Friedreich Ataxia

Typical Friedreich Ataxia

Neurologic manifestations. Individuals with typical Friedreich ataxia (FRDA) develop progressive ataxia with onset from early childhood through to early adulthood, starting with poor balance when walking, followed by slurred speech and upper-limb ataxia. The mean age at onset of symptoms is ten to 15 years [Delatycki et al 1999b]; onset can be as early as age two years and as late as the eighth decade. Gait ataxia, caused by a combination of spinocerebellar degeneration and loss of joint-position sense (proprioception), is the earliest symptom in the vast majority. The poor balance is accentuated when visual input is eliminated, such as in darkness or when the eyes are closed (Romberg sign). Ankle and knee jerks are generally absent, and plantar responses are up-going.

Within five years of symptom onset, most individuals with FRDA exhibit "scanning" dysarthria, lower-extremity weakness, and diminished or absent joint-position and vibration sense distally ‒ neurologic manifestations that result from progressive degeneration of the dorsal root ganglia, posterior columns, corticospinal tracts, dorsal spinocerebellar tracts of the spinal cord, and cerebellum. Involvement of peripheral sensory and motor neurons results in a mixed axonal peripheral neuropathy.

Muscle weakness is often present and is most prominent in hip extensors and abductors; as disease advances, distal limb muscle weakness and wasting become evident.

Spasticity in the lower limbs is common and can be significant, affecting foot plantar flexors and inverters to a greater extent than dorsiflexors and everters. Thus, in the late stages of disease, equinovarus deformity is commonly seen [Delatycki et al 2005] and may result in contractures ‒ more commonly in nonambulatory affected individuals [Milne et al 2016] ‒ and significant morbidity. Pes cavus is common (55%) but generally causes little problem for affected individuals. Restless leg syndrome is common in individuals with Friedreich ataxia, affecting 32%-50% of individuals in two studies [Frauscher et al 2011].

Scoliosis is present in approximately two thirds of individuals with FRDA when assessed clinically and 100% when assessed radiographically. A study found that 49 of 77 individuals with FRDA had scoliosis; ten were treated with a brace and 16 required spinal surgery [Milbrandt et al 2008].

Autonomic disturbance becomes more common with disease progression. The most common manifestations are cold, cyanosed feet; bradycardia is less common.

Electrodiagnostic findings. Nerve conduction studies generally show a motor nerve conduction velocity of greater than 40 m/s with reduced or absent sensory nerve action potential with an absent H reflex.

Central motor conduction time is abnormal after transcranial magnetic stimulation [Brighina et al 2005].

Speech. Dysarthria, present in the majority of individuals with FRDA, is generally of three types: mild dysarthria, increased velopharyngeal involvement manifest as hypernasality, and increased laryngeal dysfunction manifest as increased strained-strangled vocal quality [Folker et al 2010]. Dysarthria becomes worse as the disease progresses with the main changes seen over time being in speaking rate and utterance duration [Rosen et al 2012].

Mild dysphonia characterized by hoarseness (combined roughness and breathiness), increased strain, and altered pitch variability is also seen [Vogel et al 2017].

Swallowing. Dysphagia is common in FRDA with 92% of individuals reporting issues with swallowing [Vogel et al 2014]. Dysphagia in FRDA relates to oropharyngeal incoordination, weakness, and spasticity.

Hypertrophic cardiomyopathy, defined as increased thickness of the interventricular septum, is present in about two thirds of individuals with FRDA [Delatycki et al 1999a]. Echocardiographic evaluation may reveal left ventricular hypertrophy that is more commonly asymmetric than concentric [Dutka et al 2000, Bit-Avragim et al 2001, Koc et al 2005]. When more subtle cardiac involvement is sought by methods such as tissue Doppler echocardiography, an even larger percentage of individuals have detectable abnormalities [Dutka et al 2000, Mottram et al 2011]. Between 12% and 20% of individuals have reduced ejection fraction [Regner et al 2012a, Weidemann et al 2012] and longitudinal strain is commonly reduced [St John Sutton et al 2014].

Later in the disease course, the cardiomyopathy may become dilated. Progressive systolic dysfunction is common [Kipps et al 2009] and reduction in left ventricular wall thickness is often seen as the disease progresses [Rajagopalan et al 2010]. A longitudinal study identified two groups; a "low risk" group (approximately 80%) with normal ejection fraction that declined slowly and remained in the normal range and a "high risk" group (approximately 20%) in whom ejection fraction declined into the abnormal range and was associated with high mortality [Pousset et al 2015]. Those in the "high risk" group had longer GAA expansions on the shorter allele. The degree of neurologic impairment did not predict whether an affected individual would have stable or rapid progression of cardiomyopathy.

Electrocardiography (ECG) is abnormal in the vast majority, with T wave inversion, left axis deviation, and repolarization abnormalities being most commonly seen [Dutka et al 1999].

Symptoms related to cardiomyopathy usually occur in the later stages of the disease [Dutka et al 1999] but in rare instances may precede ataxia [Alikaşifoglu et al 1999, Leonard & Forsyth 2001]. Quercia et al [2010] established the diagnosis of FRDA in a young child evaluated for sudden death. Subjective symptoms of exertional dyspnea (40%), palpitations (11%), and anginal pain may be present in moderately advanced disease. Arrhythmias (especially atrial fibrillation) and congestive heart failure frequently occur in the later stages of the disease and are the most common cause of mortality [Tsou et al 2011]. Coronary artery disease may occur and should be considered if there is angina and/or sudden deterioration in cardiac function [Giugliano & Sethi 2007].

Urinary issues. Bladder symptoms including urinary frequency and urgency were reported by 41% of individuals in one study [Delatycki et al 1999a]. A study of 158 individuals with FRDA revealed lower urinary tract symptoms in 82% with impact on quality of life in 22% of those [Musegante et al 2013]. Of 28 who underwent urodynamic studies, all had normal serum creatinine and four had upper urinary tract dilatation.

Sleep-disordered breathing. Sleep-disordered breathing and sleep apnea are more prevalent in those with FRDA than in the healthy population. There is a minimum prevalence of 21% of obstructive sleep apnea compared to an incidence of about 5% in the general population [Corben et al 2013].

Diabetes mellitus occurs in up to 30% of individuals with FRDA [Cnop et al 2013]. Impaired glucose tolerance is seen in up to an additional 49% [Ristow 2004, Cnop et al 2012]. Non-diabetic individuals with FRDA demonstrate high insulin responsiveness to oral glucose testing and low insulin sensitivity [Isaacs et al 2016].

Ophthalmic manifestations. Optic nerve atrophy, often asymptomatic, occurs in approximately 25% of individuals with FRDA. Reduced visual acuity was found in 13% in one study [Dürr et al 1996]. Study of the anterior and posterior visual pathways in FRDA by visual field testing and optical coherence tomography, pattern visual evoked potentials, and diffusion-weighted imaging revealed that all individuals studied had optic nerve abnormalities, but only 5/26 (19%) had related symptoms [Fortuna et al 2009]. Progressive diminution of contrast acuity is typical with disease progression [Seyer et al 2013].

Abnormal extraocular movements include irregular ocular pursuit, dysmetric saccades, saccadic latency, square wave jerks, ocular flutter, and marked reduction in vestibulo-ocular reflex gain and increased latency [Fahey et al 2008]. Horizontal and vertical gaze palsy does not occur.

Hearing loss. Sensorineural hearing loss occurs in 13% of individuals with FRDA [Dürr et al 1996]. Auditory neuropathy may occur and difficulty hearing in background noise is common [Rance et al 2008].

Cognitive skills. While cognition is generally not impaired in FRDA, motor and mental reaction times can be significantly slowed [Wollmann et al 2002, Corben et al 2006]. Motor planning is markedly impaired [Corben et al 2010, Corben et al 2011]. The intelligence profile of individuals with FRDA is characterized by concrete thinking and poor capacity in concept formation and visuospatial reasoning with reduced speed of information processing [Mantovan et al 2006]. Problems with attention and working memory have also been demonstrated [Klopper et al 2011]. Motor overflow is also more prevalent in FRDA than in controls [Low et al 2013]. Those with earlier onset and larger FXN intron 1 GAA repeats tend to have more severe cognitive difficulties than those with later onset and smaller GAA repeats [Nachbauer et al 2014]. Impairment of inhibition and cognitive flexibility was identified in individuals with FRDA on the Haylings Sentence Completion Task [Corben et al 2017].

Bone mineral density. A study of 28 individuals with FRDA identified that six (21.4%) had reduced bone mineral density for age in at least one site assessed [Eigentler et al 2014]. There was a negative correlation between disease severity and femoral neck bone density. Females were more likely to have clinical fractures than males but no association was found between bone mineral density and fracture occurrence. In fact, all fractures occurred in those with a Z-score better than -2.

Other. Inflammatory bowel disease and growth hormone deficiency are more common in individuals with FRDA than the general community [Shinnick et al 2016].

Progression. The rate of progression of FRDA is variable. The average time from symptom onset to wheelchair dependence is ten years [Dürr et al 1996, Delatycki et al 1999a]. A number of studies have found that progression is more rapid in those with earlier disease onset [Reetz et al 2015, Tai et al 2015, Patel et al 2016].

In a large study conducted in the early 1980s, the average age at death was 37 years [Harding 1981]. In a more recent study, the mean and median age of death was 36.5 years and 30 years, respectively [Tsou et al 2011]. Survival into the sixth and seventh decades has been documented. The most common cause of death was cardiac (38/61), with the remainder (17/61) being non-cardiac (most commonly pneumonia) or unknown cause (6/61) [Tsou et al 2011].

Pregnancy. A study of 65 pregnancies in 31 women with FRDA found no increase in the rate of spontaneous miscarriage, preeclampsia, prematurity, or cesarean section delivery [Friedman et al 2010]. Worsening, improving, or unchanged symptoms during pregnancy were each reported by approximately one third of women with FRDA.

Neuroimaging. MRI is often normal in the early stages of FRDA. With advanced disease, atrophy of the cervical spinal cord and cerebellum may be observed [Bhidayasiri et al 2005]. Atrophy of the superior cerebellar peduncle, the main outflow tract of the dentate nucleus, may also be seen [Akhlaghi et al 2011]. Cervical spinal cord size correlates with disease severity as measured by the Friedreich Ataxia Rating Scale [Chevis et al 2013].

A voxel-based morphometry study showed a symmetric volume loss in the dorsal medulla, infero-medial portions of the cerebellar hemispheres, rostral vermis, and dentate region [Della Nave et al 2008]. No volume loss in cerebral hemispheres was observed. Lower fractional anisotropy, higher mean diffusivity, and increased radial diffusivity compared to controls have been found in the dentatorubral, dentatothalamic, and thalamocortical tracts in individuals with FRDA [Akhlaghi et al 2014].

Reduced N-acetylaspartate in the cerebellum has been demonstrated by ¹H-MRS [Iltis et al 2010] and increased diffusion weighted imaging may be present in a number of brain white matter tracts [Rizzo et al 2011].

Atypical Presentations

Approximately 25% of individuals homozygous for full-penetrance GAA expansions in FXN have atypical findings [Dürr et al 1996] that include the following.

Late-onset FRDA (LOFA) and very late-onset FRDA (VLOFA). In approximately 15% of individuals with FRDA, onset is later than age 25 years. In individuals with LOFA, the age of onset is 26-39 years; and, in VLOFA, onset is after age 40 years [Bidichandani et al 2000, Bhidayasiri et al 2005]. The oldest reported age of onset among individuals homozygous for the GAA expansion is 80 years [Alvarez et al 2013].

A study of 44 individuals with LOFA and 30 individuals with VLOFA found that dysarthria, absent tendon reflexes, extensor plantar reflexes, weakness, amyotrophy, ganglionopathy, cerebellar atrophy, scoliosis, cardiomyopathy, and functional disability were milder, and GAA expansion on the smaller allele shorter, than in individuals with typical-onset FRDA [Lecocq et al 2016]. Another study of 18 individuals with LOFA reported similar findings [Martinez et al 2017].

• FRDA with retained reflexes (FARR) accounts for approximately 12% of individuals who are homozygous for the GAA expansion [Coppola et al 1999]. Some individuals with FARR show brisk tendon reflexes that can be accompanied by clonus. Tendon reflexes may be retained for more than ten years after the onset of the disease. FARR usually has a later age of onset and lower incidence of secondary skeletal involvement and cardiomyopathy [Coppola et al 1999].
• FRDA in Acadians. Montermini et al [1997b] showed that Acadians with FRDA have a later age of onset (on average 3.0 years later than those with typical FRDA) and of wheelchair confinement, and a much lower incidence of cardiomyopathy (48% vs 82%).

Spastic paraparesis without ataxia. Individuals who have biallelic full-penetrance GAA expansions may rarely present with spastic gait disturbance without gait or limb ataxia. These individuals usually have hyperreflexia and a later age of onset (on average 5.8 years later than those with typical FRDA); they develop ataxia with time [Montermini et al 1997b, Gates et al 1998, Castelnovo et al 2000, Lhatoo et al 2001, Badhwar et al 2004].

Other rare presentations of FRDA

• Chorea and pure sensory ataxia [Berciano et al 1997, Hanna et al 1998, Zhu et al 2002]
• Apparently isolated cardiomyopathy, with ataxia only becoming evident some time later [Leonard & Forsyth 2001]

Biallelic Pathogenic GAA Repeat Expansions

GAA repeat size. The age of onset, presence of leg muscle weakness/wasting, duration until wheelchair use, and prevalence of cardiomyopathy, pes cavus, and scoliosis have all shown statistically significant inverse correlations with the size of the expanded GAA repeat [Dürr et al 1996, Filla et al 1996, Monrós et al 1997, Montermini et al 1997b]. The size of the shorter of the two expanded pathogenic GAA repeat alleles shows better correlation than the larger repeat allele and accounts for approximately 50% of the variation in age of onset [Filla et al 1996].

La Pean et al [2008] found that age at diagnosis is a better predictor of disease severity – including disease progression and association with scoliosis and cardiomyopathy. This suggests that factors other than the repeat length (e.g., other genetic, epigenetic, and environmental variables) play a role in determining the severity of disease.

A longitudinal natural history study using a large heterogeneous cohort stratified by the size of the shorter of the two expanded alleles showed that individuals with fewer than 300 GAA repeats progressed more slowly compared to individuals with longer repeat sizes [Regner et al 2012b]. Similarly, Metz et al [2013] found that the rate of disease progression as a function of the length of the shorter of the two expanded alleles was most prominent with alleles containing fewer than 600 GAA repeats.

Late-onset FRDA (LOFA) and very late-onset FRDA (VLOFA)

• Individuals with LOFA (i.e., age of onset >25 years) frequently exhibit fewer than 500 GAA repeats in at least one of the expanded alleles [Bhidayasiri et al 2005].
• Individuals with VLOFA (i.e., age of onset >40 years) usually have fewer than 300 GAA repeats in at least one of the expanded alleles [Bidichandani et al 2000, Berciano et al 2005]. However, Bidichandani et al [2000] reported an individual with VLOFA who had biallelic expansions with greater than 800 GAA repeats on each allele, underscoring the inability to predict the clinical outcome in each individual.
  In the full penetrance range, there are uncommon FXN alleles that are interrupted by other nucleotides thereby disrupting a section of the long tract of tandem GAA repeats (see Molecular Genetics). Counting only the number of GAA repeats in the uninterrupted section, such alleles tend to be shorter in length (equivalent in length to alleles of 100-300 GAA repeats), and are often associated with LOFA/VLOFA. Stolle et al [2008] reported six people with such interrupted alleles (with a conventional expanded GAA repeat variant containing >600 repeats in the other FXN allele) whose onset ranged from age 34 to 75 years. It is not clear if the milder FRDA phenotype results from the interruptions per se, or the fact that interrupted alleles are often short, or both.

FRDA in Acadians. Despite the milder phenotype in this population, no significant differences were found either in the size of the GAA expansions or in the pathogenic sequence variants of FXN compared to individuals with typical FRDA [Montermini et al 1997b]. This finding supports the existence of other genetic modifiers of disease severity.

Spastic paraparesis without ataxia may be seen in those with smaller expanded alleles [Berciano et al 2002], or in association with the p.Gly130Val missense variant [McCabe et al 2002].

Cardiomyopathy is more frequently seen in individuals with a higher number of GAA repeats [Dürr et al 1996, Filla et al 1996, Monrós et al 1997]:

• Isnard et al [1997] found echocardiographic evidence of left ventricular hypertrophy in 81% of those with FRDA with GAA repeat lengths greater than 770 repeats and in only 14% of those with repeat lengths of fewer than 770 repeats.
• Significant correlation is seen between the size of the GAA expansion and various diastolic parameters [Mottram et al 2011] as well as the thickness of the interventricular septum and left ventricular wall [Isnard et al 1997, Dutka et al 1999, Bit-Avragim et al 2001]. Pousset et al [2015] found that longer GAA expansions on the shorter allele were associated with greater progression to low ejection fraction and poorer resultant survival.
• Montermini et al [1997b] and Delatycki et al [1999b] showed that the presence of cardiomyopathy correlated with disease severity as defined by age of onset.
• Cuda et al [2002] described an individual with particularly severe early childhood-onset cardiac hypertrophy that preceded the onset of ataxia; the individual had biallelic large GAA expansions and additionally had a pathogenic variant in TNNT2, the gene encoding cardiac troponin T.

Diabetes mellitus or abnormal glucose tolerance does not show a clear-cut correlation with the size of the GAA expansion. Filla et al [1996] found that individuals with diabetes mellitus tend to have larger repeat lengths; in a larger cohort, however, Dürr et al [1996] did not find significant correlation either with the size of the GAA expansion or with disease duration. Despite the lack of correlation with the GAA expansion size, Delatycki et al [1999b] found a correlation between the incidence of diabetes mellitus and earlier age at onset. A study of glucose metabolism in individuals with FRDA identified a correlation between longer GAA repeat length on the shorter allele and higher serum glucose and hemoglobin A1C concentrations [Greeley et al 2014].

Compound Heterozygotes for a GAA Expansion and an Intragenic Inactivating Pathogenic Variant or Deletion

A study of 111 compound heterozygotes identified three subgroups based on the pathogenic non-expansion variant: (1) null variant (no frataxin produced), (2) moderate/strong impact on frataxin function and (3) minimal impact of frataxin function [Galea et al 2016]. Compared to those with biallelic GAA expansions, subgroup 1 had earlier onset and higher incidence of diabetes whilst those with biallelic GAA expansions had a higher rate of cardiomyopathy than any of the three compound heterozygous subgroups. Another study found an almost tenfold increase in diabetes in compound heterozygotes compared to those with biallelic GAA repeats [Greeley et al 2014].

Compound Heterozygotes for a Full-Penetrance GAA Expansion and a Borderline "Mutable" Allele

Individuals with somatically unstable, borderline alleles often have LOFA/VLOFA, mild and gradually progressive disease, and normal reflexes/hyperreflexia [Sharma et al 2004].