A study of up to 12 years of follow-up of Friedreich ataxia utilising four measurement tools

RESEARCH PAPER

A study of up to 12 years of follow-up of Friedreichataxia utilising four measurement toolsGeneieve Tai,1 Louise A Corben,1,2 Lyle Gurrin,3 Eppie M Yiu,1,4,5

Andrew Churchyard,2 Michael Fahey,6 Brian Hoare,2 Sharon Downie,2

Martin B Delatycki1,2,4,6,7

For numbered affiliations seeend of article.

Correspondence toProfessor Martin B Delatycki,Bruce Lefroy Centre for GeneticHealth Research, MurdochChildrens Research Institute,Flemington Rd, Parkville,VIC 3052, Australia;[email protected]

Received 5 March 2014Revised 24 June 2014Accepted 16 July 2014

To cite: Tai G, Corben LA,Gurrin L, et al. J NeurolNeurosurg PsychiatryPublished Online First:[please include Day MonthYear] doi:10.1136/jnnp-2014-308022

ABSTRACTObjective To explore the progression of Friedreichataxia by analysing the change in scores of four clinicalmeasures (the Friedreich Ataxia Rating Scale (FARS), theInternational Cooperative Ataxia Rating Scale (ICARS),the Functional Independence Measure (FIM) and theModified Barthel Index (MBI)) over a period of up to12 years, to ascertain the effects of clinical variables onperformance of these measures, and to determine themost sensitive rating scale for measuring diseaseprogression.Methods We measured the disease progression of upto 147 individuals against disease duration grouped into5-year intervals. Additional subgroups were created tostudy the effects of the size of the smaller FXN intron 1GAA repeat size (GAA1) and onset age on rating scaleperformance.Results Both the FARS and ICARS demonstratedgreater change in the first 20 years post disease onsetthan in the subsequent 20 years during which there waslittle change in the mean score. While the FIM and MBIcontinued to deteriorate beyond 20 years post diseaseonset, floor effects were noted. As measured by theFARS, individuals with a larger GAA1 repeat were foundto progress more quickly in the first 20 years of disease.Conclusions Individuals with larger GAA1 repeat sizesand earlier ages of disease onset were shown todeteriorate at a faster rate and were associated withgreater FARS and ICARS scores and lower FIM and MBIscores, which are indicative of greater disease severity.

INTRODUCTIONFriedreich ataxia (FRDA) is the most commoninherited ataxia with a prevalence of about 1 in29 0001 and is characterised by progressive ataxia,dysarthria, scoliosis, hypertrophic cardiomyopathyand diabetes mellitus.2 Disease onset usually occursbefore 25 years of age, with affected individualsrequiring full-time wheelchair use on average15 years later.3

FRDA is caused by mutations in FXN.Approximately 93% of affected individuals arehomozygous for an expanded GAA triplet repeat inintron 1 of the gene with the remainder being com-pound heterozygous for a GAA expansion and apoint mutation/deletion.4 The size of the smallertriplet repeat (GAA1) has been found to be themain predictor of disease outcome. An inverse cor-relation between the GAA1 repeat length and a

number of clinical variables including age at diseaseonset has been demonstrated.3 5–7

At present, there is no therapy proven to delayor reverse the progression of FRDA; however, anumber of potential therapeutic agents have beenidentified, including antioxidants, iron chelatorsand agents that increase frataxin expression.8 It isthus critical that assessment tools used to measurethe efficacy of such treatments are accurate and sen-sitive to even the smallest clinical change.The aim of the current study was to explore the

natural history of FRDA by analysing the change inannual scores of the Friedreich Ataxia Rating Scale(FARS), the International Cooperative AtaxiaRating Scale (ICARS), the Functional IndependenceMeasure (FIM) and the Modified Barthel Index(MBI) over a period of up to 12 years.

METHODSParticipantsOne hundred and forty-seven individuals withFRDA due to homozygosity for an expansion of aGAA repeat in intron 1 of FXN were examinedover a period of up to 12 years. Examinations gen-erally occurred 12 months apart. Participants wererecruited from a dedicated FRDA clinic in Victoria,Australia.

Standard protocol approval, registration andpatient consentThis study had approval from the Southern HealthHuman Research Ethics Committee (ProjectNumber 02114A). All participants gave theirinformed, written consent in accordance with the1964 Declaration of Helsinki.

Measurement toolsThe scales used in this study to examine the naturalhistory of FRDA were the FARS, ICARS, MBI andFIM.The FARS consists of three subscales measuring

ataxia, activities of daily living (ADL) and a neuro-logical examination. It is scored out of 167. HigherFARS scores indicate worse disease.9 The ICARSconsists of 19 items divided into four subscaleswith a maximum score of 100.8 10 11 The subscalesassess posture and gait, limb movement, speech andoculomotor disturbance. Higher ICARS scores indi-cate greater disease severity.The Barthel Index (BI) and FIM are two of the

most widely used generic measures of functional

Tai G, et al. J Neurol Neurosurg Psychiatry 2014;0:1–7. doi:10.1136/jnnp-2014-308022 1

Neurodegeneration JNNP Online First, published on August 11, 2014 as 10.1136/jnnp-2014-308022

Copyright Article author (or their employer) 2014. Produced by BMJ Publishing Group Ltd under licence.

group.bmj.com on August 13, 2014 - Published by jnnp.bmj.comDownloaded from

ability.12–14 The original BI was modified to improve its sensitiv-ity and reliability.14 While the MBI consists of the same testitems and is also scored out of 100, each item is given a differ-ent weighting to the original BI.15 Higher MBI scores indicategreater independence in performing ADL.12–14

The 18-item FIM is considered a more comprehensive andresponsive measure of functional status in comparison to theMBI.13 16 17 The FIM consists of six subscales evaluating self-care, sphincter control, mobility, locomotion, communicationand social cognition. It results in a score of 18–126 with higherscores indicative of better function. The FIM has added sensitiv-ity compared to the MBI in that it acknowledges the require-ment and use of assistive equipment in completing ADL.14

All four tools have good inter-rater reliability and satisfy trad-itional psychometric criteria. They were administered by trainedclinicians according to published protocols.

FXN GAA repeat size analysisThe size of the intron 1 GAA repeat size was analysed in onelaboratory, as previously described.18

Data analysisDisease progression of participants was measured against diseaseduration (calculated as the difference between the participant’sage at review and age at disease onset) grouped into 5-yearintervals (1–5, 6–10, 11–15, 16–20, 21–25, 26–30, 31–35 and36–40 years). Age at disease onset was defined as the age atwhich symptoms relating to FRDA were first noted, either bythe individual or a parent. Since participants were followed-upannually, some contributed multiple data points to a single5-year interval. To accommodate these repeated measures, weused a multiple linear regression model with indicator variablesfor each 5-year interval and robust SEs to ensure correct infer-ences. Such an approach makes optimal use of the data even ifthe numbers of repeated measures on participants are not com-pletely balanced between groups.19–21

In order to further explore the effects of disease duration onperformance and discern the population trend for the meanmeasurement scores in the four rating scales, the mean outcomeof the four scales for each 5-year interval was comparedbetween subgroups defined in two ways. The first subgroup ana-lysis was based on the GAA repeat size categorised as above orbelow the median repeat number. The second subgroup wasbased on age at disease onset categorised as above or below themedian. Some studies indicate that individuals with an earlierage of onset progress more quickly than those with a later ageof onset.6 The time trend for each measurement tool for eachsubgroup was estimated by generating the sample mean and SEof the mean for each 5-year time interval after disease onset.The estimated mean difference between groups for each 5-yearinterval was calculated by including an interaction between a(binary) group variable and the 5-year time interval in theregression model described above. This model was comparedusing likelihood ratios (a general measure of the goodness-of-fitof a model to the data) to the model where the mean differencebetween groups is assumed to be constant over time, that is, amodel where the magnitude of the association between groupand outcome is not allowed to vary over time.

The rate of change of outcome measures was comparedbetween the first 20 years and the subsequent 20 years post-diagnosis by fitting linear regression models that allowed for aseparate rate of change in each of these two time periods. Each ofthese rates of change and their difference, with correspondingSEs, was derived from the output of the linear regression proce-dures. All regression models used robust SEs to accommodatecorrelation between repeated measurements on the same individ-ual participant. These time periods were chosen because first wehave data on up to 40 years disease duration, and second becausethe FARS and ICARS have a noticeable slowing in the rate ofchange beyond 20 years’ disease duration (see figure 1).

A standardised effect size was defined as the rate of change forthe period 1–20 years post-diagnosis divided by the approximate

Figure 1 Trends in the FARS, ICARS, FIM and MBI according to the GAA1 repeat size. Note: ‘ALL’ indicates the trends in the four scales for thetotal cohort. FARS, Friedreich Ataxia Rating Scale; FIM, Functional Independence Measure; ICARS, International Cooperative Ataxia Rating Scale;MBI, Modified Barthel Index.

2 Tai G, et al. J Neurol Neurosurg Psychiatry 2014;0:1–7. doi:10.1136/jnnp-2014-308022

Neurodegeneration

group.bmj.com on August 13, 2014 - Published by jnnp.bmj.comDownloaded from

SD (the SE multiplied by the square root of a nominal samplesize of n=100); the same calculation was performed for theperiod 21–40 years post-diagnosis. A sample size was calculatedto determine the number of participants required to distinguish areduction of 50% in the rate of change of the outcome in a subse-quent period of similar length for both the period 1–20 yearspost-diagnosis and the period 21–40 years post-diagnosis.

All statistical analyses were performed using Stata StatisticalSoftware V.10.0 (Stata Corp., 2007, College Station, Texas, USA).

RESULTSParticipant demographics are shown in table 1. The mean age ofthe 147 participants at initial examination was 29.0 years (SD12.7), and 79 (53.7%) were male. The mean age at disease onsetwas 13.7 years (SD 7.2) and the mean disease duration was15.3 years (SD 10.4). Not every participant completed all fourassessments at each visit. Participants were followed-up for up to12 years, with a mean follow-up duration of 3.5 years (SD 2.5).

Figures 1 and 2 indicate the trends in all four measurementtools according to the GAA1 repeat size and age of onset,respectively, in 5-year disease duration intervals, for up to40 years. A plateau in the FARS and ICARS is evident fromdisease duration greater than 20 years; this is less apparent inthe FIM and the MBI. The GAA1 repeat size and age of onsetappeared to influence the rate of change in the scale scores andthis was thus explored in more detail.

Subgroup 1: GAA repeat sizeThe GAA1 repeat size has been found to correlate more oftenand to a greater degree with various measures of disease severitythan GAA21 and was thus used in this subgroup analysis. Westudied the effects of the GAA2 repeat size against clinical para-meters in our cohort but found no significant correlations (datanot shown). The groups were then characterised according tobeing above or below the median GAA1 repeat number, whichwas 669.

There was a significant difference between the two GAA1repeat groups for the FARS scores at years 6–10 (7.5 points,95% CI (0.9 to 14.2)), 11–15 (8.6 points, 95% CI (2.0 to15.1)) and 16–20 (19.3 points, 95% CI (11.3 to 27.4)) andICARS scores at years 6–10 (7.4 points, 95% CI (2.7 to 12.2)),11–15 (5.1 points, 95% CI (0.4 to 9.6)) and 16–20 (13.9points, 95% CI (8.0 to 19.7)) (figure 1, table 2). Examinationof the trends for MBI and FIM (figure 1, table 2) reported anopposite trend to that seen for the FARS and ICARS. Lowerscores in the MBI and FIM indicate worsening disease whilelower scores in the FARS and ICARS indicate improvement.There was no significant difference between groups for any timeinterval for the MBI. There was a significant difference at years31–35 for the FIM (8.6 points, 95% CI (16.3 to 0.8)).Regardless of the GAA1 repeat size, individuals in the first yearsof the disease tended to have maximum FIM and MBI scores,indicating total independence with performing ADL. Individualswith larger GAA1 repeat sizes deteriorated in their ability toperform ADLs at a quicker rate, particularly in the third decadeof disease, perhaps indicative of disease progression.

Subgroup 2: age of onsetIndividuals with an age of disease onset greater than the cohortmedian of 14 years demonstrated lower FARS and ICARS scoresthroughout the duration of the disease, indicating less severedisease when compared to the subgroup of participants with anearlier age of disease onset (figure 2). A significant difference inFARS scores occurred in the second decade of disease, particu-larly at years 11–15 after disease onset (−16.8 points, 95% CI(−23.3 to −10.3)) and years 16–20 (−13.1 points, 95% CI(−21.0 to −5.2)) (table 3). Significant differences were alsodemonstrated at years 26–30 (−11.6 points, 95% CI (−20.6 to−2.6)) and 36–40 (−15.0 points, 95% CI (−29.5 to −0.4)). Asimilar trend is noted for the ICARS, with significant differencesbetween groups at years 11–15 (−11.7 points, 95% CI (−16.4to −7.1)), 16–20 (−11.0 points, 95% CI (−16.8 to −5.3)) and26–30 (−8.4 points, 95% CI (−14.8 to −2.0)) (table 3).

As with the GAA1 repeat size subgroup, there was little differ-ence in functional ability as shown in MBI and FIM perform-ance scores between groups at disease onset. However, as thedisease progressed, those with a later age of onset had a slowerdecline in the ability to perform ADL independently; individualswith a disease onset later than 14 years demonstrated higherscores at each time interval (figure 2). Significant differenceswere observed for the MBI at the second, third and first half ofthe fourth decades: years 11–15 (10.9 points, 95% CI (5.6 to16.2)), 16–20 (8.4 points, 95% CI (2.4 to 14.5)), 21–25 (6.7points, 95% CI (0.0 to 13.3)), 26–30 (8.8 points, 95% CI (2.1to 15.5)) and 31–35 (20.5 points, 95% CI (12.0 to 29.0)).Similar results were noted in the FIM, with the greatest dispar-ities between groups noted at years 6–10 (6.3 points, 95% CI(1.4 to 11.3)), 11–15 (8.0 points, 95% CI (3.4 to 12.5)), 16–20(8.0 points, 95% CI (2.7 to 13.3)), 31–35 (19.0 points, 95% CI(11.6 to 26.4)) and 36–40 (9.4 points, 95% CI (0.3 to 18.4)).

Table 1 Participant characteristics

Characteristics (n=147)

Age at initial examination (years)Mean (SD) 29.0 (12.7)Range 8–70

GenderMale (%) 79 (53.7)

GAA1 repeat sizeMean (SD) 668.3 (221.5)Range 56–1298

GAA2 repeat sizeMean (SD) 889.0 (189.5)Range 182–1345

Age of onset (years)Mean (SD) 13.7 (7.2)Range 2–39

Disease duration (years)Mean (SD) 15.3 (10.4)Range 0.67–42.3

FARS score at initial examination (n=107)Mean (SD) 83.9 (31.2)Range 20.5–137

ICARS score at initial examination (n=104)Mean (SD) 47.6 (21.7)Range 1–83

FIM score at initial examination (n=124)Mean (SD) 107.7 (18.0)

Range 57–126MBI score at initial examination (n=125)Mean (SD) 78.2 (21.5)Range 21–100

FARS, Friedreich Ataxia Rating Scale; FIM, Functional Independence Measure; GAA1repeat size, GAA repeat size of smaller FXN allele; GAA2 repeat size, GAA repeat sizeof the larger FXN allele; ICARS, International Cooperative Ataxia Rating Scale; MBI,Modified Barthel Index.

Tai G, et al. J Neurol Neurosurg Psychiatry 2014;0:1–7. doi:10.1136/jnnp-2014-308022 3

Neurodegeneration

group.bmj.com on August 13, 2014 - Published by jnnp.bmj.comDownloaded from

How do the scales perform in measuring change over time?We initially examined the annual rate of change for the fouroutcome measures from the first 20 years and the subsequent20 years post disease onset (table 4). Both the FARS and ICARSdemonstrated differences in the rate of change over the twotime intervals; the first 20 years reported a greater rate ofchange with a difference of −2.6 points, 95% CI (−3.8 to −1.5)and −1.8 points, 95% CI (−2.6 to −1.1), respectively. Changesin the MBI and FIM over the two time intervals, on the otherhand, were not significantly different, with mean differences of0.6 points, 95% CI (−0.3 to 1.5) and 0.2 points, 95% CI (−0.6to 1.1), respectively.

Examination of differences in subgroups over time revealed asignificant difference in the rate of change in the FARS score inthe 0–20 year time interval (1.5 points, 95% CI (0.1 to 2.8))between the GAA1 repeat size subgroups whereby individualswith larger GAA1 repeat sizes demonstrated a quicker rate ofchange (table 5). There were no other significant differences inthe rate of change in the 0–20 or 21–40 year time periods postdisease onset for any of the four outcome measures between ageof onset or GAA1 subgroups.

Power calculations revealed that fewer subjects are required ifthe FARS and ICARS are used to measure clinical severity in thefirst 20 years of disease onset where the GAA1 repeat size is

Figure 2 Trends in the FARS, ICARS, FIM and MBI according to age of onset. Note: ‘ALL’ indicates the trends in the four scales for the totalcohort. FARS, Friedreich Ataxia Rating Scale; FIM, Functional Independence Measure; ICARS, International Cooperative Ataxia Rating Scale; MBI,Modified Barthel Index.

Table 2 Differences in mean outcome scores between GAA1 repeat size subgroups (GAA1 repeat ≤669 vs GAA1 repeat >669) for each timeinterval

Time interval sincedisease onset (years)

FARSSE (95% CI)

ICARSSE (95% CI)

MBISE (95% CI)

FIMSE (95% CI)

1–5 −7.95.5 (−18.8 to 3.0)

−4.63.9 (−12.3 to 3.2)

1.24.8 (−8.2 to 10.6)

0.64.1 (−7.5 to 8.7)

6–10 7.53.4 (0.9 to 14.2)

7.43.2 (2.7 to 12.2)

−1.13.0 (−7.0 to 4.8)

−0.72.6 (−5.8 to 4.5)

11–15 8.63.3 (2.0 to 15.1)

5.12.4 (0.4 to 9.6)

−2.52.8 (−8 to 2.9)

−3.22.4 (−7.9 to 1.5)

16–20 19.34.1 (11.3 to 27.4)

13.93.0 (8.0 to 19.7)

−4.73.2 (−11.0 to 1.7)

−4.32.8 (−9.8 to 1.2)

21–25 5.44.8 (−4 to 14.7)

4.83.4 (−1.9 to 11.6)

−4.43.6(−11.4 to 2. 7)

−3.33.1 (−9.4 to 2.8)

26–30 −3.94.7 (−13.1 to 5.2)

−1.43.3 (−7.9 to 5.1)

2.33.6 (−4.9 to 9.4)

3.73.2 (−2.5 to 9.9)

31–35 0.97.0 (−12.9 to 14.7)

2.54.9 (−7.2 to 12.2)

−7.74.6 (−16.7 to 1.2)

−8.64.0 (−16.3 to −0.8)

36–40 9.68.7 (−7.4 to 26.7)

6.86.1 (−5.2 to 18.8)

−11.35. 8 (−22.6 to 0.9)

−8.95.0 (−18.6 to 0.8)

FARS, Friedreich Ataxia Rating Scale; FIM, Functional Independence Measure; ICARS, International Cooperative Ataxia Rating Scale; MBI, Modified Barthel Index.

4 Tai G, et al. J Neurol Neurosurg Psychiatry 2014;0:1–7. doi:10.1136/jnnp-2014-308022

Neurodegeneration

group.bmj.com on August 13, 2014 - Published by jnnp.bmj.comDownloaded from

above the median (table 5). Age of onset did not impact samplesize in the same way as did the GAA1 repeat size.

DISCUSSIONThis study sought to describe the natural history of FRDA bystudying the trend in performance scores in the FARS, ICARS,MBI and FIM. The effect of two clinical variables, age of onsetand GAA1 repeat size, on performance scores was compared,and the ability of the four measurement tools in measuringchange over time was examined. In this study, individuals withlarger GAA1 repeat sizes had, on average, greater FARS andICARS scores, and lower FIM and MBI scores at similar diseaseduration compared to those with smaller GAA1 repeat sizes,reflecting greater impairment in the former group. Conversely,later age of disease onset was associated with lower FARS andICARS scores and greater FIM and MBI scores at similar diseaseduration. The influence of the GAA1 repeat size on the FARSand ICARS was most apparent in the earlier years of thedisease, whereas the influence of the GAA1 repeat size onthe functional measures was more marked in the latter years of

the disease. Age of disease onset also influenced the progressionof the disease, particularly in the earlier years; however, after20 years beyond disease onset, scores were similar for bothgroups. This presumably coincides with the requirement of full-time wheelchair use and ceiling effects of the scales.

Both the FARS and ICARS require fewer individuals withFRDA to detect a slowing of disease progression of 50%; hence,they are best suited for use in studies of intervention. Less studyparticipants are required if those recruited are within 20 yearsof disease onset compared to participants with a disease dur-ation beyond 20 years. Recruitment of those individuals with alarger GAA1 repeat size may also reduce the sample sizerequired. If the study requires recruitment of participants fromall stages of the disease, the MBI appears to be the best scale touse as it requires the least number of subjects with greater than20 years disease duration. As is the case in the current study,other studies have also shown the FARS to demonstrate ceilingeffects, and thus there is strong evidence that the FARS is oflimited utility in measuring change in the latter stages ofdisease.22 23 Conversely, the FIM and MBI continue to show

Table 3 Differences in mean outcome scores between age of onset subgroups for each time interval (age of onset <14 years vs age of onset≥14 years)

Time interval sincedisease onset (years)

FARSSE (95% CI)

ICARSSE (95% CI)

MBISE (95% CI)

FIMSE (95% CI)

1–5 −8.75.5 (−19.4 to 2.1)

−4.43.9 (−12.0 to 3.3)

1.14.7 (−8.2 to 10.4)

3.54.1 (−4.6 to 11.6)

6–10 −6.53.3 (−13.0 to 0.1)

−3.62.4 (−8.3 to 1.1)

4.62.9 (−1.0 to 10.2)

6.32.5 (1.4 to 11.3)

11–15 −16.83.3 (−23.3 to −10.3)

−11.72.4 (−16.4 to −7.1)

10.92.7 (5.6 to 16.2)

8.02.3 (3.4 to 12.5)

16–20 −13.14.0 (−21.0 to −5.2)

−11.02.9 (−16.8 to −5.3)

8.43.1 (2.4 to 14.5)

8.02.7 (2.7 to 13.3)

21–25 −4.04.7(−13.3 to 5.3)

−4.33.4 (−11.0 to 2.3)

6.73.4 (0.0 to 13.3)

3.43.0 (−2.4 to 9.2)

26–30 −11.64.6 (−20.6 to −2.6)

−8.43.3 (−14.8 to −2.0)

8.83.4 (2.1 to 15.5)

3.83.0 (−2.1 to 9.7)

31–35 −6.06.8 (−19.4 to 7.4)

3.94.8 (−13.3 to 5.6)

20.54.3 (12.0 to 29.0)

19.03.8 (11.6 to 26.4)

36–40 −15.07.4 (−29.5 to −0.4)

−9.65.2 (−19.9 to 0.6)

9.05.4 (−1.7 to 19.7)

9.44.6 (0.3 to 18.4)

FARS, Friedreich Ataxia Rating Scale; FIM, Functional Independence Measure; ICARS, International Cooperative Ataxia Rating Scale; MBI, Modified Barthel Index.

Table 4 Estimated rate of change and sample size calculations for the four outcome measures per year from years 1 to 20 and years 21 to 40post disease onset and power calculations estimating the number per group required for a 12 month study to detect a reduction of diseaseprogression of 50%

FARSSE; SD (95% CI)

ICARSSE; SD (95% CI)

MBISE; SD (95% CI)

FIMSE; SD (95% CI)

Years 1–20 3.10.4; 3.9 (2.3 to 3.9)

2.10.3; 2.7 (1.6 to 2.7)

−1.70.2; 2.0 (−2.1 to −1.3)

−1.10.2; 1.7 (−1.4 to −0.8)

Years 21–40 0.50.4; 3.8 (−0.3 to 1.2)

0.30.2; 2.3 (−0.1 to 0.8)

−1.10.4; 3.6 (−1.8 to −0.4)

−0.90.4; 3.5 (−1.6 to −0.2)

Difference in rate of change betweenthe two time intervals

−2.6(−3.8 to −1.5)

−1.8 (−2.6 to −1.1) 0.6 (−0.3 to 1.5) 0.2 (−0.6 to 1.1)

p Value <0.001 <0.001 0.199 0.586Effect size (years 1–20) 0.799 0.808 −0.872 −0.653Power calculation: participants per group 99 97 83 147Effect size (years 21–40) 0.124 0.142 −0.318 −0.244Power calculation: participants per group 4084 3097 621 1055

FARS, Friedreich Ataxia Rating Scale; FIM, Functional Independence Measure; ICARS, International Cooperative Ataxia Rating Scale; MBI, Modified Barthel Index.

Tai G, et al. J Neurol Neurosurg Psychiatry 2014;0:1–7. doi:10.1136/jnnp-2014-308022 5

Neurodegeneration

group.bmj.com on August 13, 2014 - Published by jnnp.bmj.comDownloaded from

significant change beyond 20 years post disease onset; however,both measures have a major floor effect with little change fromthe maximum score early in the disease course. Significant dif-ferences in the rate of change of scores were not seen for any ofthe measures assessed between those whose onset was earlierthan the median compared to those whose onset was later. Theimplications for design of clinical trials are that recruitment ofindividuals with larger GAA1 repeat sizes in the first 20 yearsafter disease onset will maximise study power; however, in thisregard, the age of disease onset is less important. It should benoted, however, that the results of our study are based on mul-tiple annual data points for many of the individuals (approxi-mately four on average per individual), indicating acorresponding loss of power for study designs where patientscontribute only two data points.

The natural history of FRDA has been studied and comparedwith other degenerative ataxias including early onset cerebellarataxia and autosomal dominant cerebellar ataxias.24

Klockgether et al24 used progression to full-time wheelchair useas a proxy measure of disease severity in various ataxias andfound that individuals with an earlier disease onset progressedto full-time wheelchair use at a quicker rate. Similarly, thecurrent study found that those individuals with an earlierdisease onset progressed more quickly to severe disease withassociated functional decline;however, the two groups reached asimilar level by 20 years after disease onset as measured by thescales used. This effect could be a function of the diseaseprocess and/or result from the floor/ceiling effects of the meas-uring tools used.

Other measurement tools that have been advocated for use inFRDA clinical trials are the Scale for the Assessment and Ratingof Ataxia (SARA)25 26 and functional composites.7 The SARA hasbeen found to change over a mean study period of 1.84 years

and unlike the current study, change was not found to be influ-enced by any clinical parameters including age of onset, diseaseduration and GAA1 repeat length.27 The Friedreich AtaxiaFunctional Composite (FAFC) consists of the 25-foot walk, the9-hole peg test and the Sloan Low Contrast Letter Chart. A studyof the rate of disease progression over 1 and 2 years using theFARS and FAFC found that the FAFC had less ceiling and flooreffects and a more linear trend over the 2 years when comparedto the FARS.22 Both the SARA and the FAFC, however, are morerecently published scales and were not therefore available for usewhen data for this study started being collected. Moreover, theSARA had not initially been used to study FRDA as it wasdeemed unsuitable due to FRDA being characterised as a sensorytype of ataxia.26 The three measures included in the FAFC arealso susceptible to floor and ceiling effects, and may be best usedin specific FRDA populations in order for clinical change to bedemonstrated. There were substantial floor effects shown withthe use of the Z2 composite in individuals with more severedisease. Lynch et al23 thus suggest using the Z3 compositemeasure where these effects are reduced.

In conclusion, we have shown that the ICARS and FARS aremost sensitive to change in the first 20 years of disease in indivi-duals with FRDA. Moreover, those with an earlier age of onsetand a larger GAA1 repeat size generally have a more severedisease profile as measured by all four tools over the entiredisease course. In clinical trials where measurement of the rateof disease progression is a primary outcome, participants withlarger GAA1 repeat sizes and in the first 20 years of disease pro-gression should be considered ideal candidates for participation.

Author affiliations1Bruce Lefroy Centre for Genetic Health Research, Murdoch Childrens ResearchInstitute, Parkville, Victoria, Australia

Table 5 Comparison of rate of change over the first 20 years of disease between the GAA1 repeat size and age of onset subgroups

FARSSE; SD (95% CI)

ICARSSE; SD (95% CI)

MBISE; SD (95% CI)

FIMSE; SD (95% CI)

GAA1 repeat size ≤669 2.40.5; 5.0 (1.4 to 3.4)

1.70.4; 3.6 (1.0 to 2.5)

−1.50.3; 2.8 (−2.1 to −1.0)

−0.90.2; 2.3 (−1.4 to −0.5)

GAA1 repeat size >669 3.90.5; 4.5 (3.0 to 4.8)

2.60.3; 3.0 (2.0 to 3.2)

−2.00.3; 2.6 (−2.5 to −1.5)

−1.30.2; 2.3 (−1.8 to −0.9)

Difference 1.50.7; 6.8 (0.1 to 2.8)

0.90.5; 4.7 (−0.1 to 1.8)

−0.50.4; 3.8 (−1.2 to 0.3)

−0.40.3; 3.3 (−1.0 to 0.3)

p Value* 0.031 0.063 0.226 0.239Effect size (GAA1 repeat size ≤669) 0.481 0.486 −0.538 −0.413Power calculation: participants per group 273 266 217 370Effect size (GAA1 repeat size >669) 0.859 0.873 −0.769 −0.562Power calculation: participants per group 86 83 106 199

Age of onset <14 years 3.40.6; 6.0 (2.2 to 4.6)

2.50.4; 4.2 (1.7 to 3.4)

−2.00.4; 3.8 (−2.8 to −1.3)

−1.20.3; 3.1 (−1.8 to −0.6)

Age of onset ≥14 years 2.90.5; 4.8 (1.9 to 3.8)

1.90.3; 3.2 (1.2 to 2.5)

−1.40.2; 1.7 (−1.8 to −1.1)

−0.90.1; 1.4 (−1.2 to −0.7)

Difference −0.50.8; 7.6 (−2.1 to 1.0)

−0.70.5; 5.3 (−1.7 to 0.4)

0.60.4; 4.1 (−0.2 to 1.4)

0.30.3; 3.4 (−0.4 to 0.9)

p Value† 0.482 0.796 0.152 0.424Effect size (age of onset <14 years) 0.570 0.595 −0.533 −0.395Power calculation: participants per group 194 177 222 405Effect size (age of onset ≥14 years) 0.595 0.580 −0.839 −0.671Power calculation: participants per group 177 187 90 140

*Comparison of the estimated rate of change between those with GAA1 greater to those with GAA1 less than the median value (669).†Comparison of the estimated rate of change between those with age of onset greater to those with age of onset less than the median value (14 years).FARS, Friedreich Ataxia Rating Scale; FIM, Functional Independence Measure; ICARS, International Cooperative Ataxia Rating Scale; MBI, Modified Barthel Index; SD, approximate SD.

6 Tai G, et al. J Neurol Neurosurg Psychiatry 2014;0:1–7. doi:10.1136/jnnp-2014-308022

Neurodegeneration

group.bmj.com on August 13, 2014 - Published by jnnp.bmj.comDownloaded from

2Monash Health, Clayton, Victoria, Australia3Department of Molecular, Environmental, Genetic and Analytic Epidemiology,University of Melbourne, Parkville, Victoria, Australia4Department of Paediatrics, University of Melbourne, Parkville, Victoria, Australia5Children’s Neuroscience Centre, Royal Children’s Hospital, Parkville, Victoria,Australia6Department of Paediatrics, Monash University, Clayton, Victoria, Australia7Department of Clinical Genetics, Austin Health, Heidelberg, Victoria, Australia

Acknowledgements The authors thank the participants for their involvement inthis study.

Contributors GT was involved in the acquisition of data, analysis andinterpretation of data, and drafting of the manuscript. LAC was involved in the studyconcept and design, acquisition of data, analysis and interpretation of results anddrafting of the manuscript. LG was involved in the analysis and interpretation ofdata and contribution to the manuscript. EMY was involved in the acquisition,analysis and interpretation of data and contribution to the manuscript. AC, MF, BHand SD were involved in the acquisition of data and contribution to the manuscript.MBD was involved in the study concept and design, acquisition of data, analysis andinterpretation of results, drafting of the manuscript and study supervision. He is theguarantor for the paper.

Funding Funding for the study was from the Friedreich Ataxia Research Association(Australasia) and the Friedreich Ataxia Research Alliance (USA). This study wassupported by the Victorian Government’s Operational Infrastructure SupportProgram.

Competing interests LAC and EMY are Early Career Fellows and LG is a CareerDevelopment Fellow, all of the National Health and Medical Research Council.

Ethics approval Southern Health Human Research Ethics Committee.

Provenance and peer review Not commissioned; externally peer reviewed.

Data sharing statement The data from this study are available to others to viewon request. There are no additional unpublished data.

REFERENCES1 Delatycki MB, Williamson R, Forrest SM. Friedreich ataxia: an overview. J Med

Genet 2000;37:1–8.2 Delatycki MB, Corben LA. Clinical features of Friedreich ataxia. J Child Neurol

2012;27:1133–37.3 Delatycki MB, Paris DBBP, Gardner RJM, et al. Clinical and genetic study of

Friedreich ataxia in an Australian population. Am J Med Genet 1999;87:168–74.4 Pandolfo M. Friedreich Ataxia: the clinical picture. J Neurol 2009;256:3–8.5 Monrós E, Moltó MD, Martínez F, et al. Phenotype correlation and intergenerational

dynamics of the Friedreich ataxia GAA trinucleotide repeat. Am J Hum Genet1997;61:101–10.

6 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 Genet1996;59:554–60.

7 Lynch DR, Farmer JM, Wilson RL, et al. Performance measures in Friedreich ataxia:potential utility as clinical outcome tools. Mov Disord 2005;20:777–82.

8 Delatycki MB. Evaluating the progression of Friedreich ataxia and its treatment.J Neurol 2009;256:36–41.

9 Subramony SH, May W, Lynch DR, et al. Measuring Friedreich ataxia: interraterreliability of a neurologic rating scale. Neurology 2005;64:1261–2.

10 Trouillas P, Takayanagi T, Hallett M, et al. International Cooperative Ataxia RatingScale for pharmacological assessment of the cerebellar syndrome. The AtaxiaNeuropharmacology Committee of the World Federation of Neurology. J Neurol Sci1997;145:205–11.

11 Storey E, Tuck K, Hester R, et al. Inter-rater reliability of the InternationalCooperative Ataxia Rating Scale (ICARS). Mov Disord 2004;19:190–2.

12 Ravaud J-F, Delcey M, Yelnik A. Construct validity of the Functional IndependenceMeasure (FIM): questioning the unidimensionality of the scale and the ‘value’ ofFIM scores. Scand J Rehabil Med 1999;31:31–41.

13 Sangha H, Lipson D, Foley N, et al. A comparison of the Barthel Index and theFunctional Independence Measure as outcome measures in stroke rehabilitation:patterns of disability scale usage in clinical trials. Int J Rehabil Res 2005;28:135–9.

14 Cohen ME, Marino RJ. The tools of disability outcomes research functional statusmeasures. Arch Phys Med Rehabil 2000;81:S21–9.

15 Shah S, Vanclay F, Cooper B. Improving the sensitivity of the Barthel Index forstroke rehabilitation. J Clin Epidemiol 1989;42:703–9.

16 Ring H, Feder M, Schwartz J, et al. Functional measures of first-stroke rehabilitationinpatients: usefulness of the functional independence measure total score with aclinical rationale. Arch Phys Med Rehabil 1997;78:630–5.

17 Hobart JC, Lamping DL, Freeman JA, et al. Evidence-based measurement: whichdisability scale for neurologic rehabilitation? Neurology 2001;57:639–44.

18 Evans-Galea M, Corben L, Hasell J, et al. A novel deletion–insertion mutationidentified in exon 3 of FXN in two siblings with a severe Friedreich ataxiaphenotype. Neurogenetics 2011;12:307–13.

19 Huber PJ. The behaviour of maximum likelihood estimates under non-standardconditions. In: Fifth Berkeley Symposium in Mathematical Statistics and Probability.Berkeley, CA: University of California Press, 1967:221–33.

20 White H. Maximum likelihood estimation of misspecified models. Econometrica1980;50:1–25.

21 Royall RM. Model robust confidence intervals using maximum likelihood estimators.Int Stat Rev 1986;54:221–6.

22 Friedman LS, Farmer JM, Perlman S, et al. Measuring the rate of progression inFriedreich ataxia: implications for clinical trial design. Mov Disord 2010;25:426–32.

23 Lynch DR, Farmer JM, Tsou AY, et al. Measuring Friedreich ataxia: complementaryfeatures of examination and performance measures. Neurology 2006;66:1711–16.

24 Klockgether T, Ludtke R, Kramer B, et al. The natural history of degenerative ataxia:a retrospective study in 466 patients. Brain 1998;121:589–600.

25 Schmitz-Hubsch T, du Montcel ST, Baliko L, et al. Scale for the assessment andrating of ataxia: development of a new clinical scale. Neurology 2006;66:1717–20.

26 Bürk K, Mälzig U, Wolf S, et al. Comparison of three clinical rating scales inFriedreich ataxia (FRDA). Mov Disord 2009;24:1779–84.

27 Marelli C, Figoni J, Charles P, et al. Annual change in Friedreich’s ataxia evaluatedby the Scale for the Assessment and Rating of Ataxia (SARA) is independent ofdisease severity. Mov Disord 2012;27:135–9.

Tai G, et al. J Neurol Neurosurg Psychiatry 2014;0:1–7. doi:10.1136/jnnp-2014-308022 7

Neurodegeneration

group.bmj.com on August 13, 2014 - Published by jnnp.bmj.comDownloaded from

doi: 10.1136/jnnp-2014-308022 published online August 11, 2014J Neurol Neurosurg Psychiatry

 Geneieve Tai, Louise A Corben, Lyle Gurrin, et al. tools

measurementFriedreich ataxia utilising four years of follow-up of A study of up to 12

http://jnnp.bmj.com/content/early/2014/08/10/jnnp-2014-308022.full.htmlUpdated information and services can be found at:

These include:

References http://jnnp.bmj.com/content/early/2014/08/10/jnnp-2014-308022.full.html#ref-list-1

This article cites 26 articles, 3 of which can be accessed free at:

P<P Published online August 11, 2014 in advance of the print journal.

serviceEmail alerting

the box at the top right corner of the online article.Receive free email alerts when new articles cite this article. Sign up in

CollectionsTopic

(459 articles)Spinal cord   � (612 articles)Brain stem / cerebellum   �

 Articles on similar topics can be found in the following collections

Notes

(DOIs) and date of initial publication. publication. Citations to Advance online articles must include the digital object identifier citable and establish publication priority; they are indexed by PubMed from initialtypeset, but have not not yet appeared in the paper journal. Advance online articles are Advance online articles have been peer reviewed, accepted for publication, edited and

http://group.bmj.com/group/rights-licensing/permissionsTo request permissions go to:

http://journals.bmj.com/cgi/reprintformTo order reprints go to:

http://group.bmj.com/subscribe/To subscribe to BMJ go to:

group.bmj.com on August 13, 2014 - Published by jnnp.bmj.comDownloaded from