of those children evaluated between 23.5 and 27.9 months corrected age.
Anthropometric measures, including body weight, length, and head circumference, were recorded at birth by members of the clinical staff in the intensive care nursery. The research-trained examiner (neurologist, development/behavior pediatrician, or psychologist) measured the head circumference at age 2 years corrected age .
The developmental assessment included the Bayley Scales of Infant Development – Second Edition (BSID) , a neurologic exam , and the Vineland Adaptive Behavior Scales (VABS). Those who conducted these assessments were not provided with any information about the child prior to the examination.
Motor impairment was classified with the Gross Motor Function Classification System (GMFCS) . 1018 children had the full Bayley Scales of Infant Development assessment.
One set of analyses eliminated 122 children who had a major neurosensory limitation. This was defined as an inability to walk even with assistance (GMFCS 1), a significant hearing impairment (a parent report that the child required a hearing aid or needed special services for hearing impairment) or a significant vision impairment (a parental report that the child was considered legally blind in at least one eye).
We evaluated two separate hypotheses. The first hypothesis postulated that compared to children who were AGA, those who were classified as B-SGA and those classified as F-SGA were not at increased risk of developmental dysfunctions at 24 months corrected age. The second postulated that the relationships between intrauterine growth restriction and development dysfunctions at 24 months corrected age were not influenced by the child’s sex or neurosensory impairment.
We created multivariable logistic regression models for each sex and each outcome that compared B-SGA and F-SGA infants to the referent group of AGA infants. Potential confounders, defined as variables associated with growth restriction and with an outcome were included in each model. The strengths of association of the SGA variables are presented as risk ratios with 95% confidence intervals.
We created the same models in the sample of children who had no major neurosensory disturbance (defined as a GMFCS score < 1, and no appreciable hearing or vision impairment).
Infants born of multi-fetal gestations were included in multivariable analyses that controlled for multiple gestations. We also conducted separate multivariable analyses of singletons only, which provided similar findings, but with a smaller sample.
Because cerebral palsy diagnoses are mutually exclusive and each is compared to the same referent group, we created multinomial multivariable logistic regression models for these diagnoses.
In the total sample of 1506 newborns, 183 (12%) were classified as B-SGA (data not shown). An additional 109 (7%) infants, were classified as F-SGA only . These 292 infants represent 19% of the total study population.
In the sample of girls who had an MDI and PDI, 12% (N = 57) were B-SGA, 9% (N = 45) were F-SGA, 79% (N = 390) were AGA, while in the sample of boys who had an MDI and PDI, 8% (N = 42) were B-SGA, 7% (N = 35) were F-SGA, 85% (N = 449) were AGA. This is the sample for Table 4.
Among the girls in the restricted sample of children who had an MDI and PDI, but who did not have a major neurosensory limitation, 11% (N = 49) were B-SGA, 9% (N = 41) were F-SGA, and 79% (N = 346) were AGA. Among the boys in this restricted sample, 8% (N = 38) were B-SGA, and 7% (N = 34) were F-SGA, 84% (N = 388) were AGA. This is the sample for Table 5.
Race was the only maternal characteristic that was associated with SGA. Multi-fetal gestation was the only characteristic that was under-represented among B-SGA infants. Both SGA groups had higher Cesarean delivery rates than the AGA group. Preeclampsia and fetal indication were the disorders leading to preterm delivery for 9% of the AGA infants, but together they were the reason for delivery for 64% of B-SGA infants and 45% of F-SGA infants.
AGA infants were more likely than infants in both SGA groups to be born at 23–24 weeks. F-SGA infants had median head circumferences and birth weights intermediate between AGA and B-SGA infants.
No appreciable differences were noted in Apgar scores, delivery room resuscitation or SNAP–II scores (data not shown). The rate of death in the entire sample during the first 28 postnatal days was higher among B-SGA infants (22%) than among F-SGA (10%) and AGA infants (13%). The rate of death in the NICU after the 28th postnatal day was also higher among B-SGA infants (12%) than among F-SGA (6%) and AGA infants (4%).
AGA boys were more likely than AGA girls to have had moderate/severe ventriculomegaly on an ultrasound scan of the brain, and to a much smaller extent, an echolucent lesion. B-SGA boys were also more likely than B-SGA girls to be in the highest quartile of days on the ventilator and in the highest quartile of days receiving supplemental oxygen. F-SGA boys were more likely than F-SGA girls to have had bacteremia during the first postnatal week, and also surgical necrotizing enterocolitis. On the other hand, B-SGA girls were more likely than B-SGA boys to have had late bacteremia.
Since girls tend to differ from boys in their developmental trajectories [10, 32], we evaluated SGA relationships with developmental outcomes in girls separately from boys. Among girls, rates of MDI <70, PDI <70, and head circumference Z-score < −2 were highest in those who were B-SGA, while the rates of those who were F-SGA were only slightly higher than those who were AGA. Among boys, those born B-SGA and F-SGA had similarly elevated rates of MDI <70 and PDI <70 while only B-SGA boys were at increased risk of microcephaly.
At 24 months corrected age, 37% of B-SGA girls had an MDI below 70, while 40% of B-SGA boys had such a low MDI. This lack of male-female difference is in contrast to the rates of low MDIs among F-SGA and AGA children. 24% of F-SGA girls had an MDI below 70, while 40% of F-SGA boys had such a low MDI. Similarly, only 18% of AGA girls had an MDI <70, while 31% of AGA boys had such a low MDI.
B-SGA girls and boys tended to have low rates of quadriparesis and diparesis, but relatively high rates of hemiparesis. Both SGA groups of both sexes were more likely than their AGA peers to have a weight at age 2 years more than 2 standard deviations below the expected mean. B-SGA girls and boys had the highest rates of microcephaly at age 2 years, while F-SGA girls and boys tended to have rates that more closely approximated those of their AGA, rather than their B-SGA, peers.
For all multivariable analyses, we adjusted for race, gestational age, multi-fetal gestation, maternal and fetal indication for delivery, and route of delivery. The models for a head circumference Z-score <−2 also adjusted for head circumference <−2 at birth.
Compared to AGA girls, B-SGA girls were at significantly increased risk of a PDI < 70 [OR = 2.6 (1.4, 5.0)] and a 24-month weight Z-score < −2 [OR =12 (2.1, 69)]. These girls were at a nearly significantly increased risk of an MDI < 70 [OR = 2.0 (1.00, 3.9)]. In contrast, B-SGA boys were at a significantly greater risk only of a head circumference Z-score < −2 [OR = 4.8 (1.7, 13)].
To evaluate to what extent what we saw in Table 4 reflected the influence of major neurosensory limitations, we again created multivariable models, but this time among children who were able to walk independently, who by parent report did not use a hearing aid or receive special services for the hearing impaired, and who were not considered legally blind in either eye. This resulted in a loss of 56 girls (43 because of motor limitation, 15 for visual limitation and 10 for hearing limitation), and 66 boys (54 because of motor limitation, 14 for visual limitation and 12 for hearing limitation), some with multiple limitations.
In this sample, B-SGA girls were at increased risk of a score < 70 on both the MDI and PDI. In contrast, while they were not at increased risk of a PDI < 70 in the entire sample, both B-SGA boys and F-SGA boys were at increased risk in this restricted sample [B-SGA: OR = 2.6 (1.1, 5.8); F-SGA: OR = 2.2 (1.02, 4.7)]. The body weight and head circumference findings in Table 4 persisted. In addition, however, the modestly increased risk of a weight Z-score < −2 seen previously among B-SGA boys achieved statistical significance in this restricted sample [OR = 8.2 (1.6, 41)]
Three of our findings are worthy of comment. First, among all girls, severe growth restriction (B-SGA) is associated with delayed development while less severe growth restriction (F-SGA) is not. Second, among all boys, neither severe, nor less severe fetal growth restriction is associated with developmental delay. Third, major neurosensory limitations appeared to obscure increased risks of low MDI in girls with severe growth restriction, and low PDI in boys with severe and less severe growth restriction.
The weaknesses of our study are those of all observational studies. We are unable to distinguish between causation and association as explanations for what we found. We were not able to assign growth-restricted infants to groups of uniform risk of developmental adversities based on the etiology of the growth restriction, or reason for delivery.
Our study has several strengths. We selected infants based on gestational age, and not birth weight. This minimizes confounding due to factors related to fetal growth restriction . We collected all of our data prospectively. Also, at follow-up, examiners were not aware of the medical histories of the children they examined, thereby minimizing “diagnostic suspicion bias” . We minimized observer variability as best we could in the assessments and classifications of neurodevelopmental functions. We studied more homogenous, and therefore more informative, outcomes than studies that have composite outcomes, such as any neurodevelopmental impairment (defined as impaired early cognitive function or cerebral palsy or sensory impairment). Our sample is sufficiently large to appreciate relatively small increases in the odds ratios in subgroups. For example, the 34 F-SGA boys were approximately twice as likely as their 388 AGA peers to have a PDI < 70 (odds ratio = 2.2, 95% CI: 1.02, 4.7). Finally, attrition was modest, with 92% of survivors having had a developmental assessment at approximately two years post-term equivalent.
Among children in the EPIPAGE Study born between the 29th and 32nd weeks of gestation, SGA children and “mildly”-SGA children (birth weight was at or above the10th centile, but below 20th centile) were at comparably increased risk for minor cognitive and behavioral difficulties at age 5 years . No gradient was found. SGA and mildly-SGA had similar increased risks of cognitive and behavioral limitations.
Among EPIPAGE Study children born between the 24th and 28th weeks of gestation, however, neither being born SGA nor “mildly”-SGA was associated with increased risks of cognitive “deficiency” . Although a small percentage of our sample was born before 24 weeks, all the rest were born between the 24th and 27th weeks of gestation. Unlike the EPIPAGE Study, we evaluated girls separately from boys. Thus, our finding that B-SGA girls, but not F-SGA girls, are at increased risk of developmental delay should not be viewed as discrepant with the findings of the EPIPAGE Study. Indeed, our finding that SGA boys (regardless of whether B-SGA or F-SGA) are at not at increased risk of developmental delay at 2 years can be seen as similar to the EPIPAGE Study findings of no increased risk of cognitive deficiency among SGA and mildly-SGA children at 5 years.
Several commentators have suggested that infants whose birth weight falls below the 10th centile for fetuses of the same gestational age, but above the 10th centile for live-born infants of the same gestational age be re-classified as F-SGA [1, 2, 35]. Our data for infants < 28 weeks indicate that this reclassification does not add much predictive power about neurodevelopmental outcomes. Unlike B-SGA infants, F-SGA infants were not more likely to die in the neonatal period, nor to suffer more neurodevelopmental impairments. Perhaps it is time to approach the definition of growth restriction in relation to the points along the birth weight distribution where adverse outcomes begin to increase, rather than at an arbitrary point that falls a certain distance below the mean [36–38].
In our cohort, B-SGA females were at a substantially increased risk of MDI < 70, PDI < 70 and 24-month weight Z-score < −2. Intrauterine growth did not predict neurodevelopment in male infants. It appears that the neurodevelopmental measures most influenced by restricted by processes associated with reduced in-utero growth are MDI and PDI, and that the smallest female infants are most affected by phenomena associated with disordered growth. In this cohort of infants 52% of survivors are male, whereas 57% of deaths were males. Consequently, preferential mortality is unlikely to account for much of what we see.
The difference in the association of growth restriction and poor neurodevelopmental outcome in girls and boys is not well understood. Another study found that this discrepancy could not be attributed to differences in neonatal disorders such as intraventricular hemorrhage, periventricular leukomalacia or bronchopulmonary dysplasia . We, on the other hand, found that AGA boys were more likely than AGA girls to have ultrasound lesions indicative of focal and diffuse white matter damage were more common in AGA boys than in AGA girls, as well as impaired development at age 2 years. Others have also found that brain development in the male is more vulnerable that that in females .
Because these brain lesions are associated with heightened risk of developmental dysfunctions , the dysfunction in AGA boys probably has little, if anything, to do with intra-uterine growth restriction. Consequently, AGA boys, the most appropriate referent group for both B-SGA and F-AGA boys, might have been more similar to B-SGA and F-SGA boys, than AGA girls were to B-SGA and F-AGA girls. In part, this might have accounted for the relatively small effect on developmental function attributable to intra-uterine growth restriction in boys.
If the brain damage attributed to processes associated with fetal growth restriction was accompanied by major neurosensory limitations, then excluding children with major neurosensory limitations should result in our finding no association between growth restriction and cognitive impairment. Our excluding children with major neurosensory limitations, however, first revealed an association between growth restriction and PDI < 70 among boys that was not seen in the entire sample of boys.
A likely explanation is that low PDI among boys associated with a neurosensory limitation, most probably major motor limitation (i.e., GMFCS 1), is not associated with fetal growth restriction. Indeed that is just what we found. 6% of B-SGA, 7% of B-SGA and 19% of AGA boys who had a PDI <70 were not able to walk independently, while all boys (regardless of fetal growth status) who had a PDI 70 could walk independently.
In summary, our data support the concept that growth restriction is far more complicated than simply identifying a ‘small’ infant or fetus. Gestational age, the degree of growth restriction, and now sex and neurosensory limitations, all need to be considered when appreciating the long-term impact of altered in-utero growth . Unmeasured factors contribute to disordered growth, and these factors, not growth restriction itself, might influence development. Questions persist about how best to define growth restriction, especially in light of evidence that morbidity and mortality vary with birth weight above the 10th decile [36–38], and evidence that growth restriction effects/correlates vary with gestational age at birth [36, 41].
The study sponsors did not participate in, nor did they influence, the collection, analysis or interpretation of the data, the writing of the manuscript; or the decision to submit this manuscript for publication.
This study was supported by a cooperative agreement with the National Institute of Neurological Diseases and Stroke (5U01NS040069-05) and a center grant award from the National Institute of Child Health and Human Development (5P30HD018655-28). Dr. Streimish was supported by training grants from Eunice Shriver National Institute of Child Health and Human Development (T32 HD07094) and the Agency for Healthcare Research and Quality (2T32HS00006D-16). The authors gratefully acknowledge the contributions of their subjects, and their subjects’ families, as well as those of their colleagues.
Children’s Hospital, Boston, MA
Kristen Ecklund, Haim Bassan, Samantha Butler, Adré Duplessis, Cecil Hahn, Catherine Limperopoulos, Omar Khwaja, Janet S. Soul
Baystate Medical Center, Springfield, MA
Bhavesh Shah, Frederick Hampf, Herbert Gilmore, Susan McQuiston
Beth Israel Deaconess Medical Center, Boston, MA
Camilia R. Martin, Jane Share,
Brigham & Women’s Hospital, Boston, MA
Linda J. Van Marter, Sara Durfee
Massachusetts General Hospital, Boston, MA
Robert M. Insoft, Sjirk J. Westra, Kalpathy Krishnamoorthy
Floating Hospital for Children at Tufts Medical Center, Boston, MA
Cynthia Cole, John M. Fiascone, Roy McCauley, Paige T. Church, Cecelia Keller, Karen J. Miller
U Mass Memorial Health Care, Worcester, MA
Francis Bednarek, Jacqueline Wellman, Robin Adair, Richard Bream, Alice Miller, Albert Scheiner, Christy Stine
Yale University School of Medicine, New Haven, CT
Richard A. Ehrenkranz, Cindy Miller, Nancy Close, Elaine Romano, Joanne Williams
Wake Forest University Baptist Medical Center and Forsyth Medical Center, Winston-Salem, NC
T. Michael O’Shea, Barbara Specter, Deborah Allred, Robert Dillard, Don Goldstein, Deborah Hiatt, Gail Hounshell, Ellen Waldrep, Lisa Washburn, Cherrie D. Welch
University Health Systems of Eastern Carolina, Greenville, NC
Stephen C. Engelke, Ira Adler, Sharon Buckwald, Rebecca Helms, Kathyrn Kerkering, Scott S. MacGilvray, Peter Resnik
North Carolina Children’s Hospital, Chapel Hill, NC
Carl Bose, Lynn A. Fordham, Lisa Bostic, Diane Marshall, Kristi Milowic, Janice Wereszczak
Helen DeVos Children’s Hospital, Grand Rapids, MI
Mariel Poortenga, Bradford W. Betz, Steven L. Bezinque, Joseph Junewick, Wendy Burdo-Hartman, Lynn Fagerman, Kim Lohr, Steve Pastyrnak, Dinah Sutton
Sparrow Hospital, Lansing, MI
Ellen Cavenagh, Victoria J. Caine, Nicholas Olomu, Joan Price
Michigan State University, East Lansing, MI
Nigel Paneth, Padmani Karna
University of Chicago Medical Center, Chicago, IL
Michael D. Schreiber, Kate Feinstein, Leslie Caldarelli, Sunila E. O’Connor, Michael Msall, Susan Plesha-Troyke
William Beaumont Hospital, Royal Oak, MI
Daniel Batton, Karen Brooklier, Beth Kring, Melisa J. Oca, Katherine M. Solomon
Arkansas Children’s Hospital
Joanna J Seibert
Children’s Hospital of Atlanta
The authors have indicated that they have no financial relationships relevant to this article to disclose, nor do they have any conflict of interest to disclose.