 |
|
 |
|
 |
 |
|
 |
 |
 |
 |
 |
 |
 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
PEDIATRICS Vol. 108 No. 1 July 2001, pp. 142-151
; Michael J. O'Callaghan, MBBS, FRACP
, and
From the * Department of Paediatrics and Child Health,
University of Queensland; and the Objective. To examine the
relationship between child maltreatment and cognitive development in
extremely low birth weight infants, adjusting for perinatal and
parental risk factors.
Methods. A total of 352 infants with birth weight of
<1000 g were followed prospectively for 4 years. The data were
analyzed with regard to perinatal and parental risk factors and
referrals for suspected child maltreatment to government agencies.
Perinatal risk factors included birth weight, gestation, gender,
periventricular hemorrhage, ventricular dilation, home oxygen
requirement, and necrotizing enterocolitis. Parental risk factors
included maternal age, race, marital status, education, and hospital
insurance status. Cognitive z scores were calculated at
1, 2, and 4 years, and head circumference z scores were
calculated at birth, 2 years, and 4 years.
Results. Fifteen percent of infants were referred to child
protective services for suspected child maltreatment. The adjusted
general cognitive index at 4 years was significantly reduced in infants who were referred for neglect ( Conclusions. Childhood neglect is associated significantly
with delayed cognitive development and head growth. Addressing risk
factors antenatally and in early childhood may improve
outcomes.
Growth and Development Research
Unit and the § Child Protection Unit, Mater Children's Hospital,
Brisbane, Queensland, Australia.
ABSTRACT
Top
Abstract
Methods
Results
Discussion
References
17.6; 95% confidence interval: 3.3,
31.9). Infants whose neglect was substantiated had a progressive decline in their cognitive function over time (cognitive
z scores: 0.97, 1.37, and 2.05 standard deviations
at 1, 2, and 4 years, respectively), compared with non-neglected
infants (z scores: 0.04 to 0.36). They had a
significantly smaller head circumference at 2 and 4 years but not at
birth (adjusted z score at 4 years: 0.812; 95%
confidence interval: 0.167, 1.458). Perinatal risk factors and
physical disability were not related to maltreatment referral; only
parental factors were independent predictors.
There are many potential risk factors associated with
cognitive delay in extremely preterm infants. These include perinatal factors such as periventricular hemorrhage, ventricular
dilation,1 periventricular leukomalacia,2
chronic neonatal lung disease, necrotizing enterocolitis (NEC), male
gender,1 and being small for gestational age
(SGA).3 However, during recent years there has also been
recognition of the association between sociodemographic factors This has led to the question of how sociodemographic variables actually
influence cognitive development. A recent report of the National
Research Council and Institute of Medicine of the National
Academies6 provided extensive evidence that childhood
experience has a substantial impact on brain development. Child abuse
and neglect More than a decade ago, Leonard et al8 demonstrated that
among very low birth weight infants who were followed to 4.5 years of
age, a referral to child protective services was more significant in
predicting cognitive outcome than severe intraventricular hemorrhage,
chronic neonatal lung disease, or low socioeconomic status (SES).
Despite this finding, few, if any, subsequent studies have included
child maltreatment as a potential outcome predictor, and these results
have not been replicated in other high-risk infant groups. Although
Leonard's study identified child maltreatment as an important direct
risk factor, it did not adjust for other potential confounders, such as
ventricular dilation, NEC, maternal age, or marital status. The
potential impact of different types of maltreatment, such as childhood
neglect, also was not assessed.
In a review of the past decade, Kaplan et al9 reported
that child maltreatment was consistently associated with impaired cognitive ability within a childhood population. A recent
population-based study showed a strong association between cognitive
disability and maltreatment, including neglect, although the direction
of association could not be determined.10 The role of
disability, as either a maltreatment risk factor or an outcome, is
still unclear.
Although higher rates of child abuse and neglect have been reported
among preterm infants,7 no prospective studies within this
group have evaluated potential risk factors. Although one study
described the antecedents of abuse and neglect among neonatal intensive
care graduates,11 the mean birth weight and gestation was
significantly lower in the referred group. The question of the relative
contribution of perinatal and parental risk factors to the subsequent
maltreatment of preterm infants has not been addressed adequately in
the literature. The answer to this question is important in formulating
effective intervention strategies.
The purpose of this study was to examine potential risk factors and
cognitive outcomes associated with child maltreatment, within a highly
vulnerable infant group Patients
The study cohort consisted of 353 infants who were born at
<1000 g and admitted to the Mater Mothers' Hospital, Brisbane, Australia, between 1983 and 1993 inclusive, and survived to at least 4 months' corrected age. One child with Down's syndrome was excluded.
At 4 years of age, 269 (76%) of these children were cognitively
assessed at the Growth and Development Research Unit of the Mater
Children's Hospital (mean corrected age: 4.19 years; standard
deviation [SD]: 0.57). Of the total study cohort, 314 children (89%)
also had been assessed at 12 months of age and 336 (93%) had been
assessed at 2 years. A comparison of those who had been cognitively
assessed at 4 years with those who had not been assessed is presented
in Table 1.
TABLE 1
such
as poverty, low maternal education, and single parenthoodand the
cognitive development of preterm infants.4,5
including a deprived learning environment and physical or
emotional neglect or abuseoften coexist with many of these
sociodemographic factors,7 potentially confounding their
association with cognitive outcomes.
those born with extremely low birth weight
(ELBW; <1000 g). With a large prospective cohort of ELBW infants, we
were able to adjust for potential confounding perinatal and parental
variables while measuring associated cognitive changes over time.
It was hypothesized that among ELBW infants, parental risk
factorsrather than perinatal factorswould predict subsequent
maltreatment and that child maltreatment would be associated independently with cognitive delay.
METHODS
Top
Abstract
Methods
Results
Discussion
References
Comparison of Cases Assessed With GCI at 4 Years and Those Not Assessed
Perinatal Risk Factors
Ten perinatal risk factors, for which data were available consistently, were identified as being predictive of cognitive outcomes from a review of the literature.1,3,812-14 They were assessed prospectively in relation to cognitive development, as well as maltreatment referral. These included birth weight, gestation, SGA status, gender, multiple births, the requirement for home oxygen, grade 3 to 4 periventricular hemorrhage, moderate to severe ventricular dilation, NEC (equivalent to modified Bell's staging IIA or higher15), and retinopathy of prematurity (any stage, as identified by an ophthalmologist). SGA was defined as birth weight less than the third percentile for gestation on standardized growth charts.16 Periventricular hemorrhage was measured as grade 0 (nil), grade 1 (subependymal hemorrhage), grade 2 (hemorrhage filling <50% of ventricle), grade 3 (hemorrhage filling >50% of ventricle), and grade 4 (hemorrhage with parenchymal involvement).17 Ventricular dilation was classified as nil, mild, moderate, or severe, as described previously.17
Apgar scores also were recorded at 1 and 5 minutes. Assessments of periventricular leukomalacia and sepsis were not included in the analysis because of incomplete data collection.
Parental Risk Factors
Identified parental risk factors included maternal age, race, marital status (married, cohabiting couple, or single ["single" included all divorced, separated, widowed, or never-married mothers]), maternal education, and hospital insurance status.4,5,18 In Australia, public hospital care is readily available; only 48% of the population had private hospital insurance at the time of this study.19 Maternal racial origins were classified as white, Australian aboriginal, Asian, or other. These factors also were assessed in relation to cognitive outcomes and the risk of maltreatment referral.
Referral for Child Maltreatment
Referrals for suspected child abuse or neglect between 1983 and July 1999 were obtained from the government agency, Families, Youth and Community Care Queensland (FYCCQ), and from interstate registries and overseas, where indicated. This included the date of referral, whether the referral was substantiated on subsequent investigation, and the type of maltreatment referred or substantiated (physical, emotional, sexual, and neglect). Although medical practitioners are the only mandatory reporters by law in Queensland, referrals may be received from any member of the public, after being screened by FYCCQ workers.
Substantiated referrals included cases in which, after departmental
investigation, there was "reasonable cause to believe that the child
had been, was being, or was likely to be abused or
neglected."20 Although these referrals were more likely
to represent the more significant caseswith confirmatory evidence
availablethis study considered any referral to indicate a degree of
risk. All referrals were used in regression analyses. Physical abuse
was defined as any nonaccidental physical injury inflicted by a person
who had care of the child. The definition of emotional abuse included
any act resulting in a child's suffering any kind of emotional
deprivation or trauma. Childhood neglect was defined as a "failure to
provide conditions that were essential for the healthy physical and
emotional development of a child." Finally, the definition of sexual
abuse included exposing a child to or involving a child in
inappropriate sexual activities.20 Referrals that occurred
before the time of the 4-year cognitive assessment were distinguished
from all recorded referrals, to ensure that infants who had been born
earlier in the study were not subject to reporting bias.
Confidentiality was maintained with the use of a confidential identification number. Names and dates of birth along with the confidential number were sent directly to FYCCQ, who then linked referral data to the number only, returning this data to the researchers directly. Researchers then obtained anonymous child and maternal data to link to the referral records. Thus, researchers and Growth and Development Research Unit staff had no information to link referrals to particular infants. Ethical approval was obtained from both the Mater Mothers' and Mater Children's Hospitals Research Ethics Committees.
Outcome Measures
Surviving ELBW infants were followed at the multidisciplinary
Growth and Development Research Unit of the Mater Children's Hospital,
where they were medically assessed and seen by a developmental psychologist. The psychologist assessed cognitive development using the
Griffiths general quotient21 (GQ; mean: 100; SD: 12) at 1- and 2-year visits and the McCarthy general cognitive index22 (GCI; mean: 100; SD: 15) at 4 years, with
z scores calculated to compare scores across time.
"Borderline intellectual disability" was defined as IQ from 1 to
2 SD below the mean, and "intellectual disability" was defined as
IQ below 2 SD. Medical assessment included the recording of clinical
data by a pediatrician, as well as plotting height, weight, and head
circumference on standardized growth charts at birth, 2 years, and 4 years of age. Standardized z scores, independent of
gestation and corrected age, were calculated on the basis of the 1990 British growth data.16 Assessment of physical disability
was completed by pediatricians in consultation with physical
therapists, ophthalmologists, and audiologists, as relevant. A
diagnosis of cerebral palsy included diplegia, spastic hemiplegia,
quadriplegia, or other related forms of motor dysfunction. All study
infants had hearing assessed by brainstem evoked response before
discharge and by behavioral response and tympanometry at 8 to 12 months. Deafness was defined as sensorineural hearing impairment
requiring amplification. Vision was assessed by an ophthalmologist,
with blindness defined as absent or minimal light perception in both
eyes at last assessment.
Statistical Analysis
The 
2 test was used to compare
differences in categorical variables and analysis of variance for
differences in mean scores. Multiple linear regression analysis was
used to determine the independent predictors of GCI and head
circumference. Logistic regression analysis was used to assess the
independent predictors of referral for child maltreatment. A 2-tailed
P value <.05 was considered significant. Statistics were
performed using SPSS for Windows (Release 9.0; SPSS Inc, Chicago, IL).
| |
RESULTS |
|---|
|
Top
Abstract
Methods
Results
Discussion
References
|
|---|
Fifty-two (15%) of the 352 ELBW children were referred to child protective services for suspected child maltreatment; 32 children (9%) were referred before the 4-year assessment. Fifty percent of these 32 infants were referred before the age of 5.5 months (median corrected age: 7.6 months; interquartile range: 0.4, 30.0 months).
Twenty-seven (52%) of the 52 referred children were reported on >1 occasion, often with >1 type of maltreatment reported on each occasion. Thirty-one children (60% of those referred) had 1 or more substantiated report, including 19 children (37%) with substantiated neglect. Overall, 167 maltreatment reports were received on 117 occasions, of which 80 (48%) were substantiated. Neglect was the most frequently reported type of maltreatment (N = 71), followed by emotional abuse (N = 51), physical abuse (N = 39), and sexual abuse (N = 6). The impact of sexual abuse was not evaluated because of insufficient numbers. Both neglect and emotional abuse were seen in 22 children (42% of referred children), neglect and physical abuse were seen in 14 children (27% of referred children), and neglect and physical and emotional abuse were seen in 13 children (25% of referred children).
Of the 352 study infants, 269 (76%) were cognitively assessed at 4 years of age, including 32 (62%) of the 52 referred infants (Table 1). Of the 83 children without cognitive scores, just more than half were lost to follow-up (N = 43); almost one third of these families subsequently were reported for suspected child maltreatment (ie, 14 referred children). However, 27 of the remaining children attended the clinic appointment but were unable to be cognitively assessed because of disability (N = 13; none referred), behavior problems (N = 12; 4 referred), or being from a non-English-speaking family (N = 2; none referred). In addition, 1 child had died at 2 years of age (with no reports recorded), 4 had moved interstate (1 referred), and 8 were assessed at a later age with the use of other psychometric tests (1 referred). Of the 13 infants who could not be assessed because of disability, 10 had had ventricular dilation and 12 had cerebral palsy, bilateral blindness, or deafness, as determined at age 4.
Those who were not assessed at 4 years were more likely to have had lower intelligence scores on previous assessments and to have been born to young, unmarried, poorly educated mothers who were more often reported for child maltreatment. They also were more likely to have had periventricular hemorrhage and dilation and physical disability at 2 years (Table 1). Ten of the 16 infants who had moderate to severe ventricular dilation and who were not assessed had cerebral palsy, bilateral blindness, or deafness identified at 4 years, precluding their cognitive testing. All of the infants who had grade 3 to 4 periventricular hemorrhage and who were not assessed also had ventricular dilation. However, none of these medical factors was associated with a significant increase in the rate of referral (Table 2) or of referral for neglect.
|
Parental risk factors were far more likely to predict referral for child maltreatment than perinatal factors; only NEC and retinopathy of prematurity reached statistical significance. Physical disability at 2 years was not predictive of referral at any time. On univariate analysis, all parental risk factors were associated strongly with maltreatment referral (Table 2). The same factors also were significant in predicting referral for neglect.
To determine the independent contribution of risk factors on maltreatment referral, we included parental and significant perinatal risk factors simultaneously in a logistic regression model. Although aboriginal race was associated strongly with referral (with 82% of aboriginal families referred), because of small numbers and a strong association with other parental variables, these cases were excluded from the analysis. After adjustment for other potential confounding variables, the association between maltreatment referral and the perinatal risk factors became nonsignificant, as did maternal age, single marital status, and maternal education. The only independent predictors of referral among ELBW infants were public hospital status (adjusted odds ratio [adj OR]: 3.5; 95% confidence interval [CI]: 1.1, 10.6) and unmarried cohabitation (adj OR: 3.3; 95% CI: 1.4, 7.7). None of the factors included in the model was associated independently with referral for neglect.
In examining the perinatal predictors of GCI outcome, child maltreatment and parental risk factors were more strongly associated with cognitive delay than were perinatal factors. The only perinatal factors that were associated significantly with lowered scores were male gender (male: 93; female: 99; P = .013) and ventricular dilation (GCI: 86 vs 97; P = .037). However, all parental variables were highly predictive of reduced GCI (Table 3). Similarly, all categories of child maltreatment referral, except substantiated physical abuse, were associated strongly with cognitive delay (Table 4); the most pronounced differences were seen in cases of substantiated neglect (GCI: 98 vs 69 in maltreated infants; P < .001).
|
|
To control for possible confounding, we used a multiple linear regression model with GCI at 4 years as the dependent variable. Perinatal and parental risk factors were entered simultaneously in the model, together with a single variable for all child maltreatment referrals (adjusted R2: 0.18; F = 4.7 [df 16, 251]; P < .001). As 5 of the 11 aboriginal mothers were lost to follow-up (of whom 4 were reported), aboriginal cases were excluded from the regression analysis. The only significant independent perinatal risk factors were low birth weight (<750 g) and male gender (Fig 1). Of the parental factors, limited maternal education, public hospital status, and referral for any form of child maltreatment also were associated significantly with cognitive delay (Fig 2).
|
|
The most dramatic results were seen after adjustment for the 3 individual maltreatment subtypes (physical abuse, emotional abuse, and
neglect). Factors that were found to be significant in the previous
model (ie, male gender, birth weight, incomplete high school education,
and public hospital status) were added to the regression model. Sexual
abuse cases and aboriginal families were excluded, again because of
limited numbers. After adjustment for confounding variables, neglect
was found to be the only maltreatment subtype that was associated
independently with cognitive delay, with a difference of 17.6 GCI
points (Fig 3; adjusted R2: 0.15; F = 7.67 [df 7, 249]; P < .001). In addition, the association
between cognitive outcome and hospital insurance status was no longer
statistically significant.
|
Cognitive assessments were evaluated from 12 months and 2 years of age
with the use of the GQ. This showed a progressive decline in the mean
cognitive z score in infants who were referred for neglect,
reaching the level of "intellectual disability" in the substantiated group. However, IQ levels in infants who were not referred for neglect remained relatively constant (Fig
4). All differences between referred and
nonreferred groups were statistically significant. This remained true
after inclusion of only cases that were referred before the 4-year
assessment. A similar decline also was seen in neglected infants after
adjustment for perinatal and parental risk factors (as previously
described in Fig 3; adjusted z scores: 0.16, 0.34, and
1.01 at 1, 2, and 4 years, respectively), in all referred cases
(cognitive z scores: 0.42, 0.83, and 1.13), and in
substantiated cases (cognitive z scores 0.58, 1.00, and
1.20).
|
Infants whose neglect was substantiated also had a significantly
smaller head circumference at 2 and 4 years, after adjustment for birth
weight of <750 g, SGA, ventricular dilation, Apgar score 
5 at 5 minutes, and cerebral palsy at 2 or 4 years, respectively (adjusted
R2: 0.146; F = 9.36 [df 6, 288]; P < .001 at 4 years). This was despite
having similar measurements at birth to those not referred for neglect
(adjusting for factors above, except cerebral palsy; Fig
5). Other types of child maltreatment, including physical abuse, were not associated with a reduced head circumference at 4 years. There also was no significant association between neglect and infant weight or height at 2 or 4 years, even after
adjustment for factors that were significant on univariate analysis
(including birth weight of <750 g, gender, SGA, cerebral palsy, and
multiple births).
|
| |
DISCUSSION |
|---|
|
Top
Abstract
Methods
Results
Discussion
References
|
|---|
In assessing the relative impact of perinatal and parental variables on an ELBW infant's cognitive development, it is evident that both play an important role. However, this study also has shown that childhood neglect is a highly significant independent factor, associated with lower intellectual abilities and reduced head circumference.
Although physical disability is the most visible adverse outcome of extreme prematurity, cognitive delay is by far the most common.23 In matching school kindergarten records with statewide birth records, Resnick et al18 found that whereas perinatal factors were most predictive of severe disability, sociodemographic factors most strongly correlated with cognitive delay, including learning disability, emotional disability, and academic problems. The authors concluded, after considering differences in prevalence, that the impact of sociodemographic factors on 5-year educational outcomes was more than 10 times that of very low birth weight (<1500 g).
Predictors of Cognitive Delay
After adjustment for social risk, the only 2 perinatal factors that remained significantly associated with cognitive delay were low birth weight (<750 g) and male gender. Other studies also have shown that long-term cognitive delay is associated significantly with birth weight of <750 g23,24 and male gender.25 Ventricular dilation also has been shown to predict intellectual impairment.1,13 Although our study demonstrated this association on univariate analysis, it was not significant after adjustment for potential confounding variables. However, more than 60% of the neonates who had ventricular dilation and who were not cognitively assessed at 4 years had cerebral palsy, blindness, or deafness, which precluded testing. It is highly likely that ventricular dilation is, in fact, a true predictor of cognitive delay.
Numerous studies5,18,24,26 also have shown an association
between cognitive delay in preterm infants and sociodemographic
variables, including SES, low maternal education, single marital
status, and nonwhite raceall of which were significant on univariate
analysis in our study. However, after adjustment for childhood neglect
and other significant factors, maternal education was the only
independently significant parental risk factor. Numbers were
insufficient to analyze further the observed association with
aboriginal race, although this issue requires further investigation.
Public hospital status, a broad marker of SES, was marginally below the
specified level of significance.
Escalona,26 in prospectively following a neonatal intensive care population and adjusting for SES, suggested that environmental deficits and stressors might impair early cognitive development, particularly in preterm infants. Although that study used a standardized measure of SES that was not incorporated into our study, the potential for child maltreatment was not considered. Although Leonard et al8 found low SES to be associated with lowered cognitive outcomes, the individuals in the subgroup that was not referred for abuse were more likely to score within the normal range. Providing financial resources without recognizing issues related to child maltreatment is unlikely to have a significant impact on the problem of resource utilization, as observed elsewhere.27
Because families that are at social risk also are at increased risk of child maltreatment,7 sociodemographic factors may be surrogate measures for the quality of parent-child interaction. As reviewed by Zeanah et al,28 poverty and socioeconomic disadvantage may be associated with fewer resources, affecting nutrition, shelter, and health. Maternal education may have an impact on the mother's problem-solving abilities, material resources, or understanding of normal developmental milestones, which in turn may have an impact on the parent-child relationship. As Kalmar and Boronkai29 demonstrated, intellectual stimulation of both preterm and term infants is associated significantly with IQ outcomes, and this persists after adjustment for other biological factors over time.
In effect, the parental risk factors may represent inadequacies in the
parents' abilities to provide the physical and emotional resources
required for a child's optimal developmenta less identifiable but
significant form of "neglect." Other studies that have measured home environment factors directly in preterm infants also have shown a
significant relationship with cognitive
development.24,30,31 "Referral for child maltreatment"
is an even more specificalthough probably much less sensitivemarker
of adverse parent-child interaction. The main disadvantage is its
potential to underestimate prevalence vastly. Without a non-ELBW infant
control group, however, the present study could not determine whether
the parental factors measured were specific to this ELBW population or
were applicable to other high-risk infants.
Another limitation of this study was that it did not examine child
protective interventions, such as the provision of additional family
support or out-of-home care, which may have affected cognitive outcomes. However, a majority of studiesbut not
all32have shown improved long-term cognitive outcomes in
children who receive intervention.33-35 Without such
intervention, one would expect the cognitive differences between
referred and nonreferred infants to be even more pronounced.
Predictors of Maltreatment Referral
The present study also demonstrated that parental risk factors are
more significant than perinatal factors in predicting which ELBW
infants are at risk for being maltreated. Although disability in
general has been implicated as a risk factor for child
maltreatment,10 in our study, physical disabilitydefined
as cerebral palsy, blindness, or deafnesswas not associated with a
higher rate of referral. Although Sullivan and Knutson's
work10 showed that referral rates of children with
physical disabilities were higher than the control population without
disabilities, the rates of maltreatment were still lower than for
children with cognitive-based or behavioral disabilities. With more
than 80% of their study population representing children with
cognitive or behavioral disability, it is possible that many cases of
disability may represent an outcome of rather than a risk factor for
child maltreatment. This question deserves further study.
Socioeconomic variables also have been implicated commonly as risk factors for child maltreatment.11 Unmarried cohabitation, however, rarely has been examined; most studies have failed to distinguish this family type from married or single status. One other study demonstrated a higher relative risk for abuse in cohabiting couples compared with married or single parents, although preterm infants specifically were not examined.36 Although the mechanisms behind this finding are uncertain, it may be related to relationship instability or characteristics of the perpetrator of abuse. More specific risk factors for child abuse and neglect, such as family violence, drug addiction, and maternal depression,7 were not assessed. Although our data are useful in formulating intervention strategies, the findings cannot be extrapolated to the general birth population without a term infant control group.
Neglect, Cognition, and Head Growth
In seeking to understand how childhood neglect may be associated with head growth, neuroscience research provides valuable insight. During fetal development and early childhood, brain growth is regulated by genes but is critically influenced by sensory stimulation and experience. Neuronal connections are formed and modified by repetitive, patterned stimulation of the neural system in a "use-dependent" manner.37 Forty years of animal research has demonstrated unequivocally the anatomic plasticity of the brain in response to sensory stimulation or, conversely, to deprivation. Animals that are reared in stimulus-deprived environments tend to have a reduced cerebral weight and length and cortical depth. This finding correlates with a variety of histologic changes, including decreased neuron perikaryonal and nuclear size and dendritic branching and reduced numbers of neuroglia and synaptic connections, as reviewed by Walsh38 and Glaser.39
More recent human studies have shown that a small head circumference in infancy is associated with cognitive delay in very low birth weight 40,41 and ELBW42 infants. In addition, the English and Romanian Adoptees Study Team43 demonstrated a 2-SD reduction in the mean head circumference of late-adopted Romanian orphans who had experienced profound emotional and physical neglect. However, in contrast to our study, the authors were unable to distinguish the effects of malnutrition on head growth from those of emotional or cognitive deprivation. The present study demonstrated that a reduced head circumference and cognitive delay both are associated with childhood neglect, independent of other growth parameters, and that this association seems to become more pronounced over time and with severity of exposure (see Figs 4 and 5). Although our study did not distinguish differing types of neglect, other researchers have suggested that global sensory deprivation is more strongly correlated with head size than "chaotic" neglect.44 Distinguishing subtypes of neglect would be useful in future research.
Using anatomic magnetic resonance imaging, De Bellis et al45 observed significant differences in the cerebral volumes of maltreated children compared with matched controls. They postulated that this could be related to elevated catecholamine and cortisol levels often found in traumatized children46,47 or to early sensory deprivation, both of which may result in altered neuronal differentiation and synaptic proliferation.
With these studies linking early childhood neglect, reduced head circumference, and cognitive delay, one possible causal pathway is proposed. However, without parental IQ or head circumference data, genetic factors cannot be excluded. It is conceivable that some ELBW children may have a genetic predisposition for slowing of head growth in early infancy, with associated cognitive delay. These children with developmental delay may then be more likely to come to the attention of child protection authorities.
Addressing Childhood Neglect
In 1994, the Advisory Board on Child Abuse and Neglect declared a "state of emergency" with respect to child maltreatment, because of the widening gap between the extent of the problem and resources allocated to address it.48 The National Institutes of Health recently identified childhood neglect as a "serious public health, justice, social services, and education problem," with a paucity of research addressing the issue.49
Despite advances in neonatal intensive care technology, which result in increased survival,50 extremely preterm infants remain at high risk for cognitive delay.12,51 Our failure to recognize or address adequately the risk factors for abuse and neglect may be a contributory factor. As Rosenblatt observed, "The basic incongruity in ... perinatal care lies in our superb ability to care for the individual patient and our dismal failure to address the problems of the larger society."52
One of the most promising means of addressing our "dismal failure" is through home visitation programs, focusing on parents of ELBW infants who are at the highest risk of maltreatment. Several programs have been shown to improve cognitive outcomes in preterm infants34,53 as well as in infants with nonorganic failure to thrive, presumably related to neglect.54 Another study55 demonstrated that home visitation was associated with a significantly reduced incidence of child abuse and neglect. In contrast, a longitudinal study of preterm infants that focused more on center-based child intervention failed to demonstrate sustained long-term results.56
Although our study identified possible predictors and cognitive
consequences of child maltreatment in ELBW infants, we now need to
secure resources to intervene effectively. Only then can we hope to
improve significantly the cognitive developmentand overall quality of
lifeof this vulnerable infant group.
| |
ACKNOWLEDGMENTS |
|---|
This research was supported by Golden Casket funding.
We thank Yvonne Rogers for assistance in coordinating the follow-up of children through the Growth and Development Research Unit and managing the database. We also thank Families, Youth and Community Care Queensland for permitting access to its database, and Mary Greenwood for collating the child maltreatment referral data.
| |
FOOTNOTES |
|---|
Received for publication Feb 18, 2000; accepted Jan 17, 2001.
Reprint requests to (L.S.) Department of Pediatrics, Baylor College of Medicine, Meyer Center for Developmental Pediatrics, Texas Children's Hospital MC 3-2335, Houston, TX 77030-2399. E-mail: lxstrath{at}texaschildrenshospital.org
| |
ABBREVIATIONS |
|---|
NEC, necrotizing enterocolitis; SGA, small for gestational age; SES, socioeconomic status; ELBW, extremely low birth weight; SD, standard deviation; FYCCQ, Families, Youth and Community Care Queensland; GQ, general quotient; GCI, general cognitive index; Adj OR, adjusted odds ratio; CI, confidence interval.
| |
REFERENCES |
|---|
|
Top
Abstract
Methods
Results
Discussion
References
|
|---|
a review.
J
Child Psychol Psychiatry
2000;
41:97-116 [CrossRef][Medline]This article has been cited by other articles:
 |
|
 |
|
  K. M. Beaver, J. P. Wright, and M. Delisi Self-Control as an Executive Function: Reformulating Gottfredson and Hirschi's Parental Socialization Thesis Criminal Justice and Behavior, October 1, 2007; 34(10): 1345 - 1361. [Abstract] [PDF]
|
|
|
|
 |
|
 V. Lee and P. N. S. Hoaken Cognition, Emotion, and Neurobiological Development: Mediating the Relation Between Maltreatment and Aggression Child Maltreat, August 1, 2007; 12(3): 281 - 298. [Abstract] [PDF]
|
|
|
|
 |
|
 J. V. Browne and A. Talmi Family-Based Intervention to Enhance Infant-Parent Relationships in the Neonatal Intensive Care Unit J. Pediatr. Psychol., December 1, 2005; 30(8): 667 - 677. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. D. De Bellis The Psychobiology of Neglect Child Maltreat, May 1, 2005; 10(2): 150 - 172. [Abstract] [PDF]
|
|
|
|
|
|
H. Dubowitz, R. R. Newton, A. J. Litrownik, T. Lewis, E. C. Briggs, R. Thompson, D. English, L.-C. Lee, and M. M. Feerick Examination of a Conceptual Model of Child Neglect Child Maltreat, May 1, 2005; 10(2): 173 - 189. [Abstract] [PDF]
|
|
|
|
|
|
H. Dubowitz, S. C. Pitts, and M. M. Black Measurement of Three Major Subtypes of Child Neglect Child Maltreat, November 1, 2004; 9(4): 344 - 356. [Abstract] [PDF]
|
|
|
|
|
|
M. Jonson-Reid, B. Drake, J. Kim, S. Porterfield, and L. Han A Prospective Analysis of the Relationship Between Reported Child Maltreatment and Special Education Eligibility Among Poor Children Child Maltreat, November 1, 2004; 9(4): 382 - 394. [Abstract] [PDF]
|
|
|
|
 |
|
 L. Strathearn Long-term Cognitive Function in Very Low-Birth-Weight Infants JAMA, May 7, 2003; 289(17): 2209 - 2209. [Full Text] [PDF]
|
|
|
|
 |
|
 L. R. Ment, H. S. Bada, P. Barnes, P. E. Grant, D. Hirtz, L. A. Papile, J. Pinto-Martin, M. Rivkin, and T. L. Slovis Practice parameter: Neuroimaging of the neonate: Report of the Quality Standards Subcommittee of the American Academy of Neurology and the Practice Committee of the Child Neurology Society Neurology, June 25, 2002; 58(12): 1726 - 1738. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
