[Skip to Content]
Access to paid content on this site is currently suspended due to excessive activity being detected from your IP address 54.147.238.168. Please contact the publisher to request reinstatement.
Sign In
Individual Sign In
Create an Account
Institutional Sign In
OpenAthens Shibboleth
[Skip to Content Landing]
Download PDF
Figure 1.
Survival Curve in 133 Patients With Friedreich Ataxia
Survival Curve in 133 Patients With Friedreich Ataxia

The mean age at death was 39 years, and survival was 88.7% (118 of 133) after 10 years of follow-up.

Figure 2.
Trajectory Groups of Left Ventricular Ejection Fraction During Follow-up in 103 Patients With Friedreich Ataxia
Trajectory Groups of Left Ventricular Ejection Fraction During Follow-up in 103 Patients With Friedreich Ataxia

Trajectories of left ventricular ejection fraction (LVEF) during follow-up (bold lines) were calculated using the model’s coefficient estimates and 95% CIs (dotted lines). Two cardiac evolutions were distinguished, including a low-risk cardiac group (78.6% [81 of 103]), with a normal ejection fraction at baseline that declined slightly over time but remained in the normal range, and a high-risk cardiac group (21.4% [22 of 103]), which started with a lower ejection fraction and had a progressively decreasing ejection fraction.

Table 1.  
Clinical and Echocardiographic Characteristics at Baseline of 133 Patients Having Friedreich Ataxia (FRDA) With and Without Cardiac Hypertrophy
Clinical and Echocardiographic Characteristics at Baseline of 133 Patients Having Friedreich Ataxia (FRDA) With and Without Cardiac Hypertrophy
Table 2.  
Survival Analysis in Friedreich Ataxia and Factors Influencing Survivala
Survival Analysis in Friedreich Ataxia and Factors Influencing Survivala
Table 3.  
Characteristics of Trajectory Groups Among 103 Patients Having Friedreich Ataxia With at Least 2 Follow-up Visitsa
Characteristics of Trajectory Groups Among 103 Patients Having Friedreich Ataxia With at Least 2 Follow-up Visitsa
1.
Harding  AE, Hewer  RL.  The heart disease of Friedreich’s ataxia: a clinical and electrocardiographic study of 115 patients, with an analysis of serial electrocardiographic changes in 30 cases. Q J Med. 1983;52(208):489-502.
PubMed
2.
Durr  A, Cossee  M, Agid  Y,  et al.  Clinical and genetic abnormalities in patients with Friedreich’s ataxia. N Engl J Med. 1996;335(16):1169-1175.
PubMedArticle
3.
Raman  SV, Phatak  K, Hoyle  JC,  et al.  Impaired myocardial perfusion reserve and fibrosis in Friedreich ataxia: a mitochondrial cardiomyopathy with metabolic syndrome. Eur Heart J. 2011;32(5):561-567.
PubMedArticle
4.
Tsou  AY, Paulsen  EK, Lagedrost  SJ,  et al.  Mortality in Friedreich ataxia. J Neurol Sci. 2011;307(1-2):46-49.
PubMedArticle
5.
Campuzano  V, Montermini  L, Moltò  MD,  et al.  Friedreich’s ataxia: autosomal recessive disease caused by an intronic GAA triplet repeat expansion. Science. 1996;271(5254):1423-1427.
PubMedArticle
6.
Pandolfo  M, Pastore  A.  The pathogenesis of Friedreich ataxia and the structure and function of frataxin. J Neurol. 2009;256(suppl 1):9-17.
PubMedArticle
7.
Reetz  K, Dogan  I, Costa  AS,  et al.  Biological and clinical characteristics of the European Friedreich’s Ataxia Consortium for Translational Studies (EFACTS) cohort: a cross-sectional analysis of baseline data. Lancet Neurol. 2015;14(2):174-182.
PubMedArticle
8.
Isnard  R, Kalotka  H, Dürr  A,  et al.  Correlation between left ventricular hypertrophy and GAA trinucleotide repeat length in Friedreich’s ataxia. Circulation. 1997;95(9):2247-2249.
PubMedArticle
9.
Filla  A, De Michele  G, Cavalcanti  F,  et al.  The relationship between trinucleotide (GAA) repeat length and clinical features in Friedreich ataxia. Am J Hum Genet. 1996;59(3):554-560.
PubMed
10.
Lodi  R, Rajagopalan  B, Blamire  AM,  et al.  Cardiac energetics are abnormal in Friedreich ataxia patients in the absence of cardiac dysfunction and hypertrophy: an in vivo 31P magnetic resonance spectroscopy study. Cardiovasc Res. 2001;52(1):111-119.
PubMedArticle
11.
Lamarche  J, Shapcott  D, Côté  M, Lemieux  B. Cardiac iron deposits in Friedreich’s ataxia. In: Lechtenberg  R, ed. Handbook of Cerebellar Diseases. Boca Raton, FL: CRC Press; 1993:453-457.
12.
Michael  S, Petrocine  SV, Qian  J,  et al.  Iron and iron-responsive proteins in the cardiomyopathy of Friedreich’s ataxia. Cerebellum. 2006;5(4):257-267.
PubMedArticle
13.
Rötig  A, de Lonlay  P, Chretien  D,  et al.  Aconitase and mitochondrial iron-sulphur protein deficiency in Friedreich ataxia. Nat Genet. 1997;17(2):215-217.
PubMedArticle
14.
Perdomini  M, Belbellaa  B, Monassier  L,  et al.  Prevention and reversal of severe mitochondrial cardiomyopathy by gene therapy in a mouse model of Friedreich’s ataxia. Nat Med. 2014;20(5):542-547.
PubMedArticle
15.
Trouillas  P, Takayanagi  T, Hallett  M,  et al; Ataxia Neuropharmacology Committee of the World Federation of Neurology.  International Cooperative Ataxia Rating Scale for pharmacological assessment of the cerebellar syndrome. J Neurol Sci. 1997;145(2):205-211.
PubMedArticle
16.
Schmitz-Hübsch  T, du Montcel  ST, Baliko  L,  et al.  Scale for the Assessment and Rating of Ataxia: development of a new clinical scale. Neurology. 2006;66(11):1717-1720.
PubMedArticle
17.
Lang  RM, Bierig  M, Devereux  RB,  et al; Chamber Quantification Writing Group; American Society of Echocardiography’s Guidelines and Standards Committee; European Association of Echocardiography.  Recommendations for chamber quantification: a report from the American Society of Echocardiography’s Guidelines and Standards Committee and the Chamber Quantification Writing Group, developed in conjunction with the European Association of Echocardiography, a branch of the European Society of Cardiology. J Am Soc Echocardiogr. 2005;18(12):1440-1463.
PubMedArticle
18.
Ganau  A, Devereux  RB, Roman  MJ,  et al.  Patterns of left ventricular hypertrophy and geometric remodeling in essential hypertension. J Am Coll Cardiol. 1992;19(7):1550-1558.
PubMedArticle
19.
Henry  WL, Gardin  JM, Ware  JH.  Echocardiographic measurements in normal subjects from infancy to old age. Circulation. 1980;62(5):1054-1061.
PubMedArticle
20.
Satterthwaite  FE.  Synthesis of variance . Psychometrika. 1941;6:309-316.Article
21.
Nagin  DS, Odgers  CL.  Group-based trajectory modeling in clinical research. Annu Rev Clin Psychol. 2010;6:109-138.
PubMedArticle
22.
Leone  M, Rocca  WA, Rosso  MG, Mantel  N, Schoenberg  BS, Schiffer  D.  Friedreich’s disease: survival analysis in an Italian population. Neurology. 1988;38(9):1433-1438.
PubMedArticle
23.
Mottram  PM, Delatycki  MB, Donelan  L, Gelman  JS, Corben  L, Peverill  RE.  Early changes in left ventricular long-axis function in Friedreich ataxia: relation with the FXN gene mutation and cardiac structural change. J Am Soc Echocardiogr. 2011;24(7):782-789.
PubMedArticle
24.
St John Sutton  M, Ky  B, Regner  SR,  et al.  Longitudinal strain in Friedreich ataxia: a potential marker for early left ventricular dysfunction. Echocardiography. 2014;31(1):50-57.
PubMedArticle
25.
Dedobbeleer  C, Rai  M, Donal  E, Pandolfo  M, Unger  P.  Normal left ventricular ejection fraction and mass but subclinical myocardial dysfunction in patients with Friedreich’s ataxia. Eur Heart J Cardiovasc Imaging. 2012;13(4):346-352.
PubMedArticle
26.
Weidemann  F, Rummey  C, Bijnens  B,  et al; Mitochondrial Protection With Idebenone in Cardiac or Neurological Outcome (MICONOS) Study Group.  The heart in Friedreich ataxia: definition of cardiomyopathy, disease severity, and correlation with neurological symptoms. Circulation. 2012;125(13):1626-1634.
PubMedArticle
27.
Puccio  H, Simon  D, Cossée  M,  et al.  Mouse models for Friedreich ataxia exhibit cardiomyopathy, sensory nerve defect and Fe-S enzyme deficiency followed by intramitochondrial iron deposits. Nat Genet. 2001;27(2):181-186.
PubMedArticle
28.
Kipps  A, Alexander  M, Colan  SD,  et al.  The longitudinal course of cardiomyopathy in Friedreich’s ataxia during childhood. Pediatr Cardiol. 2009;30(3):306-310.
PubMedArticle
29.
Hawley  RJ, Gottdiener  JS.  Five-year follow-up of Friedreich’s ataxia cardiomyopathy. Arch Intern Med. 1986;146(3):483-488.
PubMedArticle
30.
Rajagopalan  B, Francis  JM, Cooke  F,  et al.  Analysis of the factors influencing the cardiac phenotype in Friedreich’s ataxia. Mov Disord. 2010;25(7):846-852.
PubMedArticle
31.
Corben  LA, Lynch  D, Pandolfo  M, Schulz  JB, Delatycki  MB; Clinical Management Guidelines Writing Group.  Consensus clinical management guidelines for Friedreich ataxia. Orphanet J Rare Dis. 2014;9:184.
PubMedArticle
Original Investigation
November 2015

A 22-Year Follow-up Study of Long-term Cardiac Outcome and Predictors of Survival in Friedreich Ataxia

Author Affiliations
  • 1Department of Cardiology, Assistance Publique–Hôpitaux de Paris, Groupe Hospitalier Pitié-Salpêtrière Charles-Foix, Paris, France
  • 2Department of Genetics and Cytogenetics, Assistance Publique–Hôpitaux de Paris, Groupe Hospitalier Pitié-Salpêtrière Charles-Foix, Paris, France
  • 3Institut du Cerveau et de la Moelle Epinière, Sorbonne Universités, Université Pierre et Marie Curie, University Paris 06, Unité Mixte de Recherche (UMR) 1127, Institut National de la Santé et de la Récherche Médicale (INSERM) Unité 1127, Centre National de la Recherche Scientifique UMR 7225, Groupe Hospitalier Pitié-Salpêtrière Charles-Foix, Paris, France
  • 4Sorbonne Universités, Université Pierre et Marie Curie, University Paris 06, UMR 1166, INSERM, Paris, France
  • 5Department of Neurology, Université Libre de Bruxelles, Hôpital Erasme, Brussels, Belgium
  • 6Institute for Cardiometabolism and Nutrition, Paris, France
  • 7Sorbonne Universités, Université Pierre et Marie Curie, University Paris 06, UMR 1136, Institut Pierre Louis d’Epidémiologie et de Santé Publique, Paris, France
  • 8INSERM, UMR 1136, Institut Pierre Louis d’Epidémiologie et de Santé Publique, Paris, France
  • 9Biostatistics Unit, Assistance Publique–Hôpitaux de Paris, Groupe Hospitalier Pitié-Salpêtrière, Paris, France
JAMA Neurol. 2015;72(11):1334-1341. doi:10.1001/jamaneurol.2015.1855
Abstract

Importance  Friedreich ataxia (FRDA) is the most common genetic sensory ataxia, and myocardial involvement is a major determinant of survival.

Objective  To assess FRDA survival and cardiac outcome to adapt future therapeutic trials.

Design, Setting, and Participants  In a longitudinal follow-up study, all patients with genetically confirmed FRDA seen in the reference center and referred for cardiac evaluation (standard 12-lead electrocardiogram and transthoracic echocardiography) to the cardiology department were enrolled and followed up from April 27, 1990, to July 31, 2013. The setting was the French National Reference Center for Rare Diseases and the Department of Cardiology, Salpêtrière University Hospital, Paris, France. In total, 138 patients with FRDA were followed up. Among 133 patients homozygous for expanded GAA repeats, the mean (SD) age was 31 (10) years (age range, 11-62 years), with a mean (SD) age at disease onset of 16 (8) years (age range, 3-50 years) and a mean (SD) age at first wheelchair use of 26 (9) years (age range, 11-64 years). Cardiac hypertrophy was present in 57.9% (77 of 133), and electrocardiography was normal in 6.8% (9 of 133).

Main Outcomes and Measures  Long-term cardiac outcome and predictors of survival in FRDA.

Results  After a mean (SD) follow-up of 10.5 (5.5) years (range, 0.6-23.0 years), the 10-year survival rate was 88.5%. In 80.0% of patients (12 of 15), death was due to cardiac causes. Predictors of survival were a shorter GAA repeat length on the smaller allele of the frataxin gene (hazard ratio [HR], 1.85; 95% CI, 1.28-2.69), left ventricular ejection fraction (HR, 0.42; 95% CI, 0.20-0.89), and left ventricular mass index (HR, 1.19; 95% CI, 1.04-1.36). Two cardiac evolutions were distinguished with a group-based trajectory model, including a low-risk cardiac group (78.6% [81 of 103] with normal ejection fraction at baseline that declined slightly over time but remained within the normal range) and a high-risk cardiac group (21.4% [22 of 103] in which the ejection fraction progressively declined during follow-up). The patients with the worse cardiac evolution had longer GAA repeats. Neurological impairment was not predictive of cardiac change over time.

Conclusions and Relevance  Survival in FRDA is determined by cardiac complications, which are dependent on the mutation (ie, the size of the expanded GAA repeat). Patients with progressive decline of the left ventricular ejection fraction had a worse prognosis. This finding demonstrates that cardiac follow-up is important in FRDA to identify individuals at risk for further cardiac complications.

Introduction

Friedreich ataxia (FRDA) is the most common autosomal recessive ataxia. The phenotype is reminiscent of mitochondrial disease, with cerebellar signs (ataxic gait, dysarthria, coordination difficulties, and ocular fixation instability), sensory neuropathy (sensory ataxia, weakness, and deep sensory loss), pyramidal signs (extensor plantar response and leg spasticity), visual and hearing impairment, and nonneurological features, such as hypertrophic cardiomyopathy and diabetes mellitus. The onset is around puberty, sometimes later in life, and, in rare cases, during early childhood.

Although the results of small studies have suggested that myocardial involvement is a major determinant of survival, to our knowledge, no large study has yet investigated survival, causes of death, and predictors of evolution in FRDA. Almost all patients with FRDA show abnormalities on the electrocardiogram,1,2 and most show evidence of hypertrophic cardiomyopathy. Fibrosis on magnetic resonance imaging has been shown to be an early manifestation of cardiomyopathy in the absence of significant hypertrophy.3 Clinical signs of heart involvement usually appear late in FRDA, but one study4 reported an excess of cardiac deaths before age 40 years.

The mutated gene in FRDA, FXN (OMIM 606829), maps to the 9q13 chromosome5 and encodes frataxin, a mitochondrial protein involved in the biosynthesis of iron-sulfur clusters. Most patients are homozygous for an expanded trinucleotide GAA repeat in the first intron of the FXN gene, whereas others are compound heterozygotes with an expanded GAA repeat and an FXN point mutation or deletion.2,5,6 The expanded GAA repeat inhibits expression of the FXN gene through epigenetic mechanisms, resulting in reduced levels of frataxin. Longer repeats cause a greater loss of frataxin. Its level mostly depends on the length of the shorter of the 2 GAA repeat alleles.6 Accordingly, a larger number of GAA repeats on the shorter allele is associated with earlier symptom onset, increased neurological severity, and more severe cardiomyopathy.2,710 Impaired mitochondrial oxidative phosphorylation, deficient iron-sulfur cluster enzymes, and mitochondrial iron overload are observed in the myocardium of patients with FRDA as a consequence of the frataxin deficiency.1113 However, the correlation between residual frataxin levels and left ventricular (LV) hypertrophy and function remains poorly understood. Although the neurological condition deteriorates over time, the progression of cardiac disease is variable. In light of developments in the prevention and treatment of cardiomyopathy by gene therapy in a mouse model of FRDA,14 we assessed survival and cardiac outcome in a large cohort of patients with FRDA. Our objectives were to describe survival and cardiomyopathy to identify predictors of poor cardiac evolution.

Methods
Study Design

We included clinical data from all patients with genetically confirmed FRDA seen from April 27, 1990, to July 31, 2013, in the French National Reference Center for Rare Diseases and referred for cardiac evaluation to the Department of Cardiology, Salpêtrière University Hospital, Paris, France. The Institut national de la santé et de la recherche médicale (INSERM) approved the study. Informed written consent was given in accord with the law and local ethical regulations. All patients gave consent for genetic and modifier studies. Fifty patients were included since 2010 as part of a prospective European FRDA registry (European Friedreich’s Ataxia Consortium for Translational Studies; http://www.e-facts.eu),7 which is coordinated by one of us (M.P.).

Information about events was gathered from September 1, 2012, to December 31, 2015, by direct examination, telephone, or mail. When contact with a patient or the family was not possible, the date of death was obtained from the French national registry of deaths.

Neurological examination was performed within 6 months of cardiac evaluation, including 2 neurological rating scales for ataxia. These were the International Cooperative Ataxia Rating Scale15 (ICARS), with a maximum score of 100, and the Scale for the Assessment and Rating of Ataxia16 (SARA), with a maximum score of 40.

Cardiac evaluation included clinical examination, standard 12-lead electrocardiography, and transthoracic echocardiography. Cardiac arrhythmia, heart failure, and stroke were recorded. Electrocardiograms were interpreted retrospectively by a cardiologist (L.L.) masked to the echocardiography results. An electrocardiogram was considered abnormal in the presence of T-wave inversion (≥0.1 mV) or flat T wave in at least 2 leads except for V1 or V2 in the absence of a bundle branch block. Transthoracic echocardiographic studies were performed using an initial device (Sigma 1; Kontron Instruments) and subsequently using other systems (Sequoia C256; Acuson and Vivid 7; General Electric Company). Left ventricular end diastolic diameter (LVEDD) and end systolic diameter (LVESD), LV end diastolic interventricular septal wall thickness (SWT), and posterior wall thickness (PWT) were measured using standard M-mode from a parasternal long-axis view.17 Left ventricular contractility was evaluated by LV ejection fraction (LVEF) using the biapical Simpson disk method.17 Systolic dysfunction was defined as LVEF below 50%. Calculated morphologic parameters included the ratio of SWT to PWT and relative wall thickness, defined as (SWT + PWT) / LVEDD.18 Pathological concentric LV remodeling was defined by a relative wall thickness exceeding 0.42. Left ventricular mass was assessed using the following formula:

LV mass = 0.8 × [1.04 × (LVEDD + PWT + SWT)3 − LVEDD3] + 0.6 g.

LV mass was indexed to body surface area as an LV mass index (LVMI).18 Because the patients were younger than 40 years, nomograms by Henry et al19 were used to adjust thickness parameters for age and body surface area. Cardiac hypertrophy was defined when SWTs or PWTs exceeded the 95% predicted value for age and body surface area.8 Mitral valve inflow was measured by Doppler to obtain the ratio of early to late diastolic flow velocity.

Statistical Analysis

Data are expressed as the mean (SD) and range for quantitative variables except where otherwise indicated. Relationships between continuous variables were evaluated by Pearson product moment correlation. Clinical and echocardiographic characteristics of patients with and without cardiac hypertrophy were compared using Pearson χ2 test for qualitative variables and analysis of variance for quantitative variables. Survival was defined as the interval between the first cardiac visit and death or last news. Cox proportional hazards univariate analysis, followed by multivariable models with a stepwise procedure, were used to evaluate risk factors for death. Evolution of echocardiographic parameters was analyzed using linear mixed models, with time as both a fixed and random effect (df was estimated using approximation by Satterthwaite20).

To identify subgroups of participants exhibiting different trajectories of LVEF, we used a semiparametric mixture group-based trajectory model.21 This models the link between time and a variable of interest as a polynomial relationship. For each trajectory group, the model defines the shape of the trajectory and the estimated proportion of the population belonging to the group. It also determines the probability of belonging to the different trajectory groups (called posterior membership probabilities) for each individual, with the highest probability group determining group assignment. To select the best fit, we ran models with 2 to 6 trajectories, with a base specification requiring all shapes to be quadratic, and selected the model with the largest significant trajectory order estimates and a maximized Bayesian information criterion. We calculated the mean individual posterior membership probabilities for each group to examine the model fit. Finally, baseline characteristics associated with the members of each trajectory group were compared using logistic regression analysis.

All reported P values are 2-tailed. A type I error rate of 5% was used. Analyses were performed with a statistical software package (SAS, version 9.3; SAS Institute Inc).

Results

We included 138 patients with genetically confirmed FRDA from 122 families. Five patients with heterozygous FXN point mutations were excluded from the statistical analysis to avoid introducing a severity bias. Among them, 4 had cardiac hypertrophy, 4 experienced no cardiac events during follow-up, and 1 patient underwent heart transplantation (eTable 1 and eTable 2 in the Supplement).

In total, 103 of 133 patients homozygous for expanded GAA repeats were evaluated at least twice. Overall, there were 498 cardiac evaluations, a median interval between 2 cardiac evaluations of 1.7 years (range, 0.3-19.8 years), and a median number of 4 (range 2-12) cardiac evaluations per patient. The neurological and cardiac characteristics of these 133 patients are listed in Table 1.

At the first cardiac visit, 3.0% (4 of 133) of the patients had dyspnea, 9.0% (12 of 133) had chest pain, and 9.8% (13 of 133) had palpitations, while no patients had signs of heart failure. Two patients had diabetes mellitus, 7 patients had hypertension, 6 patients had previous atrial fibrillation, and 1 patient had previous symptoms of heart failure.

Only 6.8% (9 of 133) of the patients had normal electrocardiograms, and 18.8% (25 of 133) had a right bundle branch block. The most common abnormal feature was the presence of abnormal repolarization (87.9% [116 of 132]), with negative (89.7% [104 of 116]) or flat (10.3% [12 of 116]) T wave in inferior (7.7% [9 of 116]), lateral (15.5% [18 of 116]), or both (16.3% [19 of 116]) leads. The mean (SD) heart rate was 76 (12) beats/min, and all patients were in sinus rhythm except for 2 who had atrial fibrillation. No patients had LVEFs less than 50%. Left ventricular hypertrophy was detected in 77 of 133 patients (57.9%), with SWTs or PWTs exceeding the 95% predicted value for age and body surface area, as previously defined. Among 133 patients, LVMI was significantly elevated in 52 (39.1%), including 22 men (LVMI >115 g/m2) and 30 women (LVMI >95 g/m2). Only 17 of 133 (12.8%) had severe cardiac hypertrophy, with SWT exceeding 15 mm. Among 56 patients with no cardiac hypertrophy, 18 (32.1%) had normal SWT or PWT.

Two-thirds of patients with normal echocardiography had abnormal T waves on electrocardiograms. This finding suggested myocardial involvement even without detectable hypertrophy.

There was a strong correlation between a shorter GAA repeat length and age at onset (r = −0.61, P < .001), as well as with the following parameters: SWT (r = 0.24, P = .005), LVEDD (r = −0.32, P < .001), transmitral A wave (r = −0.27, P = .004), and the ratio of early to late diastolic flow velocity (r = 0.26, P = .007). The length of the longer GAA repeat also correlated with age at onset (r = −0.42, P < .001), SWT (r = 0.20, P = .02), and PWT (r = 0.21, P = .01).

Overall, patients with cardiac hypertrophy were younger, had longer GAA repeats on both alleles, experienced earlier disease onset, and initiated wheelchair use at an earlier age. Fifty-four percent of patients with no cardiac hypertrophy had abnormal LV remodeling. Patients with cardiac hypertrophy had a higher ratio of early to late diastolic flow velocity, suggesting diastolic dysfunction. The LVEF was similar in the 2 groups with and without cardiac hypertrophy.

Survival Analysis

Among 133 patients, 15 (11.3%) died at a mean (SD) age of 39 (10) years (age range, 27-61 years) after a mean (SD) follow-up of 10.5 (5.5) years (range, 0.6-23.0 years) (Figure 1). Eight patients died of cardiovascular causes. Of those, 6 (age range, 27-49 years) died of progressive heart failure, and 2 had atrial fibrillation and progressive heart failure and died after cardioembolic stroke at ages 32 and 43 years. Two patients died of noncardiac causes and one of respiratory disease (at age 33 years), while the death of the other was from suicide (at age 28 years). Five deaths (age range, 33-61 years) were of unknown origin, but 4 of these 5 patients were known to have atrial fibrillation.

Cardiac events included supraventricular arrhythmias in 16.5% (22 of 133) of patients, LVEFs below 50% in 9.8% (13 of 133), and heart failure in 8.3% (11 of 133) (eTable 3 in the Supplement). One patient with a low LVEF received a prophylactic implantable defibrillator with no shock, and 1 patient had a pacemaker for bradyarrhythmia. In addition, 4 patients developed hypertension, and 9 patients had diabetes mellitus.

Patients with hypertrophy had more cardiac events than patients without hypertrophy, with 16 vs 6 having supraventricular arrhythmia (P = .12), 11 vs 2 having LVEF less than 50% (P = .04), 10 vs 1 having episodes of heart failure (P = .03), and death occurring in 12 vs 3 (P = .06). For 18 patients with normal hearts, neither death nor cardiac events were reported.

In the univariate analysis, mortality was associated with longer GAA repeats on both alleles, earlier age at onset, higher SWT and LVESD, LVMI, and smaller LVEF (Table 2). In the multivariable analysis, independent predictors of mortality were the length of the shorter GAA repeat (hazard ratio [HR], 1.85; 95% CI, 1.28-2.69 per 100 additional repeats), LVEF (HR, 0.42; 95% CI, 0.20-0.89 per 10% decrease), and LVMI (HR, 1.19; 95% CI, 1.04-1.36 per increase of 10 g/m2).

We estimated the neurological evolution of 69 patients. Among 34 patients, the mean (SD) SARA score worsened from 24 (7) (range, 9-38) to 27 (6) (range, 13-36) after a mean (SD) follow-up of 4 (2) years. Among 55 patients, the mean (SD) ICARS score worsened from 54 (20) (range, 7-90) to 62 (20) (range, 10-98). The SARA score increased a mean (SD) of 4 (5) points (P < .001), and the ICARS score increased a mean (SD) of 8 (10) points (P < .001). Neither the SARA score nor the ICARS score at baseline was predictive of cardiac change over time.

All cardiac parameters evolved over time, with a mean (SE) decrease in LVEF of 0.68% (0.10%) (P < .001) and with the mean (SE) decreases in SWT of −0.055 (0.021) mm (P = .01) and PWT of −0.11 (0.02) mm (P < .001). The mean (SE) increases were 0.27 (0.05) mm (P < .001) for LVEDD and 0.33 (0.06) mm (P < .001) for LVESD. The LVMI decreased by a mean (SE) of −0.42 (0.21) g/m2 (P = .05).

Trajectory Analysis

The best-fitting model included 2 trajectories: a large low-risk group (78.6% [81 of 103]) with normal LVEF at baseline that declined slightly over time but remained in the normal range and a smaller high-risk group (21.4% [22 of 103]) with a progressively decreasing LVEF (Figure 2). The mean (SE) LVEFs at inclusion were 70% (6%) in the low-risk group and 61% (5%) in the high-risk group (P < .001). The mean posterior probabilities that a participant belonged to his or her assigned group were above 0.5, indicating a good model fit. The probability of belonging to the high-risk group depended on the length of the shorter GAA repeat, age at onset, and LVESD at inclusion (Table 3). In the multivariable analysis, only a shorter GAA repeat length and LVESD remained predictors of a decreased LVEF. The LVESD and LVEF were significantly correlated at baseline (r = −0.55, P < .001).

The initiation of wheelchair use as an indicator of severity did not differ between the 2 trajectory groups, with the mean (SD) age at wheelchair use being 26.6 (9.5) years for normal LVEF vs 22.4 (8.6) years for low LVEF (P = .94). The time between the onset of disease and wheelchair use did not indicate a faster neurological evolution in the group with a poor cardiac trajectory, with the mean (SD) periods being 10.0 (6.0) years for normal LVEF vs 7.4 (5.7) years for low LVEF (P = .17).

Discussion

To our knowledge, this study is the first survival analysis and long-term follow-up of a large cohort of patients with FRDA. The 10-year survival rate of 88.5% indicates a better prognosis than was previously reported.17 We found that the mean age at death was 39 years, which was slightly older than previously reported.2 We confirm that the major causes of mortality in patients with FRDA are cardiovascular, primarily progressive heart failure, with a low LVEF. When the cause of death was classified as unknown, 4 of 5 patients had previous cardiac events, so these deaths could be attributed to cardiac causes.

An important finding of this study is that the length of GAA repeats on both alleles of the FXN gene (ie, the residual amount of frataxin) is the best predictor of mortality in FRDA. The 2 other independent predictors of mortality were LVEF (reflecting systolic function) and LVMI (reflecting hypertrophy).

As previously reported,2224 we found that 57.9% (77 of 133) of patients had moderate cardiac concentric hypertrophy, with a small left ventricle without obstruction. Hypertrophy is mostly dependent on the size of the shorter GAA repeat,7,23,24 but the length of the longer GAA repeat also correlated with cardiac wall thickness. The weak relationship suggests that additional factors modify cardiac phenotypic expression. Several definitions of hypertrophy are used in the literature.25 Thirty-nine percent of our patients had hypertrophy based on LVMI. We used nomograms by Henry et al19 because morphologic parameters are influenced by age and body surface area. Although this method might overestimate the degree of hypertrophy, it was found to be a good compromise between magnetic resonance imaging and echocardiographic categorization in the Mitochondrial Protection With Idebenone in Cardiac or Neurological Outcome (MICONOS) study.26 Patients with hypertrophy had a worse prognosis and more cardiac events than patients without hypertrophy, but survival was similar despite a follow-up of 10 years, which is perhaps too short to produce a sufficient number of events in each group. In frataxin cardiac conditional knockout mice, hypertrophy is followed by dilated cardiomyopathy, which appears in parallel with mitochondrial respiratory chain dysfunction and iron accumulation.27 In FRDA, cardiac involvement may progress to LV systolic dysfunction and heart failure. Two patients herein had severe dilated cardiomyopathy at ages 16 and 23 years. We observed a small decrease in LVEF, suggesting that cardiac function deteriorated slightly with time as previously shown,28 and 21.4% (22 of 103) of patients herein developed significant alterations of LVEF. Evolution of LVEF depended on the size of the shorter GAA repeat and was independent of cardiac hypertrophy. We observed a regression of hypertrophy over time and an increase in the size of the left ventricle. Previous reports suggested contradictory cardiac changes over time, with either an increase29 or a decrease30 in hypertrophy.

As recommended,31 cardiac follow-up in FRDA is necessary to identify early systolic LV dysfunction because clinical signs of cardiac dysfunction develop only later in the course of the disease. New echocardiography techniques, such as speckle tracking imaging, can detect alterations of contractility before a decreased LVEF.2325 Follow-up studies will be necessary to evaluate the usefulness of these new techniques in patients with FRDA.

Until specifically tailored treatments are developed, early treatment of cardiomyopathy by β-blockers, angiotensin-converting enzyme inhibitors, or antagonists of mineralocorticoid receptors should be proposed in patients with LVEF below 50% because of the expected progression of cardiac involvement.31 However, controlled studies are necessary to evaluate whether earlier cardiac intervention can modify the progression of cardiomyopathy in FRDA. Recently, a heart conditional Fxn knockout mouse model, which develops rapidly fatal cardiomyopathy, was rescued by administration of a frataxin-expressing viral vector after the onset of heart failure, with complete reversal of cardiomyopathy at the functional, cellular, and molecular levels.14 This finding suggests that gene therapy may be a promising approach to treat cardiomyopathy in patients with FRDA and a low LVEF.

Our study had some limitations. Recruitment at our center mainly comprises adults, although 10 of 133 patients (7.5%) herein were younger than 18 years at baseline. However, age at onset was younger than 18 years for 94 of 133 patients (70.7%). We could have missed patients with early onset and severe evolution who did not come to our attention. Therefore, the findings can be generalized only to adults with FRDA.

We were unable to correlate evolution of neurological scores to cardiac measures. However, we showed that SARA scores at baseline were not predictive of survival, and age at first wheelchair use was not earlier in the worse cardiac trajectory group.

Idebenone has been identified as a potential treatment for FRDA, with a possible effect on cardiac disease.26 Forty patients herein received idebenone at baseline, but intermittent use and incomplete prescription records prevented our assessment of the effect of idebenone on longitudinal cardiac outcome.

Conclusions

Survival is shortened in FRDA as a result of cardiac complications. We identified a low-risk cardiac group (78.6% [81 of 103]) that included most patients and a high-risk cardiac group (21.4% [22 of 103]) with a progressively decreasing LVEF. This finding demonstrates the importance of cardiac follow-up in FRDA and the necessity for patient stratification in therapeutic trials.

Back to top
Article Information

Accepted for Publication: June 17, 2015.

Corresponding Author: Alexandra Durr, MD, PhD, Institut du Cerveau et de la Moelle Epinière, Hôpital de la Salpêtrière, 47 Blvd de l’Hôpital, 75651 Paris CEDEX 13, France (alexandra.durr@icm-institute.org).

Published Online: September 28, 2015. doi:10.1001/jamaneurol.2015.1855.

Author Contributions: Dr Durr had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis. Drs Tezenas du Montcel and Durr contributed equally to this work and are co–last authors.

Study concept and design: Pousset, Durr.

Acquisition, analysis, or interpretation of data: Pousset, Legrand, Ewenczyk, Charles, Isnard, Tezenas du Montcel, Durr.

Critical revision of the manuscript for important intellectual content: All authors.

Statistical analysis: Tezenas du Montcel.

Conflict of Interest Disclosures: None reported.

Funding/Support: This study was supported by Frame Project 7 grant HEALTH-F2-2010-242193 from the European Commission.

Role of the Funder/Sponsor: The funding source had no role in the design and conduct of the study; collection, management, analysis, or interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.

Additional Contributions: Sandra Benaich and Fabien Lesne (Department of Genetics and Cytogenetics, Assistance Publique–Hôpitaux de Paris, Groupe Hospitalier Pitié-Salpêtrière Charles-Foix) assisted with data handling. We thank the patients who participated in the study.

References
1.
Harding  AE, Hewer  RL.  The heart disease of Friedreich’s ataxia: a clinical and electrocardiographic study of 115 patients, with an analysis of serial electrocardiographic changes in 30 cases. Q J Med. 1983;52(208):489-502.
PubMed
2.
Durr  A, Cossee  M, Agid  Y,  et al.  Clinical and genetic abnormalities in patients with Friedreich’s ataxia. N Engl J Med. 1996;335(16):1169-1175.
PubMedArticle
3.
Raman  SV, Phatak  K, Hoyle  JC,  et al.  Impaired myocardial perfusion reserve and fibrosis in Friedreich ataxia: a mitochondrial cardiomyopathy with metabolic syndrome. Eur Heart J. 2011;32(5):561-567.
PubMedArticle
4.
Tsou  AY, Paulsen  EK, Lagedrost  SJ,  et al.  Mortality in Friedreich ataxia. J Neurol Sci. 2011;307(1-2):46-49.
PubMedArticle
5.
Campuzano  V, Montermini  L, Moltò  MD,  et al.  Friedreich’s ataxia: autosomal recessive disease caused by an intronic GAA triplet repeat expansion. Science. 1996;271(5254):1423-1427.
PubMedArticle
6.
Pandolfo  M, Pastore  A.  The pathogenesis of Friedreich ataxia and the structure and function of frataxin. J Neurol. 2009;256(suppl 1):9-17.
PubMedArticle
7.
Reetz  K, Dogan  I, Costa  AS,  et al.  Biological and clinical characteristics of the European Friedreich’s Ataxia Consortium for Translational Studies (EFACTS) cohort: a cross-sectional analysis of baseline data. Lancet Neurol. 2015;14(2):174-182.
PubMedArticle
8.
Isnard  R, Kalotka  H, Dürr  A,  et al.  Correlation between left ventricular hypertrophy and GAA trinucleotide repeat length in Friedreich’s ataxia. Circulation. 1997;95(9):2247-2249.
PubMedArticle
9.
Filla  A, De Michele  G, Cavalcanti  F,  et al.  The relationship between trinucleotide (GAA) repeat length and clinical features in Friedreich ataxia. Am J Hum Genet. 1996;59(3):554-560.
PubMed
10.
Lodi  R, Rajagopalan  B, Blamire  AM,  et al.  Cardiac energetics are abnormal in Friedreich ataxia patients in the absence of cardiac dysfunction and hypertrophy: an in vivo 31P magnetic resonance spectroscopy study. Cardiovasc Res. 2001;52(1):111-119.
PubMedArticle
11.
Lamarche  J, Shapcott  D, Côté  M, Lemieux  B. Cardiac iron deposits in Friedreich’s ataxia. In: Lechtenberg  R, ed. Handbook of Cerebellar Diseases. Boca Raton, FL: CRC Press; 1993:453-457.
12.
Michael  S, Petrocine  SV, Qian  J,  et al.  Iron and iron-responsive proteins in the cardiomyopathy of Friedreich’s ataxia. Cerebellum. 2006;5(4):257-267.
PubMedArticle
13.
Rötig  A, de Lonlay  P, Chretien  D,  et al.  Aconitase and mitochondrial iron-sulphur protein deficiency in Friedreich ataxia. Nat Genet. 1997;17(2):215-217.
PubMedArticle
14.
Perdomini  M, Belbellaa  B, Monassier  L,  et al.  Prevention and reversal of severe mitochondrial cardiomyopathy by gene therapy in a mouse model of Friedreich’s ataxia. Nat Med. 2014;20(5):542-547.
PubMedArticle
15.
Trouillas  P, Takayanagi  T, Hallett  M,  et al; Ataxia Neuropharmacology Committee of the World Federation of Neurology.  International Cooperative Ataxia Rating Scale for pharmacological assessment of the cerebellar syndrome. J Neurol Sci. 1997;145(2):205-211.
PubMedArticle
16.
Schmitz-Hübsch  T, du Montcel  ST, Baliko  L,  et al.  Scale for the Assessment and Rating of Ataxia: development of a new clinical scale. Neurology. 2006;66(11):1717-1720.
PubMedArticle
17.
Lang  RM, Bierig  M, Devereux  RB,  et al; Chamber Quantification Writing Group; American Society of Echocardiography’s Guidelines and Standards Committee; European Association of Echocardiography.  Recommendations for chamber quantification: a report from the American Society of Echocardiography’s Guidelines and Standards Committee and the Chamber Quantification Writing Group, developed in conjunction with the European Association of Echocardiography, a branch of the European Society of Cardiology. J Am Soc Echocardiogr. 2005;18(12):1440-1463.
PubMedArticle
18.
Ganau  A, Devereux  RB, Roman  MJ,  et al.  Patterns of left ventricular hypertrophy and geometric remodeling in essential hypertension. J Am Coll Cardiol. 1992;19(7):1550-1558.
PubMedArticle
19.
Henry  WL, Gardin  JM, Ware  JH.  Echocardiographic measurements in normal subjects from infancy to old age. Circulation. 1980;62(5):1054-1061.
PubMedArticle
20.
Satterthwaite  FE.  Synthesis of variance . Psychometrika. 1941;6:309-316.Article
21.
Nagin  DS, Odgers  CL.  Group-based trajectory modeling in clinical research. Annu Rev Clin Psychol. 2010;6:109-138.
PubMedArticle
22.
Leone  M, Rocca  WA, Rosso  MG, Mantel  N, Schoenberg  BS, Schiffer  D.  Friedreich’s disease: survival analysis in an Italian population. Neurology. 1988;38(9):1433-1438.
PubMedArticle
23.
Mottram  PM, Delatycki  MB, Donelan  L, Gelman  JS, Corben  L, Peverill  RE.  Early changes in left ventricular long-axis function in Friedreich ataxia: relation with the FXN gene mutation and cardiac structural change. J Am Soc Echocardiogr. 2011;24(7):782-789.
PubMedArticle
24.
St John Sutton  M, Ky  B, Regner  SR,  et al.  Longitudinal strain in Friedreich ataxia: a potential marker for early left ventricular dysfunction. Echocardiography. 2014;31(1):50-57.
PubMedArticle
25.
Dedobbeleer  C, Rai  M, Donal  E, Pandolfo  M, Unger  P.  Normal left ventricular ejection fraction and mass but subclinical myocardial dysfunction in patients with Friedreich’s ataxia. Eur Heart J Cardiovasc Imaging. 2012;13(4):346-352.
PubMedArticle
26.
Weidemann  F, Rummey  C, Bijnens  B,  et al; Mitochondrial Protection With Idebenone in Cardiac or Neurological Outcome (MICONOS) Study Group.  The heart in Friedreich ataxia: definition of cardiomyopathy, disease severity, and correlation with neurological symptoms. Circulation. 2012;125(13):1626-1634.
PubMedArticle
27.
Puccio  H, Simon  D, Cossée  M,  et al.  Mouse models for Friedreich ataxia exhibit cardiomyopathy, sensory nerve defect and Fe-S enzyme deficiency followed by intramitochondrial iron deposits. Nat Genet. 2001;27(2):181-186.
PubMedArticle
28.
Kipps  A, Alexander  M, Colan  SD,  et al.  The longitudinal course of cardiomyopathy in Friedreich’s ataxia during childhood. Pediatr Cardiol. 2009;30(3):306-310.
PubMedArticle
29.
Hawley  RJ, Gottdiener  JS.  Five-year follow-up of Friedreich’s ataxia cardiomyopathy. Arch Intern Med. 1986;146(3):483-488.
PubMedArticle
30.
Rajagopalan  B, Francis  JM, Cooke  F,  et al.  Analysis of the factors influencing the cardiac phenotype in Friedreich’s ataxia. Mov Disord. 2010;25(7):846-852.
PubMedArticle
31.
Corben  LA, Lynch  D, Pandolfo  M, Schulz  JB, Delatycki  MB; Clinical Management Guidelines Writing Group.  Consensus clinical management guidelines for Friedreich ataxia. Orphanet J Rare Dis. 2014;9:184.
PubMedArticle
×