. . . the lax and pliable state of a child's head in parturition, the bones of the cranium having no sutures at that time, was such - that by force of the woman's effort, which in strong labor pains, was equal, upon an average, to a weight of 470 pounds avoirdupois acting perpendicularly upon it; it so happened that, in 49 instances out of 50, the said head was compressed and molded into the shape of an oblong conical piece of dough, such as a pastry cook generally rolls up in order to make a pie of.
- Good God! cried my father, what havoc and destruction must this make in the infinitely fine and tender texture of the cerebellum! Lawrence Sterne in The Life and Opinions of Tristram Shandy, Gentleman, 1760.
The notion that perinatal events are major causes of brain damage has been so prevalent for so long that it was a fit subject for parody more than 200 years ago. And yet it is only recently that longitudinal studies have clarified the relationship between events surrounding birth and neurologic development. The most important clarification, emerging from several sources, but most especially the analyses of the Collaborative Perinatal Project data by Nelson and Ellenberg, is that of the many forms of neurologic dysfunction that can arise in children, only cerebral palsy bears a consistent relationship to perinatal adversity.1
The term cerebral palsy encompasses a number of clinical neurologic syndromes of quite heterogeneous etiology. These syndromes are grouped together not because they are assumed to have a common cause but because they share characteristics that are important in clinical management. In fact, a wide variety of adverse influences in the perinatal period, acting by themselves or in concert, can injure the brain in such a way as to produce the motor impairments characteristic of cerebral palsy. Some authors therefore prefer to use the plural form the "cerebral palsies."2
As investigation into the etiology of cerebral palsy develops, it may be possible to someday classify this disorder etiologically. At present, however, only a very small proportion of cases of cerebral palsy can confidently be assigned to an etiologic category based on clinical findings alone. The combination of choreoathetosis and sensorineural deafness virtually always implies kernicteric brain damage attributable to bilirubin encephalopathy. Pure congenital ataxia (without associated diplegia) is very commonly familial. Infants with spastic diplegia frequently, but by no means always, have a history of very pre-term birth. Aside from these three examples, only the first two of which exhibit a high degree of specificity, it is not possible to reliably associate a specific cause with a specific subtype of cerebral palsy.
Although it is tempting to do so, clinicians should be wary of assigning a cause to a case of cerebral palsy simply because of the presence of adverse factors in the perinatal period, since perinatal complications are very common in the general population. Sixty-two percent of pregnancies registered in the Collaborative Perinatal Project experienced one or more late obstetric complications. 3 When a patient with cerebral palsy is found to have experienced nuchal cord or meconium passage, it is all to easy to assume that these complications have played a causal role in the production of the neurologic impairment. But such a conclusion should be tempered by the realization that 25% of births experience nuchal cord, and 18% meconium passage.3
Before concluding that a perinatal factor plays an etiologic role in cerebral palsy, there must be evidence from well-designed studies showing that an excess risk of cerebral palsy exists when the factor is considered independently. Nelson and Ellenberg's data show, for example that infants with nuchal cord have no higher risk of cerebral palsy than do infants without this complication, and infants who have passed meconium in labor have only a small increase in risk, which is limited to those infants who also have depressed Apgar scores.3 On the other hand, it was found that infants delivered by the breech have a considerably higher risk of cerebral palsy than do infants delivered vertex, even when the lower birthweight of breech deliveries is taken into account. Yet it is unlikely that in all cases of breech delivery, any neurologic abnormality that appears is attributable to a stressful delivery. Fetuses with neurologic impairment existing prior to delivery (eg, congenital myotonia) are more likely to be in the breech position.4
DEFINITION OF CEREBRAL PALSY
Since vascular disease or trauma can cause a spastic hemiplegia at any age, inherent in the definition of cerebral palsy is the notion of congenital or almost congenital onset. Onset here refers to the timing of the insult, not the diagnosis of the illness, as the final neurologic picture does not generally appear until 1 or 2 years of life, reflecting the developmental changes in the infantile nervous system. In this article follow Kurland's suggestion5 that the insult must have occurred between conception and one month of life. In most series of cerebral palsy cases about 10% are noted to have been acquired postnatally from causes such as head trauma, bacterial meningitis, lead poisoning, etc. (Table 1). In such cases, the cause is usually apparent, and will not consider such postnatal cerebral palsy in this review. Children with obvious congenital malformations (eg, neural tube defects) are excluded from the cerebral palsy rubric, although it is likely that in some cases of cerebral palsy an underlying undetected brain malformation is present.
PREVALENCE-RATES OF CEREBRAL PALSY (PER 1,000) FOUND IN POPULATION STUDIES
FREQUENCY OF CEREBRAL PALSY
Cerebral palsy is not common, but among handicapping conditions it is prominent. The prevalence among children at school entry is consistently about two per thousand live births with little variation across developed countries (Table I).6 Perhaps surprisingly, there is no firm evidence that this figure is on the decline. Although both Hagberg et al in Sweden7 and Stanley et al in Australia8 documented declines in cerebral palsy in well-defined populations between the 1950s and 1960s, both authors have recently reported rises in the prevalence of spastic diplegia.9,10 These authors have also provided evidence that the improved survival of very low birthweight infants may be contributing to this recent increase.
The only recent United States data on the subject are from Rochester, Minnesota and are based on small numbers.11 In that study, Kudrjavcev found a decline in cerebral palsy prevalence between the 1950s and 1960s and a leveling off since then.
MAJOR RISK FACTORS FOR CEREBRAL PALSY
The two major risk factors for cerebral palsy are low birthweight and birth asphyxia. Data from the Collaborative Perinatal Project indicate that the risk of cerebral palsy is increased at least 20-fold in infants weighing 1,500 g or less,12 and in normal weight infants the risk of cerebral palsy for the most severe category of clinical asphyxia (Apgar score at 20 minutes or three or less), is elevated 250-fold. 1 No other perinatal factor comes close to exerting such a degree of influence on the risk of cerebral palsy as do these two.
Figure 1. Birth-prevalence rates for the cerebral palsies by birthweight from three countries (drawn from Hagberg and Hagberg. Alberman, and Stanley; all unpublished data). Reprinted with permission from Stanley F, Alberman E: Birthweight, gestational age and the cerebral palsies. Clin Dev Med 1984; 87:57-68.
Many factors that are statistically associated with cerebral palsy are so only because they are related to either low birthweight or asphyxia. Nelson and Ellenberg have recently shown that virtually all obstetrical risk factors for cerebral palsy in normal weight infants, including meconium passage, toxemia, and middle and high forceps use, increase the risk of cerebral palsy only when they are also associated with a depressed 5minute Apgar score.3 In the absence of such evidence of birth asphyxia, these factors carry no increased risk.
Although very low birthweight and extreme asphyxia are associated with high risks of cerebral palsy, such conditions are not common, and by no means account for all cases. Fifty-five percent of cerebral palsy arises in infants who had no evidence of birth asphyxia (as indicated by a 1-minute Apgar score of 7 or above), and 67% arises in infants weighing 2,500 g or more.1,12
The risk of cerebral palsy rises steadily as birthweight declines (Figure 1). In the Collaborative Perinatal Project, the risk of cerebral palsy was found to be 3.4 per thousand in infants 2,500 g and over, 13.9 per thousand in infants weighing 1 , 501 to 2 , 500 g and 90.4 per thousand in infants weighing less than 1,500 g.12
Cerebral palsy in low birthweight infants most typically manifests as spastic diplegia and is often associated with normal intelligence or mild mental retardation. Although an excess of virtually all forms of cerebral palsy is found in the low birthweight infant, the spastic syndromes (including hemiplegias and quadriplegias) are more overrepresented than are the dyskinetic and ataxic syndromes.13 Pre-term delivery and impaired fetal growth each contribute to cerebral palsy. In both the Collaborative Project and in Hagbergs large series of cases from Sweden13 either preterm delivery or low birthweight was found in about 40% of cerebral palsy and both together in about 25%. Infants with spastic diplegia born at term commonly have a history of impairment of fetal growth.14
Does the fact of pre-term delivery indicate an infant already prenatally damaged, or does the pre-term infant encounter brain-damaging insults during delivery or postnatally? This chicken-or-egg argument has gone on for almost a century and there is evidence to support both sides. The availability of new brain imaging techniques in recent years has led to better understanding of the changes in the brain that occur after birth in very low birthweight infants. It is now clear that perhaps one-half of all very low birthweight infants experience brain hemorrhage postnatally.
Moreover, it is likely, though not yet proven, that there is a strong relationship between the cerebral palsy that arises in pre-term infants, and the special susceptibility of such infants to germinal matrix and intraventricular bleeding and associated periventricular ischemia. Recent authors have particularly emphasized that the ischemic damage may be more important than the hemorrhage itself. 1S A specific ischemic lesion found commonly at autopsy in very low birthweight infants is periventricular leukomalacia. Volpe has argued convincingly that this is the pathologic lesion that corresponds to spastic diplegia in life16 but other authorities do not agree.17
Recent observations seem then to point to a postnatal etiology of brain damage mediated through an asphyxic/ischemic mechanism in very small or preterm infants. And yet it is often difficult to distinguish, on the basis of the perinatal clinical history, which low birthweight infant will develop cerebral palsy.18
Unlike birthweight, asphyxia is difficult to define with precision. Most recent studies have taken the Apgar score as a reasonable clinical approximation of birth asphyxia. This assumption is not without its critics, and it should be noted that anesthetic drugs, hypovolemia, infection, etc. can all cause low Apgar scores. Nevertheless, the relationship between low Apgar scores and cerebral palsy risk is striking (Figure 2).
The asphyxiated infant with cerebral palsy is likely to have spastic quadriplegia, sometimes with a superimposed dyskinesia. Severe mental retardation is more common than in the pre-term infant with spastic diplegia. One of the few well-established neuropathologic correlations in cerebral palsy is the finding of "status marmoratus" or "marbling" of the thalamus and basal ganglia in children who exhibited choreoathetotic cerebral palsy in life and who had been severely asphyxiated at birth.16
It apparently takes a great deal of birth asphyxia to cause cerebral palsy. As can be seen in Figure 2, even infants with a 5-minute Apgar score of three or less have only a 5% probability of developing cerebral palsy. Moreover, many infants with cerebral palsy had little or no evidence of asphyxia at birth. Seventy-five percent of all cases of cerebral palsy in the Collaborative Perinatal Project had 5-minute Apgar scores of 7 or higher.1
What role does mechanical trauma during birth play in the genesis of cerebral palsy? Until quite recently, Tristram Shandy's father would have been right to be concerned about the "fine and tender texture of the cerebellum" in its passage through the birth canal. In 1922, Capon found that 20% of autopsied neonates in a London hospital had subdural hematomas, indicating that about 2% of all infants died with this lesion.19
However, emphasis on mechanical trauma as an etiologic factor has waned since the case-control studies of Lilienfeld et al in the 1950s showed that indicators of asphyxia at birth were more strongly associated with later cerebral palsy than were indicators of mechanical injury.20 In a recent series of 100 cases of hypoxic-ischemic injury in term infants in St. Louis, the etiology was judged to have been traumatic in 25%, as evidenced by the presence of factors such as precipitous or prolonged labor, transverse arrest, use of high forceps, difficult breech extraction or difficult manual or forceps rotation.16 Even in some of these situations, however, mechanical injury is unlikely to have played the only role.
Figure 2. Relationship of Apgar Score to cerebral palsy in the Collaborative Perinatal Project. (Derived from Reference 1.)
LESS IMPORTANT RISK FACTORS FOR CEREBRAL PALSY
Prenatal Asphyxiai or Ischemic Insults The clinical effects of asphyxia occurring during the birth process or in the neonatal period can generally be observed in the infant. By contrast, prenatal asphyxie events which do not kill the fetus are almost impossible to detect. Yet several lines of evidence suggest that they do play a role in the genesis of some cases of cerebral palsy. The first line of evidence is the finding in many series of infants with cerebral palsy of an overrepresentation of factors such as abnormal uterine bleeding, placental infarction, and other maternal disorders likely to interfere with fetal nutrition-oxygénation. Hagberg has grouped these abnormalities together as "prenatal deprivation of supply" and has found that they are risk factors for cerebral palsy.21 A second line of evidence derives from neuropathologic studies showing that some infants with cerebral palsy have cystic degenerative lesions (porencephaly, hydranencephaly, cystic encephalomalacia) in the absence of obvious perinatal insults. I7 Such lesions are presumed to have arisen from vascular accidents or ischemic events occurring during gestation, but after the development of the major cerebral structures (as the remainder of the brain appears intact). It is not known at present what proportion of cerebral palsy may be attributable to such intrauterine events.
After asphyxia, the most important metabolic cause of cerebral palsy is bilirubin encephalopathy. The special importance of this condition lies in the fact that it produces both a distinct clinical picture and a distinct neuropathological picture. With the advent of Rh immunization, and clinical management of severe hyperbilirubinemia through exchange transfusion and phototherapy, bilirubin levels high enough to cause kernicteric brain damage are rarely encountered. A decline in the prevalence of dyskinetic cerebral palsy attributable to bilirubin encephalopathy has been documented by several investigators.7,8
It is important to remember that kernicterus (and perhaps by inference, choreoathetotic cerebral palsy as well) can occasionally occur in infants with only mild elevations of serum bilirubin, when factors such as low birthweight, acidosis or infection are present.22
Severe hypoglycemia in the neonatal period is occasionally followed by motor handicap, particularly when seizures have occurred.23 However, it is often difficult to establish that the motor impairment is not due to the asphyxia which so commonly accompanies neonatal hypoglycemia. Amino acid and urea cycle disorders are extremely rare causes of cerebral palsy.
Specific genetic syndromes make only a small contribution to overall cerebral palsy prevalence, except perhaps in communities where consanguinity is very common. Gustavson has found that in Sweden familial clustering of cerebral palsy in infants with normal perinatal histories is found only for the ataxic syndromes (congenital ataxia and ataxic diplegia), especially congenital ataxia.24 Among 12 cases of congenital ataxia found in Gustavson's large survey, ten exhibited familial clustering. The inheritance is usually autosomal recessive, the children are generally retarded and chromosome and metabolic studies are usually normal. This genetic form of ataxic cerebral palsy can be distinguished from the hereditary degenerative ataxias by its non-progressive nature.
Methylmercury ingestion prenatally has produced spastic quadriplegia and mental retardation. In Minamata, Japan, the exposure derived from eating fish that had ingested industrial waste.25 In an Iraqi outbreak, the source was fungicide-contaminated bread.26
Postnatal lead intoxication can certainly produce cerebral palsy, generally manifest as a spastic hemiplegia, but other spastic and even dyskinetic syndromes have been reported.27 To date, however, no cases of cerebral palsy attributable to prenatally acquired lead exposure have been described.
In a report from Sweden,28 one case of hemiplegia and three of congenital ataxia were found among 48 children born to 15 chronic alcoholic women, raising the possibility that prenatal alcohol exposure may on occasion cause cerebral palsy.
An important cause of spastic diplegia in several parts of the world is iodine deficiency. The full-blown condition, referred to as neurological cretinism, seen only where endemic goiter is very prevalent, also includes sensorineural deafness and mental retardation. This disorder is the most preventable form of cerebral palsy as it can be abolished via a program of preconceptional iodized oil administration.29 Although vitamin deficiency has been suggested as a cause of neural tube defects, there is no evidence for its role, or that of any other micronutrient other than iodine, in the genesis of cerebral palsy.
Cases of bacterial meningitis or viral encephalitis (eg, herpes) occurring in the early neonatal period can cause motor deficits in survivors which will generally be classed as congenital cerebral palsy.
APPROXIMATE DISTRIBUTION OF CAUSES IN CEREBRAL PALSY
Congenital rubella is on occasion associated with cerebral palsy. Among 1,624 cases of cerebral palsy in Western Australia and Sweden reviewed by Stanley,30 seven were thought attributable to this viral infection. Data reviewed by Stanley further suggested that about one in 10,000 births will have cerebral palsy attributable to congenital cytomegalovirus infection. Given a cerebral palsy prevalence of two per thousand live births, congenital CMV infection may account for as much as 5% of cerebral palsy. Congenital toxoplasmosis may cause cerebral palsy, particularly when hydrocephalus is present.
Although, as noted earlier, infants with clear-cut congenital anomalies are not usually classified under the cerebral palsy rubric even if they have non-progressive motor deficits, careful study will sometimes uncover developmental brain abnormalities in children with cerebral palsy. Among 681 cases studied by Hagberg, 17 were diagnosed as having a cerebral malformation such as absence of the corpus callosum, aqueductal stenosis, cerebellar hypoplasia, etc.13 Clinicians should be alert to dysmorphic features, abnormal dermatoglyphic patterns and other stigmata of congenital maldevelopment in children with cerebral palsy.
A variety of insults can cause cerebral palsy, but the dominant mechanism of damage is ischemic and/or asphyxiai. Table 2 provides a rough estimate of the relative contribution of each of the several risk factor groups to the total burden of cerebral palsy. This table is only approximate both because of our lack of knowledge, and because risk factors often interact with one another. Cerebral palsy is frequently multifactorial in nature. For example, the small-for-gestational age infant is both more likely to experience labor asphyxia, and is also more susceptible to its effects.
The numerically largest etiologic grouping in cerebral palsy consists of pre-term/low birthweight infants, many of whom have experienced ischemic damage perinatally. The second largest grouping is infants born at term experiencing severe perinatal asphyxia. Congenital infections, and metabolic conditions such as hyperbilirubinemia certainly play some role in the genesis of cerebral palsy but genetic conditions as such rarely cause cerebral palsy. Some infants, if carefully studied, will prove to have a congenital brain malformation. The role of intrauterine ischemic events is at present not well understood, but is probably significant.
1. Nelson KB. Ellenberg J: Apgar scores as predictors of chronic neurologic disability. Pediatrics 1981;68:36-44.
2. Stanley F. Albcrman E: The epidemiology of the cerebral palsies. CIm Dev Med 1984; Volume 87.
3. Nelson K, Ellenberg JL: Obstetric complications as risk factors for cerebral palsy or seizure disorders. 7AMA 1984; 2S t : 1843- 1848.
4. Braun FHT. Jones KL, Smith DW: Breech presentation as an indicator of fetal abnormality. J Pediair 1975; 86:419-421.
5. Kurland LT: Definitions of cerebral palsy and their role in epidemiologic research. Neurology 1957; 7:641-654.
6. Paneth N. Kiely JL: The frequency of cerebral palsy: A review of population -rujies from industrialized nations since 1950. Clin Dev Med 1984; 87:46-56.
7. Hagberg B, Hagberg O. Olow 1: The changing panorama of cerebral palsy in Sweden. 1954-1970. I. Analysis of the general changes. Acta Paediair Scani 1975;64:187-192.
8. Stanley FJ: An epidemiologic study of cerebral palsy in Western Australia 1956-1975. L Changes in total incidence of cerebral palsy and associated factors. Dev Med ChIi Neurol 1979;21:710-713.
9. Hagberg B. Hagberg O. Olow I: The changing panorama of cerebral palsy in Sweden. IV. Epidemiological trends 1959-1978. Acuì Paeduur Scarni 1984; 73:433-440.
10. Atkinson S, Stanley FJ: Spastic diplegia among childten of low and normal birthweight. Dn Med Child Neurol 1983; 25:693-708.
11. Kudrjavcev T. Schoenberg BS. Kurland LT et al: Cerebral palsy - (rends in incidence and changes in concurrent neonatal mortality. Rochester. Minnesota, 1950-1976. Neurology 1984; 33:1433.
12. Ellenberg J, Nelson KB: Birthweight and gestational age in children with cerebral palsy or seizure disorders. AmJ Des Child 1979; 133:1044-1048.
13. Hagberg B. HagbergG: Prenatal and perinatal risk factors in a survey of 681 Swedish cases. Clin Dev Med 1984; 57: 116-1 34.
14. Stanley F. Alberman E: Birthweight, gestational age and the cerebral patsies. CIm Dev Med 1984; 87:57-68.
15. Volpe JJ. Herscovitch P. Perlman JM, et al: Positron emission tomography in the newborn: Extensive impairment of regional cerebral blood flow with intraventricular hemorrhage and hemorrhagic intracerebral involvement. Pediatrics 1983; 72:589-601.
16. Volpe JJ: Neurology <>/' the Newborn. Philadelphia. W. B. Saunders, 1981. ? 204, 215, 217.
17. GilliesF: Neuropathologic indicators of abnormal development, in Fteeman JH (ed): Prenatal and Perinatal Factors Associated with Brain Disorders USDHEW. NlH publication #85-1149, Bethesda, 1985, ? 84.
18. Bennett FC. Chandler LS. Robinson NM, et al: Spastic diplegia in premature infants. Etiologic and diagnostic considerations. Am] Dis Child 1981; 135:732-737.
19. Capon NB: Intracranial traumara in the newborn. )nl Oh G-vn Bnt Empire 1922; 29:572-590.
20. Liltetifcld AM. Pasamanick B-. The association of maternal and fetal factors with the development of cerebral palsy and epilepsy Am ./ ( ihstct ( ivnecol 1955; 70:93-101.
21. Hagberg G, Hagberg B. Olow I: The changing panorama ot cerebral palsy in Sweden 1954-1970. III. The importance of fetal deprivation of supply. Acta Paediair Scand 1976; 65:403-408.
22. Pearlman MA, Gartner LM, Lee KS. et al: The association of kernicterus with bacterial infecrion in the newborn. Pediatrics 1980; 65:26-29.
23. Koivisto M. Blanco-Sigucros M, Krause O: Neonatal symptomatic and asymptomatic hypoglycemia. A follow-up study. Dev Med Child Neurol |972; 14:603-614.
24. Gustavson KH. Hagberg B, Samver C: Identical syndromes of cerebral palsy in the same family. Acta Paediair Scand 1969; 58:330-340.
25. Murakami U: Organic mercury problems affecting intrauterine life, in Klingberg MA, Abramovic A. Clark J (eds): Drugs and Fetal Deielopment Proceedings of an International Symposium on the Effects ot Prolonged Drug Usage on Fetal Development. NewYork: Plenum Press. 1972.
26. Amin-Zaki L, Majeed MA, Elhassani SB. et al: Prenatal methvl-mercury poisoning. Am ) Dis Child 1979; 133:172-177.
27. Perlstern MA. Attola R: Neurologic sequelae of plumbism in children. Clin Pediatr 1966; 5:292-297.
28. Olegard R. Sabel KG. Aronsson M. et al: Effects on the child of alcohol abuse during pregnancy: Retrospective and prospective studies. Acta Paeduur Scand 1979, 275(suppl):ll2-121.
29. Pharoah POD. Buttfield IH. Herzel BS: Neurological damage to the fetus resulting from severe iodine deficiency during pregnancy Loncel 1971; 1:308-310.
30. Stanley F: Prenatal risk factors in the -tudv of the- cerebral palsies. CIm I Vv Med 1984; 87:87-97.
PREVALENCE-RATES OF CEREBRAL PALSY (PER 1,000) FOUND IN POPULATION STUDIES
APPROXIMATE DISTRIBUTION OF CAUSES IN CEREBRAL PALSY