Prevalence, Clinical Features, Neuroimaging, and Genetic Findings in Children With Ataxic Cerebral Palsy in Europe

Abstract

Background and Objectives

To report on prevalence, associated impairments, severity, and neuroimaging findings in children with ataxic cerebral palsy (CP).

Methods

In children coded as having ataxic CP in the Central database of Joint Research Center–Surveillance of Cerebral Palsy in Europe (JRC-SCPE) and born during 1980–2010, birth characteristics, severity profiles including associated impairments, neuroimaging patterns, and the presence of syndromes were analyzed. Definitions were according to validated SCPE guidelines. Prevalence over time was estimated using Poisson regression.

Results

In total, 679 children with ataxic CP were identified in 20 European CP registers. The proportion with ataxic CP was 3.8% and varied from 0% to 12.9%. Prevalence over time showed no significant trend. Approximately 70% of children with ataxic CP were able to walk, and 40% had severe intellectual impairment and a high impairment index. Children with ataxic CP were mostly born at term (79%) and with normal birth weight (77%). Neuroimaging patterns revealed normal findings in 29%, brain maldevelopments in 28.5%, miscellaneous findings in 23.5%, and brain injuries in 19%, according to the SCPE classification. Genetic syndromes were described in 9%.

Discussion

This register-based multicenter study on children with ataxic CP provides a large sample size for the analysis of prevalence, severity, and origin of this rare CP subtype. Even with strict inclusion and classification criteria, there is variation between registers on how to deal with this subtype, and diagnosis of ataxic CP remains a challenge. Ataxic cerebral palsy differs from other CP subtypes: children with ataxic CP have a disability profile that is more pronounced in terms of cognitive than gross motor dysfunction. They are mostly term born and the origin rarely suggests acquired injuries. In addition to neuroimaging, a comprehensive genetic workup is particularly recommended for children with this CP type.

Introduction

Ataxic cerebral palsy (CP) is considered a distinct subgroup in most population-based registers and studies of CP surveillance groups. It constitutes the smallest clinical subgroup of CP and accounts for 4% of cases in Europe,^1,2 5.3% in Australia,³ and 2% in the Canadian CP registry.⁴ Several challenges are encountered when studying the characteristics of this subgroup. First, large multicenter studies are needed, and second, the definition of the characterizing neurologic features are not commonly agreed upon (e.g., in the Canadian registry, this subtype is called ataxic-hypotonic⁴).

The inclusion of ataxic forms within the CP concept has been questioned in the past. Pathogenic mechanisms of ataxic CP are poorly known and its nosologic interpretation is controversial.⁵ Terms such as “nonprogressive congenital ataxia” have been proposed.^6,7 These authors described imaging findings in children with nonprogressive cerebellar ataxia. In contrast to the other CP types, only exceptionally was a lesional pattern revealed, and in more than 50% of cases, CT or MRI showed a normal result; in approximately 30%–40% of cases, there was cerebellar hypoplasia of variable degree. Studies on systematically collected neuroimaging findings in CP showed a similar pattern.^3,4,8

The overall prevalence of cerebral palsy in high-income countries has recently been reported to be decreasing,^9-11 mainly driven by a decrease in CP in preterm-born children.^9,10,12 There are only a few population-based studies reporting on the prevalence of ataxic CP over a long period. In Sweden^13,14 the trend in time for ataxic CP showed no significant change and a variation between 3% and 5% of total cases with CP since the 1970s and 4% in the period 2011–2014. Bottos et al. reported on ataxic and ataxic-diplegic CP in Italy 1965–1989, showing an inconsistent trend in this period.¹⁵

The common database of the Surveillance of Cerebral Palsy in Europe (SCPE) comprises 20 registers across Europe. Not only its size/magnitude but also several quality criteria make it most suitable to analyze such a small group as ataxic CP, with special clinical-neurologic characteristics. The SCPE has spent considerable time and effort to develop a common language of definition, inclusion and exclusion criteria, and a reliable classification. CP definition is based on phenomenology (clinical picture and history) because criteria such as imaging and laboratory results, which explain etiology or pathogenesis, are not available to the same extent in all countries and over time.¹ SCPE subtypes are defined as unilateral spastic, bilateral spastic (formerly diplegia and/or quadriplegia/tetraplegia), dyskinetic, and ataxic CP, accounting for 33.5%, 55%, 7.8%, and 3.7% of cases, respectively.² The “classification tree” of the SCPE is being widely recommended to facilitate reproduceable classification.¹ A teaching tool “Reference and training manual”¹⁶ was developed to promote a shared understanding and describe the clinical, functional, and neurologic features of CP and pitfalls in diagnosis and classification. Interrater reliability studies of CP diagnosis, neurologic subtype, and gross motor function were performed, and regular quality monitoring has been established.¹⁷

According to the SCPE, ataxic CP is characterized by an abnormal pattern of movement and posture, accompanied by the loss of orderly muscular coordination, so that movements are performed with abnormal force, rhythm, and accuracy.¹ Typical features are as follows: trunk and gait ataxia (disturbed balance), past pointing (overshooting and undershooting of goal-directed movements), tremor (mainly a slow intention tremor), and low muscle tone.

It has been suggested to bring together data on phenomenology, neuroimaging, and genetics, and use them in an epidemiologic approach to better understand the role of ataxic CP within the overall CP context.¹⁸ This needs a larger collaboration, which is ideally given by the SCPE network. The aim of this analysis of ataxic CP within the SCPE common database was to:

• look at trends in prevalence over time with the hypothesis of a constant trend, assuming a mainly nonlesional origin of ataxic CP, thus, not affected by changes in neonatal care;

• analyze the severity profile of children with ataxic CP with the hypothesis that the association of motor severity and additional impairment is low in accordance with the assumption that ataxic CP is mainly nonlesional;

• investigate additional evidence for a genetic origin, also in terms of neuroimaging findings; and

• analyze further the neuroimaging patterns, especially the miscellaneous subgroup.

Methods

Definitions

CP and ataxic CP were defined according to the definitions of the SCPE.¹ Muscular hypotonia as a sole neurologic finding is considered as an exclusion criterion for CP, according to this definition.

Source of Data

The collaboration of European Cerebral Palsy registers has been previously described.^1,19 The SCPE network was established in 1998 by 14 registers. These registers first agreed on a standardized definition and classification of children with CP. Then a common database was created and has been updated since, with an annual data submission process. A guideline for data submission describes all the variables and their definition. We used data from 20 population-based registers participating in the SCPE network, covering either a part or their whole country. These registers provide yearly pseudonymized data on children with CP to the SCPE central database. Population data (live births the same year in the same catchment area) are available from the census or other official data population source.

Study Population

Children with CP were eligible for this study if they (1) were born between 1980 and 2010, (2) were classified as having the ataxic CP subtype according to the SCPE definition and classification, and (3) had a CP of prenatal/perinatal origin, that is, children with postneonatal CP were excluded. In prevalence analyses, cases were included if they were born (or, for 2 registers, were living at age of registration) in the geographical areas covered by the registers and if denominators were available for the corresponding birth year/register.

Data Collection

Motor Impairment and Associated Impairments

Severity of gross motor function was described either using the Gross Motor Function Classification System (GMFCS)²⁰ or walking ability divided into 3 categories. Thus, these 2 variables were combined to describe the gross motor function: “unaided walking” corresponded to GMFCS levels I–II, “walking with assistive mobility device” to GMFCS level III, and “unable to walk” to GMFCS levels IV–V. Fine motor function was described according to the Bimanual Fine Motor Function (BFMF) classification²¹ or the Manual Ability Classification System (MACS).²² BFMF and MACS levels were also combined to describe fine motor function in 3 categories: I–II (corresponding roughly to limitations in more advanced fine motor skills/manual abilities do not usually restrict independence in daily activities), III (limitations in fine motor skills/handles objects with difficulty), and IV–V (severely limited abilities, requires continuous support)—the BFMF also accounting for asymmetries, the MACS being centered on abilities.

Intellectual impairment was classified into 3 categories of IQ, in accordance with earlier analyses of the SCPE: IQ ≥ 70 (normal or near-normal intellect further defined as attendance of a regular school with or without support), IQ 50–69 (mild intellectual impairment; basic reading, calculating and writing abilities in a modified school curriculum), and IQ < 50 (severe intellectual impairment; no reading, writing, and calculation abilities). IQ classification was revised by the SCPE from birth year 1990, introducing a subdivision of the group with IQ between 50 and 85 (prior classification). Cases where it was not possible to distinguish between those with an IQ below or above 70 were excluded from analyses involving intellectual impairment. The evaluation was performed by formal psychometric testing or by clinician's estimate based also on parental report.

Visual and hearing impairments were considered as the presence or absence of any such impairment; severe visual impairment was defined as blind or no useful vision in both eyes (after correction) and severe hearing impairment as a loss greater than 70 dB (before correction, on the better ear).

Epilepsy was defined as a history of 2 unprovoked seizures, excluding febrile or neonatal seizures; active epilepsy was defined as receiving antiepileptic treatment at the age of registration. Speech performance was described using the Viking Speech Scale (VSS)²³: VSS levels were grouped into 2 categories, I–II (understandable speech) and III–IV (not understandable speech).

Impairment Index

A compound measure (impairment index), comprising 3 categories, was developed to characterize severity and combinations of impairments according to previous definitions.²

Low impairment index was defined as being able to walk without aids (GMFCS levels I–II), IQ ≥ 70, no visual impairment, no hearing impairment, and no epilepsy. High impairment index was defined as inability to walk (GMFCS levels IV–V) and/or severe intellectual impairment (IQ < 50) with or without 1 or more of the following impairments: severe visual impairment, severe hearing impairment, and active epilepsy. Moderate impairment included all other levels of impairment not defined as low or high level of impairment.

Data on Birth Characteristics

According to usual description in epidemiologic studies, children were categorized into 3 birth weight groups: less than 1,500 g (corresponding to very low birth weight), 1,500–2,499 g (moderately low birth weight), and greater than 2,499 g (normal birth weight). Gestational age was categorized as ≥37 weeks (term), 32–36 weeks (moderate preterm), and <32 weeks (very premature birth).

Data on MR Imaging

Only images performed postneonatally were considered, that is, when MR imaging was performed after the age of 4 weeks. Each register enters MRI according to MRI classification system (MRICS),²⁴ and the entered data then are checked at each annual submission by a group of 3 pediatric neurologists with specific expertise in MRI (K.H., V.H., and I.K.M.). In case of incongruity, additional information may be requested from the registers and the category corrected accordingly. The MRICS classification is as follows: A. maldevelopments, B. predominant white matter injury, C. predominant gray matter injury, D. miscellaneous findings, and E. normal findings.²⁴ In the registers, either MR images directly or written reports are classified, both of which have been validated.²⁴ In addition, the SCPE common database includes age at imaging and records the description of MRI findings in a text field. Category D (miscellaneous findings) refers to imaging patterns that cannot be allocated to the malformations (A) or to white or gray matter lesions (B, C). For the purpose of this study, miscellaneous findings were further analyzed according to the text description and grouped into 5 subgroups following the method previously performed in a larger group.²⁵ In addition, we specifically searched for cerebellar pathology in the text description of MRI classified as A (i.e., cerebellar hypoplasia) or D (i.e., cerebellar lesion, cerebellar atrophy). MRI patterns in children with imaging performed before vs after 2 years of age were compared. Children who had MRI performed were compared with those without MRI in terms of associated impairment and birth characteristics, to check on the representativeness of data.

Data on Genetic Analysis

The SCPE data form systematically collects data on syndromes (either ICD-9 or ICD-10, along with text description) and provides the possibility to give free text on additional findings such as genetic findings. Children with syndromes, chromosomal anomalies, or developmental brain anomalies are included if they meet the clinical criteria of the agreed definition of CP and excluded if they do not.¹ The SCPE uses guidance for the definition of syndromes related to reference documents, especially Eurocat syndrome's guide and the conclusions of the international consensus reached by Smithers-Sheedy et al. 2014.²⁶

Data Availability

Due to privacy/ethical restrictions and the sensitive nature of the research, supporting data are not publicly available.

Statistical Analysis

The sample description used usual statistics: counts and percentages calculated on nonmissing data. Prevalences were estimated per 100,000 live births (LBs) and presented with exact 95% CIs.

We analyzed trends in prevalence over time using Poisson regression considering the number of children with CP by registry and by year as the dependent variable and the number of live births per year in each registry as offset term. We tested nonlinearity of the trend using polynomial terms for birth years up to third order. Nested models were tested using likelihood ratio χ² tests. The trend over time was represented by a smoothed curve, estimated using locally weighted scatterplot smoothing.

To test for the association between severe intellectual impairment and gross motor function level, the χ² test for trend was used. To test for the association between MRI findings and gestational age groups, the Fisher exact test was used.

The threshold selected for statistical significance was p < 0.05. Analyses were generated using Stata Statistical Software: Release 15 (StataCorp LLC., College Station, TX).

Results

Contribution of Participating Registers

Table 1 summarizes the number of children with ataxic CP in the participating registers and the birth years with available data in the period 1980–2010. A total of 679 children with ataxic CP were registered. These cases accounted for 3.8% (varying between 0% and 12.9% within centers) of all children with CP in the registers. There was 1 register with no child with ataxic CP in this period.

Table 1.

Numbers and Prevalence Rates of Ataxic CP in Areas Covered by the Participating SCPE Registers (Birth Years 1980–2010)

Location of the registry	Birth years available	No. of live births	No. of children with ataxic CP	No. of children with ataxic CP included in prevalence analysis	No. of children with CP	% of children with ataxic CP	No. of children with missing CP type	Prevalence of ataxic CP by 100,000/live births (95% CI)
C01, RHEOP, France	1980–2010	576,102	73	73	925	8.4	54	12.7 (9.9–15.9)
C02, RHE31, France	1980–2010	387,836	42	42	676	6.6	45	10.8 (7.8–14.6)
C05, Northern Ireland	1981–2010	744,913	29	27	1,725	1.8	129	3.6 (2.4–5.3)
C06, Western Sweden	1980–2010	674,404	64	60	1,391	4.6	0	8.9 (6.8–11.4)
C07, Counties of Dublin, Kildare and Wicklow, Ireland	1980–2010	684,953	72	70	1,639	4.4	19	10.2 (8.0–12.9)
C11, Merseyside and Cheshire, UK	1980–1989	323,853	34	34	723	4.7	0	10.5 (7.3–14.7)
C12, Denmark	1980–2007	1,309 677	78	78	2,860	2.8	26	6.0 (4.7–7.4)
C13, Viterbo, Italy	1980–2010	56,333	9	9	207	4.3	7	16.0 (7.3–30.3)
C15, Norway	1991–2010	839,417	82	75	2,086	4.1	97	8.9 (7.0–11.2)
C18, Madrid, Spain	1991–1999	61,701	2	2	93	2.2	0	3.2 (0.4–11.7)
C19, Slovenia	1999–2010	224,599	9	9	468	2.0	11	4.0 (1.8–7.6)
C21, Portugal	1996–1997 2001–2010	1,072,522	87	78	1,936	5.0	181	7.3 (5.7–9.1)
C22, Riga, Latvia	2000–2008	61,210	4	4	93	4.3	1	6.5 (1.8–16.7)
C23, South West Hungary	1999–2010	104,984	25	24	194	12.9	1	22.9 (14.6–34.0)
C25, Iceland	1990–2010	92,866	8	7	213	3.8	1	7.5 (3.0–15.5)
C27, Belgium	1999–2010	563,635	28	28	875	3.3	18	5.0 (3.3–7.2)
C28, Croatia	2003–2010	280,926	5	5	588	0.9	0	1.8 (0.6–4.1)
C29, St Gallen, Switzerland	1995–2010	77,902	11	11	144	7.6	0	14.1 (7.0–25.3)
C31, Attica (Athens Metropolitan area), Greece	1999–2010	473,661	17	16	735	2.6	89	3.4 (1.9–5.5)
C32, Sunderland, UK	1990–2010	57,636	0	0	257	0	0	0 (0–6.5)
Total		8,669,130	679	652	17,828	3.8	679	7.5 (6.9–8.1)

Abbreviation: CP = cerebral palsy..

Prevalence

A total of 652 children with ataxic CP were included in prevalence analyses (Table 1). Figure 1 presents the aggregate prevalence per 100,000 live births and trend over time over the studied period 1980–2010. The overall prevalence was 7.5/100,000 LB (95% CI 6.9–8.1), showing no significant trend over the period (p = 0.19) ranging from 7.7/100,000 LBs in 1980 to 7.3/100,000 LBs in 2010. The annual percent change in prevalence was a decrease of 0.7% (95% CI −1.8% to +0.4%). The overall prevalence per register varied from 0 (95% CI 0–6.5) to 22.9 (95% CI 14.6–34.0) per 100,000 LBs.

Figure 1. Time Trend for Ataxic CP Prevalence per 100,000 Live Births (p = 0.19) in the SCPE Common Database (Birth Years 1980–2010) .

CP = cerebral palsy; SCPE = Surveillance of Cerebral Palsy in Europe.

Walking Ability/GMFCS and Associated Impairments

Table 2 describes the clinical characteristics of children with ataxic CP. Approximately 70% of children were able to walk unaided. Severe motor impairment was comparatively rare, below 13%. The functional involvement of the upper extremities seemed to be also mostly mild. However, almost 40% of children presented with a severe intellectual impairment, another 20% with IQ 50–69 (mild-moderate intellectual impairment). Severe visual impairment and severe hearing impairment were rare (4.7% and 2.9%, respectively). Active epilepsy was present in approximately 20% of cases. More than half of children had a relevant speech performance problem. The analysis of impairment index showed a low index (e.g., walking without aids and no other additional impairment) only for 20% and a high impairment index in 40% of children.

Table 2.

Clinical Characteristics of Children With Ataxic CP in the SCPE Common Database (Birth Years 1980–2010)

Characteristics	All children with ataxic CP N = 679 (%)	Children with MRI result N = 221 (%)	Children without MRI result N = 458 (%)	p Value
GFMCS				0.18
I–II, able to walk	451 (69.4)	139 (64.7)	312 (71.7)
III, walk with aids	117 (18.0)	45 (20.9)	72 (16.6)
IV–V, unable to walk	82 (12.6)	31 (14.4)	51 (11.7)
Missing	29	6	23
BFMF (MACS for 26 cases)				0.34
I–II	248 (75.4)	137 (72.5)	111 (79.3)
III	34 (10.3)	21 (11.1)	13 (9.3)
IV–V	47 (14.3)	31 (16.4)	16 (11.4)
Missing	350	32	318
Intellectual impairment				0.17
Severe (IQ < 50)	202 (38.5)	78 (41.1)	124 (37.0)
Mild (IQ 50–69)	105 (20.0)	43 (22.6)	62 (18.5)
Near-normal of normal (IQ > 69)	218 (41.5)	69 (36.3)	149 (44.5)
Missing	154	31	123
Visual impairment				0.001
No	399 (63.1)	111 (53.4)	288 (67.9)
Not severe	203 (32.1)	86 (41.3)	117 (27.6)
Severe	30 (4.7)	11 (5.3)	19 (4.5)
Missing	47	13	34
Hearing impairment				0.53
No	541 (93.1)	190 (92.7)	351 (93.4)
Not severe	23 (4.0)	7 (3.4)	16 (4.3)
Severe	17 (2.9)	8 (3.9)	9 (2.4)
Missing	98	16	82
Epilepsy				0.23
No	437 (71.5)	144 (67.3)	293 (73.8)
Not active	44 (7.2)	18 (8.4)	26 (6.5)
Active	130 (21.3)	52 (24.3)	78 (19.6)
Missing	68	7	61
Viking speech scale				0.81
I–II	80 (48.2)	63 (47.7)	17 (50.0)
III–IV	86 (51.8)	69 (52.3)	17 (50.0)
Missing	513	89	424
Impairment index				0.07
High	220 (41.6)	83 (44.6)	137 (38.2)
Moderate	210 (39.0)	74 (39.8)	136 (37.9)
Low	115 (19.4)	29 (15.6)	86 (24.0)
Missing	134	35	99

Abbreviations: BFMF = bimanual fine motor function; CP = cerebral palsy; GMFCS = Gross Motor Function Classification System; MACS = Manual Ability Classification System; SCPE = Surveillance of Cerebral Palsy in Europe.

The severity of intellectual impairment increased with the severity of motor impairment (Figure 2, p < 0.001). The proportion of mild-to-moderate and severe intellectual impairment, however, proved to be high, even in children who were able to walk: 19.8% and 27.7%, respectively.

Figure 2. Intellectual Impairment by Severity of Motor Impairment in Ataxic CP (p < 0.001) in the SCPE Common Database (Birth Years 1980–2010).

CP = cerebral palsy; GMFCS = Gross Motor Function Classification System; SCPE = Surveillance of Cerebral Palsy in Europe

Birth Characteristics: Birth Weight, Gestational Age

Sixty percent of the children were male individuals. Children with ataxic CP were mostly born at term (79%) and had normal birth weight (77%). Ten percent were very preterm or of very low birthweight (for additional information on neonatal characteristics, see eTable 1, links.lww.com/WNL/D220).

Neuroimaging Findings

Data for MRICS were available for 221 children with ataxic CP (32.5%). This classification was collected systematically for children born after 1999. Distribution of MRICS patterns of the 72 children with MRI performed before the age of 2 years did not differ significantly from the 132 children with MRI performed after the age of 2 years (p = 0.44). Age at MRI was unknown for 17 children. The group of children with MRI performed was representative for the whole group of ataxic CP: associated impairments and birth characteristics did not differ significantly when compared with the group without MRI. There were only more male individuals in the group with MRI (p = 0.04), more children tended to have been admitted in a neonatal care unit (p = 0.06) and severity index tended to be higher (p = 0.07) (Table 2 and eTable 1, links.lww.com/WNL/D220).

The analysis, as listed in Table 3 and presented in Figure 3, revealed a brain malformation in almost 30% (N = 63) of the children. Of them, most of them (66%, N = 42) were malformations other than disorders of cortical formation (A2); cerebellar hypoplasia was the most frequent (N = 21). Predominant white and gray matter injuries (e.g., patterns B and C) accounted for only 19% of the cases (N = 42). In 29% of cases (N = 64), MRI was normal.

Table 3.

MRI Findings of Children With Ataxic CP in the SCPE Common Database (Birth Years 1980–2010), Specifying Cerebellar Pathology in A and D (Cerebellar and Posterior Fossa Pathology Indicated by *)

MRI classification system (MRICS), n (%)	n = 221
A. Maldevelopments	63 (28.5)
A1. Disorders of cortical formation	8
A2. Other maldevelopments	42
Cerebellar hypoplasia*	21
Dandy-Walker malformation*	5
Pontocerebellar hypoplasia*	3
Other posterior fossa anomalies, septo-optic dysplasia, molar tooth sign*	8
Corpus callosum agenesis	5
A. Not otherwise specified	13
B. Predominant white matter injury	29 (13.1)
B1. Periventricular leukomalacia, PVL	9
B2. Sequelae of intraventricular hemorrhage or periventricular hemorrhagic infarction	3
B3. Combination of PVL and IVH sequelae	1
B. Not otherwise specified	16
C. Predominant gray matter injury	13 (5.9)
C1. Basal ganglia/thalamus lesions	4
C2. Corticosubcortical lesions only	2
C3. Arterial infarctions	2
C. Not otherwise specified	5
D. Miscellaneous	52 (23.5)
Cerebellar atrophy*	19
Cerebral atrophy	10
Cerebellar lesions*	2
Infections or hemorrhage not covered under B or C	3
Brain injury not further specified	3
Ventricular dilatation, calcifications, other nonspecific descriptions	13
No description	2
E. Normal	64 (29.0)

Figure 3. Comparison of MRI Findings in Term-Born and Preterm-Born Children (p = 0.08) With Ataxic CP in the SCPE Common Database (Birth Years 1980 -2010).

CP = cerebral palsy; GA = gestational age; SCPE = Surveillance of Cerebral Palsy in Europe.

Miscellaneous findings accounted for 23.5% of cases (N = 52), and description was available for 96,15% (N = 50) of them. 29 of 50 (58%) indicated atrophy: cerebellar atrophy (n = 19) and cerebral atrophy (n = 10).

Cerebellar pathology taken together as indicated in the text fields of categories A (maldevelopments) and D (miscellaneous) accounted for 23% of cases (50). Lesions were only rarely reported (in 2 term born children); maldevelopments were the most frequent finding (29 or 13%), followed by cerebellar atrophy (19 or 8%).

A total of 46 MRIs in children with ataxic CP born before 37 weeks gestational age (GA) were described. We observed more often predominant white matter injury for children born before 32 weeks in comparison with children born at term, but the distribution of the MRICS patterns according to GA groups showed no significant difference (p = 0.08). The analysis of cerebellar pathology resulted in the following: the 2 cerebellar lesions were reported in term-born children; 5 of the 19 children with cerebellar atrophy were preterm-born children (24, 25, 26, 34, and 36 weeks of gestation); 6 of the 29 children with cerebellar malformations of variable degree described earlier within A2 (N = 42) were preterm born (25, 2 × 26, 28, 34, and 35 weeks of gestation).

Data on Etiology (Syndromes)

One register did not collect data on syndromes (C 11). Of the 645 children included in the other registers, 56 (8.7%) were diagnosed with a genetic syndrome. Of these, Angelman syndrome was the most common (22 cases), followed by Joubert syndrome (10 cases) and other distinct genetic syndromes (24 cases), such as Aicardi syndrome, Cri du Chat, duplication 17q12, deletion 1p36, Prader-Willi, and Smith-Lemli-Opitz.

Children with a specific genetic syndrome presented with significantly more severe motor and intellectual impairment and high impairment index compared with those without a genetic syndrome (Table 4).

Table 4.

Comparison of Impairment and MRI Characteristics Between Children With a Syndrome and Children With No Described Syndrome in the SCPE Common Database (Birth Years 1980–2010)

	Syndrome N = 56	Not described with a genetic syndrome N = 623	p Value
GMFCS IV–V*	15 (30.6%) 7 missing	67(11.1%) 22 missing	<0.001
Severe intellectual impairment*	33 (73.3%) 11 missing	169 (35.2%) 143 missing	<0.001
High impairment index*	34 (79.0%) 13 missing	186 (37.0%) 121 missing	<0.001
Available MRI results	N = 26	N = 195
Maldevelopments (A)	11 (42.3%)	52 (26.7%)	0.38
Predominant white matter injury (B)	4 (15.4%)	25 (12.8%)
Predominant gray matter injury (C)	1 (3.8%)	12 (6.1%)
Miscellaneous (D)	6 (23.1%)	46 (23.6%)
Normal (E)	4 (15.4%)	60 (30.8%)

Children with a syndrome diagnosis were more severely impaired. Their MRI more often indicated a maldevelopment and was less often normal. Percentages were calculated for available data.

Abbreviations: GMFCS = Gross Motor Function Classification System; SCPE = Surveillance of Cerebral Palsy in Europe.

Discussion

Data from 20 geographically defined registers across Europe contributing to the common database of the SCPE yielded 679 cases born between 1980 and 2010: a solid basis for detailed analysis of ataxic CP, the subtype with the lowest prevalence. This is by far the largest study on ataxic CP. As we hypothesized, there was no significant trend over time, the prevalence being in the range of approximately 7.5/100,000 live births. Birth characteristics showed that nearly 80% of children with ataxic CP were born at term or with normal birth weight, which is higher than in the more frequent spastic CP types. Indeed, this proportion is below 50% in bilateral spastic CP and below 70% in unilateral spastic CP.² Keeping in mind that total CP prevalence decreased in this period driven by the decrease of CP occurrence in preterm-born children,⁹ it is obvious that ataxic CP did not contribute to this decrease.

The proportion of ataxic CP among all included children with CP varied between 0% and 12.9% in the contributing registers. The reason for this lack of harmonization is probably 2-fold: clinicians may have difficulties with what defines nonprogressive ataxic features, which may even slowly change over time, and register holders may have an inconsistent approach regarding how to deal with syndromes that are relatively frequently seen in children with ataxic CP. SCPE has developed important teaching tools to facilitate a shared understanding of definition and classification criteria²⁷; interrater reliability exercises were undertaken, and regular quality monitoring is being performed. These differences between centers indicate that despite these efforts within the network, inconsistencies in inclusion and exclusion criteria between centers may exist and show that the clinical diagnosis of this particular subtype may be difficult.

Data indicate that the clinical profile of children with ataxic CP differs from other CP subtypes. Almost 70% of children with ataxic CP were able to walk unaided. However, approximately half of these children had an intellectual impairment, corresponding to an IQ < 70. Altogether, only 20% of these children had a low impairment index (meaning that they were able to walk unaided without any other additional impairment). This clinical profile differs to that of unilateral spastic and bilateral spastic and dyskinetic subtypes. In these subtypes, the association between unaided walking and normal or near-normal intellect proved to be much stronger and was highest for spastic CP.² Severe intellectual impairment was rare in children with spastic and dyskinetic CP who were able to walk unaided (GMFCS levels I–II), that is, 7.7% in comparison with 27% in children with ataxic CP; and among children with spastic and dyskinetic CP who walked with mobility device (GMFCS level III), 25.1% had severe intellectual impairment in comparison with 52.2% among those with ataxic CP. The disability profile more pronounced in terms of cognitive than gross motor dysfunction has been observed also by others when describing children with nonprogressive cerebellar ataxia.^6,7,28 Of interest, the findings of the Canadian register when analyzing the ataxic-hypotonic CP subtype did not show a significant difference to the other CP types in terms of GMFCS or the presence of comorbidities, which might be due to the smaller number or different definition.⁴

Difference to other CP subtypes is most striking when looking at neuroimaging findings. In spastic and dyskinetic CP, neuroimaging patterns indicating a lesional origin were found in almost 70% of cases,⁸ contrasting with the 19% found in children with ataxic CP. Because cerebral white and gray matter injuries do not specifically explain ataxic neurology, this is not surprising. But, when analyzing cerebellar pathology as specified in the categories A and D, it seems of note that it accounted only for approximately 20% of cases. More than half were maldevelopments followed by cerebellar atrophy (8%), whereas lesions were only very rarely reported. Atrophy accounted, in fact, for more than half of the miscellaneous findings and was specified in two-thirds as cerebellar and a third as cerebral atrophy. This means that slowly progressive disorders with early onset may be mimicking ataxic CP in some cases.

The heterogeneity of neuroimaging findings with little lesional pathology explains why in ataxic CP the association of motor severity and additional impairment is rather low, contrary to what is found for other CP subtypes where cerebral lesions, their extent, and topography are related to functional outcome. A more extensive lesion does not only affect the motor system more severely but also risks affecting other domains causing additional impairments (such as cognitive or visual).⁵ This also has important implications for diagnostic workup. The predominance of normal (29%) and miscellaneous (23.5%) patterns indicate rather than an acquired lesional pathology that genetic factors may be in the background. Within the SCPE network, the role of genetic testing has been recently discussed, and recommendations have been given using neuroimaging findings, family history, and missing causative factors to stratify decisions concerning genetic testing. Following these recommendations, a broad genetic workup is indicated in most cases of ataxic CP.²⁵

The discussed characteristics of ataxic CP (disability profile more pronounced in terms of cognitive than gross motor dysfunction, mostly term-born children, a probably mostly genetic origin and little evidence for acquired injuries) are very similar to what has been called nonprogressive congenital or early-onset cerebellar ataxia. It has been described in series of children analyzing their clinical and/or neuroimaging profiles^6,7 or in reports of single cases usually reporting a specific genetic origin^29,30 or allocation to a specific syndrome such as Joubert syndrome.³¹ Some series describe early-onset and later-onset forms,³² and it has been suggested that acquired origin is to be excluded when using the terminology of nonprogressive congenital ataxia (NPCA),³³ which is clearly different from the CP concept. Nevertheless, there is certainly an overlap in series describing children with ataxic CP and series with NPCA, and our findings could stimulate the discussion how to deal with these concepts in the future.

Concerning the aspect of acquired origin, it is increasingly recognized that cerebellar injuries play a role in outcomes for preterm-born children, especially in very premature children.³⁴ Thus, it could have been expected that cerebellar injury contributes to ataxic CP and even a prevalence increase (with increasing survival of these children). Because these injuries occur during cerebellar development, they lead to cerebellar hypoplasia rather than cerebellar atrophy.³⁵ Our findings illustrate that cerebellar pathology can be captured well if free-text information is requested alongside MRICS categorization. This was mostly found in term-born children with ataxic CP, whereas it was rare in children born at <32 weeks gestation. This is in line with the observation that isolated cerebellar lesions in preterm-born children play a role in their cognitive, learning, and behavioral dysfunctions but are not specifically associated with ataxic CP.^36,37

To consider in detail any indicators of genetic background, we analyzed the presence of syndromes and genetic findings. The exhaustivity of these findings is difficult to judge on and is the topic of an ongoing analysis of SCPE data covering all CP subtypes. Despite this fact, the proportion of cases with syndromes described in children with ataxic CP was clearly higher than in the other CP subtypes, that is, 8.7% in comparison with 3.7% in the total CP cohort, as previously published in a linkage study.³⁸ Of interest, children with ataxic CP and diagnosed with a syndrome had a higher impairment load overall, indicating that mainly syndromes with a high functional severity profile were recorded. These results also point toward a genetic origin in this subtype and underline the importance of genetic analysis as discussed earlier in terms of neuroimaging findings.

This analysis of ataxic CP, the least common CP subtype, relies on the largest cohort of children reported up to now. The main results were that prevalence was stable over time, the disability profile was more pronounced in terms of cognitive rather than gross motor dysfunction, and there was a low percentage of preterm-born children. Neuroimaging findings rarely suggested acquired injuries, indicating a high likelihood of genetic etiology.

Thus, ataxic CP is different from the other CP subtypes, where neuroimaging findings indicate a mostly lesional background, leading to an impairment profile with pronounced motor dysfunction. Of note is also that cerebellar injury was not a common finding in very preterm-born children diagnosed with ataxic CP

Ataxic CP turned out to be challenging regarding the reliability of neurologically based diagnosis. Even with the strict criteria that SCPE has introduced and validated, a large variation was observed in the reporting of cases of ataxic CP between registers, suggesting different interpretations of the SCPE guidelines and different approaches regarding syndromic disorders.

Thus, the diagnosis of ataxic CP remains difficult, and further effort has to be undertaken within the SCPE network and beyond to harmonize the approach, for example, by improving guidance regarding associated syndromes. The results suggest that ataxic CP is worth considering separately when analyzing large cohorts of children with CP (although the subtype is so rare that inconsistencies in inclusion and exclusion probably do not affect very much on overall results). The findings could also stimulate discussion on an international basis whether to include ataxic CP within the phenomenologically defined CP concept or rather describe it separately as nonprogressive congenital ataxia.

Glossary

BFMF

bimanual fine motor function

CP

cerebral palsy

GMFCS

Gross Motor Function Classification System

LBs

live births

MACS

Manual Ability Classification System

MRICS

MRI classification system

SCPE

Surveillance of Cerebral Palsy in Europe

VSS

Viking Speech Scale

Appendix. Authors

Name	Location	Contribution
Veronka Horber, MD	Department of Paediatric Neurology, University Children's Hospital Tübingen	Drafting/revision of the article for content, including medical writing for content; study concept or design; analysis or interpretation of data
Guro L. Andersen, MD, PhD	Norwegian Quality and Surveillance Registry for Cerebral Palsy, Vestfold Hospital Trust, Tønsberg	Drafting/revision of the article for content, including medical writing for content; major role in the acquisition of data
Catherine Arnaud, MD	CERPOP, UMR 1295 Toulouse University, Inserm, Paul Sabatier University; Clinical Epidemiology Unit, University Hospital of Toulouse	Drafting/revision of the article for content, including medical writing for content; major role in the acquisition of data
Javier De La Cruz, MD	Imas12, Hospital Universitario 12 de Octubre, RedSAMID, Madrid	Drafting/revision of the article for content, including medical writing for content; analysis or interpretation of data
Ivana Dakovic, MD	Department of Pediatrics, Children's Hospital, University of Zagreb	Major role in the acquisition of data
Andra Greitane, MD	Association Rehabilitation Center, Riga	Major role in the acquisition of data
Owen Hensey, MB, FRCPI	The Central Remedial Clinic, Dublin	Major role in the acquisition of data
Kate Himmelmann, MD, PhD	Department of Pediatrics, Clinical Sciences, Sahlgrenska Academy, University of Gothenburg; Regional Rehabilitation Centre, Queen Silvia Children's Hospital, Gothenburg	Drafting/revision of the article for content, including medical writing for content; major role in the acquisition of data
Katalin Hollody, MD, PhD	Department of Pediatrics, Faculty of Medicine, University of Pecs	Major role in the acquisition of data
Karen Horridge, MB, MSc	Childhood Disability and Development, University of Sunderland	Drafting/revision of the article for content, including medical writing for content; major role in the acquisition of data
Christoph T. Künzle, MD	Zentrum für Kinderneurologie, Entwicklung und Rehabilitation, Ostschweizer Kinderspital, St. Gallen	Major role in the acquisition of data
Marco Marcelli, MD	Developmental age mental health and rehabilitation unit, ASL (local health institution Viterbo), Viterbo	Major role in the acquisition of data
Els Ortibus, MD, PhD	Department of Development and Regeneration, KU Leuven	Major role in the acquisition of data
Antigone Papavasiliou, MD, PhD	Iaso Children's Hospital, Athens	Major role in the acquisition of data
Oliver Perra, MD, PhD	Queen's University Belfast	Major role in the acquisition of data
Mary J. Platt, MD	Norwich Medical School, University of East Anglia, Norwich	Drafting/revision of the article for content, including medical writing for content; major role in the acquisition of data
Gija Rackauskaite, MD, PhD	Department of Pediatrics and Adolescent Medicine, Aarhus University Hospital	Drafting/revision of the article for content, including medical writing for content; major role in the acquisition of data
Solveig Sigurdardottir, MD, PhD	Counselling and Diagnostic Centre	Major role in the acquisition of data
Anja Troha Gergeli, MD	Department of Child and Adolescent & Developmental Neurology, Children´s Hospital, University Medical Centre Ljubljana	Major role in the acquisition of data
Daniel Virella, MD, MScEpi	PVNPC, Programa de Vigilância Nacional da Paralisia Cerebral, Departamento de Epidemiologia, Instituto Nacional de Saúde Doutor Ricardo Jorge, Lisboa	Drafting/revision of the article for content, including medical writing for content; major role in the acquisition of data
Ingeborg Krägeloh-Mann, Prof. Dr.	Department of Paediatric Neurology, University Children´s Hospital Tübingen	Drafting/revision of the article for content, including medical writing for content; study concept or design; and analysis or interpretation of data
Elodie Sellier, MD, PhD	Grenoble Alpes University, CNRS, Grenoble INP, CHU Grenoble Alpes, TIMC-IMAG, Grenoble; Registre des Handicaps de l'Enfant et Observatoire Périnatal, Grenoble	Drafting/revision of the article for content, including medical writing for content; major role in the acquisition of data; and analysis or interpretation of data

Study Funding

University of Tübingen (TÜFF Nr. 2622-0-0).

Disclosure

V. Horber received funding for this project from the University of Tübingen (TÜFF Nr. 2622-0-0). G.L. Andersen, C. Arnaud, J. de la Cruz, I. Dakovic, A. Greitane, O. Hensey, K. Himmelman, K. Hollódy, K. Horridge, C.T. Künzle, M. Marcelli, E. Ortibus, A. Papavasiliou, O. Perra, M.J. Platt, G. Rackauskaite, S. Sigurdardottir, A. Troha Gergeli, D. Virella, I. Krägeloh-Mann, and E. Sellier report no disclosures relevant to the manuscript. Go to Neurology.org/N for full disclosures.