- J L Hutton,
- T Cooke,
- P O D Pharoah
- Departments of Statistics, Computational Mathematics, and Public Health, University of Liverpool, Liverpool L69 3BX Department of Public Health, University of Liverpool, Liverpool L69 3BX.
- Correspondence to: Professor Pharoah.
- Accepted 8 June 1994
Objective : To determine life expectancy of children with cerebral palsy.
Design : Cohort analysis, by means of register compiled from multiple sources of ascertainment, of all children with cerebral palsy born during 1966-84 to mothers resident in Mersey region. Status of children was determined by flagging through NHS central register.
Subjects : 1258 subjects with idiopathic cerebral palsy, of whom 1251 were traced and included in analysis.
Main outcome measures : Effect of functional ability (ambulation, manual dexterity, and mental ability), sex, birth weight, and gestational age on survival.
Results : 20 year survival for whole cohort was 89.3% for females and 86.9% for males. For subjects with no severe functional disabilities 20 year survival was 99% (95% confidence interval 98% to 100%), while subjects severely disabled in all three functional groups had 20 year survival of 50% (42% to 58%). Subjects with birth weight <=2500 g had 20 year survival of 92% (89% to 95%), while those with birth weight <2500 g had survival of 87% (84% to 89%). Subjects with gestational age of <37 weeks had 20 year survival of 93% (91% to 96%), while those with gestational age >=37 weeks had survival of 85% (83% to 88%). Birth weight and gestational age were less predictive of survival than functional disability. Best statistical model used gestational age and number of severe functional disabilities as predictors.
Conclusions : Life expectancy of this cohort of children with cerebral palsy was greater than has been suggested in some previous studies. This has important implications for social, educational, and health services.
Public health implications
Public health implications
Cerebral palsy is the commonest cause of severe physical disability in children, and its prevalence is increasing because of improved survival of low birthweight babies
Little information is available on life expectancy, which is needed to plan for adequate provision of services for disabled people
We studied life expectancy of all children with cerebral palsy born in Mersey region during 1966-84
Survival rates (to 20 years of age) of mild and moderately disabled children were not much lower than those of unaffected children, and about half of severely disabled children survived to age of 20
These results have important implications for provision of social, educational, and health services
Although cerebral palsy is the commonest cause of severe physical disability in children, current routine health information systems do not allow the monitoring of prevalence. Several reports from ad hoc registers of people with cerebral palsy have recorded an increase in prevalence among low birthweight infants.*RF 1-7* Because statistics on cerebral palsy are not routinely collected there is little information on life expectancy. Such information is needed to counsel parents about a child's prognosis; to plan for the provision of medical, educational, and social services; and for medicolegal purposes relating to litigation (even though perinatal cause and effect is not proved).
Numerous studies of survival in children with cognitive and motor disabilities have been carried out, but most are of limited use for assessing current life expectancy of children with cerebral palsy. Data may have been collected many years ago, since when life expectancy is likely to have altered.8 Alternatively, only children in long term institutional care have been assessed.8,9 This is liable to bias the estimate of life expectancy not only because children admitted to institutions tend to be more severely affected but because institutional care itself might prejudice survival. Motor disability may have a significant effect on life expectancy, but some studies relate to children in whom mental retardation without concomitant motor disability is the main feature.10 Others are based on small numbers and a short follow up.11,12
One recent study has reported on survival in a population based cohort of children with cerebral palsy born during 1970-9 in the south east of England.13 We report on the survival of children with cerebral palsy born during 1966-84 to mothers resident in the Mersey region and examine the effect of cognitive and motor disability, birth weight, and gestational age on life expectancy. We hope that the data will be useful in planning for the future needs of children with cerebral palsy.
Subjects and methods Cerebral palsy register
The cohort consists of all children with cerebral palsy born between 1966 and 1984 to mothers whose area of residence at the time of birth was in the area administered by Mersey Regional Health Authority. The region has a population of about 2.5 million with 25 000-30 000 births a year. A register of cases was compiled from multiple data sources to ensure completeness of ascertainment as described previously.1,2 All the subjects on the register were flagged at the National Health Service Central Register of the Office of Population Censuses and Surveys. The individuals were traced, and notifications of deaths were provided with copies of the death certificates.
The 250 cases of “acquired” cerebral palsy - where the impairment was considered to have arisen more than 28 days after birth - were excluded from the analysis. Problems arise in separating idiopathic and acquired cerebral palsy, particularly in some low birthweight infants, among whom adverse conditions may extend well into the postnatal period. Also, making a firm diagnosis early in infancy is often not possible. Only if an infant's progress was normal during the neonatal period and there was a clear cut adverse event after the 28th day did we consider the cerebral palsy to be acquired.
Once ascertained, hospital and community records were abstracted for details of the clinical type and severity of the cerebral palsy.
Classification of functional ability
Information was sought for ambulatory ability, manual dexterity, and mental ability. Children who were too young or who died before they could be assessed, has functional ability recorded as missing.
Ambulatory ability was classified as minimal disability - able to run and climb stairs but with less than usual dexterity and speed (category 1); limited - difficulty experienced in climbing stairs, able to walk only short distances without a rest and at a slow pace (category 2); walking aids required during normal everyday activities (category 3); wheelchair required, which can be propelled by child (category 4); and wheelchair required and assistance is needed for propulsion (category 5). Manual dexterity was classified as normal - upper limbs unaffected (category 1); mild - some clumsiness of fine movements but able to feed and dress without assistance (category 2); moderate - able to feed and dress with difficulty (category 3); and severe - unable to feed and dress without assistance (category 4). Mental ability was categorised according to the intelligence quotient (IQ) as normal - IQ<=85 (category 1); mild learning disability - IQ 70-84 (quotient 2); moderate learning disability - IQ 50-69 (category 3); and severe learning disability - IQ <50 (category 4).
Most of the subjects had had a formal assessment of IQ by a variety of standard tests. Usually the assessment had been made by an educational psychologist to determine the child's need for special education. If there had been more than one assessment, the most recent one was used in the analysis. A few children (mainly normal or with mild learning disability) were classified according to a subjective assessment of the child's paediatric records.
We calculated the subjects’ ages in days: age at death for those who died and age at 31 July 1993 for those who were still alive. The most recent receipt of notifications of deaths was on 31 March 1994, and we therefore assumed notification of all deaths up to 31 July 1993 to be complete.
We produced tables of survival according to grouped categories of functional disability, birth weight, and gestational age by means of the life table method with intervals of one year. A log-logistic accelerated life model, which is a fully parametric model of survival time, was applied to determine which covariates to use. Accelerated life models based on the log-logistic distribution were significantly better than those using exponential, Weibul, or Gamma models.14
Likelihood ratio tests15 were used to determine whether the categories of functional disability could be grouped. Categories not significantly different at the 5% level were combined. The explanatory values of the combined categories for the functional disability groups, with birth weight and gestational age as continuous variables, were then studied. The comparison of non-nested models (for example, one with three covariates indicating severe ambulatory, manual, and mental disabilities and one in which three covariates indicated the number of severe disabilities) was done by means of Akaike's information criterion.16 The maximum number of individuals with complete data was included for each of the analyses, which were performed with the SAS software package.17
Altogether 1508 cases of cerebral palsy (1258 idiopathic and 250 acquired) were submitted to the National Health Service Central Register for flagging. Of the idiopathic cases, seven were excluded (three children had emigrated and four could not be traced). Of the 1251 children included in the analysis, 1101 were still alive, 143 had died, and seven were assumed to be alive because, although they were not on the National Health Service Central Register, notification would have been made of any death.
Table I summarises the characteristics of the 1251 subjects. The ratio of males to females was 1.3:1 (709:541), 437 of the subjects had a birth weight of 2500 g or less, and 432 had a gestational age of 37 weeks or less. About a third (424) of the subjects had a quadriplegia, another third (397) had a hemiplegia, and a quarter (276) had a diplegia. The remainder had other conditions, such as dyskinesias and ataxias. Almost a quarter (274) had severe ambulatory disability, over a fifth (255) had severe manual disability, and a third (398) had an IQ of 50 or less. Deaths were concentrated in these groups with the most severe disabilities.
Table II shows the types of functional ability by birth weight. The group with normal birth weight (<2500 g) contained the highest proportion of subjects with disabilities categorised as severe in each functional ability group: 24% (182/743) were confined to a wheelchair that was not self propelled (compared with 17% (22/131) of those with a birth weight <=1500 g); 24% (174/736) had severe manual disability (compared with 13% (17/130) of the lowest birthweight group); and 37% (268/726) had severe mental disability (compared with 26% (33/129) of the lowest birthweight group).
Table III shows the subjects’ sex-specific survival to 1 year of age and subsequent ages. About 85-90% of the subjects survived to 20 years of age compared with the 97% 20 year survival reported for the population of England and Wales in 1970-2.18
Ambulatory ability could be reduced from five to two categories for predicting survival. Mildly affected subjects (categories 1-4) had a 97.5% probability of surviving to the age of 20, while only 56.4% of those with severe ambulatory disability (category 5) reached the age of 20 (fig 1).
Manual dexterity could also be reduced to two groups (categories 1-3 and category 4): 54.8% of the subjects with severe manual disability (category 4) survived to the age of 20 compared with 97.1% of the remainder (categories 1-3) (fig 2).
Mental ability was less useful for prognosis: the subjects with IQ <50 (category 4) had a 72.5% probability of surviving to the age of 20 whereas the subjects with IQ >=50 (categories 1-3) had a 98.0% probability, which is similar to that of the normal population (fig 3).
The clinical type of cerebral palsy was considerably less informative for prognostic purposes than the categories of functional disability. However, both birth weight and gestational age were strongly associated with survival. Figure 4 shows the survival for low (<=2500 g) and normal (<2500 g) birthweight babies; the 20 year survival rates were 91.6% and 86.4% respectively. Comparison of the two gestational age groups gave similar results: 20 year survival for those born at 37 weeks or less was 93.2% and for those born at more than 37 weeks was 85.3%. There was no evidence of a birth cohort trend in survival during 1966-84: 10 year survival was 92.1% for those born before 1980 and 92.2% for those born from 1980 onwards.
Analysis of the survival of the 1179 subjects (of whom 115 had died) with complete data on ambulatory ability, manual dexterity, and mental ability showed that ambulation in three categories was most informative and manual dexterity in two categories was almost as informative. Combining the three types of functional disability increased the ability to predict survival. Four combination categories were defined depending on whether the subjects were classed as having no, one, two, or three severe functional disabilities (that is, ambulation category 5, manual dexterity category 4, and mental ability category 4). This model was simpler though not significantly better than one that analysed the three severe types of functional disability separately.
Table IV shows the survival of the 1179 subjects with complete data on disability function. For those who had no functional disability in the severe category 20 year survival was 98.8%, decreasing to 93.5% for those with one, 84.7% for those with two, and 50.3% for those with all three disability functions in the severe category. About 40% (28) of the 77 children for whom data on at least one functional disability were missing died before the age of 4 years.
Analysis of the 1054 subjects with complete data on functional disability, birth weight, and gestational age showed that birth weight and gestational age had substantially and significantly less explanatory value than functional disability. Birth weight added no information when gestational age was known. The best model used gestational age and number of severe disabilities as predictors.
Our analysis shows that several aspects of disability could be used to predict the life expectancy of children with cerebral palsy. As expected, the severely disabled children had poorer survival.
Some of the missing data relate to children who had died at an early age before there had been any assessment of their functional disability. These missing values are unlikely to be independent of the severity of disability; the 28 infants who died aged between 1 and 40 months probably included a disproportionate number who were more severely affected. This may have led to an overestimation of life expectancy of the more severely disabled groups.
Comparison with other studies
It is reassuring that the results presented here are similar to those reported from the south of England.13 The greater survival of girls compared with boys is confirmed, though the difference was not significant at 5% with a likelihood ratio test. For the southern England cohort also, it was found that the most important predictors of mortality were mobility and mental ability. Manual dexterity was not assessed as a separate variable in that study. The survival of mildly and even moderately affected children approached that of normal, unaffected children. Each severely disabled children in both studies had about 50% probability of surviving to 20 years of age.
In contrast, an analysis from the United States of people with developmental disabilities who received services from the Californian Department of Developmental Services appeared to show a much shorter life expectancy.19,20 These results have been extensively used for medicolegal purposes, but there are problems in their interpretation and in comparisons with our results. The people included in the Californian study were in receipt of services, and the sample was assumed to include almost all people with severe mental retardation and to exclude about half of those who were moderately retarded and 90% of those who were mildly retarded. The authors did not discuss these assumptions or the potential biases that might have arisen.
The description of the statistical analysis that was performed is limited and the methods may be inappropriate. For example, it is not clear whether age, which was used to define the groups for which survival analyses were given, was age at entry to the study, age at presentation to the services, or current age.
Survival and birth weight
Our observation that children with cerebral palsy who were of low birth weight and low gestational age had longer life expectancy than those who were full term or of normal birth weight is contrary to intuition. In our cohort of children with cerebral palsy the low birthweight group had a lower proportion with severe disability, which accounts for their longer survival. This was probably due to the more severely affected low birthweight infants dying before the cerebral palsy was recognised. If this is so, as mortality among low birthweight infants improves, more severely affected infants will survive long enough for the diagnosis of cerebral palsy to be made. Consequently, future cohorts of low birthweight infants with cerebral palsy will show a higher proportion with severe disability, and life expectancy will approach that of normal birthweight children with cerebral palsy.
Survival and birth cohort
It is perhaps surprising that we found no effect of birth cohort on survival. One possible explanation is that the cerebral palsy register was started in 1980. For the earlier period (1966-79) the register was compiled by retrospective study of several sources of information. It is possible that, among the children with cerebral palsy born in this earlier period, some may have died and their names have been deleted from the information sources. They would, therefore not have been recorded in our register. A small number of such unrecorded deaths would have a substantial impact on the estimated life expectancy of children born in the earlier period.
Implications of results
The life expectancies of children with cerebral palsy that we report here show that, even for moderately affected children, survival rates were not very different from those of normal children, at least for the first 20 years. Even severely disabled children had a probability of about 50% of surviving 20 years or more. This has considerable implications for social, educational, and health services. It is imperative that an adequate routine system of monitoring morbidity must accompany the monitoring of mortality, particularly among low birthweight infants.
We gratefully acknowledge computing help from Mr Chris West and grant support from the Mersey Region Health Authority and from Children Nationwide.
Full details of the accelerated parametric life model are available from the authors.