GM2 gangliosidosis in a UK study of children with progressive neurodegeneration: 73 cases reviewed
This article is commented on by Maegawa on pages 104105 of this issue.
Dr Nicholas Smith at Department of Medicine, University of Cambridge, Box 157, Addenbrooke’s Hospital, Cambridge CB2 0QQ, UK. E-mail: email@example.com
Aim To report the demographic, phenotypic, and time-to-diagnosis characteristics of children with GM2 gangliosidosis referred to the UK study of Progressive Intellectual and Neurological Deterioration.
Method Case notification is made via monthly surveillance card, administered by the British Paediatric Surveillance Unit to all UK-based paediatricians; children with GM2 gangliosidosis were identified from cases satisfying inclusion in the UK study of Progressive Intellectual and Neurological Deterioration and analysed according to phenotypic and biochemical categories.
Results Between May 1997 and January 2010, 73 individuals with GM2 gangliosidoses were reported: 40 with Tay–Sachs disease, 31 with Sandhoff disease, and two with GM2 activator protein deficiency. Together they account for 6% (73/1164) of all diagnosed cases of progressive intellectual and neurological deterioration. The majority (62/73) were sporadic index cases with no family history. Children of Pakistani ancestry were overrepresented in all subtypes, particularly juvenile Sandhoff disease, accounting for 10 of 11 notified cases. Infantile-onset variants predominated (55/73); the mean age at onset of symptoms was 6.2 and 4.7 months for infantile-onset Tay–Sachs and Sandhoff disease respectively, and 26.2 and 34.7 months for the corresponding juvenile-onset variants. Time to diagnosis averaged 7.4 months and 28.0 months in infantile- and juvenile-onset disease respectively.
Interpretation GM2 gangliosidosis is a significant cause of childhood neurodegenerative disease; timely diagnosis relies upon improved clinical recognition, which may be increasingly important as specific therapies become available. There is a potential benefit from the introduction of screening programmes for high-risk ethnic groups.
- GM2 APD
GM2 activator protein deficiency
Progressive intellectual and neurological deterioration
What this paper adds
- • Each year in the UK, approximately four children with GM2 gangliosidosis are reported to the Progressive Intellectual and Neurological Deterioration study.
- • The majority are index presentations with prominent diagnostic delay, emphasizing the need for increased clinical awareness of this disorder which, with few exceptions, is readily confirmed by enzymatic assay.
- • The study identified a higher proportion of children of Pakistani origin with GM2 gangliosidosis than would be expected from UK population figures.
The GM2 gangliosidoses (Tay–Sachs disease [TSD; OMIM 272800], Sandhoff disease [OMIM 268800], and GM2 activator protein deficiency [GM2 APD; OMIM 272750]) are neurodegenerative diseases that typically present in childhood and are transmitted as autosomal recessive conditions. Characterized by impaired lysosomal breakdown of the glycosphingolipid GM2 ganglioside, which accumulates within the organelle in association with related glycoconjugates, these disorders result from impaired activity of the hydrolase β-hexosaminidase (EC 220.127.116.11) or, rarely, defects in the GM2 activator protein responsible for presentation of GM2 ganglioside to the enzyme. Two principal isoforms of β-hexosaminidase exist: β-hexosaminidase A, a heterodimer of non-identical subunits (α and β), and β-hexosaminidase B, a homodimer of the β-subunit. Consequently, pathological mutations in either of the cognate genes, HEXA, encoding the α-subunit, or HEXB, encoding the β-subunit, or in the gene encoding the activator protein (GM2A) give rise to TSD, Sandhoff disease, and GM2 APD respectively.1,2
Clinically indistinguishable with the exception of subtle visceral and skeletal manifestations in some individuals with Sandhoff disease,2,3 these disorders can present at almost any age with neurodegeneration proceeding at a variable rate. Individuals with classical onset disease in infancy have prominent storage in retinal ganglion cells, giving rise to a characteristic ‘cherry-red spot’; these children undergo rapid neurodegeneration and die in early childhood. Less commonly, juvenile-onset variants occur, typically with progressive cerebellar dysfunction, spasticity, and dementia. Rarely, adult-onset forms present; however, the substantive burden of disease exists in childhood.4–8
The GM2 gangliosidoses have a wide ethnic distribution with an estimated birth prevalence in the order of one in 200 000 and 1 in 385 000 for TSD and Sandhoff disease respectively.9–11 Historically, TSD is overrepresented in Jews, with a carrier frequency of one in 30 individuals of Ashkenazi (eastern European) Jewish descent;8,12,13 increased disease frequency is also seen within Irish Americans, Canadians from south-eastern Quebec, and the Cajun people of Louisiana.14–17 Ethnic associations are less well known in Sandhoff disease, this disease being more common in several communities in which founder mutations of the HEXB gene occur at a high frequency.16–20
In the UK, systematic knowledge of prevalence, disease burden, and natural history are not available. However, in countries where successful application of community-based preconception screening programmes have been introduced, TSD has significantly decreased in the Ashkenazi population, with most new diagnoses occurring in children of non-Jewish descent.21–23
We present a population-based review of all individuals with GM2 gangliosidosis reported by paediatricians to the UK-wide study of Progressive Intellectual and Neurological Deterioration (PIND). Our findings indicate that lysosomal diseases generally, and the GM2 gangliosidoses in particular, constitute a significant and under-recognized subgroup of paediatric neurodegenerative disease within the population of the UK.
The British Paediatric Surveillance Unit has studied rare childhood disorders since 1986 via cards mailed to all paediatricians in the UK requesting notification of conditions under surveillance (or confirmation of no cases seen); 2800 to 2900 cards are sent out each month and over 90% are returned. The PIND study has been engaged in the scheme since May 1997.24
The case definition for PIND is ‘Any child (<16y of age at onset of symptoms) who fulfils the following criteria: progressive deterioration for more than 3 months with loss of already attained intellectual or developmental abilities and development of abnormal neurological signs’. This definition includes presymptomatic individuals identified with diagnosed neurodegenerative conditions. Children with static (non-progressive) intellectual loss are excluded. Clinical information is obtained via a standardized, structured, and free-text questionnaire completed by the notifying paediatrician via telephone interview or site visit. Anonymized clinical data are reviewed by an independent expert group of nine neurologists and one geneticist to confirm and classify diagnoses.
Notifications of all GM2 gangliosidoses confirmed by β-hexosaminidase activity assay or molecular analysis and reported to the PIND study as of 31 January 2010 were reviewed in detail; an additional questionnaire was sent to referring paediatricians requesting updated information about seizure disorder and mortality data in notified cases.
Disease was classified according to biochemical variant and further subtyped into infantile-onset disease (first reported symptoms ≤12mo of age) and juvenile-onset disease (first reported symptoms between 12mo and 16y of age). Adult-onset disease was not surveyed.
Demographic and clinical features are reported; ethnicity categorization was adopted from the PIND questionnaire (based on categories used by the Communicable Disease Surveillance Centre, Colindale, London, UK); where available, the presence or absence of Jewish ancestry is reported. Clinical features attributed to comorbid conditions were excluded. Age at diagnosis, confirmed by β-hexosaminidase activity, is reported as the individual’s age at which the diagnostic assay was performed; time to definitive diagnosis was determined relative to the age at symptom manifestation, excluding those identified following diagnosis in a sibling.
Ethical consent for the PIND study was obtained from the Cambridge Local Research Ethics Committee (ref. 97/010), the Public Health Laboratory Service Ethics Committee, and the national Patient Information Advisory Group (British Paediatric Surveillance Unit 2–10(c)/2005).
Between May 1997 and January 2010, 1483 notifications met the criteria for PIND in childhood. A definitive diagnosis was established in 1164 cases (78%), constituting 146 distinct disorders. Of these 1164, lysosomal diseases accounted for 45% (527/1164), of which 14% (73/527) were GM2 gangliosidoses; thus the GM2 gangliosidoses represent 6% (73/1164) of all individuals with a confirmed diagnosis reported to the PIND study.
Demographics of the cohort
Forty individuals with TSD, 31 with Sandhoff disease, and two with GM2 APD were notified (Table I); recorded birth dates ranged from June 1983 to June 2007. Infantile-onset disease accounted for 33 of the 40 (82.5%) individuals with TSD, 20 of the 31 (64.5%) individuals with Sandhoff disease, and both children with GM2 APD; the remaining seven (17.5%) children with TSD and 11 (35.5%) with Sandhoff disease experienced juvenile onset. Six of the children with juvenile TSD were further classified as having early juvenile disease (onset between 12mo and 8y of age) and the remaining individual as having late juvenile disease (onset between 8 and 15y).
Table I. Cohort demographicsa
|Index casec||32 (33)||1 (2)||18 (20)||5 (6)||0 (1)||6 (11)|
|Consanguinity||3 (29)||0 (1)||5 (20)||0 (3)||1 (1)||11 (11)|
| White||27 (31)||1 (2)||13 (20)||5 (5)||0 (1)||0 (11)|
| Pakistani||3 (31)||1 (2)||2 (20)||0 (5)||1 (1)||10 (11)|
| Black||0 (31)||0 (2)||0 (20)||0 (5)||0 (1)||0 (11)|
| Indian||0 (31)||0 (2)||1 (20)||0 (5)||0 (1)||0 (11)|
| Bangladeshi||0 (31)||0 (2)||1 (20)||0 (5)||0 (1)||0 (11)|
| Other||1 (31)d||0 (2)||3 (20)e||0 (5)||0 (1)||1 (11)f|
The majority ethnicity within the infantile-onset cohort was ‘white’, accounting for 41 of 53 (77%) notifications, with six of 53 (11%) children having Pakistani ancestry. In juvenile-onset disease, Pakistani heritage dominated, accounting for 11 of 17 (65%) cases. Inquiry as to Jewish ancestry is not specifically surveyed by the PIND in childhood questionnaire; however, this was confirmed voluntarily in three of 10 children with TSD.
Index cases without family history accounted for 36 of 40 (90%) individuals with TSD, 24 of 31 (77%) individuals with Sandhoff disease, and a single child with GM2 APD at diagnosis; there was a positive family history in 11 families, eight of which were of Pakistani ancestry. Although response rates were incomplete, consanguinity was reported in 20 of 65 (31%) individuals.
Overall clinical course (Table II)
Table II. Overall clinical coursea
|Mean age [range] at symptoms onset (mo)||6.2 [3–11] (31)||(i) 8; (ii) 3||4.7 [2–6] (17)||26.2 [14–36] (6)||132 (1)||34.7 [16–72] (6)|
|Mean age [range] at developmental regression (mo)||6.6 [4–11] (32)||(i) 8; (ii) 6||5.1 [3–6] (16)||25.6 [14–36] (5)||– (0)c||40.3 [16–74] (7)|
|Mean age [range] at diagnosis (mo)d||12.9 [6–29] (30)||(i) –; (ii) 11||13.7 [10–18] (12)||53.6 [33–85] (5)||176 (1)||69.7 [36–115] (6)|
|Mean age [range] at seizure onset (mo)||14.6 [4–24] (28)||(i) 15; (ii) 14||16.1 [11–30] (16)||37.8 [25–49] (4)||0 (1)||168 (1)|
|Number deceased at data analysis||27 (27)||– (0)c||13 (14)||2 (3)||– (0)c||4 (8)|
|Mean age [range] at death (mo)||38.1 [18–81] (27)||– (0)c||32.3 [18–55] (13)||(i) 78; (ii) 80 (2)||– (0)c||155.8 [141–173] (4)|
The first symptoms in infantile-onset disease arose at a mean age of 6.2 months (range 3–11mo) in individuals with TSD and 4.7 months (range 2–6mo) in individuals with Sandhoff disease; three individuals presenting at birth were excluded from analysis as first symptoms were attributed to comorbid pathology. The two children with GM2 APD also presented in infancy, one at 8 months and one at 3 months of age. In juvenile-onset disease, symptoms arose at an overall mean of 3 years 5.4 months (range 14mo–11y) within the cohort with TSD; the majority (6/7) were of an early-juvenile phenotype with symptom onset at a mean age of 26.2 months (range 14–36mo). A single individual with late-juvenile TSD presented at 11 years of age. Age at symptom onset in juvenile-onset Sandhoff disease was specified exactly in only six of 11 notifications with a mean age of 34.7 months (range 16mo–6y); in the remaining five individuals, three were reported to manifest during ‘primary school’ and two in ‘early childhood’.
Developmental regression was reported at a mean age of 6.6 months (range 4–11mo) and 5.1 months (range 3–6mo) in children with infantile-onset TSD and Sandhoff disease respectively; at a mean age of 25.6 months (range 14–36mo) in children with early juvenile-onset TSD; and at a mean age of 3 years, 4.3 months (range 16mo–6y 2mo) in juvenile-onset Sandhoff disease; age at onset of regression was not specified in the single individual with late-juvenile TSD.
After the follow-up questionnaire information about survival was available in 41 of 55 individuals with infantile onset. Death was reported in 40 of the 41 and occurred at a mean age of 36.3 months (range 18mo–6y 9mo); pneumonia and/or respiratory failure was the terminal event in 23 (58%) children, with cause of death unspecified in the remaining 17 (42%) children. Mortality data are more limited in juvenile-onset disease; one individual with early-juvenile TSD survived to 78 months and one to 80 months of age, with death attributed to pneumonia in the latter case. Death in four individuals with juvenile-onset Sandhoff disease occurred at a mean of 12 years 11.8 months (range 11y 9mo–14y 5mo); cause of death was given as pneumonia in two cases and not specified in the others; five children with juvenile-onset disease were alive at the time of writing.
Presenting features within the infantile-onset cohort were cited, singularly or in combination with other features, as developmental delay or regression in 38 of 55 (69%) individuals, strabismus or perceived visual impairment in nine of 55 (16%), and hyperacusis with excessive startle in four of 55 (7%); this last feature was not specifically surveyed by the PIND questionnaire, but relied upon notification of presence or absence by referring clinicians in the free-text part of the questionnaire. One child was diagnosed by chance detection of lysosomal vacuolation upon rectal biopsy for investigation of presumed Hirschsprung disease at 3 months of age; seizures, poor feeding, and constipation were reported in individual cases. Among those with juvenile-onset disease, the first symptoms reported included gait abnormalities or problems with balance (including ataxia) in eight of 18 (44%), general motor or cognitive regression in nine of 18 (50%), and delayed speech or language concerns in six of 18 (33%). Obsessive–compulsive behaviours were prominent in the single child with late-juvenile TSD.
Overall clinical features
Infantile-onset disease: Overall clinical features in infantile-onset disease included hyperacusis with excessive startle response, independently volunteered in 34 individuals, and increased myotatic reflexes noted in 34 of 40 (85%) (Table III). No children with infantile-onset disease achieved independent walking. A pattern of combined axial hypotonia with limb spasticity was seen in 31 of 36 (86%) individuals for whom the relevant data were provided (not shown in Table III). Visual deterioration was noted in 42 of 47 (89%) individuals and retinal storage resulting in a visible cherry-red spot upon fundoscopy in 43 of 47 (91%). Seizures occurred in 46 of 55 (84%) children with infantile-onset disease at a mean age of 15.1 months (range 4–30mo); multiple seizure types were noted in the majority of children with generalized convulsive seizures (21 of 40) and myoclonic seizures (16 of 40) being most frequent; status epilepticus was reported in six children.
Table III. Overall clinical featuresa
|Hyperacusisc||24||1|| 9||1||0|| 1|
|Hypotonia||30 (30)||2 (2)||18 (20)||1 (4)||0 (1)|| 1 (4)|
| Axial||23||2||15||0|| || 0|
| Generalized|| 3||0|| 0||1|| || 1|
| Not specified|| 4||0|| 3||0|| || 0|
|Hypertonia||21 (23)||1 (2)||12 (12)||3 (6)||0 (1)|| 0 (4)|
| Limb||21||1||12||2|| || |
| Generalized|| 0||0|| 0||0|| || |
| Not specified|| 0||0|| 0||1|| || |
| Increased||23 (26)||2 (2)|| 9 (12)||4 (5)||0 (1)|| 6 (8)|
| Normal|| 3 (26)||0 (2)|| 2 (12)||1 (5)||1 (1)|| 2 (8)|
| Decreased|| 0 (26)||0 (2)|| 1 (12)||0 (5)||0 (1)|| 0 (8)|
| Independent gait achieved|| 0 (31)||0 (2)|| 0 (20)||5 (6)||1 (1)|| 9 (11)|
| Gait abnormal |
| Ataxia|| || || ||4||0|| 9|
| Not specified|| || || ||2||1|| 2|
|Visual loss||29 (31)||0 (1)||13 (15)||2 (6)||0 (1)|| 0 (6)|
|Cherry-red spot||28 (31)||2 (2)||13 (14)||0 (3)||0 (1)|| 0 (7)|
|Seizures||28 (33)||2 (2)||16 (20)||4 (6)||0 (1)|| 1 (11)|
Juvenile-onset disease: Clinical features were more variable in juvenile-onset disease; all individuals walked independently, with ataxia being a notable feature in 13 of 18 (72%). Motor signs in individuals with early-juvenile TSD included hypertonia in 50% (3/6) and increased myotatic reflexes in 80% (4/5); the single child with late-juvenile TSD had a normal motor examination aside from reported gait abnormalities. The majority of children with juvenile Sandhoff disease had relatively few findings upon motor examination, with the exception of increased myotatic reflexes in six of eight (75%) reported cases and a single child with generalized hypotonia. Two of the six children with early juvenile-onset TSD were reported to have visual deterioration, with normal vision noted in the remaining individuals with TSD. None of the six children with juvenile Sandhoff disease for whom information was available suffered visual loss. Fundal examination was documented as normal in 11 of 18 individuals with juvenile onset; none showed visible retinal storage.
Generalized convulsive seizures featured in four of six children with early-juvenile TSD with a mean age at onset of 3 years 1.8 months (range 25mo–4y 1mo) of age; seizures were not reported in the single child with late-juvenile TSD. Only one child with juvenile Sandhoff disease had infrequent, generalized convulsive seizures documented from 14 years of age.
Time to diagnosis
The mean age at diagnosis, by determination of β-hexosaminidase activities, was 12.9 months (range 6–29mo) and 13.7 months (range 10–18mo) in children with infantile-onset TSD and Sandhoff disease respectively. Age at diagnosis was 11 months in one individual with GM2 APD and not recorded in the other. In the early-juvenile TSD subgroup, diagnosis was made at a mean age of 4 years 5.6 months (range 33mo–7y 1mo), while the sole individual with late-juvenile disease had a diagnostic assay at 14 years 8 months. By comparison, children with juvenile-onset SD received a diagnosis at a mean age of 5 years 9.7 months (range 36mo–9y 7mo).
Time to diagnosis was calculated for individuals where data were available. A delay of 6.8 months (range 1–25mo) and 7.8 months (range 2–12mo) from symptom onset to diagnosis was seen in infantile-onset TSD and Sandhoff disease respectively (overall average of 7.4mo). Two individuals with manifestations typical of infantile-onset TSD had normal β-hexosaminidase activity in routine diagnostic assays, as expected in GM2 APD. In one, diagnosis was confirmed by molecular analysis, and in the second after detection of an elevated GM2 ganglioside concentration in cerebrospinal fluid almost 12 months after symptom onset. In juvenile-onset disease, a comparatively longer diagnostic delay of 28.0 months was seen, occurring at 30.7 months (range 10–18mo) and 22.7 months (range 20–25mo) in TSD and Sandhoff disease respectively. Analysis was restricted to index cases, excluding children identified subsequent to diagnosis in an older sibling.
The GM2 gangliosidoses are an important, often overlooked, cause of neurodegenerative disease; historical perceptions perpetuate the belief that these disorders are restricted to the Jewish community, thus underestimating their true burden in the contemporary UK population and potentially compounding diagnostic delay.
Accurate determination of disease frequency within the UK population has not been previously established, partly because data from independent diagnostic laboratories and specialist management centres have not been systematically collated and also because universal population screening is not employed. In the ethnically similar Australian population, the majority white population, among whom GM2 gangliosidosis is most prevalent within the UK, shares a common European ancestry and includes a similar proportion with Jewish ancestry.25 The birth prevalence of TSD and Sandhoff disease within the Australian population is reported as 0.5 and 0.26 per 100 000 live births respectively9– as they depend upon clinical recognition and referral for biochemical testing, these figures are probably an under-representation. By extrapolation, it might be expected that, with average annual births in the order of 633 000,26 approximately three children with TSD and one or two with Sandhoff disease will present each year in the UK. The PIND study referrals by birth date averaged 2.6 cases of TSD and 1.5 cases of Sandhoff disease annually, which concur with the estimates based on Australian data and provide reassurance about the representative accuracy of this study; however, we recognize that study notification is referrer dependent and may not identify all cases of GM2 gangliosidosis in UK children.
Nevertheless, this study provides an overall picture of the distribution of GM2 gangliosidosis presenting with progressive intellectual and neurological disease, without the selection bias inherent in case series from specialist centres. However, the PIND study methodology has disadvantages, reflecting the design concessions necessary to provide an ethical and practical surveillance model; we acknowledge that notification bias exists, and clearly the quality of data obtained in each case is reliant upon the referring clinician, introducing the potential for underreporting of symptom prevalence in those who are notified.
Collectively, most notified individuals were of white ancestry, reflecting the majority ethnic group within the UK population. However, individuals of Pakistani descent are overrepresented, accounting for four of 37 (10.8%) respondents with TSD, 12 of 31 (38.7%) with Sandhoff disease, and one of two children with GM2 APD, compared with a UK population percentage of 1.7%,27 most probably reflecting the presence of community founder mutations propagated by consanguineous marriages (present in 16/17 Pakistani children notified), as we have reported previously.28 Religious and cultural beliefs may also negatively influence uptake of antenatal diagnosis and technologies such as pre-implantation genetic screening in future pregnancies.
The efficacy of preconceptual screening programmes is supported by data from the Ashkenazi Jewish population.12,22,23 However, Jewish ancestry is not specifically surveyed in the PIND study, necessitating caution when interpreting the independently volunteered history in three individuals with TSD.
Detailed clinical characterization of disease is limited by the data recorded; broad congruence within subtypes is clear within this series and, although a degree of genotype–phenotype correlation has been recognized,6,29,30 possibly reflecting residual hexosaminidase activity of the mutant enzyme in situ,31,32 molecular analysis of the HEXA and HEXB genes was not available in this study and cannot be commented upon.
The principal difficulty in diagnosing the GM2 gangliosidoses is the non-specific nature of clinical features. Although no clinical feature is pathognomonic, the presence of hyperacusis, characterized by excessive startle in response to auditory stimuli, is highly suggestive of infantile-onset disease; in our experience this is a prominent early feature which may pass unmentioned unless specifically sought. The characteristic cherry-red spot is an almost invariable feature of infantile-onset disease and reflects foveal prominence juxtaposed against glycosphingolipid storage within retinal ganglion cells.33 All children under investigation for suspected neurodegenerative disease require fundal examination – with pupillary dilatation if necessary – by an experienced clinician. In this series, a cherry-red spot was reportedly absent in less than 10% of children with PIND with infantile-onset disease; while poor technique may account for this, attrition of retinal ganglion cells with advanced disease may lead to the disappearance of this feature.34 Classical fundal changes are uncommon in juvenile-onset disease, although visual loss with optic atrophy and retinitis pigmentosa-like changes may develop.5,8 Moreover, while juvenile-onset disease is typically subtle at onset and insidious in its progression, ataxia in children with speech and language difficulties, particularly dysarthria, should heighten diagnostic suspicion.
The numerous causes of PIND may explain the prolonged time to definitive diagnosis for many children in this study; indeed, the shorter time to diagnosis in index individuals from the unique cohort of juvenile-onset Sandhoff disease in Pakistani families reflects awareness of the disorder amongst paediatricians with experience of this community.
Even in the absence of established treatments for the GM2 gangliosidoses, early identification is critical; not only may it inform reproductive choice for parents and relatives of the proband, but it clarifies management and obviates the need for further, often invasive, investigations. In the long term, the promise of definitive therapies based on specific molecular understanding of the disease provides a strong impetus for detailed study and timely diagnosis; clinical trials of pharmacological chaperones in juvenile- and adult-onset disease35–37 provide early encouragement, and gene transfer techniques to introduce functional copies of the β-hexosaminidase gene are in development in the light of positive preclinical findings.38–40 While UK population screening for the GM2 gangliosidoses (utilizing enzymatic assay and molecular analysis of common mutations) is currently limited to TSD within the Jewish community,21 the relatively high rates of disease in Pakistani families evident in this study suggest a potential benefit from similar approaches within this ethnic group.
The GM2 gangliosidoses remain an under-recognized cause of neurodegeneration in children from the UK. Timely diagnosis depends on diagnostic vigilance and an awareness of disease patterns in the changing ethnic make-up of the contemporary population. Key clinical features should be sought and, once suspected, prompt assessment of β-hexosaminidase activities undertaken.
Dr Smith is supported by the UK NIHR Cambridge Biomedical Research Centre. This is an independent report commissioned and funded by the Policy Research Programme in the Department of Health, UK. The views expressed in the publication are those of the authors and not necessarily those of the Department of Health. Many thanks to all the paediatricians who report cases, to Mr Richard Lynn, the BPSU Scientific Co-ordinator, and the PIND Expert Group: Professor Jean Aicardi, Dr Peter Baxter, Professor Yanick Crow, Dr Carlos de Sousa, Dr John Livingston, Dr Michael Pike, Professor Richard Robinson, Professor Robert Will, Dr John Wilson, Dr Evangeline Wassmer, and Dr Sameer Zuberi.