 |
|
 |
|
 |
 |
|
 |
 |
 |
 |
 |
 |
 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
* Department of Child Health, University of Bristol, Bristol, United Kingdom
Cardiovascular Genetics, University College London, London, United Kingdom
Neonatal Intensive Care Unit, University of Bristol Medical School, Southmead Hospital, Bristol, United Kingdom
|| School of Human Development, University of Nottingham, Nottingham, United Kingdom
 |
ABSTRACT |
|---|
 TOP ABSTRACT METHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
Methods. Genotype was determined from dried blood spots that were taken for neonatal screening tests 7 days or more after birth; outcome was independently assessed as part of a longitudinal study of children of 
32 weeks’ gestational age.
Results. CC genotype was associated with worse intensive care indices. Significant hemorrhagic brain injuries occurred in 5 (19%) of 27 children with CC genotype compared with 7 (6%) of 121 children with GC or GG genotype, and images consistent with white matter damage (ventriculomegaly or cystic periventricular leukomalacia) occurred in 9 (26%) of CC patients compared with 9 (7%) in GC/GG children. Disability occurred significantly more often in CC children: 8 (31%) compared with 16 (13%). A similar trend was also noted in children with cerebral palsy (15% compared with 7%, respectively). Developmental, cognitive, and motor scores at 2 years and 5.5 years were independent of genotype among children with or without disability.
Conclusions. In a population of surviving children who were born at 32 weeks’ gestational age, variation of the gene that may increase IL-6 synthesis is associated with disabling brain injury but not cognitive development despite association with worse early critical care indices.
Key Words: interleukin-6 • polymorphism • infant premature • neurodevelopment
Abbreviations: PVL, periventricular leukomalacia • CP, cerebral palsy • IL-6, interleukin-6 • WMD, white matter damage • IVH, intraventricular hemorrhage • APIP, Avon Premature Infant Project • BAS-II, British Ability Scales, Second Edition • OR, odds ratio • CI, confidence interval
Serious disability is common in preterm children and is inversely related to gestational age. The prevention of disability after premature birth remains a major challenge for neonatal care. As many as one quarter of those who survive after birth at 25 weeks’ gestational age or younger have a severe disability at 2.5 years of age, and an additional 25% have other disabilities.1 Furthermore, very preterm children without disability remain at risk for a range of motor, cognitive, behavioral, and psychological deficits during childhood.2 In addition, persistent educational problems are found even in children who are born at 32 to 35 weeks.3
Neurologic deficits in surviving very preterm infants are associated with hemorrhagic brain injuries (periventricular-intraventricular hemorrhage, hemorrhagic parenchymal infarction) or injury to the periventricular white matter (periventricular leukomalacia [PVL]).4,5 Both motor dysfunction and abnormality of cognitive function may be related to observed brain injuries,6,7 but this association is far from universal, as some children with brain injury and subsequent motor disability from cerebral palsy (CP) have normal intellect, and many children who develop disability do not have abnormal cranial imaging.1
The magnitude of the inflammatory response may be causally associated with worse outcome in critically ill patients.8–10 The development of respiratory distress syndrome after preterm birth is accompanied by local (lung) and systemic inflammatory responses.11–13 The risk of preterm brain injury is increased with the severity of the early respiratory illness and its cardiovascular effects and effects on cerebral perfusion.14–16 In addition, the development of cerebral injury and disabling motor problems, such as CP, may be directly related to the concentrations of the proinflammatory cytokines, such as interleukin-6 (IL-6), to which the preterm or term infant is exposed.17–23
A functional polymorphism in the IL-6 gene promoter has recently been identified: a C>G change at position –174. In vitro IL-6 production in lipopolysaccharide–stimulated neonatal monocytes is higher among those of CC genotype than among G-allele carriers.24 Similar trends in IL-6 level are identified in vivo in response to neonatal delivery24 and cardiopulmonary bypass25 and among those with abdominal aortic aneurysm,26 whereas the C allele is associated with poorer endothelial function in smokers,27 raised C-reactive protein levels,28,29 and myocardial infarction.29,30 Such data suggest a robust association of the C allele with IL-6 gene responsiveness.
We previously reported that IL-6 GG genotype is associated with an elevated rate of sepsis in preterm newborns.31 Although septicemia is associated with white matter injury, it is probably the total exposure to the proinflammatory mediators that affects the white matter that is important (in terms of absolute IL-6 levels and duration of exposure).20–22 Thus, the development of sepsis may not be as important a factor in determining white matter damage (WMD) as the degree to which IL-6 is induced. If increased IL-6 levels are causally related to an enhanced inflammatory response, then we might expect the IL-6 CC genotype to be associated with increased illness severity in the early neonatal period and ultrasound evidence of severe intraventricular hemorrhage (IVH) and periventricular WMD and the development of later CP, disability, or impaired cognitive function. We have taken the opportunity afforded by a detailed outcome evaluation, The Avon Premature Infant Project (APIP),32 to test this hypothesis in a group of white children who were born prematurely at 32 weeks.
|
METHODS |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
Patients
Patients were white children who were recruited to APIP. These individuals were born at 32 weeks’ gestational age or less in 1 of 2 Bristol (United Kingdom) hospitals during a 30-month period. Care in the 2 centers was supervised by the same group of 3 neonatologists using shared protocols, ensuring uniformity of management. The ventilatory style used an "open lung" strategy with inspiratory times between 0.35 and 0.45 seconds with the aim of maintaining normocapnia (arterial pressure of carbon dioxide between 4.5 and 6.0 kPa). The minimal acceptable blood pressure on day 1 was equivalent to the gestational age of the infant in mm Hg. Neonatal data were coded prospectively for computer analysis. Observations were made in the first 12 hours of the maximum and minimum fraction of inspired oxygen (FIO2; the highest FIO2 required to maintain arterial oxygen concentrations [arterial oxygen pressure] between 5 and 10 kPa in the first 12 hours after birth). All participated in a prospective, randomized controlled trial of Portage (an early education program focused on the child) or social support (parent adviser scheme focused on the mother) intervention started at discharge and continued for up to 2 years, as previously described.32 The original study32 demonstrated a small benefit (an improvement in Griffith Developmental Quotient33 of 3–4 points) for developmental and parental support programs after accounting for social and family influences. Follow-up has continued to 5.5 years.
Cranial Injury and Developmental Progress
Cranial ultrasound was performed on the day of birth and then at least weekly until discharge. Images were reviewed prospectively (by N.M.). Hemorrhagic ultrasound findings were classified as subependymal, intraventricular, or intraparenchymal, IVH being graded as small or large (the latter where the clot distended the lateral ventricle). Severe hemorrhage was defined as a large IVH or an intraparenchymal hemorrhage. Cystic PVL was defined as echo-lucent cysts of 
2 mm diameter in the periventricular white matter34; ventriculomegaly was present when the ventricular size on the last scan was >97th percentile.35 No children developed progressive posthemorrhagic hydrocephalus in this cohort. For analysis, cystic PVL and ventriculomegaly were combined as WMD, consistent with the view that most ventricular enlargement is attributable to WMD.36–38 The worst grade of ultrasonic findings was used to describe hemorrhagic injuries and images consistent with focal and diffuse WMD.
CP, defined as a disorder of movement or posture, including hypertonia, associated with disability, was formally assessed at both ages (2 years corrected for prematurity and 5.5 years’ chronological age). Disability was defined, as any disability, using published descriptions from the Oxford Health Authority/National Perinatal Epidemiology Unit,39 by outcome over neuromotor, vision, hearing, and cognitive domains and included those with disabling CP only and not those with abnormality of tone without significant functional deficit. At 2 or 5.5 years, children who were not ambulatory, had developmental scores <70, were blind, or had profound deafness were considered severely disabled.
At 2 years’ corrected age, a psychologist performed the Griffiths Scales of Mental Development,33 and a neurologic assessment was performed by a single pediatrician (N.M.) and a physiotherapist. At 5.5 years an additional psychologist performed the British Ability Scales, Second Edition (BAS-II),40 and a research nurse performed the Movement ABC.41 These assessments were blind to the child’s neonatal course and progress over the intervening period. The Griffiths Scales comprise 5 subscales from which is derived an overall developmental quotient. Griffiths quotient was standardized originally to a mean of 100, with a standard deviation of 15, but secular drifts in population scores have resulted in a higher population mean. Thus, for severe disability, a score of 70 (–2 standard deviations) was chosen to indicate severe disability instead of the conventional 55 (–3 standard deviations). The BAS-II was standardized in the early 1990s and was used to compute general cognitive ability together with visuospatial, verbal, and nonverbal subscales. The Movement ABC scales assesses manual dexterity, ball skills, and balance over 10 tests. Scores of each component are summed to produce a score ranging from 0 to 40, with high scores indicating a more impaired motor skills and 0 indicating normal skills.
IL-6 Genotyping
DNA was extracted from the blood spots that were taken from the stored newborn metabolic screening (Guthrie) cards by boiling in sterile distilled water after heavy metal ion chelation.42 IL-6 genotypes were resolved by polymerase chain reaction amplification of a 190-bp fragment of the IL-6 gene promoter region, application of the restriction endonuclease NlaIII, and separation of the digestion products using a microtiter array diagonal electrophoresis gel stained with 0.1% ethidium bromide, as previously described.25,26 Two independent staff members who were blinded to patient data performed analysis. All tests were taken on or after the seventh day as part of the routine screening for biochemical and endocrine disorders.
Statistical Analysis
Outcomes were examined for association with IL-6 promoter polymorphism genotypes –174 CC and GC+GG, consistent with previous study24–26 as the CC genotype in particular enhances circulating IL-6 concentrations. Analysis was performed using SPSS for Windows v9.0 (SPSS Inc, Chicago, IL). Categorical data were analyzed by 
2 or Fisher exact test where appropriate. Normally distributed data were analyzed by t test. Mann-Whitney U test was used when the distribution of data was skewed or numbers in 1 group for comparison were <30. P < .05 was considered statistically significant for this exploratory study.
|
RESULTS |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
Characteristics of those who were evaluated for genotype did not differ from those of the whole cohort in terms of birth weight, gestation, use of antenatal or postnatal steroids, duration of ventilation, or presence of ultrasound brain injury. Of the 148 children, 48 (32%) had GG genotype, 73 (49%) had GC genotype, and 27 (18%) had CC genotype. This distribution was compatible with Hardy Weinberg equilibrium and was not significantly different from that found for other populations of white individuals in the United Kingdom in work by this group25,26 or in term infants.24 Table 1 shows the distribution of perinatal and social variables between the 2 groups, for which there were no significant differences, although the CC group had fewer boys and more randomized to developmental intervention groups in the original trial. In addition, there was a slight trend toward more infants from twin pregnancy in the CC group (Table 1), although, because of the random nature of isolating the Guthrie spots, there were only 3 twin pairs in the study. Twin pregnancy, compared with singleton pregnancy, was not associated with neurologic outcome: severe IVH (twin, 2, [9%]; singletons, 10 [8.5%[; P = .97), PVL (twin, 3 [13%]; singleton, 13 [11%]; P = .78), CP (twin, 3 [13%]; singleton, 10 [8%]; P = .53), or disability (twin, 4 [16%]; singleton, 20 [16%]; P = .96).
|
Compared with children with CG+GG genotype, children who were homozygous for the C allele had higher frequency of severe neonatal hemorrhagic lesions (large IVH or hemorrhagic parenchymal injury) and images consistent with WMD (Table 2). CC genotype increased the risk of the development of severe hemorrhagic lesions (odds ratio [OR]: 3.5; 95% confidence interval [CI]: 1.0–12.2; P = .038) and WMD (OR: 4.1; 95% CI: 1.4–12.2; P = .008). At follow-up, 4 (15%) children with the CC genotype had CP, compared with 9 (7%) of the CG/GG group (OR: 2.2; 95% CI: 0.6–7.6; P = .224). However, disability was present in 8 (30%) children in the CC group compared with 16 (13%) of the CG/GG group (OR: 2.8; 95% CI: 1.04–7.4; P = .046; Table 2).
|
|
|
DISCUSSION |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
Although our data support the theory that the –174 CC genotype is causally responsible for enhanced IL-6 gene expression and IL-6 activity, within the circulation and/or the brain, studies of the role of IL-6 genotype and its effects (if causal) in different disease states are contradictory. The –174 C allele or CC genotype apparently upregulates IL-6 production in most conditions24–30 but not all.43 Furthermore, the IL-6 gene haplotype (the contribution of other IL-6 promoter polymorphisms), which we have not evaluated, may be important.44 We have simply associated CC genotype with worse brain outcomes, and we have no data to inform us as to the effect of genetic variation on circulating or tissue (brain or lymphocyte) cytokine activity in the healthy or ill preterm infant. A candidate gene association study, however, obviates the need for IL-6 levels particularly when complex gene–environment interactions (of unpredictable timing, scale, and duration) are anticipated; here, for instance, IL-6 levels would have needed to be assayed ideally every 6 hours and perhaps more frequently when clinical changes occurred, as a result of the volatility of cytokine expression.25
Although the developing oligodendrocytes of the periventricular white matter might be particularly vulnerable to the neurocytopathogenic effects of IL-6 after ischemia or hypoxic or free radical injury, to date, in vivo evidence in the human to support a causal role for IL-6 for the transduction of the neuropathic effects of inflammatory stimuli has relied on measurements of amniotic fluid, blood, cerebrospinal fluid cytokine levels, or histology. The association between genotype and the occurrence of ultrasound images consistent with WMD noted here supports the proposition that cytokine (IL-6)-mediated inflammation may affect the developing brain and influence neurologic outcome. However, we cannot determine whether it is systemic IL-6 production or secondary local cerebral IL-6 synthesis that is neuropathogenic45 or whether it is the inflammatory response itself that is the primary insult. In addition, it must be noted that little is known about the constitutive regulation of the IL-6 gene and IL-6 production in different centers of the human brain or after exposure to ischemia, free radical injury, endotoxin, or inflammatory mediators of systemic origin.
The developmental study of this cohort provided a unique opportunity to have independently and prospectively collected data with which to evaluate genotype. No significant developmental effect was demonstrated as part of the randomized trial (APIP) of which they formed a part.32 The study group allocation did not confound any of the analyses reported above. However, it should be stressed that this was a cohort of surviving children who had had neonatal biochemical screening tests performed, and we are not aware of the genotype distribution among the deaths in this population. The genotype distribution among survivors was typical of other white populations that we have studied, although we might expect the IL-6 –174 CC genotype to be associated with perinatal risk in its own right. Mortality is known to be higher among boys, and it is noted that there were numerically fewer boys in the CC genotype group than in the GC+GG group (48% vs 65%; Table 1). However, the finding of Hardy-Weinberg equilibrium and allele frequencies consistent with other studies in adults and term children25–27,43 or in term infants24 suggest that no preadmission survival bias had occurred. Even if such a phenomenon had existed, however, it could in no way have influenced the allele-associated illness propensities identified.
Sample size for the CC group was small (n = 27), and this reduces the power of the comparisons performed. Although CP was found twice as often in this group, this did not reach conventional significance. In contrast, the hybrid functionally defined group of children with disability was found significantly more often in association with CC genotype. This observation may suggest widespread IL-6–mediated cerebral injury. In contrast, we were unable to demonstrate an association between IL-6 genotype and cognitive performance at either age, despite the associations with severe IVH, PVL, and disability. We examined the comparisons post hoc for a range of potential social confounders within the original APIP study, but this did not alter the lack of association found (Table 3). Although preterm infants with periventricular WMD have reduced cortical volume,46 the pathologic processes that mediate the development of the gray matter in this group are poorly understood. Abnormalities in structure of gray or white matter or other brain regions in preterm children detected in the second decade, using current cranial magnetic resonance imaging techniques, do not seem to be closely associated with cognitive, behavioral, or psychological performance measures.47
The suggestion that acute neonatal respiratory illness may be more severe in children with the CC genotype provides a possible independent association that may have bearing on the risk of neuroinjury, for which respiratory illness is a major risk factor. The lack of association between genotype and the use of treatment to support blood pressure may be attributable to the difficulty in defining a "normal" blood pressure for a preterm infant. Alternatively, this latter finding may demonstrate the relative unimportance of blood pressure per se in comparison with cardiac output, blood flow, and end-organ (eg, cerebral) perfusion.14,15 Consistent with this is our finding that infants with the CC genotype have worse base excess. Therefore, it is tempting to speculate that the CC genotype through worsening the severity of early cardiorespiratory status and cerebral perfusion could increase the risk of severe IVH or WMD.
The lack of association between genotype and chronic lung disease of prematurity, a condition that may share some common inflammatory mediators with WMD and itself a risk factor for poor developmental progress after preterm birth,48 may simply reflect the small number of patients with chronic lung disease and the relative maturity of the population studied. It is also possible, however, that early pulmonary inflammatory affects are not as crucial to the development of chronic lung disease as subsequent insults such as pulmonary infection.49 Given the links between chronic lung disease and the fetal inflammatory response50 and also PVL, CP, and the fetal inflammatory response,21,22 it was interesting to note the slight trend toward a lower gestational age at birth in the CC group and a lower birth weight in the GG+GC group (Table 1). This would be consistent with CC infants’ being at greater risk of inflammatory phenomena, whereas infants of the GG+GC genotype may be less likely to suffer from intrauterine inflammation. However, both premature rupture of the fetal membranes and spontaneous preterm labor perhaps occurred less frequently in pregnancies with CC genotype infants (Table 1), although, again, the differences did not approach statistical significance. Additional larger, prospective study is required to explore these possibilities as precedent for genetically determined proinflammatory effects on preterm birth are known.51,52
|
CONCLUSIONS |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
|
ACKNOWLEDGMENTS |
|---|
|
FOOTNOTES |
|---|
Reprint requests to (D.H.) Department of Child Health, St Michael’s Hospital, Bristol, BS82JZ, UK E-mail: david.harding{at}bristol.ac.uk
|
REFERENCES |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
This article has been cited by other articles:
 |
|
 |
|
  V. Degos, G. Loron, J. Mantz, and P. Gressens Neuroprotective Strategies for the Neonatal Brain Anesth. Analg., June 1, 2008; 106(6): 1670 - 1680. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 D. Harding Impact of common genetic variation on neonatal disease and outcome Arch. Dis. Child. Fetal Neonatal Ed., September 1, 2007; 92(5): F408 - F413. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. Wilson and H. Montgomery Impact of genetic factors on outcome from brain injury Br. J. Anaesth., July 1, 2007; 99(1): 43 - 48. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
D R Harding, S E Humphries, A Whitelaw, N Marlow, and H E Montgomery Cognitive outcome and cyclo-oxygenase-2 gene (-765 G/C) variation in the preterm infant Arch. Dis. Child. Fetal Neonatal Ed., March 1, 2007; 92(2): F108 - F112. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
 |
||
|
||
Home | My Pediatrics | Journal Information | Current Issue | Past Issues | Subscriptions & Services | Contact Us Click here for faster international access © 2004 American Academy of Pediatrics. All rights reserved. |
||
|
||
