WCO/18688; Total nos of Pages: 7;
WCO 18688
Is cerebral palsy preventable? Karin B. Nelsona,b and Taeun Changb a
National Institute of Neurological Disorders and Stroke, Bethesda, Maryland, USA and bDepartment of Neurology, Children’s National Medical Center, Washington, DC, USA Correspondence to Karin B. Nelson, MD, National Institutes of Health, Building 31, Room 8A03, Bethesda, MD 20892-2540, USA Tel: +1 301 594 9486; fax: +1 301 496 2358; e-mail:
[email protected]
Current Opinion in Neurology 2008, 21:000–000
Purpose of review To use evidence of good medical quality to update information on strategies for prevention of cerebral palsy, and on the success of these preventive efforts to date. Recent findings Causes of cerebral palsy, and therefore promising approaches to prevention, differ by gestational age group and by clinical subtype. Neuroimaging and neuropathology indicate the importance of white matter disorders and of ischemic stroke in cerebral palsy; birth asphyxia, congenital malformations, placental pathology, and genetic variants also contribute to cerebral palsy risk. Multiplicity of risk factors markedly increases risk. Recent studies indicate that mild hypothermia lessens cerebral palsy risk in term infants with moderate neonatal encephalopathy, and the possibility that administration of magnesium sulphate to women in preterm labor may aid in primary prevention of cerebral palsy in very preterm infants. Summary Past efforts to prevent cerebral palsy have not had the benefits sought, but recent results provide new hope and new challenges. Keywords cerebral palsy, congenital hemiparesis, electronic fetal monitoring in labor, prematurity, quadriplegia, spastic diplegia Curr Opin Neurol 21:000–000 ß 2008 Wolters Kluwer Health | Lippincott Williams & Wilkins 1350-7540
Introduction Considerable effort has gone into the prevention of cerebral palsy (CP), most of it to date ending in disappointment. Techniques such as continuous electronic monitoring of the fetus in labor have not had the anticipated benefits. Many neuroprotective strategies once considered for infants have been tried in adult stroke, and most have failed. In premature infants, an increase in survival without a decrease in prevalence added more healthy citizens but also more disabled children to the population. Now, however, a drop in prevalence of CP in one sector of preterm infants may be occurring. Until recently, strategies for primary or secondary prevention of CP were apparently ineffective [1], but of late there is evidence that some proportion of CP may be preventable.
Trends in cerebral palsy prevalence Has the rate of CP declined? That is, is CP being prevented? For very preterm infants, evidence from population-based registries is chiefly but not consistently encouraging, a number of studies reporting a decline of CP in infants less than 1500 g [2,3], while in Finland [4] and in a Canadian province [5] no such decline was evident. In term infants, there was no change in CP frequency over 1350-7540 ß 2008 Wolters Kluwer Health | Lippincott Williams & Wilkins
a 16-year period [6], although Himmelman et al. [7] report decline in severe bilateral CP. Interestingly, the form of CP most related to birth asphyxia, dyskinetic CP [8], has not declined [9]. To the extent that there has been a change in CP rate in very preterm births, what accounts for it? Wilson-Costello [10] indicates more use of prenatal steroids, less use of steroids postnatally, and decreases in neonatal sepsis and intraventricular hemorrhage. Comparisons across nations must be cautious as differing patterns of care may alter the balance between mortality and long-term morbidity.
Is etiology, and therefore strategy for prevention, similar in all children with cerebral palsy? The pathobiology of CP differs considerably in different gestational age groups and in different CP subtypes [11]. In premature infants, spastic diplegia is the dominant form of CP, and its basis as indicated by neuroimaging [12] and by neuropathologic investigation [13] is most often white matter disorder. White matter disorder may also underlie CP in infants delivered at or near term. While this lesion is sometimes described as ischemic, the most prominent risk factors identified for white matter
WCO/18688; Total nos of Pages: 7;
2 Developmental disorders
disorders are inflammatory conditions and cytokine exposure [14]. Overexpression of inflammatory cytokines has been observed in areas of leukomalacia [15]. There is potential interaction between hypoxic-ischemic and inflammatory pathologies [16]. In a neuroimaging study of congenital hemiplegic CP, the most common causes identified in term infants were perinatal stroke and congenital malformation [17]. Injury to neck vessels can be a factor in perinatal stroke but is seldom documented. Asymmetric white matter disorder may be a more common cause of spastic diplegia in premature infants, although stroke can also occur in infants born before term [18]. Quadriplegic CP, especially with dyskinesia, is the subtype of CP sometimes related to acute asphyxia during the process of birth [8]. Other potential causes of dyskinetic CP include marked neonatal hyperbilirubinemia. Dystonia, sometimes with chorea, can be a result of a treatable metabolic disorder, dopa-responsive dystonia. Unlike the classical later-onset Sagawa disease [19], which must be considered in an older child with dystonia, there is an early onset form with hypotonia including bulbar involvement and cognitive delay, commonly with diurnal variation [20]. Treatment with L-dopa can produce dramatic though incomplete response in the motor disorder but does not alter cognitive status.
Etiology factors and their preventability It is reasonable to suppose that preventability would be different in different etiologies; we therefore examine preventability of CP according to its apparent causal factors. Birth asphyxia
Severe asphyxia at birth can lead to CP, but controlled studies in defined populations over two decades make it clear that this is not a common cause of CP. Determining when birth asphyxia is present is not easy as there is no direct measurement and definitions tend to depend upon Apgar scores or respiratory depression, both of which are clearly results of antecedent pathology rather than causes. The best available indicator is probably metabolic acidosis, yet acidosis too can have other causes. Features of electronic fetal monitoring during labor, such as decelerations, are not specific for asphyxial conditions. A question that is key for development of strategies for prevention but seldom addressed is, if hypoxia or ischemia is present, when did it occur? Was it the initiating pathology, or a downstream consequence of some other instigating problem? This distinction is crucial for development of strategies for prevention.
Primary prevention of cerebral palsy due to birth asphyxia: fetal monitoring in labor
Continuous fetal heart rate monitoring in labor (EFM) was in the hope of preventing CP. This approach was based on the assumption that most CP is due to birth asphyxia and that early recognition of impending asphyxia could alert caregivers to take action that would prevent neurologic injury. EFM was widely disseminated in the 1970s, and initial reports indicated that its use was attended with higher median pH values and Apgar scores. Differences within the ranges of pH and Apgar score, however, are not associated with substantial change in risk of CP, so this initial dependence on surrogate measures for evaluation of the utility of EFM was misleading. Subsequently randomized clinical trials have demonstrated no decrease in CP – and in one [21], a statistically significant increase – in children whose births were monitored continuously by electronic means compared with those monitored by intermittent auscultation. A recent summary [22] describes this disappointing saga. An earlier review of the topic [23] had the informative title, ‘Birth can be a hazardous journey: electronic fetal monitoring does not help’. Recently, investigators aware that most CP does not arise from birth asphyxia argued reasonably that a fairer test would be to examine the association of specific findings on EFM as predictors of marked fetal acidosis or hypoxic– ischemic encephalopathy [24]. Their conclusion was that while such associations did exist, ‘EFM is not a precise tool in the identification of metabolic acidosis or HIE’. Thus evidence of good medical quality does not indicate that acting on EFM findings makes it possible to prevent CP. In the decades that followed the widespread use of EFM, the CP prevalence in term and near-term babies did not decline. Importantly, all the relevant clinical trials found that use of EFM was associated with increased rates of obstetric intervention, commonly surgical delivery which, especially during active labor as tends to occur with EFM, poses risks to mother and infant. The addition of fetal pulse oximetry did not reduce the surgical delivery [25]. Use of fetal electrocardiography in labor was associated with less severe acidosis and fewer infants with encephalopathy, but not with fewer low Apgar scores or admissions for special neonatal care [26]. Could even wider use of cesarean section lead to better outcomes? In infants delivered preterm for premature rupture of membranes, surgical delivery did not improve neonatal outcomes [27], nor was elective section associated with lower mortality or fewer cerebral abnormalities on cranial ultrasonography [28]. Respiratory morbidity is higher with elective section at term but stillbirth rate lower; surgical delivery is thus associated with competing
WCO/18688; Total nos of Pages: 7;
Is cerebral palsy preventable? Nelson and Chang 3
risks and benefits for the infant, and with risks for the mother. Signore et al. [29] conclude that impact of more cesarean delivery on the fetus/neonate would be mixed and any improvement small.
Prevention of preterm birth
Obstetric catastrophes can occur, and when they do, often arise suddenly and unpredictably. Apparently we do not yet have reliable tools for preventing harm to babies when such catastrophes strike, without risk to other mothers and babies. The moral of this disappointing saga is, as Adrian Grant, an obstetrician epidemiologist, commented, more ‘information’ does not necessarily lead to better outcomes, and that new diagnostic interventions, however promising these may look, should be tested in randomized trials before they become widely disseminated.
Prevention of cerebral palsy among infants born before term
Secondary prevention of cerebral palsy due to birth asphyxia: hypothermia
Therapeutic hypothermia for term neonatal encephalopathy has been associated with a lower rate of CP in infants with moderate encephalopathy in two major clinical trials, one with selective head cooling [30], the other with whole-body hypothermia [31]. Hypothermia must be initiated within 6 h of insult (or birth) and maintained for 48–72 h, with gradual rewarming [32]. A metaanalysis of eight randomized controlled hypothermia trials (regardless of method) of over 600 term newborns [33] showed a significant reduction in the combined outcome of infant mortality and major neurodevelopmental disability at 18 months of age. No clinical trial has yet compared the two primary methods of cooling. Long-term neurologic follow up is pending in the two large studies. There is experimental evidence that hypothermia may be beneficial in inflammation, as well as in hypoxia or ischemia. Trials of therapeutic hypothermia are underway for newborns and infants with other indications, including cardiac arrest. Future studies will probably incorporate hypothermia as an adjunct to other therapies for neural protection [34]. Prematurity
Birth before term is a major risk factor for CP. The rate of preterm birth in the US is now 12.7% and rising [35]. Causes of preterm birth include infection, congenital maldevelopment [36], and genetic factors [37]. A substantial minority arise from iatrogenic interventions for complications in infant or mother; not surprisingly, the conditions presenting risk to mother or infant may themselves have consequences for the infant. Many very preterm infants have bouts of definite or suspect sepsis in the newborn period, and these add to risk of CP [38]. Complexity of early history in these fragile infants also makes it difficult to track the relationship of maternal and pregnancy factors with later outcome.
Despite efforts to prevent prematurity, available strategies have not yet been consistently successful at decreasing the frequency of preterm birth in the US [39,40].
CP in preterm infants is dominantly stable disease, which is especially linked to inflammation and inflammatory mediators. There has been no effort to block cytokines and their receptors in human newborns, as these cytokines are multifunctional, related to development and repair as well as to injury. Nitric oxide is important in vasomotor regulation and, among its wide range of other actions, increases blood flow to the lungs after birth. In an observational study [41] comparing administration of inhaled nitric oxide with 100% oxygen to very preterm infants with persistent pulmonary hypertension, nitric oxide administration was reported to be associated with a significantly lower rate of CP at 3 years of age. This conclusion disagreed with results of two prior randomized trials, and disagreed also with findings of a recent trial of inhaled nitric oxide in preterm infants with severe respiratory failure [42]. A systematic review [43] of 11 randomized trials of nitric oxide in preterm infants that did not include the two studies just mentioned concluded that inhaled nitric oxide may decrease serious brain injury in mildly but not in severely ill preterm infants. In term infants, however, nitric oxide use was not associated with less disability later on [44]. Animal data and an observational clinical study [45] suggested that administration of magnesium sulphate to women in very preterm labor may provide neuroprotection for the infant, and a lower rate of motor handicap. A systematic review of randomized trials to date [46] indicated no major maternal or fetal complications, and ‘a significant reduction in the rate of substantial gross motor dysfunction’ (relative risk 0.56, 95% confidence interval 0.33–0.97), based on two trials, although the diagnosis of CP was only marginally significantly reduced in a further four trials. A large American multicenter trial has recently been completed and results are undergoing analysis as of this writing. Early developmental intervention programs were associated with better early but not later cognitive performance and no improvement in motor outcome [47]. Intrauterine exposure to infection/inflammation
A succession of reports has linked maternal fever or placental inflammation (chorioamnionitis) with adverse neurologic outcome [48]. The fetal inflammatory response and cytokine cascade, and the effect of these especially on oligodendrocytes, has been implicated in
WCO/18688; Total nos of Pages: 7;
4 Developmental disorders
the white matter damage that underlies much of CP [49]. The association of fever or inflammation with CP has been consistent in term infants, less consistent in very preterm infants although recent reports are dominantly positive. Episodes of sepsis or suspected sepsis, common in the early days and weeks of life in very preterm neonates, are also linked with CP risk [38]. Not all maternal fevers in labor reflect infection, and separation of hyperthermia itself from the causes and consequences of the hyperthermia has not yet been demonstrated. For example, lengthy epidural anesthesia induces fever in experimental animals, and human hyperthermia in labor associated with epidural administration is related to lower Apgar scores and more neonatal seizures [50]. In this study, however, placental histology was not included and thus chorioamnionitis could not be excluded. The association of maternal fever has been consistent, and strong enough to make it advisable to adjust for maternal fever in evaluating prognosis in the neonate. Inflammatory changes in placenta have a graver relationship to outcome if vascular pathology is also present [51,52]; as often noted, the concomitant presence of multiple factors for CP greatly increases risk. Prevention of cerebral palsy related to infection/inflammation
To date, there is no evidence-based approach to prevention of CP or other adverse long-term neurologic outcome, other than that related to meningitis or encephalitis, by means of administration of antibiotics or other therapy directed at infection or inflammation. Increasing exposure of pregnant women to antibiotics has potential risks as well as benefits. Stroke
Perinatal ischemic stroke has in recent years been recognized to be a common cause of CP, especially in term infants [53,54]. Neonates with strokes do not present clinically with hemiparesis, but rather with nonspecific signs such as neonatal seizures or apneas, or no acute manifestations of neurologic injury. Hence the diagnosis of perinatal stroke can be suspected but not established without neuroimaging or neuropathology, and the recognized prevalence of perinatal stroke depends on the frequency with which imaging studies, especially MRI, are performed. The prevalence of unilateral stroke in term infants, as recognized by MRI in the newborn period, was one in 2300 [55]. Bilateral stroke occurs in a significant minority, in addition, and most cases of [56] perinatal stroke resulting in hemiplegic CP are not recognized until after the perinatal period. Perinatal stroke differs from stroke in the older child or adult in many regards, including current or recent
attachment to a mother and a placenta, the proinflammatory and procoagulant background of pregnancy and parturition, high hematocrit and blood viscosity in the fetus/newborn, placental vasculopathy, the possibility of acidosis and vascular traction in delivery, dehydration and hypotension and intravascular catheter use in the nursery period, and other features worth attention in future strategies for prevention. It seems likely that the placenta holds the key to many of the mysteries of perinatal stroke, but only a few studies have sought to relate placental pathology to neonatal neurologic disorders in the newborn period and later [57,58]. About half of infants with stroke have thrombophilias, but these same thrombophilias are often noted in asymptomatic individuals. Perinatal stroke seldom repeats in a sibship, suggesting that environmental factors are important, but only a few studies have focused on environmental risk factors [59]. It is likely that it is combinations of thrombophilic risk factors, and of these with environmental factors, that determines whether stroke will occur. Prevention of cerebral palsy due to perinatal stroke
Identification of factors leading to thrombosis of fetal or neonatal stroke is still a work in progress. No study has yet tried to established selection criteria for women to receive anticoagulation therapy in pregnancy for the purpose of avoiding stroke in the infant, and none has examined whether stroke is more or less frequent in infants following anticoagulation of the mother for maternal or pregnancy complications. There is no available evidence that neurologic morbidity due to perinatal stroke is preventable. Congenital maldevelopment
Two recent studies link population-based registries for CP and for congenital malformations [60,61], and a third compares rates with published norms [62]. Even without consistent specialized imaging, cerebral birth defects and congenital malformation of head, clefts of lip or palate, and gut atresias are more common in children with CP than in the general population, and more distant noncerebral malformations may also be more common. That congenital maldevelopment in or near cerebral structures is more common in children with CP strongly suggests the contribution of prenatal processes to some proportion of CP. Monozygotic twinning with death of a twin is a major risk factor for CP in the survivor. ‘Vanishing’ of a co-twin may be one avenue to malformation in the survivor via catastrophic vascular collapse [61]; thus some children with CP were probably twins early in gestation, although born as singletons. This theory is suggestively supported by observations in fertility clinics [63], although most twins from IVF are dizygotic.
WCO/18688; Total nos of Pages: 7;
Is cerebral palsy preventable? Nelson and Chang 5
Prevention of congenital maldevelopment, except in folate deficiency or with use of acne or antiepileptic medications, is in general not yet feasible. Multiple gestation
Twins and higher order multiple births have a higher rate of CP, accounted for in part by the commonly associated prematurity, but to an important degree also by death of a co-twin (see above). Relatively advanced maternal age and assistive reproductive technologies are associated with a higher rate of twinning and of prematurity, which may account for much or all of the increased risk of CP in babies born of assistive techniques [64]. Prevention of neurologic disability due to complications of multiple gestation is possible through lowering the rate of multiple gestation by avoidance of pregnancy in older women and by avoidance of implantation of multiple fetuses in IVF. Genes and cerebral palsy
Familial aggregation of CP is observed in groups with high consanguinity [65], and a Swedish national database indicates increased familial risk for CP [66]. Inherited thrombophilias occur in greater numbers of children with perinatal stroke than in the general population [63,64], in very preterm infants with CP, [67] and in CP overall [68]. A genetic variant of an inflammatory cytokine is associated with CP risk [69], as is an apoliprotein E genotype [70]. There are also genetic components of a number of risk factors for CP, including preterm birth, abruption of the placenta, chorioamnionitis, and preeclampsia. Only a start has been made in linking genetic vulnerabilities to the environmental factors that interact with those vulnerabilities; this is likely to be a fruitful direction of future research.
Reasons for hope Although about half of CP arises in infants born at term, and about three-quarters in term and near-term infants, study of CP arising in big babies has lagged because so many are born outside of medical facilities with specialized personnel and equipment that permit research, and because the selected subgroup of infants born at or near term who reach research facilities may not be typical of the general population of such infants. Now with computerized medical care and national and regional registries for CP, a large and representative population of babies of all gestational ages may be feasible, and studies of uncommon maternal and pregnancy conditions linkable with CP outcome, enabling studies of important but uncommon risk factors.
Web-based data entry, computerized medical records and electronic databases will improve entry and tracking. Improvements in and greater accessibility of nonsedated magnetic imaging and MRI-compatible incubators, wider use of neuroimaging in infants and children, bedside continuous video EEG and improvements in technology and remote digital technology will enable multicenter trials of treatment and of preventive measures. Development of consensus on methodology and interpretation of placental pathology, and incorporation of placental investigation into studies in obstetrics and neonatology, are likely to add important information. Recognition that CP is commonly a result of multiple interacting risk factors rather than to a single cause will permit development of more sophisticated and realistic animal models.
Conclusion There is unlikely ever to be a silver bullet for prevention of CP. As with mental retardation, the task will be to recognize individual subsyndromes that make up the totality of CP, and develop preventive strategies targeted at each. Is CP preventable? Most cases are not, as yet, but there is some reason for hope.
Acknowledgements This review was supported in part by the Intramural Research Program of the National Institute of Neurological Disorders and Stroke. The content of this publication does not necessarily reflect the views or policies of the Department of Health and Human Services, nor does mention of trade names, commercial products, or organizations imply endorsement by the US government.
References and recommended reading Papers of particular interest, published within the annual period of review, have been highlighted as: of special interest of outstanding interest Additional references related to this topic can also be found in the Current World Literature section in this issue (pp. 000–000). 1
Nelson KB. Can we prevent cerebral palsy? N Engl J Med 2003; 349:1765– 1769.
2
Platt MJ, Cans C, Johnson A, et al. Trends in cerebral palsy among infants of very low birthweight (
