This article reviews the historical background, classification, and etiology of cerebral palsy (CP), the most common motor disability of childhood. The various methods employed to measure the prevalence of CP in the population are examined. Causes of CP are numerous, and the etiology multi-factorial. Risk factors are categorized by the timing of their proposed occurrence: prenatal, perinatal, and postnatal. The leading prenatal and perinatal risk factors for CP are birth weight and gestational age. Other risk factors include neonatal encephalopathy, multiple pregnancy, infection and inflammation, and a variety of genetic factors.
History and definition
Cerebral palsy (CP) is the most common motor disability of childhood. A recent publication from the Autism and Developmental Disability Monitoring (ADDM) CP Network sponsored by the Centers for Disease Control and Prevention (CDC) reported a prevalence of 3.3 per 1000 8-year-old children from 3 sites across the United States. The history of cerebral palsy is a long one, dating back to ancient Egypt. There are at least 2 drawings of individuals from the fifth century BC with what is recognized today as spastic cerebral palsy. An orthopedic surgeon, William John Little, who himself had an equinus deformity from early childhood secondary to poliomyelitis, is credited with the first descriptions of CP in 1843. Seeking a cure for his own deformity, he was greatly influenced by the French orthopedic surgeon, Jacques Delpeche, who was interested in surgical correction of equinus deformities, and performed many tenotomies of the Achilles tendon. After successful correction of his own deformity by a German orthopedic surgeon, George Stromeyer, Little improved on Stromeyer’s surgical techniques and set up the Orthopaedic Institution in London. Little’s interest in orthopedic deformities continued and he is regarded as a pioneer in orthopedic surgery and as the first to recognize spastic paralysis. He wrote a treatise “On the influence of abnormal parturition, difficult labor, premature birth and asphyxia neonatorum on the mental and physical condition of the child”, which posited that these deformities of childhood were related to anoxia secondary to trauma occurring during labor and delivery. For many years, spastic diplegia was commonly referred to as Little’s disease.
Sir William Osler, a British physician, is believed to have coined the term “cerebral palsy” in 1889; he described 151 patients affected by the disorder. Sigmund Freud, a neurologist, but best known as a psychoanalyst, wrote many articles on CP, adding to the sparse body of knowledge on the subject. He also disagreed with Little on its cause, observing that children with CP had many other neurologic conditions, such as intellectual disabilities, visual impairment, and epilepsy. He therefore believed that CP might be caused by in utero abnormalities of brain development. He divided CP into 3 groups based on possible causes: (1) maternal and idiopathic congenital; (2) perinatal; and (3) postnatal, and devised a classification scheme with “diplegia” used to refer to all bilateral disorders of central origin.
The American Academy of Cerebral Palsy (AACP) was formed in 1947. Minear polled the membership of the Academy in 1953 and found many different definitions of cerebral palsy. The various definitions commonly acknowledged a broad syndrome of brain damage, with predominant motor dysfunction but also psychological, epileptic, and behavioral symptoms. Transient abnormalities, neoplasms, progressive disorders, and spinal cord disorders were excluded. Despite the presence of common themes, a unified definition of CP was not presented until almost 5 years later by the Little Club, an informal group of neurologists and others formed in the United Kingdom in 1957. The Little Club developed a definition aimed to facilitate sharing knowledge and research: “Cerebral palsy is a persisting qualitative motor disorder due to non-progressive interference with development of the brain occurring before the growth of the central nervous system is complete.” The Little Club classification consisted of: (1) spastic (hemiplegic, double hemiplegic, and diplegic); (2) dystonic; (3) choreoathetoid; (4) mixed; (5) ataxic; and (6) atonic CP. In the 1960s CP was redefined but there continued to be recognition of inconsistencies in terminology.
With growing interest in public health, the Spastics Society commissioned a group to define CP for epidemiologic purposes in the 1980s. A limb-by-limb classification system, which described the functioning of each limb and the head and neck separately, built on the work in Western Australia of Fiona Stanley, was proposed by Evans. This classification system also allowed the capture of information on co-occurring medical conditions such as congenital malformations and seizures. American and European CP investigators met from 1987 to 1990 and developed a common definition: “CP is an umbrella term covering a group of non-progressive, but often changing, motor impairment syndromes secondary to lesions or anomalies of the brain arising in the early stages of development.”
By 1998, there were 14 centers across Europe conducting population-based surveillance for CP; they formed a Network, the Surveillance of Cerebral Palsy in Europe (SCPE). The Network used a case definition that was a reiteration of that of Mutch et al, and developed and published standardized procedures for ascertaining and describing children with CP for registers.
An International Workshop on Definition and Classification of CP was held in Bethesda, Maryland, July 11 to 13, 2004 because of a perceived need to revisit the definition and classification of CP. The current definition, as adopted by this group, recognizes that CP is more than a motor disability and acknowledges that often other impairments accompany CP: “Cerebral palsy describes a group of permanent disorders of the development of movement and posture, causing activity limitation, that are attributed to non-progressive disturbances that occurred in the developing fetal or infant brain. The motor disorders of CP are often accompanied by disturbances of sensation, perception, cognition, communication, behavior, by epilepsy and by secondary musculoskeletal problems.”
The definitions of CP, including the most recent one cited, have 4 core components: (1) it is a disorder of movement and posture; (2) it results from an abnormality in the brain; (3) it is acquired early in life; and (4) the condition is static at the time of recognition. However, there are still many challenges with the use of all CP definitions for epidemiologic purposes because of the lack of specificity of the definition. The criteria do not address severity of the motor disability to be included; how to assure that the brain abnormality is static; age of the acquisition of the brain lesion; or the youngest age of recognition of the condition. Also, there are other conditions that do meet these stated criteria for CP that are not included. Blair and Stanley have proposed that to make the term cerebral palsy more specific, especially for epidemiologic studies, CP researchers should: (1) define the lower limit of severity using a validated measure, such as the Gross Motor Function Classification System (GMFCS); (2) specify an upper age limit for postneonatally acquired cases; (3) develop inclusion and exclusion criteria related to known chromosomal, genetic, and metabolic conditions; (4) define the age of certainty of the diagnosis beyond which one would not expect resolution or change in the diagnosis; and (5) define the minimum age of inclusion of the child in a register or surveillance system should the child die before diagnostic confirmation. Blair and Stanley also state that even if investigators do not agree on the same criteria for studies, a description of the study population according to the 5 areas as suggested would allow for comparison of results from different epidemiologic studies.
Classification
In 1956 Minear and the Nomenclature and Classification Committee of the American Academy for Cerebral Palsy presented a set of potential classification schemes that have remained pertinent over the years. This early classification system included broad clinical symptoms with categories for physiology (the nature of the motor abnormality), topography, etiology, neuroanatomic features, supplemental (associated) conditions, functional capacity (severity), and therapeutic requirements. Experts continue to address these broad categories when classifying CP.
Physiologic and Topographic Classification
CP can be divided into 2 main physiologic groups, the pyramidal (a term used somewhat inexactly to refer to cases in which spasticity is prominent) and the extrapyramidal types (chorea, athetosis, dystonia, ataxia). Spasticity is a clinical sign manifested by an increased resistance of a limb to externally imposed joint movement. The spastic types of cerebral palsy have neuromotor findings that are consistent and persistent; neurologic abnormalities remain during quiet periods and sleep, and do not vary much during the active state or when degrees of emotional stress or irritability are present. In contrast, extrapyramidal types of CP have marked variability in tone during relaxation and sleep, and especially during wakefulness when stressful situations arise. Rapid passive movement at a joint elicits spastic hypertonus. The classic descriptor of spasticity is the “clasp knife” resistance that is followed by a sudden “give.” The comparison is made to the opening or closing of a penknife. Extrapyramidal hypertonicity, in contrast, is represented by increased tone persisting throughout slow passive flexion and extension of an extremity. It is often described as “lead pipe” rigidity. Combinations of these tone patterns in the same patient are common, creating potential difficulty in finding the proper diagnostic terminology. Extrapyramidal CP has 4-limb involvement, with upper extremities typically being functionally more involved than the lower extremities. This situation precludes further useful topographic breakdown. Therefore, for practical purposes topographic classification is restricted to the spastic group.
To discriminate subgroups of spastic CP, classification systems often refer to the localization or topography of the abnormal motor function. Diplegia refers to bilateral lower extremity involvement, hemiplegia to unilateral upper and lower extremity involvement, triplegia to involvement of 3 extremities (typically both lower and one upper extremity), double hemiplegia to 4-extremity involvement with more severe spasticity of the upper extremities, and quadriplegia/tetraplegia to severe 4-extremity involvement.
There are several concerns with the physiologic and topographic schemes. The distinction between the topographic classification terms may lack sufficient reliability. How much upper extremity involvement is required to distinguish diplegia from quadriplegia? How many extrapyramidal signs are required to designate mixed CP? “Lead pipe” rigid tone is not always easily distinguished from spasticity. Alberman compared agreement on classification of CP among 6 trained clinicians and found poor reliability. Agreement on the physiologic classification of the motor dysfunction (eg, spasticity, choreoathetosis) was 40%, on the topographic classification 50%, and on severity (mild, moderate, severe) 60%. In addition to reliability concerns, the topographic and physiologic classifications do not consider functional abilities. Because brain dysfunction has diffuse manifestations in childhood, each child must be evaluated thoroughly for associated impairments in areas such as learning and cognition, vision, behavior, epilepsy, and secondary neuromuscular abnormalities. It is not possible to direct clinical assessments simply based on correlations between topography and associated dysfunctions. Finally, topographic and physiologic classifications do not aid therapy.
Etiologic Classification
Etiologic classification systems are aimed at developing prevention strategies. The association of erythroblastosis fetalis with choreoathetoid cerebral palsy served as the paradigm for this classification. However, etiologic classifications are not well developed and to date have not been successful in addressing prevention.
The Collaborative Perinatal Project helped to identify a large number of conditions that placed a child at risk for cerebral palsy. However, only a few of these conditions were found to correlate to specific motor outcome or diagnosis. Most predictors were combinations of factors present prior to onset of labor, implying that CP is not caused by a single disturbance but by the interaction of many related conditions.
Some research has also focused on discerning the mechanism of the brain damage. Because the brain has a limited number of ways to respond to insult, CP might result from a common pathophysiological mechanism. One hypothesis links inflammatory factors and white matter damage, proposing that asphyxia, maternal infection (such as urinary tract infection), and chorioamnionitis might be related to CP through a common mechanism.
Neuropathologic Classification
In the mid-twentieth century, the idea of neuropathologic classification was proposed in an effort to reflect and highlight the inability to relate brain structure to brain function. The advent of neuroimaging has not yet significantly advanced the ability to classify CP by neuropathology. Ultrasound, magnetic resonance imaging, computed tomography, and volumetric studies have not demonstrated consistent structure or functional relationships. However, as science has learned more about the developing brain, a theory of selective vulnerability has developed. Two important associations have been described: (1) periventricular leukomalacia with prematurity, and (2) basal ganglia injury with term asphyxia. Newer and functional imaging techniques with different discriminatory abilities might contribute significantly to a neuropathologic classification of cerebral palsy in the future.
Supplemental Classification and Associated Conditions
The supplemental classification describes the associated conditions or impairments found in children with CP and attempts to connect them to the physiologic and topographic classifications. The idea is to identify syndromes that have a common etiology and ultimately lead to prevention. Bilirubin encephalopathy is a prototypical example of such a syndrome, and includes choreoathetoid cerebral palsy, vertical gaze palsy, dental enamel dysplasia, and sensorineural hearing loss. It has a predictable clinical course, with extensor spells during the first few months, followed by hypotonia, then choreoathetosis, and finally dystonia during adolescence. Despite a few such examples, the associations between supplemental disorders (associated impairments) and physiology or topography generally have low sensitivity and specificity. Individuals with CP must each be evaluated for an array of associated conditions, including deficits in hearing, vision, cognition, and academic achievement.
Functional and Therapeutic Classifications
Minear and the Nomenclature and Classification Committee originally added functional and therapeutic classifications for cerebral palsy simply to be comprehensive. The functional classification addresses the degree of severity of the condition based on limitation of activity. The therapeutic classification divides cases into 4 categories: nontreatment, modest interventions, need for a cerebral palsy treatment team, and pervasive support.
Much has changed with regard to therapeutic interventions since the 1950s. The number of interventions is significantly greater. Interventions are applied not only to the primary motor dysfunction but also to associated disorders or conditions. Service delivery systems have shifted from clinical or hospital settings to schools and the community. Therefore, older therapeutic classification systems have required adaptation. Capute and colleagues interpreted CP as part of a broader syndrome of brain dysfunction, in turn suggesting that CP be part of a broader spectrum of motor dysfunction. They pointed out that in some cases, the most limiting factor is not the motor impairment, and that the treatment of CP should extend beyond the motor deficit to associated cognitive, communicative, convulsive, or behavioral conditions that affect therapeutic and functional (adaptive) success.
Interest in functional classifications has recently intensified due to a broader understanding of outcome. Newer measures of functional abilities in cerebral palsy have evolved. The World Health Organization International Classification of Functioning, Disability, and Health (ICF) articulates three categories of function: impairment (the capacity to perform), activity limitations (the ability to perform), and participation restrictions (the opportunity to function).
Cerebral Palsy Classification for Epidemiologic Surveillance
Throughout the 1960s and 1970s, issues related to the classification of CP were largely addressed from a clinical perspective. However, in the 1980s, with rising interest in monitoring CP prevalence among populations as public health markers of rapidly changing neonatal care, significant consideration was given to classification of CP from an epidemiologic perspective. Evans’ “limb-by-limb” classification method looked at central motor abnormalities based on neurologic type: hypotonia, hypertonia (including stiffness, spasticity, and rigidity), dyskinesia, and ataxia. This classification method included information on each limb, the head and neck, functional mobility and manual dexterity, as well as associated conditions (intellectual and sensory impairments, communication problems, seizures), neuroanatomy, and etiology (congenital and acquired malformations, genetics). Based on different methods and lack of reliability of subtype classification among centers, the SCPE (the previously mentioned European network of population-based surveys and CP registers) adopted a simple classification of 4 CP subtype groups: unilateral spastic, bilateral spastic, dyskinetic, and ataxic. The SCPE participants developed a classification tree and a reference and training manual in CD format that includes video examples of the different clinical patterns of neurologic signs and motor function impairments. Those useful tools have promoted a standardized way of classifying CP subtypes. Groups in other countries, including the United States (Atlanta), Western Australia, Quebec, Canada, and South-east Australia, have adopted similar classification systems. Data from these groups have shown similar distributions of CP subtypes. However, work continues toward improving reliability of this classification system. Recent advances in neuroscience and technology, as well as increasing knowledge of age-related features, have led to consideration of broader anatomic features, radiologic findings, causative factors, and timing of injury.
International surveillance systems are now using formalized methods to assess function in addition to impairment. The Gross Motor Function Measurement Scale (GMFMS, 88 or 66 items) was developed for clinical use, reduced to a 5-point scale for epidemiologic purposes, The Gross Motor Function Classification System (GMFCS) and extended and revised in 2007. More recently, similar scales for fine motor abilities have been developed: the Manual Ability Classification System (MACS) and the Bimanual Fine Motor Function (BFMF) scales. GMFCS and MACS have been validated and are available online. BFMF takes into account asymmetry and allows data to be extracted from medical records. Comparability of results across monitoring programs is greatly facilitated by the use of these measures. Cans and colleagues compared surveillance data reported by groups in South-east Australia, Norway, Sweden, and France. In the studies reviewed, the proportion of more severely impaired children (level IV/V) on either the GMFCS or BFMF was around 25% to 35% of all CP case children. They found the dyskinetic group to have the highest variability between study sites, which not surprisingly suggests difficulties in classifying mixed types and lower frequencies of dyskinetic CP.
Classification
In 1956 Minear and the Nomenclature and Classification Committee of the American Academy for Cerebral Palsy presented a set of potential classification schemes that have remained pertinent over the years. This early classification system included broad clinical symptoms with categories for physiology (the nature of the motor abnormality), topography, etiology, neuroanatomic features, supplemental (associated) conditions, functional capacity (severity), and therapeutic requirements. Experts continue to address these broad categories when classifying CP.
Physiologic and Topographic Classification
CP can be divided into 2 main physiologic groups, the pyramidal (a term used somewhat inexactly to refer to cases in which spasticity is prominent) and the extrapyramidal types (chorea, athetosis, dystonia, ataxia). Spasticity is a clinical sign manifested by an increased resistance of a limb to externally imposed joint movement. The spastic types of cerebral palsy have neuromotor findings that are consistent and persistent; neurologic abnormalities remain during quiet periods and sleep, and do not vary much during the active state or when degrees of emotional stress or irritability are present. In contrast, extrapyramidal types of CP have marked variability in tone during relaxation and sleep, and especially during wakefulness when stressful situations arise. Rapid passive movement at a joint elicits spastic hypertonus. The classic descriptor of spasticity is the “clasp knife” resistance that is followed by a sudden “give.” The comparison is made to the opening or closing of a penknife. Extrapyramidal hypertonicity, in contrast, is represented by increased tone persisting throughout slow passive flexion and extension of an extremity. It is often described as “lead pipe” rigidity. Combinations of these tone patterns in the same patient are common, creating potential difficulty in finding the proper diagnostic terminology. Extrapyramidal CP has 4-limb involvement, with upper extremities typically being functionally more involved than the lower extremities. This situation precludes further useful topographic breakdown. Therefore, for practical purposes topographic classification is restricted to the spastic group.
To discriminate subgroups of spastic CP, classification systems often refer to the localization or topography of the abnormal motor function. Diplegia refers to bilateral lower extremity involvement, hemiplegia to unilateral upper and lower extremity involvement, triplegia to involvement of 3 extremities (typically both lower and one upper extremity), double hemiplegia to 4-extremity involvement with more severe spasticity of the upper extremities, and quadriplegia/tetraplegia to severe 4-extremity involvement.
There are several concerns with the physiologic and topographic schemes. The distinction between the topographic classification terms may lack sufficient reliability. How much upper extremity involvement is required to distinguish diplegia from quadriplegia? How many extrapyramidal signs are required to designate mixed CP? “Lead pipe” rigid tone is not always easily distinguished from spasticity. Alberman compared agreement on classification of CP among 6 trained clinicians and found poor reliability. Agreement on the physiologic classification of the motor dysfunction (eg, spasticity, choreoathetosis) was 40%, on the topographic classification 50%, and on severity (mild, moderate, severe) 60%. In addition to reliability concerns, the topographic and physiologic classifications do not consider functional abilities. Because brain dysfunction has diffuse manifestations in childhood, each child must be evaluated thoroughly for associated impairments in areas such as learning and cognition, vision, behavior, epilepsy, and secondary neuromuscular abnormalities. It is not possible to direct clinical assessments simply based on correlations between topography and associated dysfunctions. Finally, topographic and physiologic classifications do not aid therapy.
Etiologic Classification
Etiologic classification systems are aimed at developing prevention strategies. The association of erythroblastosis fetalis with choreoathetoid cerebral palsy served as the paradigm for this classification. However, etiologic classifications are not well developed and to date have not been successful in addressing prevention.
The Collaborative Perinatal Project helped to identify a large number of conditions that placed a child at risk for cerebral palsy. However, only a few of these conditions were found to correlate to specific motor outcome or diagnosis. Most predictors were combinations of factors present prior to onset of labor, implying that CP is not caused by a single disturbance but by the interaction of many related conditions.
Some research has also focused on discerning the mechanism of the brain damage. Because the brain has a limited number of ways to respond to insult, CP might result from a common pathophysiological mechanism. One hypothesis links inflammatory factors and white matter damage, proposing that asphyxia, maternal infection (such as urinary tract infection), and chorioamnionitis might be related to CP through a common mechanism.
Neuropathologic Classification
In the mid-twentieth century, the idea of neuropathologic classification was proposed in an effort to reflect and highlight the inability to relate brain structure to brain function. The advent of neuroimaging has not yet significantly advanced the ability to classify CP by neuropathology. Ultrasound, magnetic resonance imaging, computed tomography, and volumetric studies have not demonstrated consistent structure or functional relationships. However, as science has learned more about the developing brain, a theory of selective vulnerability has developed. Two important associations have been described: (1) periventricular leukomalacia with prematurity, and (2) basal ganglia injury with term asphyxia. Newer and functional imaging techniques with different discriminatory abilities might contribute significantly to a neuropathologic classification of cerebral palsy in the future.
Supplemental Classification and Associated Conditions
The supplemental classification describes the associated conditions or impairments found in children with CP and attempts to connect them to the physiologic and topographic classifications. The idea is to identify syndromes that have a common etiology and ultimately lead to prevention. Bilirubin encephalopathy is a prototypical example of such a syndrome, and includes choreoathetoid cerebral palsy, vertical gaze palsy, dental enamel dysplasia, and sensorineural hearing loss. It has a predictable clinical course, with extensor spells during the first few months, followed by hypotonia, then choreoathetosis, and finally dystonia during adolescence. Despite a few such examples, the associations between supplemental disorders (associated impairments) and physiology or topography generally have low sensitivity and specificity. Individuals with CP must each be evaluated for an array of associated conditions, including deficits in hearing, vision, cognition, and academic achievement.
Functional and Therapeutic Classifications
Minear and the Nomenclature and Classification Committee originally added functional and therapeutic classifications for cerebral palsy simply to be comprehensive. The functional classification addresses the degree of severity of the condition based on limitation of activity. The therapeutic classification divides cases into 4 categories: nontreatment, modest interventions, need for a cerebral palsy treatment team, and pervasive support.
Much has changed with regard to therapeutic interventions since the 1950s. The number of interventions is significantly greater. Interventions are applied not only to the primary motor dysfunction but also to associated disorders or conditions. Service delivery systems have shifted from clinical or hospital settings to schools and the community. Therefore, older therapeutic classification systems have required adaptation. Capute and colleagues interpreted CP as part of a broader syndrome of brain dysfunction, in turn suggesting that CP be part of a broader spectrum of motor dysfunction. They pointed out that in some cases, the most limiting factor is not the motor impairment, and that the treatment of CP should extend beyond the motor deficit to associated cognitive, communicative, convulsive, or behavioral conditions that affect therapeutic and functional (adaptive) success.
Interest in functional classifications has recently intensified due to a broader understanding of outcome. Newer measures of functional abilities in cerebral palsy have evolved. The World Health Organization International Classification of Functioning, Disability, and Health (ICF) articulates three categories of function: impairment (the capacity to perform), activity limitations (the ability to perform), and participation restrictions (the opportunity to function).
Cerebral Palsy Classification for Epidemiologic Surveillance
Throughout the 1960s and 1970s, issues related to the classification of CP were largely addressed from a clinical perspective. However, in the 1980s, with rising interest in monitoring CP prevalence among populations as public health markers of rapidly changing neonatal care, significant consideration was given to classification of CP from an epidemiologic perspective. Evans’ “limb-by-limb” classification method looked at central motor abnormalities based on neurologic type: hypotonia, hypertonia (including stiffness, spasticity, and rigidity), dyskinesia, and ataxia. This classification method included information on each limb, the head and neck, functional mobility and manual dexterity, as well as associated conditions (intellectual and sensory impairments, communication problems, seizures), neuroanatomy, and etiology (congenital and acquired malformations, genetics). Based on different methods and lack of reliability of subtype classification among centers, the SCPE (the previously mentioned European network of population-based surveys and CP registers) adopted a simple classification of 4 CP subtype groups: unilateral spastic, bilateral spastic, dyskinetic, and ataxic. The SCPE participants developed a classification tree and a reference and training manual in CD format that includes video examples of the different clinical patterns of neurologic signs and motor function impairments. Those useful tools have promoted a standardized way of classifying CP subtypes. Groups in other countries, including the United States (Atlanta), Western Australia, Quebec, Canada, and South-east Australia, have adopted similar classification systems. Data from these groups have shown similar distributions of CP subtypes. However, work continues toward improving reliability of this classification system. Recent advances in neuroscience and technology, as well as increasing knowledge of age-related features, have led to consideration of broader anatomic features, radiologic findings, causative factors, and timing of injury.
International surveillance systems are now using formalized methods to assess function in addition to impairment. The Gross Motor Function Measurement Scale (GMFMS, 88 or 66 items) was developed for clinical use, reduced to a 5-point scale for epidemiologic purposes, The Gross Motor Function Classification System (GMFCS) and extended and revised in 2007. More recently, similar scales for fine motor abilities have been developed: the Manual Ability Classification System (MACS) and the Bimanual Fine Motor Function (BFMF) scales. GMFCS and MACS have been validated and are available online. BFMF takes into account asymmetry and allows data to be extracted from medical records. Comparability of results across monitoring programs is greatly facilitated by the use of these measures. Cans and colleagues compared surveillance data reported by groups in South-east Australia, Norway, Sweden, and France. In the studies reviewed, the proportion of more severely impaired children (level IV/V) on either the GMFCS or BFMF was around 25% to 35% of all CP case children. They found the dyskinetic group to have the highest variability between study sites, which not surprisingly suggests difficulties in classifying mixed types and lower frequencies of dyskinetic CP.
Methodology
Researchers have employed a variety of methods to measure the frequency of CP in the population. This frequency is measured as prevalence, which is the proportion of the number of individuals with CP among a defined population with CP at a specified period in time. In the United States, there are 5 predominant methods for obtaining prevalence data: (1) notification (reportable disease surveillance); (2) disease registries; (3) periodic population-based surveys; (4) secondary use of administrative data systems; and (5) ongoing, population-based record review. Each data collection mechanism has a different primary purpose, which for most is not estimation of CP prevalence. Therefore, although all systems provide useful information, there are strengths and limitations to each as they pertain to obtaining a complete count of the number of children with CP in a defined community at a specified period in time.
Notification (Reportable Disease Surveillance)
In the United States, all states have laws that require the reporting of selected infectious diseases to the local, district, or state health department. These passive, provider-based reporting systems rely on the receipt of individual case reports from physicians, laboratories, and health care providers, and are simple and nonburdensome. Sometimes developmental disabilities such as CP are also included in such systems. One example is the Georgia Birth Defects Reporting and Information System (GBDRIS), which provides information to the Georgia Department of Human Resources on the incidence, prevalence, trends, and epidemiology of birth defects and related conditions in children from birth to age 6 years. CP is one of the conditions monitored. Because CP is often diagnosed after birth by medical providers in a variety of health care settings, it is not easily captured through a notifiable disease-reporting system such as the GBDRIS, which relies primarily on birth hospitals for case identification.
Disease Registries
Disease registries rely on the voluntary reporting of individuals with specific diseases and are usually based on service provision. Because disease registries are often clinic based, children who do not visit the participating clinics would not be counted in any prevalence estimates produced through analysis of registry data. As a result, disease registries may not be representative of a population.
Periodic Population-based Surveys
Periodic population-based surveys involve the systematic collection of information using a standardized data collection instrument administered as an in-person interview, self-completed questionnaire, or by telephone, or mail. In the United States, The Centers for Disease Control and Prevention’s (CDC), The National Center for Health Statistics (NCHS), administers the National Health Interview Survey (NHIS) which includes a Disability Supplement (1994–1995) and Sample Child File (1997–2006) that provide information related to participants’ experiences with children and disability. Another NCHS population-based survey that provides valuable information related to developmental disabilities is the State and Local Area Integrated Telephone Survey (SLAITS), which includes the National Survey on Children with Special Health Care Needs (2001). These surveys are conducted using a large sample size and as such are believed to be representative of national characteristics. In addition, these surveys are often more timely than other active methods of data collection. The Sample Child File, for example, is produced annually. Nevertheless, administration of population-based surveys can be labor intensive and costly. Moreover, the collection of data through parental or guardian report is subject to recall bias (that is, differences in accuracy or completeness of reporting information on risk factors and behaviors, due to disparities in recall of past events or experiences between individuals with a diagnosis compared to those without such a diagnosis) and selection bias (differences in the characteristics between individuals participating in a study and those who are not). A further limitation of these surveys that may be particularly important for a population affected by developmental disabilities is that no data are collected for individuals who live in a residential treatment facility or institution.
Secondary Use of Administrative Data
Many administrative data systems with individual-level data can be used for the public health surveillance of developmental disabilities. The most common of these include hospital discharge data, health insurance and Medicaid billing data, and managed-care encounter data. Because these systems are not designed for public health surveillance, the accuracy and completeness of diagnostic information may be uncertain. Other administrative data systems rely on the use of existing aggregate rather than individual-level data. These passive surveillance systems examine federal-, state-, and county-level data for individuals receiving education or diagnostic and treatment services. One example in the United States is the Office of Special Education Programs (OSEP) Annual Reports to Congress on the Implementation of the Individuals with Disabilities Education Act (IDEA). This provider-based reporting mechanism relies on receipt of aggregate reports from each school district in the United States. This type of data collection method is simple, timely, and not burdensome. However, the system may underestimate the population prevalence because not all children with disabilities receive special education services through the school system. Prevalence estimates for some disabilities such as intellectual disabilities can be obtained using OSEP data because there are specific special education exceptionalities for these disabilities. However, it is not possible to describe the special education services of children with CP or measure prevalence of CP using OSEP Annual Reports for several reasons: (1) the program area in which significant numbers of children with CP are served (ie, orthopedic impairment) also includes children with other motor disorders; (2) many children with CP receive services under the other health impairment exceptionality, which is a program area for children with other medical conditions as well; (3) those with co-occurring intellectual disability (ID) are most often served through an ID exceptionality.
Ongoing, Population-based Record Review
Ongoing, population-based record review is an active surveillance system whereby information is systematically collected on individual children by standardized data collection instruments through review of existing records at administrative data sources. Programs using this method track the number of children identified with CP using multiple sources in the community that diagnose, treat, or serve children with developmental disabilities. Examples of this type of data collection include the CP surveillance programs in the United States and internationally. In the United States, the ADDM Network, funded by the Centers for Disease Control and Prevention, currently conducts surveillance of CP and other developmental disabilities in 4 communities. SCPE and the Australia Cerebral Palsy Register, which is comprised of numerous registers for CP surveillance across Australia, conduct record reviews and receive notification of CP cases from other reporting sources.
For this type of surveillance, participants do not need to be contacted as a part of data collection, so there is minimal burden on families affected by CP. Objective reliable methods for determining surveillance case definition are established, and extensive training and quality control measures are implemented to ensure adherence to data collection and case determination guidelines and reliable resultant prevalence estimates. Many of the surveillance programs that employ population-based record review do not depend solely on previously documented CP diagnoses to identify children, as descriptions of motor findings consistent with CP are also used to determine case status. Incorporating information from multiple health, education, and service providers rather than relying on only one facility or one type of facility to identify children allows for more complete coverage for case identification in a defined population. Because individual-level data are collected, the identified case series may also be used to address future research questions and may be linked to other databases such as birth certificate files and census data, providing even more information about individuals with CP. The surveillance programs in the United States, Europe, and Australia have been ongoing for many decades, thus affording the ability to examine prevalence estimates of CP in the same population and using the same methods for classifying CP over time. The ADDM Network is strengthened by its heterogeneous population characteristics that enable examination of various racial/ethnic subgroups.
Although this method provides a reasonably complete picture of the population affected by CP, there are some limitations. Because this is an active surveillance method using multiple sources, it is more labor and time intensive and costly to operate than most passive systems. For the ADDM Network, in particular, which relies solely on records, information within the system is dependent on the availability and quality of these records. Some of the records may not contain the necessary information to confirm case status because the system relies on information that has been collected for purposes other than public health surveillance. For the system in the United States, the prevalence of children with mild CP may be underestimated, because these children may not have come to the attention of service providers in early childhood and records of children in regular education, in private schools, or who are being home-schooled are not reviewed. Nevertheless, data from the ADDM Network indicate that these exceptions likely represent a very small proportion of children with CP.
Prevalence
Prevalence is calculated as a proportion, and careful attention is necessary when measuring the numerator and choosing the corresponding population denominator. The international community of epidemiologists, who conduct surveillance of CP, grapples with many of the same methodological issues in obtaining population-based CP prevalence estimates. Issues related to obtaining an accurate numerator include the definition of inclusion and exclusion criteria for case determination, evaluation of completeness of case ascertainment, comparison of prevalence and trends, and ensuring validity. To make appropriate comparisons across surveillance systems and over time, it is imperative that the details of these issues are well understood.
There are 5 main CP inclusion and exclusion criteria areas that differ across surveillance systems. These areas include (1) the minimum age of survival, (2) hypotonia, (3) severity, (4) postneonatally acquired CP and timing of the injury, and (5) select chromosomal anomalies, genetic syndromes, metabolic diseases, and mitochondrial disorders. A survey of international surveillance systems and registers provided data on the characteristics of these programs. Approximately half of the international surveillance registers do not have a minimum age of survival for inclusion as a CP case. Of those registers that do impose a minimum age criterion, there is considerable variation from 1 to 8 years of age. With respect to severity, many systems do not apply severity criteria to determine case inclusion. Of those that do, often a combination of neurologic signs, dysfunction, motor impairment by age 5 years, or Level 1 on the GMFCS is applied. Most surveillance programs do not include hypotonic CP. Data from the ADDM Network, which does include hypotonic CP cases in its monitoring efforts, found only 2.6% of cases had hypotonic CP. The overwhelming majority of CP registers includes postneonatally acquired CP cases and has the ability to exclude these children for specific analyses. Of the programs that define a maximum age of cerebral damage, the age varies from 2 to 8 years. Two set of criteria currently exist detailing the specific chromosomal anomalies, genetic syndromes, and metabolic and mitochondrial disorders that constitute CP. Many of the current surveillance programs operationalize the Badawi or SCPE criteria.
All surveillance programs are faced with the challenge of attaining complete ascertainment of all children with CP within a specified geographic area at a specific period in time. From the perspective of birth prevalence, one issue influencing under ascertainment is migration from the surveillance area between birth and age of identification. If it is not possible to follow the entire birth cohort to determine the CP status at the defined age, then birth prevalence will be an underestimate of the “true prevalence” because a proportion of CP cases migrate beyond geographic ascertainment. Another challenge for ascertainment is the type of source for data collection. Three sites in the ADDM Network, which relies on multiple source record review, do not have access to education records, rather only records from clinical and service providers. Although data from the Metropolitan Atlanta Developmental Disability Surveillance Program (MADDSP), one of the ADDM Network sites, indicate that prevalence of CP is not significantly affected by under ascertainment of CP cases identified uniquely from special education sources, prevalences reported from the other three sites is likely an underestimate. MADDSP can only review records of children receiving public education and therefore may miss children who are in private school or are being home-schooled. As previously mentioned, this is believed to be a small proportion of CP cases (because many are identified through clinical sources) but still remains a source of under ascertainment.
The same rigor that is applied when ascertaining the number of individuals with CP in a specified population must also be used to choose an appropriate denominator for calculation of prevalence. The most common denominator used to report CP prevalence is live births. Many CP registers also report prevalence using neonatal survivors as the denominator. Live birth and neonatal survivor denominator data are useful when examining etiologic questions. It can be argued that neonatal survivors are the more appropriate denominator as neonatal deaths do not have the potential to be ascertained as CP cases. Use of neonatal survivors is particularly important when examining CP prevalence by birth weight (BW) or gestational age, as infants of extremely low birth weight (ELBW, <1000g) very low birth weight (VLBW <1500g) or preterm birth (< 37 weeks gestation) have a higher neonatal mortality rate than those of greater birth weights or gestational ages. Therefore, at lower birth weights and earlier gestational ages, the effect of using these 2 different denominators can be significant. Paneth and colleagues stipulate that using live births as the denominator for lower birth weight groups is the only means of obtaining a picture of the net contribution of improving survival to the population prevalence of CP. The choice of denominator is one that differs across registers, most often due to ease of availability of vital statistics data. Nevertheless, the denominator must be taken into consideration when comparing prevalence across studies. A handful of surveillance programs use children as the denominator to calculate period prevalence. These data are most informative for service provision and planning. Due to differences across CP surveillance programs with respect to the aforementioned methodological issues, it is crucial that each program assess the comparability of their own program’s methods over time and account for any within-program methodological changes before examining trends. Once internal validity is established, comparison of trends across CP surveillance programs is appropriate.
Across the various surveillance programs in developed countries, estimates of CP prevalence overall using live births and neonatal survivors have been comparable, most estimates being 2.0 per 1000 ( Table 1 ). Estimates using children as the denominator have been somewhat higher, ranging from 3.1 to 4.4 per 1000. Among population-based studies of CP, males have been found to have a higher prevalence of CP than females, with sex ratios ranging from 1.1:1 to 1.5:1. Although there have been few studies that examined racial/ethnic differences in prevalence, a higher prevalence in black non-Hispanic children compared with white non-Hispanic children has been reported for 3 time periods in metropolitan Atlanta and overall from the 3 CDC ADDM Network sites in 2002. Among all studies of CP, spastic subtypes have been found to be more common, with fewer percentages of the ataxic and dyskinetic subtypes. Little is known currently about the prevalence of CP in developing countries. Whereas differences in reported prevalence may reflect accurate differences in population prevalence, variations may also reflect differences in the methodology of both the numerator and denominator. Efforts are being made to foster international communication to understand these variations and strive for comparability where possible. One of the greatest strengths of the numerous surveillance registers in existence is that they have been in operation for many decades, which affords the opportunity to examine trends in CP prevalence over time.
Related posts:
The Adult with Cerebral Palsy: A Provider-Consumer Perspective
Bone Density in Cerebral Palsy
Access to Employment and Economic Independence in Cerebral Palsy
Clinical Applications of Outcome Tools in Ambulatory Children with Cerebral Palsy
Understanding Function and Other Outcomes in Cerebral Palsy
Establishing Access to Technology: An Evaluation and Intervention Model to Increase the Participation of Children with Cerebral Palsy
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
