The Descriptive Epidemiology of Cerebral Palsy

of 17
All materials on our website are shared by users. If you have any questions about copyright issues, please report us to resolve them. We are always happy to assist you.
Related Documents
  The Descriptive Epidemiology of Cerebral Palsy  Nigel Paneth, MD, MPH a,b, T , Ting Hong, MD, MSc a  ,Steven Korzeniewski, MSc a  a   Department of Epidemiology, College of Human Medicine, B 636 West Fee Hall, Michigan State University, East Lansing, MI 48823, USA  b  Department of Pediatrics and Human Development, College of Human Medicine, B 636 West Fee Hall, Michigan State University, East Lansing, MI 48823, USA Methodologic issues in ascertaining the frequency of cerebral palsy The first question that epidemiologists ask about a disease is ‘‘How much?’’.Before resources can be allocated toward the prevention or control of any diseaseor health condition, it is essential to weigh the importance of a disease in thecontext of competing public health priorities. The importance that is assigned toCP is a function of its severity; the consequent burden that it places on affectedchildren, their families, and societies; and of its high frequency as a cause of activity limitation in childhood. This article reviews the evidence about thefrequency of CP, how this frequency varies in different places, and whether CP is becoming more or less prevalent.Measuring the frequency of a disease in a population is not an easy task. Wehave no good ongoing count of the frequency of cardiovascular disease (CVD),even though it is the major cause of death and disability in the western world.The difficulty of accurately counting the variety of manifestations of incident CVD (eg, sudden death, onset of angina, heart attack, abnormal angiogram) 0095-5108/06/$ – see front matter   D  2006 Elsevier Inc. All rights reserved.doi:10.1016/j.clp.2006.03.011 This work was supported by the Training Program in Perinatal Epidemiology at Michigan StateUniversity, Grant No.1 T32 HD046477-01A1 from the National Institutes of Health, and UnitedCerebral Palsy Grant No. R-735-02. T  Corresponding author. Department of Epidemiology, College of Human Medicine, B 636 West Fee Hall, Michigan State University, East Lansing, MI 48823.  E-mail address: (N. Paneth).Clin Perinatol 33 (2006) 251–267  requires epidemiologists to default to mortality data to monitor CVD time trendsin most populations.Counting cases of disease in a population requires having a good operationalcase definition, sound sources of information about cases, and a system for collecting and systematizing that information. At the same time, the population at risk for the disease in question must be enumerated accurately. Fortunately, because CP is such an important cause of childhood disability, and because itsclose links to pregnancy and the perinatal period have suggested to manyinvestigators that its frequency may reflect perinatal care, efforts to measure itsfrequency in populations are ongoing in several parts of the world. Case definitions in cerebral palsy CP is a clinical diagnosis; no laboratory test or tissue histology defines its presence or absence. In addition, despite the usefulness of CTand MRI scanning,no single neuroimaging pattern or group of patterns fully encompass the diag-nostic findings that are possible in CP. In fact, some children who have CP havenormal brain imaging findings. Diagnostic language to describe CP has beendeveloped by many authorities, most recently by the International Committee onCerebral Palsy Classification [1]; however, all clinical diagnoses are subject tosome degree of observer variability, no matter how carefully a clinical entity isdefined. Unless a single highly skilled observer diagnoses all cases of CP in a population—clearly an impossible task—population counts of CP must rely onexisting clinical records. It has been shown that with enough information, espe-cially if that information includes aspects of motor functioning and not just motor examination findings, experienced clinicians can classify children reliably basedon medical records [2].Some degree of motor dysfunction is essential to the concept of CP. Theskilled child neurologist may detect Achilles tendon hyperreflexia and slight hypertonia in some young children who have clumsiness but no real impairment of motor functioning. But should such children be referred to as having CP? Thisargument is not easy to settle, but from a public health perspective, the decision isclear. Such cases do not carry the personal, familial, and social burden that weassociate with CP, and, moreover, are unlikely to be recorded as having CP usingthe usual methods that are available for population enumeration. Therefore,although it can be accepted that CP can exist in a nearly subclinical state, subject to diagnosis by skilled and experienced clinicians, it is not useful for publichealth action to include such cases in the CP rubric. The CP that is described inthis article, and which is recorded in CP registries, is disabling CP (ie, CP withenough motor disability to interfere clearly with ordinary tasks, such as walking,running, jumping or climbing stairs). The useful Gross Motor Function Clas-sification Scale [3,4] would classify all children who have disabling CP at level 2or above.At some point in the future, perhaps through further advances in neuro-imaging, we may learn that such mild cases share features in common with CP as paneth  et al252  conventionally understood, and that expanding the definition of CP to includesuch children will lead to advances in understanding etiology, analogously to theway in which the concept of asymptomatic infection has clarified much about infectious disease epidemiology. But we are not there yet.Age at diagnosis is also important to case definition. A well-known phe-nomenon in child neurology is that some signs that appear to be surprisinglysevere, including hypotonia and hypertonia, can be seen in infants, but disappear with age. Such findings are especially common in very premature infants. For thisreason, the diagnosis of CP should not be made before the age of 24 months,unless the child has an unusually severe case or other supporting information(eg, severe neuroimaging abnormalities) is available. Typology An issue that is subsidiary to the subject of case definition is whether thedifferent subtypes of CP also can be enumerated well. It is likely that somesubtypes of CP come to clinical attention more readily than do others. A mildhemiplegia in which age of walking is not delayed may be missed in populationcounts, as will some cases of spastic diplegia. At the same time, severe spasticquadriplegia or dramatic choreoathetosis is unlikely to be missed. Thus, thedistribution of subtypes of CP in any population study may provide clues to thethoroughness with which CP is being ascertained in a population. A higher thanaverage proportion of quadriplegia may signal a passive reporting system. What is the appropriate denominator population for reporting the frequency of  cerebral palsy? A fundamental concept in epidemiology is that all measures of disease fre-quency in a population should, to the extent possible 1 , be denominatored to the population that is actually at risk for acquiring the disease. For example, prostateand uterine disorders are denominatored to men and women, respectively, and not to the entire population. CP is a disease that can be acquired only once in alifetime, and then only during a specific period of risk (ie, from early in preg-nancy until about age 1 or 2 years). Because newborns and infants constitutes theonly population at risk, it makes sense to use the birth cohort from whence thecase arose as the denominator, and this is the common, although not universal, practice in CP registries.If a prevalence survey of all 5- to 10-year-old children who have CP is per-formed in a community, using the total number of 5- to 10-year-old children inthe community as the denominator population seems to be a reasonable choice. It  1 In disease rates that are denominatored to populations enumerated in censuses, it surely happensthat some cases of disease did not arise from the enumerated population because of inmigration of  people with disease, and vice versa; however, the error that is introduced is generally small. descriptive epidemiology of cp  253  must be remembered that 5- to 10-year-old children are, by definition, at zero risk for developing CP (brain damage to head injury in a child of that age will not betermed CP). Moreover, using school-age children as the denominator risks caus-ing confusion, because family patterns of movement across geographic bound-aries may be related to having a child with a disability. If a region is rich in healthcare facilities and schools that are known to accommodate children with dis-abilities, families with CP may migrate to it preferentially. Conversely, having achild who has CP may preclude some movement by families from one regionto another.An illustration of this point can be seen in two CP surveys that were madeduring roughly the same period of time by the Metropolitan Atlanta CPSurveillance Program. Each reported a different prevalence of CP. When live births were used as the denominator, the prevalence of CP ranged from 1.7 to2.0 per 1000 live births between 1975 and 1991 [5]. But when 5-year-old children in Atlanta were used as the denominator, CP prevalence was reported as3.1 to 3.6 per 1000 between 1996 and 2000 [6]. It is unlikely that the higher figure represents an increase in CP over time. Comparison with population-baseddata from other Western countries (Table 1) indicates that the second figure is the outlier. Therefore, it seems likely that families with children who have CP moveto Atlanta for services, which inflates the numerator when place of birth isignored. Such children have emerged from birth cohorts outside of Atlanta, andshould be denominatored to the populations in which they were born. Should the denominator be live births or survivors? Consistent with the view that the population at risk should be the denominator,it might be argued that survivors of the neonatal period are a better denominator than are live births, because deaths are not at risk for CP. When neonatal mortalityis low (currently  b 5/1000 in the United States [7]), the difference to too small to matter; however, for VLBW infants (  1500 g at birth), the difference betweenCP rates per live births and per neonatal survivors is substantial. When survivalrates are improving rapidly, as they have been in the United States for the past 30 years, the public health is better served by examining CP rates per live birthsin VLBW infants. This is the only way to obtain a sense of the net contribution of improving survival to the prevalence of CP in the population. Rarely, CP inVLBW infants has been reported per 1000 total live births of all weights [8]. This figure gives a sense of the net contribution of VLBW births to total CP preva-lence, but does not represent the risk accurately.  Incidence or prevalence? When a cohort of births is under effective surveillance, and the appropriatediagnostic information is sought at regular intervals, one is in a position todescribe the cumulative incidence of a disease within the cohort. A registrysystem, by contrast, inevitably produces cross-sectional information—the num- paneth  et al254


Jul 25, 2017
Related Search
We Need Your Support
Thank you for visiting our website and your interest in our free products and services. We are nonprofit website to share and download documents. To the running of this website, we need your help to support us.

Thanks to everyone for your continued support.

No, Thanks