Half of the most severe cases of cerebral palsy (CP) survive to adulthood, but because this longevity is relatively recent, there is no empirical experience of their life expectancy past middle age. The last 2 decades have seen significant developments in the management of persons with CP, involving specialist services from an increasing number of disciplines that require coordination to maximize their effectiveness. This article provides an overview of CP. The author discusses definitions of CP, its epidemiology, pathologies, and range of possible clinical descriptions, and briefly touches on management and prevention.
This article provides an overview of cerebral palsy (CP). The author discusses definitions of CP, its epidemiology, pathologies, and range of possible clinical descriptions, and briefly touches on management and prevention.
CP can no longer be considered a disease of children. For the last 50 years the routine use of antibiotics has protected even the most severely impaired from the previously inevitable early death from pneumonia. Now half of the most severely impaired survive to adulthood, but because this longevity is relatively recent, there is no empirical experience of their life expectancy past middle age. The last 2 decades have seen significant developments in the management of persons with CP, involving specialist services from an increasing number of disciplines that require coordination to maximize their effectiveness. Because the role of coordinator seems increasingly to be falling to the general practitioner, a sound understanding of the concept of CP is becoming mandatory.
What is cerebral palsy?
Many publications attempt to define CP (see for example Refs. ). Definition is defined as a precise statement of the essential nature of a thing and the clear determination of the limits of anything . A definition should therefore describe what a thing is and what it is not, precisely and clearly. No publication has yet achieved this, but there is agreement that CP is an “umbrella term” covering a wide variety of clinical conditions that meet 4 criteria:
- •
Presence of a disorder of movement or posture
- •
Secondary to a cerebral abnormality
- •
Arising early in development
- •
By the time movement impairment exists, the cerebral abnormality is static.
There is no test, genetic, metabolic, immunologic, or otherwise, that demonstrates the existence or absence of CP because there is no specified cause, cerebral pathology, or even type of motor impairment—only that motor impairment exists resulting from nonprogressive cerebral pathology acquired early in life. It is not a single disease. Even as a clinical description these criteria fail in several aspects to achieve the precision required of a definition, such as specifying the age at which development is no longer considered “early.” There is no agreement on this age, but most surveillance systems distinguish cases in which motor impairment is obviously acquired postneonatally, typically following cerebral infection or head trauma. Because it is difficult to definitively differentiate between pre- and neonatally acquired brain damage, all those not postneonatally acquired are usually considered together.
The 4 criteria cannot be addressed until (a) motor development can be clearly recognized as being normal or disordered, and (b) the possibility of progressive cerebral disease can be excluded. Signs suggesting disordered motor control may be recognized very early in life, but accurate prediction has only been confirmed by trained observers in the small proportion of persons with CP born very preterm. Acquisition of the cerebral abnormality may precede recognition of the motor disorder by many months or even years. However, brain-impaired infants, particularly the most severely impaired, are at increased risk of dying before reaching an age at which the criteria for CP can be confirmed. Early death is a competing outcome. On the other hand, it is difficult to definitively exclude the possibility of progression or resolution at any age. Even if cerebral pathology is static, motor abilities change in all children over time, even if that development is grossly abnormal, making functional change an unreliable marker for progressive cerebral pathology. Conversely, a proportion of children described as CP at an early age catch up with their normally developing peers at a later age and the CP label is withdrawn. Therefore, the choice of an age that must be attained before being counted as CP, as well as the age beyond which development is no longer early , is arbitrary and depends on the interest in using the CP label. Treating clinicians are more flexible in applying the CP label, because their primary concern is to balance the psychological effects of labeling a child as having CP with the therapeutic opportunities that the label can afford. This balance can change with time. For example, the increasing frequency of children labeled as having CP of minimal severity in Western Australia is attributed to approval of botulinum toxin therapy for the release of hypertonia in lower limbs, but only for those labeled as CP. Before the availability of this therapy, there was little advantage for a minimally impaired toe walker to be labeled as CP.
By contrast, those responsible for population-based CP registers or surveillance systems need to know exactly whom to count. The compilers adhere strictly to self-imposed limits chosen to facilitate reliability over time and between observers contributing to their database. However, different registers face different problems. Registers with a long life span require primarily a constant definition over time, and this was the guiding principle of the recommendation by Badawi and colleagues that conditions historically excluded from CP (not “diagnosed” as CP on account of having another diagnosis) continue to be excluded, even if meeting the criteria for CP. By contrast, reliability between current observers is the guiding principle of the more recent multicenter surveillance system in Europe, which adopted a flow chart driven by dichotomous responses. The reality of barriers to achieving interobserver agreement of classification is demonstrated by the relatively poor agreement achieved with this flow chart, even when initial observations were standardized by presenting classifiers with written descriptions.
Epidemiology
The reported population-based prevalences of CP depend on the definition, ascertainment proportion, rates of early mortality, and choice of denominator as well as the frequency of underlying brain abnormality. Comparisons of rates cannot be assumed to reflect relative rates of brain abnormality without careful scrutiny of the methods of estimation. However, trends are not similarly dependent. A valid trend constitutes a series of frequencies (each estimated with the same methods) that vary systematically with another variable (typically calendar time). Thus 2 valid trends may be meaningfully compared, even if generated by different methods. By contrast, comparing one CP frequency with another generated by different methods may inform primarily of the consequences of the variation in methods. A serious challenge to valid comparisons is the increasing demand for registers to obtain consent of the potential registrant/carers before registration is permitted, leading to unquantifiable and undoubtedly biased underascertainment that may change with time, undermining the value of such registers. Passive consent (opt-out systems) would significantly reduce, and statutory notification status for CP eliminate, such underascertainment, but are rarely used.
Population-based prevalences of CP have been reported from several areas in the developed world with adequate population-based reporting systems of birth, death, and impairment. Recently published rates from geographically defined populations ( Table 1 ) show significant differences, due primarily to variations in methods. Variations within a reporting system over time tend to be small (see Table 1 ) without any consistent change over the last 50 years being reported by the longest standing registers ( Fig. 1 ).
| Geographic Area | Birth Cohorts | No. of Cases | Rate per 1000 |
|---|---|---|---|
| Norway | 1996–1998 | 374 | 2.1 |
| Western Sweden | 1995–1998 | 170 | 1.9 |
| BC, Canada | 1991–1995 | 2.7 | |
| UK, 4 counties | 1984–2002 | 1301 | 2.0 |
| 1984–1988 | 2.5 | ||
| 1999–2001 | 1.2 | ||
| South Australia | 1993–2000 | 344 | 2.3 b |
| 251 | 1.6 c | ||
| Victoria, Australia | 1970–1998 | 2950 | 1.61 |
| 1970–1972 | 1.4 | ||
| 1996–1998 | 1.4 | ||
| Western Australia | 1960–1999 | 2278 | 2.6 |
| 1960–1964 | 222 | 2.6 | |
| 1995–1999 | 352 | 2.8 | |
| USA, 3 areas | 2002 | 416 | 3.6 |
a Includes postneonatally acquired.
b All ascertained cases (maximum).
With CP prevalences much less than 1%, trends are susceptible to the statistical uncertainty associated with small numbers, but several trends are reported consistently. Males are at higher risk of CP, perhaps because of gender-specific neuronal vulnerabilities recently identified. The proportion of children described as CP increases with decreasing gestational age at birth. The advent of mechanical ventilation to neonatal intensive care has allowed survival of increasingly preterm births, creating a new source of high-risk neonates, and perhaps a new cause of brain damage. In most locations, gestation-specific prevalences of CP increase as each new gestational survival boundary is crossed, and then decline as gestationally appropriate neonatal management techniques are refined, but remain severalfold higher than rates observed in term and near-term births ( Fig. 2 ). Rates around 10% of survivors, the iatrogenic nature of very preterm survival, and the investment in neonatal intensive care that such survival requires has encouraged much attention to be devoted to CP in infants born before 32 weeks’ gestation. However, because births before 32 weeks contribute less than 2% of neonatal survivors, they contribute a minority, approximately 20% to 25%, of all CP in developed countries (see for example Refs. ). Most CP cases are born at term.
The risks of CP increase fourfold in twins and 18-fold in triplets, Increasing maternal age and use of assisted reproduction technologies (ART) has increased the proportion of all births that are multiple. Concomitantly, their contribution to CP has risen from 4% in the 1960s to 10% in the 1990s. The increasing proportion of the population of triplets and higher multiples were attributable exclusively to ART. This increase has been halted in many developed countries, where the number of transferred embryos in any one cycle has been statutorily limited. The exception is the United States, where the transfer of multiple embryos is encouraged by market forces, despite awareness of the attendant risks ; risks that selective fetal reduction may not reverse once pregnancy is achieved. The lower birth weights and shorter gestations associated with multiple birth contribute significantly to their higher risk of CP, but cannot be the only relevant factors because gestation-specific rates are higher for multiples than for singletons born at term or extremely preterm. A biologically plausible mechanism of cerebral damage, specific to monochorionic multiples with placental anastomoses, has been proposed to account for the increased risk of CP in a multiple conception in which one fetus dies antenatally. This mechanism requires monochorionic multiples, which must necessarily be of like gender, as is frequently but not always observed. It may be that CP following cofetal death in multichorionic multiple pregnancies is the result of a single insult that causes both the death of one twin and independently causes brain damage in the other.
A great number of additional potentially etiologic risk factors have been associated with CP in some populations including, but not limited to: parental consanguinity ; social disadvantage ; maternal thrombophilia ; prior reproductive loss ; maternal thyroid problems : pregnancy conditions including severe antepartum hemorrhage, preeclampsia, cytomegalovirus infection ; and other infections, particularly genital tract infections and infections of the fetal membranes in both very preterm infants and infants born later. When reviewing this literature, it must be remembered that any factor causing a very preterm birth lies on a potential causal path to CP. Many etiologic studies control or stratify for gestation of delivery or limit their samples to very preterm births, thus comparing only neonates of the same gestational age and masking the causal potential of any factor that reduces gestational duration. In the fetus, CP has been associated with intrauterine growth restriction, inherited thrombophilias, transient neonatal hypothyroxinemia, and congenital anomalies not only of the brain and head, eyes, and face, but also with noncerebral anomalies (in the apparent absence of cerebral anomalies), particularly of the heart and limbs and skeleton. The risk of CP also increases with the number of suboptimal factors affecting a pregnancy.
The heterogeneity of diseases grouped under the CP label is central to the following discussion.
Epidemiology
The reported population-based prevalences of CP depend on the definition, ascertainment proportion, rates of early mortality, and choice of denominator as well as the frequency of underlying brain abnormality. Comparisons of rates cannot be assumed to reflect relative rates of brain abnormality without careful scrutiny of the methods of estimation. However, trends are not similarly dependent. A valid trend constitutes a series of frequencies (each estimated with the same methods) that vary systematically with another variable (typically calendar time). Thus 2 valid trends may be meaningfully compared, even if generated by different methods. By contrast, comparing one CP frequency with another generated by different methods may inform primarily of the consequences of the variation in methods. A serious challenge to valid comparisons is the increasing demand for registers to obtain consent of the potential registrant/carers before registration is permitted, leading to unquantifiable and undoubtedly biased underascertainment that may change with time, undermining the value of such registers. Passive consent (opt-out systems) would significantly reduce, and statutory notification status for CP eliminate, such underascertainment, but are rarely used.
Population-based prevalences of CP have been reported from several areas in the developed world with adequate population-based reporting systems of birth, death, and impairment. Recently published rates from geographically defined populations ( Table 1 ) show significant differences, due primarily to variations in methods. Variations within a reporting system over time tend to be small (see Table 1 ) without any consistent change over the last 50 years being reported by the longest standing registers ( Fig. 1 ).
| Geographic Area | Birth Cohorts | No. of Cases | Rate per 1000 |
|---|---|---|---|
| Norway | 1996–1998 | 374 | 2.1 |
| Western Sweden | 1995–1998 | 170 | 1.9 |
| BC, Canada | 1991–1995 | 2.7 | |
| UK, 4 counties | 1984–2002 | 1301 | 2.0 |
| 1984–1988 | 2.5 | ||
| 1999–2001 | 1.2 | ||
| South Australia | 1993–2000 | 344 | 2.3 b |
| 251 | 1.6 c | ||
| Victoria, Australia | 1970–1998 | 2950 | 1.61 |
| 1970–1972 | 1.4 | ||
| 1996–1998 | 1.4 | ||
| Western Australia | 1960–1999 | 2278 | 2.6 |
| 1960–1964 | 222 | 2.6 | |
| 1995–1999 | 352 | 2.8 | |
| USA, 3 areas | 2002 | 416 | 3.6 |
a Includes postneonatally acquired.
b All ascertained cases (maximum).
With CP prevalences much less than 1%, trends are susceptible to the statistical uncertainty associated with small numbers, but several trends are reported consistently. Males are at higher risk of CP, perhaps because of gender-specific neuronal vulnerabilities recently identified. The proportion of children described as CP increases with decreasing gestational age at birth. The advent of mechanical ventilation to neonatal intensive care has allowed survival of increasingly preterm births, creating a new source of high-risk neonates, and perhaps a new cause of brain damage. In most locations, gestation-specific prevalences of CP increase as each new gestational survival boundary is crossed, and then decline as gestationally appropriate neonatal management techniques are refined, but remain severalfold higher than rates observed in term and near-term births ( Fig. 2 ). Rates around 10% of survivors, the iatrogenic nature of very preterm survival, and the investment in neonatal intensive care that such survival requires has encouraged much attention to be devoted to CP in infants born before 32 weeks’ gestation. However, because births before 32 weeks contribute less than 2% of neonatal survivors, they contribute a minority, approximately 20% to 25%, of all CP in developed countries (see for example Refs. ). Most CP cases are born at term.
The risks of CP increase fourfold in twins and 18-fold in triplets, Increasing maternal age and use of assisted reproduction technologies (ART) has increased the proportion of all births that are multiple. Concomitantly, their contribution to CP has risen from 4% in the 1960s to 10% in the 1990s. The increasing proportion of the population of triplets and higher multiples were attributable exclusively to ART. This increase has been halted in many developed countries, where the number of transferred embryos in any one cycle has been statutorily limited. The exception is the United States, where the transfer of multiple embryos is encouraged by market forces, despite awareness of the attendant risks ; risks that selective fetal reduction may not reverse once pregnancy is achieved. The lower birth weights and shorter gestations associated with multiple birth contribute significantly to their higher risk of CP, but cannot be the only relevant factors because gestation-specific rates are higher for multiples than for singletons born at term or extremely preterm. A biologically plausible mechanism of cerebral damage, specific to monochorionic multiples with placental anastomoses, has been proposed to account for the increased risk of CP in a multiple conception in which one fetus dies antenatally. This mechanism requires monochorionic multiples, which must necessarily be of like gender, as is frequently but not always observed. It may be that CP following cofetal death in multichorionic multiple pregnancies is the result of a single insult that causes both the death of one twin and independently causes brain damage in the other.
A great number of additional potentially etiologic risk factors have been associated with CP in some populations including, but not limited to: parental consanguinity ; social disadvantage ; maternal thrombophilia ; prior reproductive loss ; maternal thyroid problems : pregnancy conditions including severe antepartum hemorrhage, preeclampsia, cytomegalovirus infection ; and other infections, particularly genital tract infections and infections of the fetal membranes in both very preterm infants and infants born later. When reviewing this literature, it must be remembered that any factor causing a very preterm birth lies on a potential causal path to CP. Many etiologic studies control or stratify for gestation of delivery or limit their samples to very preterm births, thus comparing only neonates of the same gestational age and masking the causal potential of any factor that reduces gestational duration. In the fetus, CP has been associated with intrauterine growth restriction, inherited thrombophilias, transient neonatal hypothyroxinemia, and congenital anomalies not only of the brain and head, eyes, and face, but also with noncerebral anomalies (in the apparent absence of cerebral anomalies), particularly of the heart and limbs and skeleton. The risk of CP also increases with the number of suboptimal factors affecting a pregnancy.
The heterogeneity of diseases grouped under the CP label is central to the following discussion.
Diagnosis/clinical description
Diagnosis is defined as “the art of distinguishing one disease from another, or determining the nature of a case of disease.” The CP label groups several diseases, rather than distinguishing them, and reveals little about the nature of either the cause or prognosis of the conditions to which it is attached. It is immaterial to the acquisition of the label how the cerebral abnormality arose and, if the cerebral abnormality should progress or resolve, it is the label of CP that is withdrawn—it is static by definition, not by empirical observation. To call the CP label a diagnosis stretches the definition of diagnosis and can be misleading. As might be anticipated from its definition, diagnoses are frequently considered mutually exclusive, but the CP label is compatible with many etiologic diagnoses. Despite the uncertainty about the causal pathways to many cases of CP, several uncontroversial causes have been identified including kernicterus following maternal Rhesus isoimmunization (now rare in developed countries but still a significant cause in less developed countries), maternal methyl mercury exposure (Minamata disease), maternal iodine deficiency (cretinism), acute hypoxia following a sentinel intrapartum event, and Moyamoya disease. These and many other diagnoses will persist whether or not the individual meets criteria for the CP label. Such diagnoses are informative about etiology, pathology, and frequently about final clinical presentation. These diagnoses tend to be mutually exclusive, whereas for each person categorized as CP an additional etiologic diagnosis must exist, whether or not it has been made. The author therefore considers that the CP label refers to clinical description rather than a diagnosis.
The range of clinical descriptions covered by the CP label is very wide in terms of the type, severity, and bodily distribution of primary motor impairment, of associated nonmotor neurologic and behavioral impairments, functional deficits, and cerebral pathology. It is seldom productive to consider the group as a whole. Subcategorization is sometimes referred to as differential diagnosis, though the term differential description is more accurate. However, agreement in categorizing people with CP has proved extremely difficult.
Whatever the responsible insult, the extent and severity of brain impairment can be variable and although specific insults can target specific areas of the brain, the area targeted frequently varies with the gestational age at which the insult occurs. Although some specific syndromes exist, the impairments associated with CP may more generally be described as a series of continua of impairment in many dimensions, rather than a discrete set of syndromes.
To be useful, categorization systems must be reliable (people agree which category each person belongs to). Reliability requires that a person fit into one and only one category; this is only possible if categories are defined by a limited number of criteria and all empirically possible combinations of criteria are represented within the categorization system. In conditions that comprise discrete syndromes, clinical features may be sharply defined and the number of possible combinations of clinical features limited. Such is not the case with CP, so attempting to reliably categorize individuals with CP on the basis of the sum of their clinical features is doomed to failure. However, there are several possible reasons for categorization and each tends to focus on a different clinical feature, or limited set of features. It is therefore likely to be both more attainable and more useful to aim for reliable categorization systems for each of the clinical features associated with CP than for a reliable categorization system for individuals with CP.
A little progress has been made toward this goal. The shining example is the Gross Motor Function Classification System (GMFCS) devised for children with CP. The GMFCS classifies only gross motor function, primarily ambulatory ability and its precursors, into 5 possible categories. It is well documented with age-specific criteria, recently augmented by illustrations ( Fig. 3 ). The GMFCS has enormous success firstly because the degree of assistance required with ambulation is very important and useful for service provision, and secondly, because the limited aim allows reliable, tessellated categories defined by a single factor. It has been followed by a similar system for upper limb function, the Manual Abilities Classification System (MACS), and similar systems for communication ability are currently being developed or tested for reliability.
Related posts:
Examination of the Child with Cerebral Palsy
Management of Hip Deformities in Cerebral Palsy
The Role of Gait Analysis in Treating Gait Abnormalities in Cerebral Palsy
The Adult with Cerebral Palsy
Assessment and Treatment of Movement Disorders in Children with Cerebral Palsy
Management of Hip Deformities in Cerebral Palsy
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
