Background

Friedreich’s ataxia (FRDA) is the most common inherited ataxia affecting about 1 in 50,000 Europeans [1]. It is a rare hereditary autosomal recessive disease caused by mutations in the frataxin (FXN) gene resulting in decreased levels of functional frataxin protein [2]. The exact pathophysiologic repercussions of FXN loss are not yet fully understood. FXN is involved in mitochondrial iron homeostasis through iron-sulfur cluster synthesis and iron storage regulation [3]. FRDA is a progressive multisystem disorder presenting with limb ataxia, proprioception loss, gait and speech disturbances, myocardial involvement, and skeletal deformities with typical age of onset during puberty [4]. Moreover, diabetes mellitus, visual deficits, and hearing loss often co-exist. During early adulthood, most FRDA patients are non-ambulatory [5]. However, myocardial involvement represents the major determinant of survival [6].

Recently, omaveloxolone, a potent activator of Nrf2 which represents a key mediator of the antioxidant response in FRDA but also a NF-kB-mediated inflammatory response suppressor, a trait with unclear effects on FRDA pathophysiology, was approved by the FDA as the first agent for the therapeutic management of adults and adolescents aged 16 years and older with FRDA [7, 8]. In general, therapeutic approaches can be categorized into three groups based on their mechanism of action. The mainstay in FRDA patients’ management relies on symptomatic approaches such as drugs for cardiac arrhythmias, cerebellar symptoms, and diabetes. The second group comprises drugs that augment mitochondrial function like omaveloxolone, idebenone, EPI-743, deferiprone, riboflavin, epicatechin, coenzyme Q10, vitamin E, l-carnitine, and creatine. The third group includes drugs that may possibly increase frataxin levels such as erythropoietin, interferon gamma, resveratrol, and nicotinamide. In addition, research efforts led to the emergence of potential disease-modifying FRDA treatments mainly based on frataxin gene modulation [9].

The difficulty in demonstrating possible efficacy of various agents tested in clinical trials could be attributed, to some degree, to the lack of a suitable quantifiable biomarker detecting slow disease progression or subtle response to a possibly effective treatment in the context of a trial’s timeline. This issue is commonly encountered in the design of trials involving patients with rare, slowly progressive heterogeneous neurodegenerative disorders [10, 11]. It should be a key priority to stratify FRDA subpopulations and utilize the most effective outcome measure for each subgroup that could depict subtle disease progression and thus responsiveness to various therapies [12]. The European Friedreich’s Ataxia Consortium for Translational Studies (EFACTS) study group recently examined 602 treatment-naive FRDA patients in an effort to provide sensitive outcome measures to monitor change over time in different stages of the disease [13].

The objective of this study is to systematically review the literature and conduct a meta-analysis to summarize and evaluate the biomarkers used to assess therapeutic efficacy in Friedreich’s ataxia patients receiving any treatment.

Materials and Methods

Our protocol was pre-published online at the International Prospective Registry of Systematic Reviews (PROSPERO accession number CRD42022319196).

The reporting of this systematic review and meta-analysis follows the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) 2020 statement (online-only Supplementary material 1) [14].

Search Strategy and Eligibility Criteria

We searched MEDLINE (via PubMed), EMBASE (via Ovid), and the Cochrane Library for original studies (with more than five participants) investigating the effect of therapeutic interventions on any biomarker in patients with genetically confirmed FRDA. We excluded case reports which were defined as studies that included ≤ 4 patients based on previous literature reviews [15, 16]. We included published full-text articles that provided numerical analysis of biomarker changes measured at least during two different timepoints (with a minimum 2-month interval). The search strategy was developed based on the combination of the following key words: Friedreich’s Ataxia, therapeutics, and treatment outcome as presented at online-only Supplementary material 2. We also reviewed references of previous systematic reviews and included records. We did not apply any language or year of publication filter. We completed the literature search on March 18, 2022 and updated it on June 3, 2023.

Data Extraction (Selection and Coding)

Title and abstract screening along with duplicate record removal were performed independently by two reviewers (M.G., M.M.). Two reviewers (M.G., E.C.) examined the full texts of the remaining records. Any disagreement regarding study eligibility was resolved by a senior author (V.K.). We extracted unadjusted raw data on a standardized Microsoft Excel spreadsheet regarding study characteristics, patients’ baseline characteristics, intervention type, the mean with corresponding SD, and 95% CI for each biomarker on two different time points. Whenever studies did not report mean and SD, we calculated the mean and SD from the data provided (sample size, IQR, SEM) [17]. For records that did not report quantitative assessments of some biomarkers, we extracted relative information to conduct a narrative summary of the main findings.

Risk-of-Bias Assessment in Individual and across Studies

To ascertain the validity of the included records, two independent reviewers (M.G. and E.C.) assessed risk of bias (RoB) using the revised Cochrane Collaboration’s Risk of Bias tool for randomized controlled trials (RCTs), Newcastle–Ottawa Scale (NOS) for case–control or cohort studies, Risk Of Bias In Non-randomized Studies of Exposure (ROBINS-E) for open-label trials, and a tool proposed by Murad et al. for case series [18,19,20,21]. According to these scales, the studies were ranked as high, fair, or low risk of bias. We also aimed to evaluate the quality of evidence applying the Grading of Recommendations, Assessment, Development and Evaluations (GRADE) approach [22].

Summary Measures and Synthesis of Results

We conducted a narrative summary of the records included in this systematic review, reporting the biomarkers examined and the therapeutic effect as reported in each study using descriptive statistics. First, biomarkers were grouped as follows: (a) clinical outcome measures, (b) cardiac biomarkers, (c) biochemical biomarkers, (d) patient-reported outcome measures (PROMs), (e) imaging biomarkers, (f) neurophysiologic biomarkers, and (g) other biomarkers. Second, records were classified according to the mechanism of action of the administered intervention into three groups: (1) drugs that augment mitochondrial function, (2) drugs that increase frataxin, and (3) symptomatic treatment. The assignment of each record retrieved from literature search in each group of therapeutic approaches according to the mechanism of action of the administered intervention was based on Friedreich’s Ataxia Research Alliance (FARA) treatment pipeline [23].

The primary outcome of interest was the change from baseline score of any biomarker examined following therapeutic intervention. We also aimed to investigate the change from baseline score of any biomarker in the placebo or no treatment group separately to evaluate the ability of the biomarker to detect subtle disease progression.

We performed a quantitative synthesis for each biomarker type whenever possible using Comprehensive Meta-analysis software (version 3.0; Biostat Inc.). We calculated the standardized mean difference (SMD) with a 95% confidence interval (CI) with a significance level set at p < 0.05 whenever three or more studies with the same intervention type reported the same biomarker. We calculated the effect size for each biomarker using a random-effect model based on follow-up sample size. We used a conservative value of 0.5 as a correlation between pre- and post-treatment assessments whenever this correlation was not reported in the original record [11]. Whenever data from the same cohort were published in more than one record, we included in the analysis the most relevant study with the largest sample size to eliminate overlapping samples. Publication bias and heterogeneity were examined by visual assessment of the funnel plots and by calculating the I2, respectively [24, 25].

We planned sensitivity analyses based on risk-of-bias assessment along with subgroup analysis according to drug administered. Additional explorative analyses were planned depending on data availability according to population age (children, adolescents, or adults) or different follow-up timepoints. However, there were insufficient data to perform these subgroup analyses.

Results

Descriptive Characteristics and RoB within Studies

Among 783 unique records retrieved from the literature search, 55 records fulfilled the prespecified inclusion criteria. Study selection process is presented as a PRISMA flow diagram in Fig. 1.

Fig. 1
figure 1

Flowchart of study selection

Full size image

We included 43 studies presented at 55 records in the final qualitative synthesis. These 43 studies examined 1409 FRDA participants (age range 4–74 years) during a follow-up period of 2 to 132 months. The intervention type included nine different drugs or drug combinations that augment mitochondrial function in 25 studies (31 records), four drugs or drug combinations that increase frataxin in ten studies (16 records), or eight different symptomatic treatments in eight studies/records. The 55 records included 19 RCTs, 25 open-label trials, 10 cohort studies, and one case series. The characteristics of the included records are summarized in Table 1.

Table 1 Descriptive characteristics of 53 studies included in the qualitative synthesis
Full size table

Eleven records were deemed of low risk of bias (RoB), six of fair quality, and 38 were deemed of high RoB. The majority of non-randomized studies (31 out of 36 open-label, cohort, or case-series) presented high risk of bias mainly because of the inadequate identification of potential confounding factors. Only seven RCTs had high RoB.

Clinical Outcome Measures

The effect of drugs that augment mitochondrial function on clinical outcome measures was examined in 22 studies [26] [27,28,29,30,31, 35, 37] [8, 32, 39,40,41,42, 44,45,46,47,48,49, 51, 53,54,55] [43]. Ten studies investigated clinical outcome measures following drugs that increase frataxin [56, 58,59,60,61,62,63, 65,66,67,68,69, 71] while seven studies following symptomatic treatment [72, 74, 76,77,78,79] [75]. The clinical scales examined along with treatment effect in each study are presented in Table 2.

Table 2 Clinical outcome measures examined in each study included in the qualitative synthesis
Full size table

Drugs that Augment Mitochondrial Function

We performed a pooled analysis of studies that reported on International Cooperative Ataxia Rating Scale (ICARS) and Friedreich Ataxia Rating Scale scores [including Friedreich Ataxia Rating Scale (FARS) and modified FARS (mFARS)] changes following an intervention targeting mitochondrial function as presented in Fig. 2.

Fig. 2
figure 2

Clinical outcome measures changes (A ICARS, B Friedreich Ataxia Rating Scale scores (FARS/mFARS)) following drugs that augment mitochondrial function

Full size image

The pooled mean effect size of ten studies examining ICARS showed no statistically significant changes after 12 months of treatment (SMD = 0.03, 95% CI − 0.26 to 0.32, p = 0.8, I2 = 80%, follow-up range 2–132 months), with substantial heterogeneity that was not eliminated even at the pre-planned sensitivity and subgroup analyses of studies grouped by drug administered (Fig. 2A). The results were similar when we pooled six studies reporting on Friedreich Ataxia Rating Scale scores (FARS or mFARS) changes following treatment with drugs that augment mitochondrial function (SMD =  − 0.05, 95% CI − 0.34 to 0.25, p = 0.8, I2 = 72%) during a follow-up period of 6 to 24 months (Fig. 2B). However, sensitivity analysis using studies of low RoB revealed a statistically significant improvement (SMD =  − 0.32, 95% CI − 0.62 to − 0.02, p = 0.04, I2 = 0%) on Friedreich Ataxia Rating Scale scores (FARS/mFARS) following drugs that augment mitochondrial function (Fig. 2B). Of note, no statistically significant improvement was found when we pooled five studies reporting on FARS changes alone following treatment with drugs that augment mitochondrial function (SMD = 0.39, 95% CI − 0.55 to 1.33, p = 0.42, I2 = 95%) suggesting that the positive result was mainly driven by the omaveloxolone study in which mFARS has been used as an outcome measure (online-only Supplementary material 3).

Drugs that Increase Frataxin

The Scale for the Assessment and Rating of Ataxia (SARA) was not affected by drugs that increase frataxin (SMD =  − 0.23, 95% CI − 0.59 to 0.12, p = 0.2, I2 = 56%) according to six studies included in the quantitative synthesis during a median follow-up period of 7.5 months (range 2–18) (Fig. 3A). Moreover, the pooled effect estimates showed no significant difference between pre- and post-drugs that increase frataxin on FARS based on four studies with a median follow-up of 3 months (SMD =  − 0.37, 95% CI − 0.81 to 0.08, p = 0.1, I2 = 67%, Fig. 3B).

Fig. 3
figure 3

Clinical outcome measures changes (A SARA, B FARS) following drugs that increase frataxin

Full size image

Symptomatic Treatment

The heterogeneity of drugs used, follow-up periods along with the different outcome measures applied impeded a pooled synthesis of these seven studies (Table 2) [59, 61,62,63,64,65].

Cardiac Biomarkers

Echocardiogram parameters were used as outcome measures in 27 studies [38] [50, 73]. Descriptive characteristics are presented in Table 3.

Table 3 Cardiac biomarkers examined in each study included in the qualitative synthesis
Full size table

Drugs that Augment Mitochondrial Function

Left ventricular mass index (LVMI) was improved significantly (SMD =  − 0.34, 95% CI − 0.5 to 0.18, p < 0.001, I2 = 33%) following 28.5 months (median; range 6 to 132 months) of treatment with drugs that augment mitochondrial function. This result was based on the pooled analysis of ten studies of which eight had high risk of bias. Interestingly, the statistically significant result remained at the subgroup analysis according to the type of drug administered (SMD =  − 0.34, 95% CI − 0.47 to 0.21, p < 0.001, I2 = 33%, Fig. 4A). The remaining parameters assessed (IVS: SMD =  − 0.002, 95% CI − 0.27 to 0.27, p = 0.99, I2 = 73%; LVEF: SMD =  − 0.16, 95% CI − 0.47 to 0.15, p = 0.3, I2 = 63%; LVPW: SMD =  − 0.15, 95% CI − 0.44 to 0.15, p = 0.3, I2 = 45%; FS: SMD = 0.2, 95% CI − 0.27 to 0.67, p = 0.4, I2 = 77%) did not change after treatment as presented in Fig. 4B, C, D, and E, respectively.

Fig. 4
figure 4

Cardiac biomarker changes following drugs that augment mitochondrial function. A Left ventricular mass index (LVMI), B intraventricular septal wall (IVS), C left ventricular ejection fraction (LVEF), D left ventricular posterior wall (LVPW), E fractional shortening (FS)

Full size image

Drugs that Increase Frataxin

Only four studies assessed the effect of drugs that increase frataxin on cardiac biomarkers (Table 3). Thus, we performed a pooled analysis of three studies reporting on LVEF changes [53, 56, 58].We found that LVEF remained stable during 10.5 months (range 3–18 months) of treatment with epoetin alfa, resveratrol, or IFNγ-1b (SMD =  − 0.13, 95% CI − 0.44 to 0.18, p = 0.4, I2 = 26%).

Symptomatic Treatment

There were no sufficient data presented at the three studies of symptomatic treatment assessing cardiac biomarkers to perform a quantitative synthesis (Table 3) [60, 61, 63].

Biochemical Biomarkers

Descriptive characteristics of 17 studies that investigated biochemical biomarkers are presented in Table 4. We were not able to perform a pooled synthesis of biochemical biomarkers after treatment with drugs that augment mitochondrial function [30, 35, 46, 47, 52, 64] nor after symptomatic treatment due to the diverse biochemical biomarkers assessed.

Table 4 Biochemical biomarkers examined in each study included in the qualitative synthesis
Full size table

Drugs that Increase Frataxin

We found no significant difference between pre- and post-treatment (median follow-up period of 3 months) with drugs that increase frataxin on peripheral blood mononuclear cells’ frataxin levels based on the pooled effect estimates of four studies (SMD =  − 0.01, 95% CI − 0.52 to 0.49, p = 0.96, I2 = 74%, Fig. 5).

Fig. 5
figure 5

Changes in peripheral blood mononuclear cells’ (PBMCs) frataxin (FXN) levels following drugs that increase frataxin

Full size image

Patient-Reported Outcome Measures

PROMs were assessed in 21 studies as presented in Table 5.

Table 5 Patient-reported outcome measures (PROMs) examined in each study included in the qualitative synthesis
Full size table

Drugs that Augment Mitochondrial Function

We found that ADLs remained stable (SMD = 0.11, 95% CI − 0.21 to 0.43, p = 0.5, I2 = 66%) after 12 months (median follow-up) of treatment with drugs that augment mitochondrial function according to five studies (Fig. 6A).

Fig. 6
figure 6

Activities of Daily Living Scale (ADLs) changes following A drugs that augment mitochondrial function, B drugs that increase frataxin

Full size image

Drugs that Increase Frataxin

ADLs were also not affected by drugs that increase frataxin according to the pooled effect of three studies (SMD = 0.11, 95% CI − 0.09 to 0.31, p = 0.3, I2 = 0%) presented in Fig. 6B.

Symptomatic Treatment

The four studies examining symptomatic approaches revealed no treatment effect on PROMs based on the qualitative synthesis (Table 5).

Imaging Biomarkers

Imaging biomarkers were assessed in nine studies (11 records); five studies examined the effect of drugs that augment mitochondrial function on imaging parameters [28, 33, 34, 47, 51, 54], one investigated a symptomatic treatment [75] while the intervention of the remaining three studies was aiming to increase frataxin [57, 58, 65, 70].

Drugs that Augment Mitochondrial Function

Boddaert et al. showed that a 6-month deferiprone treatment led to reduction of iron accumulation specifically in dentate nuclei by performing brain magnetic resonance imaging (MRI) at FRDA patients [28]. The same finding was confirmed by Velasco-Sánchez et al. after 11 months of combined deferiprone and idebenone treatment [54].

Cardiac and skeletal muscle phosphorus P31 magnetic resonance spectroscopy (MRS) was used by the UCL study group [33, 34]. A significant amelioration of cardiac and skeletal muscle bioenergetics was found following combined Q10 and vitamin E treatment. P31 MRS was also examined by Schöls et al. [51]. l-Carnitine phosphocreatine recovery was improved following 4 months of l-carnitine and creatine treatment. However, no difference was found compared to placebo group.

Recently, Qureshi et al. used a variety of imaging parameters such as spinal cord and cerebellar volume measured by 3D volumetric MRI, spinal cord fractional anisotropy by diffusion tension imaging (DTI), cerebellar metabolite ratios by 3D MRS along with cardiac MRI to monitor epicatechin administration [47]. A significant reduction of LV mass index at cardiac MRI was reported. Moreover, Qureshi et al. found a significant reduction at mean cerebellar volume but without subsequent worsening among individual patients after 24 weeks.

Drugs that Increase Frataxin

Recombinant human erythropoietin effect on imaging biomarkers was assessed by two studies presented at three records [57, 58, 65]. Axial diffusivity changes were detected in cerebral hemispheres by DTI, but this finding did not correlate with any clinical outcome [58]. Santner et al. found an increase of pulvinar and the posterior parietal cortex gray matter volume after 6-month treatment with rhuEPO using voxel-based morphometry [57]. Interestingly, this observation correlated with an improvement in clinical scores. P31 MRS examination of the calf muscles did not change following 2 months of rhuEPO administration in a study by Nachbauer et al. [65].

Valva et al. reported interferon gamma treatment induced changes on advanced MRI and retinal imaging [DTI, functional MRI (fMRI), resting-state fMRI (rs-fMRI)]. Significant alterations were detected on fMRI and rs-fMRI; the former correlated with clinical outcomes [70].

Symptomatic Treatment

Two studies assessed the effect of idebenone, a drug which augments mitochondrial function, on neurophysiologic biomarkers. Naeije et al. performed a sham-controlled, crossover RCT using anodal cerebellar transcranial direct current stimulation (ctDCS) in 24 FRDA patients with a follow-up period of 3 months [75]. A reduced cSII cortex functional magnetic resonance imaging (fMRI) response was elicited by a tactile oddball stimulation following ctDCS compared with sham ctDCS probably because of the restoration of the neocortical inhibition normally exerted by the cerebellum.

Neurophysiologic and other Biomarkers

Two studies assessed the effect of idebenone, a drug which augments mitochondrial function, on neurophysiologic biomarkers [27, 53]. Electromyography, somatosensory, and visual evoked potentials parameters did not change following 12 or 24 months of treatment. However, peroneal motor nerve conduction velocity deteriorated significantly as reported by Sival et al. [53].

Other biomarkers such as exercise testing and muscle biopsy were examined by four studies (presented in five records) [8, 36, 39, 64, 66]. Idebenone or epoetin alfa treatment did not affect peak oxygen consumption per unit time or peak work rate according to Drinkard and Saccà et al., respectively [36, 66]. However, a nonsignificant improvement in peak work was observed by Lynch et al. at the MOXIe Study [8, 39]. Muscle tissue changes after administration of recombinant human erythropoietin were investigated by Nachbauer et al. [64]. FRDA patients showed reduced respiratory chain complex and citrate synthase activities in skeletal muscle compared with healthy controls but were not affected by treatment.

Biomarker Change over Time in No Treatment Group

We performed an exploratory analysis using data of biomarker trajectory over time in the untreated patient group to elucidate the ability of a biomarker to detect subtle disease progression. We found that none of the biomarkers examined (ADLs, FARS, ICARS, LVMI, mFARS, T25FW−1, 9HPT−1) changed significantly over a median follow-up period of 6 months. Data of the quantitative analyses are summarized in Table 6.

Table 6 Pooled analyses data summary of biomarkers changes over time in untreated patients
Full size table

Certainty of Evidence

We applied Grading of Recommendations Assessment, Development and Evaluation (GRADE) tool to assess quality of evidence of our estimates which was low to very low due to the observational nature, the high RoB, along with the inconsistency of included studies (online-only Supplementary material 4).

Discussion

Summary of Evidence and Implications for Practice

The present meta-analysis explored the effect of different types of interventions targeting mitochondrial function, frataxin, or patients’ symptoms on clinical, cardiac, biochemical, PROMs, imaging, or neurophysiologic biomarkers in 1409 patients with Friedreich ataxia. In the context of the 43 included studies, a large array of biomarkers was applied as outcome measures. A statistically significant improvement was detected in Friedreich Ataxia Rating Scale scores (combining FARS and mFARS as clinical outcome measures) in 205 patients after 15 months of treatment with drugs that augment mitochondrial function. Nevertheless, this result should be interpreted with caution because it was mainly driven by omaveloxolone’s positive trial and was characterized by very low quality of evidence. Low quality of evidence from ten studies (seven of which were observational) examining 261 patients supported a beneficial effect of drugs that augment mitochondrial function on cardiac structure measured by LVMI after 28.5 months. This result was driven by idebenone studies in combination with one trial of the iron chelator deferiprone. In contrast, all the remaining biomarkers examined did not change following any treatment or during the natural course of the disease. Of note, the median follow-up period for these outcome measures did not exceed 12-months.

An ideal valuable biomarker should be able to detect both subtle changes in the natural course of a slowly progressive disorder such as Friedreich ataxia and concurrently be responsive to any treatment effect. Nevertheless, it is imperative that the observed alterations are also clinically meaningful. The pooled analysis in untreated patients showed that LVMI did not change over 12 months. The observed LVMI reduction following drugs that augment mitochondrial function without any parallel change in other cardiac or clinical biomarkers should be interpreted with caution. In view of these results, we suggest the use of a biomarker toolbox (for example, a combination of Friedreich Ataxia Rating Scale scores and LVMI assessment) evaluating different aspects of this disease as a primary outcome measure in future RCTs. Notably, study duration should be least 24 months considering that trials with shorter duration are unlikely to demonstrate any clinical benefit. Finally, the clinical meaningful change of the employed biomarkers should be predefined based on natural history studies.

Evidence from other Studies

A Cochrane review was conducted in 2016 to assess the therapeutic efficacy of pharmacological treatments for Friedreich ataxia [80]. Hence, only RCTs with a minimum follow-up of 12 months were included in this Cochrane review. Furthermore, Jain et al. carried out a systematic literature review (published at 2022 before the FDA approval of omaveloxolone) to summarize the efficacy and safety of therapeutic interventions that have been investigated in Friedreich’s ataxia [81]. On the contrary, we aimed to explore the effect of the interventions on biomarkers in order to assess their response to change. Thus, we did not limit our search to RCTs and set a minimum 2-month interval between biomarker assessments that resulted in the exclusion of some records from the present meta-analysis [82,83,84]. We found no change at any biomarker examined either in treated or untreated patients except for a statistically significantly reduction in Friedreich Ataxia Rating Scale scores (combining FARS and mFARS) and LVMI following drugs that augment mitochondrial function. In line with our findings, Jain et al. conclude that the limited sample size and follow-up duration led to inconclusive evidence. Similarly, LVMI was only investigated by one RCT included in the Cochrane review in which a significant decrease was detected [80]. Nevertheless, the clinical relevance of this result was interpreted with caution in the Cochrane review due to the low quality of evidence in line with our conclusions.

We limited this review to studies examining a therapeutic intervention but also investigated the change from baseline score of any biomarker in untreated patients separately. The EFACTS study group evaluated 552 treatment-naive patients exclusively with at least 4 years of follow-up in a prospective cohort study and found an annual progression rate of 0.82 points (SE 0.05) for SARA and 0.93 points (SE 0.05) for ADL [85]. They calculated that 190 patients would be required to detect a 50% ADL reduction in a 2-year parallel-group trial. We found a non-significant ADL SMD change of 0.46 (− 0.31 to 1.23) in 147 patients followed for a median period of 6 months. The observed difference could be attributed to the limited follow-up time of the studies included in this review.

Strengths and Limitations

The observational nature of the majority of the included records in conjunction with the lack of a control group, the short median follow-up duration in most outcomes, and the heterogeneity of the examined populations represent the major limitations of this meta-analysis. Thus, we downgraded the quality of evidence of our estimates. Management of FRDA requires a multidisciplinary approach consisting of numerous interventions (such as occupational and physical therapy, speech and swallowing therapy, psychological counseling) which could not be pooled in one systematic review. Accordingly, we conducted a comprehensive literature search following PRISMA guidelines focusing on pharmacological therapies along with non-invasive neurostimulation approach, in line with our prespecified protocol. This led to the inclusion of 1409 patients in the current meta-analysis, a large population cohort considering that Friedreich ataxia is a rare disease.

Conclusions and Future Perspectives

A statistically significant improvement was detected in Friedreich Ataxia Rating Scale scores [combining Friedreich Ataxia Rating Scale (FARS) and modified FARS (mFARS), SMD =  − 0.32 (− 0.62 to − 0.02)] in 205 patients following 15 months of treatment drugs that augment mitochondrial function in a sensitivity analysis of six studies with very low quality of evidence. Moreover, low-quality evidence from ten, mainly observational, studies evaluating 261 patients suggested a significant reduction of LVMI after a median 28.5 months of treatment with drugs that augment mitochondrial function. The clinical importance of these changes remains to be clarified in well-designed RCTs in which the minimal clinically important change should be predefined. We found no significant change on any other clinical, cardiac, biochemical, PROMs, imaging, or neurophysiologic biomarker investigated in 1409 patients with Friedreich ataxia. Nevertheless, the median follow-up period for these outcomes was extremely limited considering the slow natural disease progression. Hence, a biomarker toolbox evaluating different aspects of this complex disease (for example, combining both Friedreich Ataxia Rating Scale scores and LVMI assessment) could be applied as a primary outcome measure in future RCTs of at least 24 months’ duration.