Figure 1. Flowchart of the 226 patients with autosomal dominant spinocerebellar ataxia or spastic paraplegia followed up for 36 months.
Figure 2. Disease evolution according to the genotype and the instrument used for assessment (Composite Cerebellar Functional Severity Score including the writing test [CCFSw] or Scale for the Assessment and Rating of Cerebellar Ataxia [SARA]). A and B, Quantitative assessment with CCFSw. Patients with polyglutamine (polyQ) SCA1, 2, and 3 showed continuous disease worsening, patients with polyQ SCA6 and 7 and the other SCA subtypes remained stable, and patients with spastic paraplegia type 4 (SPG4) showed improvement. C and D, Semiquantitative assessment with SARA. Disease worsened in all patients except those with SCA6, who remained stable. A and C, Evolution of patients with autosomal dominant cerebellar ataxia and spastic paraplegias according to genotype. B and D, Evolution of patients with polyQ SCA (SCA1, 2, 3, 6, and 7).
Figure 3. Sample sizes needed to detect a treatment-induced reduction of 30% to 100% in the natural rate of disease progression. Sample sizes for a 2-arm trial (sample size per trial arm) using the Composite Cerebellar Functional Severity Score including the writing test (CCFSw) (solid lines) and Scale for the Assessment and Rating of Cerebellar Ataxia (SARA) (dotted lines) are estimated for a power of 80% (α = .05). A, Analysis according to genotype: spinocerebellar ataxia type 1 (SCA1) (black), SCA2 (blue), and SCA3 (green). B, Analysis according to the presence of parkinsonism and/or dystonia (red) or their absence (blue) in patients with SCA3 only.
Tezenas du Montcel S, Charles P, Goizet C, et al. Factors influencing disease progression in autosomal dominant cerebellar ataxia and spastic paraplegia. Arch Neurol. 2012;69(4):500-508.
eTable. Clinical characteristics and disease severity at inclusion for autosomal dominant cerebellar ataxia patients with PolyQ mutations.
This supplementary material has been provided by the authors to give readers
additional information about their work.
Tezenas du Montcel S, Charles P, Goizet C, Marelli C, Ribai P, Vincitorio C, Anheim M, Guyant-Maréchal L, Le Bayon A, Vandenberghe N, Tchikviladzé M, Devos D, Le Ber I, N’Guyen K, Cazeneuve C, Tallaksen C, Brice A, Durr A. Factors Influencing Disease Progression in Autosomal Dominant Cerebellar Ataxia and Spastic Paraplegia. Arch Neurol. 2012;69(4):500-508. doi:10.1001/archneurol.2011.2713
Author Affiliations: Department of Biostatistics and Medical Informatics and Pitié-Salpêtrière Charles-Foix Clinical Research Unit (Dr Tezenas du Montcel), Department of Genetics and Cytogenetics, Medical Genetics Unit (Drs Charles, Marelli, Ribai, Vincitorio, Anheim, Tchikviladzé, Brice, and Durr), Department of Neurology (Drs Charles, Tchikviladzé, Le Ber, Tallaksen, and Brice), and Department of Genetics and Cytogenetics, Molecular and Cellular Neurogenetics Unit (Dr Cazeneuve), Pitié-Salpêtrière Hospital, Assistance Publique–Hôpitaux de Paris (AP-HP), UMR-S975, Centre de Recherche de l'Institut du Cerveau et de la Moelle épinière (Drs Marelli, Le Ber, Tallaksen, Brice, and Durr) and ER4, Modelling in Clinical Research (Dr Tezenas du Montcel), Université Pierre et Marie Curie–Paris6, Inserm, U975 (Drs Marelli, Le Ber, Tallaksen, Brice, and Durr), Cnrs, UMR 7225 (Drs Marelli, Le Ber, Tallaksen, Brice, and Durr), Paris, Department of Medical Genetics, Pellegrin Hospital, Bordeaux CHU Hospitals, and Laboratory of Human Genetics, University Victor Segalen Bordeaux 2, Bordeaux (Dr Goizet), Service of Neurology, Carémeau Hospital, CHU Nimes, Nimes (Dr Le Bayon), Departments of Neurology and Genetics and Inserm Unit 614, Rouen University Hospital, Rouen (Dr Guyant-Maréchal), Service of Neurology C, Pierre Wertheimer Neurological Hospital, Lyon Hospitals, Lyon (Dr Vandenberghe), Department of Neurology, EA4559, IFR114 IMPRT, Université Lille Nord de France, Faculté Lille 2, CHU Lille, Lille (Dr Devos), and Department of Genetics, La Timone Hospital, Marseille Hospitals, Marseille (Dr N’Guyen), France.
Objectives To evaluate disease progression and determine validity of clinical tools for therapeutic trials.
Design Prospective cohort study (36 months).
Setting Referral center.
Patients One hundred sixty-two patients with autosomal dominant cerebellar ataxia and 64 with hereditary spastic paraplegia.
Main Outcome Measures The quantitative Composite Cerebellar Functional Severity Score with the writing test (CCFSw) and Scale for the Assessment and Rating of Ataxia (SARA) score.
Results Disease worsened in patients with SCA1, SCA2, and SCA3 mutations (mean [SE] increase in CCFSw, +0.014 [0.005] to +0.025 [0.004] per year), improved in patients with SPG4 mutations (mean [SE] increase in CCFSw, −0.012 [0.003] per year; P = .02), and remained stable in patients with SCA6, SCA7, or other SCA mutations (mean [SE] increase in CCFSw, −0.015 [0.011] to +0.009 [0.013] per year) or hereditary spastic paraplegia with other SPG mutations (mean [SE] increase in CCFSw, −0.005 [0.005] per year). Progression was faster in patients with SCA2 mutations and normal alleles with 22 or fewer repeats (P = .02) and in patients with SCA3 mutations with parkinsonism and/or dystonia at baseline (P = .003). Whereas CCFSw distinguished between patients with ataxia and spasticity, SARA scores increased in both groups. A 2-arm trial with SARA score as the outcome measure would require 57 patients with SCA2 mutations, 70 with SCA1 mutations, and 75 with SCA3 mutations per group to detect a 50% reduction in disease progression (power, 80%; α = .05).
Conclusions Disease progressed faster in SCA s with polyglutamine expansions in SCA1, 2, and 3 than the other groups. Both outcome measures are suitable for therapeutic trials; SARA requires fewer patients to attain the same power, but CCFSw needs less stratification. We demonstrate that the choice of clinical outcome measure is critical for reliable evaluation of progression in neurodegenerative diseases.
Trial Registration clinicaltrials.gov Identifier: NCT00136630
Autosomal dominant cerebellar ataxias (SCA s) are neurodegenerative diseases that are clinically and genetically heterogeneous. The core syndrome is cerebellar ataxia, but impairment of other neurological functions contributes to the phenotype.1 Cerebellar degeneration is often accompanied by changes in the brainstem, basal ganglia, cerebral cortex, spinal cord, and peripheral nervous system.2- 6
Several semiquantitative tools have been developed to assess the clinical picture of these diseases. The Scale for the Assessment and Rating of Cerebellar Ataxia (SARA)7 is a semiquantitative scale that is easy to use in a multicenter setting; the Neurological Examination Score for Spinocerebellar Ataxia, validated in patients with SCA3 mutations, evaluates the general neurological burden in addition to ataxia.8 Quantitative scales, such as the Composite Cerebellar Functional Severity Score (CCFS)9,10 and the SCA Functional Index,11 have also been developed.
To identify a clinical outcome measure appropriate for clinical trials in these rare disorders, we have performed a prospective, multisite observational study of SCA s with clinical scales. To control for the disease specificity of these scales, patients with SCA s were compared with patients with autosomal dominant spastic paraplegias (HSPs), nonataxic movement disorders with very slow disease progression and a phenotype that rarely spreads to other neurological systems over time. Like the ataxias, these diseases are rare and genetically heterogeneous, although there is a frequent genetic form of HSP, spastic paraplegia type 4 (SPG4).12- 14 However, natural history studies are as yet unavailable for these diseases or have been based on cross-sectional data.
Patients recruited from 7 study sites in France (Paris, n = 173; Bordeaux, n = 25; Nimes, n = 10; Rouen, n = 8; Lyon, n = 6; Lille, n = 2; and Marseille, n = 2) were studied at baseline and followed up on a yearly basis with 2 follow-up visits, except in Paris, where some patients had 3 follow-up visits. To determine whether there was a center effect, we compared patients from Paris with patients from the other centers.
Patients were included between April 2004 and February 2008 if they (1) had a family history compatible with dominant transmission; (2) were older than 18 years; or (3) had cerebellar ataxia (SCA) or spastic paraplegia (patients with HSP). Data collection was described elsewhere.9 Patients who missed a visit were contacted by telephone to determine why and to plan the following visits. The study was approved by the local ethics committee (Pitié-Salpêtrière University Hospital, Paris, France), and written informed consent was obtained.
Two scales were used, the quantitative CCFS with the writing test (CCFSw)10 and SARA.7
The CCFSw includes 3 functional tests of the dominant hand: (1) the 9-hole pegboard test (time needed to place dowels in 9 holes); (2) the click test (time needed to perform 10 finger-pointing cycles); and (3) the writing test (time needed to write a standard sentence as fast as possible but legibly). Times were transformed into z scores expressed as the difference between observed and expected times for a given age. The CCFSw was calculated using the following formula:
where z Pegboard D = Pegboard D − (13.4 − 0.16 × Age + 0.002 × Age2), z Click D = Click D − (8 + 0.05 × Age), and z Writing D = Writing D − (8.5 + 0.05 × Age). The mean (SD) normal value of the CCFSw is 0.85 (0.05) (range, 0.65-0.96).10
The SARA has 8 items, for a total score of 0 (no ataxia) to 40 (most severe ataxia): gait, stance, sitting, speech disturbance, finger chase, nose-finger test, fast alternating hand movements, and heel-shin slide. Limb kinetic functions are rated independently on both sides; the arithmetic mean is included in the total SARA score. The mean (SD) value in controls is 0.4 (1.1) (range, 0-7.5).7
An ambulatory score (AMBUS score) was used to evaluate functional impairment. The AMBUS evaluated the distance (in meters) walked during 5 seconds with (score A) or without (score B) help (score 0 = 5 m walking distance, 1 = 4 m, 2 = 3 m, 3 = 2 m, 4 = 1 m, and 5 = 0 m). Both scores were summed (maximum best = 0, maximum worse = 10). A functional index (AMBUS index) was calculated by dividing the AMBUS score by disease duration.
System-specific groups of signs were noted: pyramidal signs (presence of pyramidal signs with patellar hyperreflexia, bilateral extensor plantar reflex, and/or lower limb or gait spasticity); posterior column dysfunction (decreased vibration sense at ankles, score ≤5 of 8); ophthalmoplegia (ophthalmoplegia and/or double vision); extrapyramidal signs (parkinsonism and/or dystonia); dysphagia; and cognitive alterations.
Patients with SCA (n = 162) and HSP (n = 64) were divided into 4 subgroups: polyglutamine SCA s (25 with SCA1, 35 with SCA2, 58 with SCA3, 5 with SCA6, and 10 with SCA7 mutations); other SCA s (1 with SCA5, 1 with SCA14, 2 with SCA21, 1 with SCA25, and 2 with SCA28 mutations and 22 unknown genotypes); SPG4 -linked HSP (n = 31); and other HSP (4 with SPG3 mutations and 29 unknown genotypes). Repeat lengths in the patients with polyglutamine SCA s were taken from their medical records.
Characteristics of the patients at inclusion were analyzed by analysis of variance followed by pairwise comparisons using the Tukey-Kramer correction for multiple testing (Pc) and the Pearson χ2 or Fisher exact test when appropriate.
Severity at baseline was first compared among the SCA and HSP groups, then among the polyglutamine SCA subgroups. To determine which factors influenced severity at baseline, univariate and multivariate linear regression analyses were performed.
Disease progression in the SCA and HSP groups and then in the polyglutamine SCA subgroups were compared using mixed models (analysis of variance with random effect) with a random effect for patients and 2 fixed effects (group and time between inclusion and clinical examination); the interaction between these 2 fixed factors was estimated to determine whether the slope of disease progression differed significantly among groups. When significant effects were found, pairs of means were compared using a Tukey-Kramer correction for the P values. To determine which factors influenced disease progression, univariate mixed models were performed. The interaction between the factor and the time between inclusion and clinical examination tested the influence of the factor on the disease progression. A multivariate analysis with backward selection was performed including the significant factors in the univariate analysis.
The factors at inclusion tested for their influence on severity and disease progression were sex, ages at onset and inclusion, disease duration, number of repeats on the expanded and normal alleles, the AMBUS score and index, and the presence of the system indicators indicated earlier.
Because of small sample size, univariate and multivariate analyses were performed only for patients with SCA1, 2, and 3 and SPG4 mutations. All reported P values are 2-tailed with a 5% type I error rate. Analyses were performed with SAS version 9.2 (SAS Institute Inc). Quantitative variables are expressed as mean and standard deviation except for analysis of variance results, which are expressed as mean and standard error.
During the first 24 months following inclusion, 38 patients (17%) dropped out. Follow-up was stopped after 2 years, except in Paris, where 99 patients had a third follow-up visit (Figure 1). Since there was no difference in clinical severity at baseline, a possible center effect was ignored (data not shown). Six hundred eighty-nine evaluations were performed: 79 patients had 4; 107, three; 25, two; and 15, one. The mean (SD) follow-up was 2.4 (0.7) years (range, 0.9-3.8 years) in all groups. No differences were found in clinical severity at baseline (data not shown) between patients with 2 and 3 follow-up visits; restricting evolution analysis to 2 follow-up visits only did not modify the results but led to a lack of statistical power. Thus, all visits were retained for analysis.
Thirteen patients died, 2 were hospitalized because of worsening disease, 5 moved away, 47 were lost to follow-up, and 11 withdrew consent. All deceased patients were from the polyglutamine SCA group: SCA1, 5 of 25; SCA2, 2 of 35; SCA3, 4 of 58; and SCA7, 2 of 10. They died at a mean (SD) age of 44 (16) years.
Patients with polyglutamine SCA s were older at onset and had the shortest mean disease durations (Table 1). The AMBUS scores were similar, but the AMBUS index tended to be higher in patients with polyglutamine SCA s than in patients with other SCA s (P = .01 between the 4 groups, Pc = .07 between patients with polyglutamine SCA s and patients with other SCA s), indicating faster disease progression. Severity did not differ significantly among patients with polyglutamine SCA s (eTable).
In patients with SCA1 mutations, independent factors associated with higher CCFSw scores were the AMBUS score (mean [SE], +0.027 [0.009] per unit increase; P < .001) and the presence of posterior column dysfunction (mean [SE], +0.307 [0.087] when present; P = .005) (univariate analysis results are reported in Table 2). In patients with SCA2 mutations, the AMBUS score (mean [SE], +0.045 [0.010] per unit increase; P < .001) and the number of repeats on the expanded allele (mean [SE], +0.0167 [0.006] per additional repeat; P = .02) were independently associated with higher CCFSw. In patients with SCA3 mutations, disease severity increased independently of the AMBUS score (mean [SE], +0.033 [0.005] per unit increase; P < .001) and the presence of parkinsonism and/or dystonia (mean [SE], +0.110 [0.039] when present; P = .007).
In patients with SCA1 mutations, the AMBUS score (mean [SE], +1.8 [0.3] per unit increase; P < .001) and disease duration (mean [SE], +0.5 [0.6] per additional year of evolution; P = .01) were associated independently with severity. In patients with SCA2 mutations, the only factor associated with severity was the AMBUS score (mean [SE], +1.5 [0.2] per additional year; P < .001), suggesting that ambulatory function, not disease duration, was correlated with severity. In patients with SCA3 mutations, the SARA score was explained by the AMBUS score (mean [SE], +1.3 [0.1] per unit increase; P < .001) and disease duration (mean [SE], +0.3 [0.1] per additional year of evolution; P = .002). In patients with SPG4 mutations, the only factor associated with severity was the AMBUS score (mean [SE], +0.9 [0.1] per additional year; P < .001), suggesting that ambulatory function, not disease duration, was correlated with severity.
The CCFSw evolved differently over time depending on the genotype. Three patterns of evolution were seen (Figure 2A): patients with polyglutamine SCA s worsened (mean [SE], +0.017 [0.003] per year; P < .001) and patients with other SCA s (mean [SE], +0.008 [0.006] per year; P = .17) and patients with other SPG s (mean [SE], −0.005 [0.005] per year; P = .37) remained stable; patients with SPG4 improved (mean [SE], −0.012 [0.003] per year; P = .01). Among patients with polyglutamine SCA s (P = .01), patients with SCA1 (mean [SE], +0.022 [0.007] per year; P = .001), SCA2 (mean [SE], +0.014 [0.005] per year; P = .008), and SCA3 (mean [SE], +0.025 [0.004] per year; P < .001) mutations worsened; patients with SCA6 (mean [SE], −0.015 [0.011] per year; P = .17) and SCA7 (mean [SE], +0.009 [0.013] per year; P = .46) mutations remained stable (Figure 2B).
The evolution of SARA scores differed among the patient groups (P < .001) (Figure 2C). Scores of patients with polyglutamine SCA s (mean [SE], +1.5 [0.1] per year) changed more rapidly than the scores of other SCA patient groups (mean [SE], +0.6 [0.2] per year; P < .001). Among the patients with polyglutamine SCA s, evolution differed significantly according to the genotype (P = .03) (Figure 2D). While scores of patients with SCA6 did not change significantly during the study (mean [SE], +0.4 [0.4] per year; P = .42), the patients with other polyglutamine SCA s worsened: mean (SE), SCA1, 1.8 (0.3) per year; P < .001; SCA2, 1.3 (0.2) per year; P < .001; SCA3, 1.7 (0.2) per year, P < .001; and SCA7, 1.6 (0.4) per year; P < .001. The SARA scores also changed significantly in the patients with noncerebellar HSP: mean (SE), patients with SPG4, +0.4 (0.2) per year; P < .001; patients with other SPG s, +0.2 (0.2) per year; P < .001.
Neither sex, age at onset, nor the number of CAG repeats on either allele of the polyglutamine genes were significantly associated with changes in the CCFSw (Table 2). Only the evolution of patients with SCA3 mutations was influenced by parkinsonism and/or dystonia present at baseline. One-third of the patients had parkinsonism and/or dystonia at baseline. The disease progressed faster (P = .003) in these patients (mean [SE] annual increase in CCFSw per unit increase, 0.051 [0.010]) than in patients without these signs (mean [SE] annual increase in CCFSw per unit increase, 0.019 [0.005]).
We predict that a 2-arm trial using the CCFSw as the outcome measure would require 183 patients with SCA2, 193 patients with SCA1, and 218 patients with SCA3 mutations per group to detect a 50% reduction in disease progression (power 80%; α = .05) (Figure 3A). A trial on patients with SCA3 mutations stratified on the presence of parkinsonism and/or dystonia would need to include 213 patients with SCA3 mutations per group without and 81 with these symptoms (Figure 3B).
The evolution of patients with SCA1 mutations was influenced by the presence of pyramidal signs at baseline. The disease progressed more slowly (P = .01) in those with pyramidal signs (mean [SE] annual increase in SARA score, 1.7 [0.2]) than in patients without these signs (mean [SE] annual increase in SARA score, 3.2 [0.6]).
For patients with SCA 2 mutation, disease evolved more rapidly in men (mean [SD] annual increase in SARA score, +0.8 [0.4] in comparison with women; P = .05), patients who were younger at onset (mean [SE] annual increase in SARA score, +0.05 [0.02] per each year younger; P = .001) and at inclusion (mean [SE] annual increase in SARA score, +0.06 [0.02] per each year younger; P < .001), and patients without posterior column dysfunction (mean [SE] annual increase in SARA score, +1.0 [0.4] in comparison with patients with these signs; P = .01) (Table 3). Interestingly, disease in patients with SCA2 with normal alleles with 22 or fewer repeats progressed faster (mean [SE] annual increase in SARA score, +1.3 [0.6]) than in those with alleles with more than 22 repeats (P = .02). In the multivariate analysis, the only significant factor was the age at inclusion. Indeed, patients with SCA2 without posterior column dysfunction were younger at inclusion than patients with these signs (mean [SD] age, 40.3 [11.4] vs 55.0 [8.2] years; P < .001); patients who were younger at onset were also younger at inclusion (correlation coefficient = 0.73; P < .001). Sex and the number of repeats seemed not to affect age at inclusion, although male patients and those with alleles with 22 or fewer repeats on the normal allele tended to be younger.
We predict that a 2-arm trial to detect a 50% reduction in disease progression (power 80%; α = .05) using the SARA score as the outcome measure would require 57 patients with SCA2, 70 patients with SCA1, and 75 patients with SCA3 mutations per group (Figure 3A).
In view of future therapeutic trials including patients with SCA or HSP, we evaluated 2 rating scales for cerebellar involvement reflecting disease severity and progression, the CCFSw9,10 and SARA.7,15 We showed that both clinical measures are suitable for tracking cerebellar disease progression. The CCFSw was more specific for follow-up of ataxia symptoms, whereas SARA reflected the progression of the disease as a whole. The CCFSw also changed in patients with noncerebellar spastic paraplegias, who improved or remained stable when measured with this scale.
Our study has certain limitations. Observer biases are a risk in all epidemiological studies. The CCFSw, a quantitative tool in which the time required to perform the functional tests is recorded, was developed to avoid observer subjectivity9; it also offers high test-retest reliability.10 Regarding a possible selection bias, 77% of our patients were included in the Paris center, while the remaining 23% were from other centers. In Paris, the patients were evaluated by 9 different clinicians, whereas only 1 clinician evaluated the patients in each of the other centers. The evaluations in the Paris center may, therefore, be more heterogeneous. However, the Paris center is a reference center for SCA s and HSPs and is thus more attractive to patients and their families; only 40% of the Paris patients came from the Paris region. At inclusion, the disease severity of patients recruited in Paris did not differ from the disease severity of patients recruited elsewhere, making a selection bias unlikely. The number of follow-up visits also differed among patients, either because they were examined in Paris, where the follow-up visits were more numerous, or because of dropout. Because restricting analysis to the first 2 follow-up visits only did not modify the significance of the results but decreased the statistical power of the analysis, we retained the results of the third follow-up visit in the analysis. The dropout rate of 17% at 24 months was in line with comparable previous studies such as the European Integrated Project on Spinocerebellar Ataxias (EUROSCA) cohort (21% at the 2-year follow-up visit)16 or lower than in a follow-up study of patients with SCA3.8
The only factor that influenced CCFSw-measured disease progression was the presence of parkinsonism and/or dystonia, which accelerated disease progression in patients with SCA3. These symptoms are not rare in SCA3 ; one-third of the patients in the current study had them at baseline. A clinical trial including patients with SCA3 with CCFSw as the outcome measure would, therefore, need a population that is homogeneous with regard to these symptoms or have to take them into account by stratifying by or adjusting for their presence or absence. The SARA score, on the contrary, was influenced by several characteristics at baseline.
In this study, we introduced a new functional test designed to measure the ability to walk: the AMBUS index. This index needs to be validated before use as an outcome in trials. The AMBUS has 2 advantages over the 25-ft walk test in the multiple sclerosis functional composite17: it allows evaluation, in the examination room, of the ability to walk; it also distinguishes between patients who walk with and without help. Since measuring walking capacity has more weight than disease duration, one of these tests, AMBUS or the 25-ft walk, should definitely be included as a clinical outcome measurement.
We have shown that the AMBUS score influences disease severity measured by both the CCFSw and the SARA. It is not surprising that a scale that measures walking ability determines the SARA score, in which gait is a major component, but it does not account for the link between the AMBUS score and the CCFSw. Furthermore, in patients with SPG4, the AMBUS score influenced disease severity measured by SARA but not the CCFSw. We hypothesize that the same factors probably influence the lower and upper limbs in SCA diseases and that the CCFSw reflects not only disease-caused dysfunction in the upper limbs but the disease as a whole.
Disease progression was faster in polyglutamine SCA s compared with other SCA s or HSPs. Among the polyglutamine SCA group, patients with SCA1, 2, and 3, but not SCA6 and 7, clearly worsened over time. The difference could be because fewer patients with SCA6 and SCA7 were followed up; however, SCA6 has already been shown to differ from SCA1, 2, and 3 in that its progression is determined more by age than by disease-related factors.15 In the EUROSCA study, evolution of patients with SCA6 during the first year was similar to ours, when averaged on a 3-year follow-up.16 For patients with SCA1, 2, and 3, the annual increase in the SARA score in our study was similar to what was found in the EUROSCA study.16
Few clinical treatment trials in patients with SCA have been published. A recent study showed an effect of riluzole, with a 5-point change in the International Cooperative Ataxia Rating Scale score after only 8 weeks.18 However, patients with SCA s of different etiologies were included in the study. Our study shows that differences in evolution are observed even among the dominant forms of ataxia.
Several factors have already been shown to influence the appearance of clinical signs in polyglutamine SCA s. An evident role has been attributed to the number of CAG repeat expansions in the responsible gene1,19- 21; larger repeats and earlier age at onset in patients with SCA3 were associated with spasticity and hyperreflexia; and long expanded repeats and early age at onset in patients with SCA2 increased the likelihood of brainstem oculomotor signs, muscle atrophy, and hyperkinetic movement disorders.15 In contrast, their effect on progression and survival was less evident and inconstant. No effect of the repeat expansion on disease progression was found using the International Cooperative Ataxia Rating Scale score as the outcome,22 but faster disease progression with larger repeats was found using the Neurological Examination Score for Spinocerebellar Ataxia in patients with SCA38 and SARA in patients with SCA2.16 Polyglutamine length had little effect in our study, possibly because of the small number of patients per group; CAG repeat size influenced SARA-measured disease progression and CCFSw-measured disease severity only in patients with SCA2. Intermediate SCA2 alleles have been associated with other conditions, such as amyotrophic lateral sclerosis and Parkinson disease, and may play a role in neurodegeneration in general.23,24 This suggests that the CAG repeat length determines the phenotypical expression of the disease but has less influence on its evolution.
Disease progression was faster in patients with polyglutamine expansions in SCA1, 2, and 3 than in the other groups of patients studied. It was influenced in patients with SCA3 by the presence of extrapyramidal signs when measured by CCFSw. When measured by SARA, it was affected by several baseline characteristics: sex, repeat length, age at inclusion, and posterior column dysfunction in patients with SCA2 and pyramidal signs in patients with SCA1. Both CCFSw and SARA are, therefore, suitable clinical outcome measures for future therapeutic trials. Fewer patients would be needed if SARA is used, but the use of the CCFSw entails less stratification. The study demonstrates that the choice of the clinical outcome measure is critical for the reliable evaluation of progression in neurodegenerative diseases.
Correspondence: Alexandra Durr, MD, PhD, Cricm, UPMC, Inserm UMR_S975/CNRS UMR 7225, ICM Bldg, 1st floor, Hôpital Pitié-Salpêtrière, 47 boulevard de l’Hôpital, 75013 Paris, France (firstname.lastname@example.org).
Accepted for Publication: October 20, 2011.
Author Contributions: Drs Tezenas du Montcel and Durr had full access to all the data and attest to the accuracy of the data analysis. Study concept and design: Tezenas du Montcel, Brice, and Durr. Acquisition of data: Charles, Goizet, Marelli, Ribai, Vincitorio, Anheim, Guyant-Maréchal, Le Bayon, Vandenberghe, Tchikviladzé, Devos, Le Ber, N’Guyen, Cazeneuve, and Tallaksen. Analysis and interpretation of data: Tezenas du Montcel, Devos, Cazeneuve, and Brice. Drafting of the manuscript: Tezenas du Montcel, Goizet, Ribai, Vincitorio, Le Bayon, Tchikviladzé, N’Guyen, and Durr. Critical revision of the manuscript for important intellectual content: Tezenas du Montcel, Charles, Goizet, Marelli, Anheim, Guyant-Maréchal, Vandenberghe, Devos, Le Ber, Cazeneuve, Tallaksen, Brice, and Durr. Statistical analysis: Tezenas du Montcel. Obtained funding: Ribai, Tchikviladzé, and Durr. Administrative, technical, and material support: Anheim, Le Bayon, Cazeneuve, and Tallaksen. Study supervision: Tezenas du Montcel.
Financial Disclosure: None reported.
Funding/Support: This study was funded by Programme Hospitalier de Recherche Clinique National grant AOM 03059 (Assistance Publique–Hôpitaux de Paris) (Dr Durr).
Additional Contributions: We sincerely thank all the patients for their participation in the study. We are grateful to Merle Ruberg, PhD, for critical reading of the manuscript. For help with the logistics of the study and data handling, many thanks to Sandra Benaïch, Lydia Guennec, Celine Jauffret, and Sylvie Forlani (Université Pierre et Marie Curie–Paris 6, Centre de Recherche de l'Institut du Cerveau et de la Moelle épinière, UPMC, Inserm UMR-S975/CNRS UMR 7225).