Idebenone Treatment In Leber's Hereditary Optic Neuropathy

This is the authors' abstract. We don't add key points for this paper.

Sir, We have read with great interest the results presented by Klopstock et al. (2011) concerning the RHODOS study on a clinical trial with idebenone in Leber's hereditary optic neuropathy (LHON) and we would like to share our own experience of idebenone therapy in LHON. Idebenone has been an approved drug (Mnesis®, Takeda Italia Farmaceutici) in Italy since the early 1990s and, after the initial report by Mashima et al. (1992) on its possible efficacy in LHON, we offered this therapeutic option to all of our new consecutive patients with LHON, almost all of whom accepted treatment. Idebenone was given after informed consent following the regulation for ‘off-label’ drug administration and was provided for free by the National Health Service, under the legislation for certified rare disorders. Patients were initially treated with 270 mg/day (Cortelli et al., 1997; Carelli et al., 1998a, b), but following the reports on idebenone treatment in Friedreich ataxia, the dosages were increased to 540–675 mg/day (Rustin et al., 1999; Kearney et al., 2009). To evaluate retrospectively the efficacy of idebenone therapy, we reviewed all of our patients with LHON, idebenone treated and untreated, after approval of the institutional Internal Review Board. Inclusion criteria for treated patients were the initiation of therapy within 1 year after visual loss in the second eye, and for all patients (treated and untreated) age at onset of at least 10 years and a follow-up of at least 5 years. We included only patients treated within 1 year after onset because this is the time frame to reach the nadir of the visual loss and the probability of spontaneous recovery of vision is highest in the following 5 years (Nikoskelainen et al., 1983; Barboni et al., 2005, 2010; Spruijt et al., 2006). We also excluded subjects with disease onset before 10 years of age (childhood cases), because of their well-established benign prognosis (Carelli et al., 2004; Barboni et al., 2006; Yu-Wai-Man et al., 2011). The untreated patients were considered as a control cohort. These patients lost their vision and were followed before idebenone treatment became an option, thus none of them was offered and declined the use of idebenone. For the sake of uniformity in assessment of visual function in all patients, we chose visual acuity of the best and the worst eye at the last available examination as our primary outcome measure. Recovery of visual acuity was considered for patients or eyes (some patients recovered vision in only one eye) with a gain of at least two lines on Snellen acuity or a change from ‘off chart’ to ‘on chart’, as previously established (Nikoskelainen et al., 1996; Barboni et al., 2005). Given the retrospective nature of this study, all visual acuity evaluations were not blinded to treatment status. For all patients, we also determined (i) age of disease onset; (ii) time lapse between loss of visual acuity in the eyes; (iii) time between disease onset and start of therapy; (iv) time between disease onset and recovery of visual acuity; (v) average therapy dosage; and (vi) therapy duration. The interval between loss of vision in the eyes was evaluated only for the patients with asynchronous onset of symptoms, which was 56% for 11778/ND4, 50% for 3460/ND1 and 67% for 14484/ND6 mutation, having excluded three outliers (time lapse of 23, 180 and 504 months, respectively), which exceeded 2 SD from the mean. The difference in sex distribution between the group of idebenone-treated patients and the group of untreated patients was evaluated using χ2 test. Student t-test was performed for comparison of the age at onset and visual acuity of the best and worst eye evaluated at the nadir of visual loss (between 6 months and 1 year after disease onset) between the two groups. To evaluate the impact of all covariates on visual recovery, we used a Cox proportional hazards model. A first model included gender, age at onset, type of mutation and therapy/no therapy as independent variables and visual recovery as outcome. In a second model, we added the time between disease onset and start of therapy, the average therapy dosage and the therapy duration as variables. In both models, a backward stepwise method was applied to obtain the best model in which all included variables had a significant P-value. Furthermore, we separately compared the interval between asynchronous disease onset in the first and the second eye in patients treated before the involvement of the second eye (‘in-between-eyes’), patients treated after the involvement of the second eye and patients not treated by using the Kruskall–Wallis test followed by a post hoc Mann–Whitney U test. For all analyses, two-sided P < 0.05 was considered statistically significant. Overall, 103 patients with LHON, carrying a primary mitochondrial DNA mutation were included in this study. Forty-four patients (37 males) were either treated within 1 year after the onset of visual acuity loss in the second eye (n = 38) or treated before the involvement of the second eye (‘in-between-eyes’; n = 6), defining the entire group as ‘idebenone-treated’. Thirty patients had the 11778/ND4, eight the 3460/ND1 and six the 14484/ND6 mutation. We also included 59 untreated patients (40 males) and this group was defined as ‘untreated’. Of these, 43 patients had the 11778/ND4, 10 the 3460/ND1 and six the 14484/ND6 mutation. The idebenone-treated and untreated groups of patients did not differ significantly for gender, age at onset, proportion of LHON mutation type and visual acuity of the best and worst eye evaluated at the nadir of visual loss (between 6 months and 1 year after disease onset) (data not shown). The clinical data, considered separately for patients who recovered visual acuity and those that did not, are summarized for all patients in Table 1, and for the 11778/ND4 patients in Table 2. Considering all patients, the proportion of patients or eyes with visual recovery was higher for idebenone-treated compared with the untreated group (Table 1). Among the idebenone-treated patients, the duration of therapy was longer for patients recovering vision and there was an earlier onset of recovery (Table 1). Noticeably, the untreated patients recovering vision were younger at disease onset than the others. Eighty per cent of patients complied with idebenone treatment for at least 18 months, and there were no serious adverse events. One patient discontinued idebenone due to insomnia attributed to treatment. Clinical data of all patients Data are expressed as mean ± SD; disease onset always refers to the first eye involved; visual acuity is expressed in decimal scale derived from Snellen charts (e.g. 20/20 = 1.0; 20/200 = 0.1) and is referred to the last examination. Clinical data of all patients Data are expressed as mean ± SD; disease onset always refers to the first eye involved; visual acuity is expressed in decimal scale derived from Snellen charts (e.g. 20/20 = 1.0; 20/200 = 0.1) and is referred to the last examination. Clinical data of 11778/ND4 patients Data are expressed as mean ± SD; disease onset always refers to the first eye involved; visual acuity is expressed in decimal scale derived from Snellen charts (e.g. 20/20 = 1.0; 20/200 = 0.1) and is referred to the last examination. Clinical data of 11778/ND4 patients Data are expressed as mean ± SD; disease onset always refers to the first eye involved; visual acuity is expressed in decimal scale derived from Snellen charts (e.g. 20/20 = 1.0; 20/200 = 0.1) and is referred to the last examination. Upon stratification by mutation, we found that in the 11778/ND4 patients (n = 73) the higher frequency of visual recovery in idebenone-treated patients or idebenone-treated eyes was even more evident (Table 2). The 11778/ND4 idebenone-treated patients also had an earlier onset of recovery and a longer duration of therapy. Cox proportional hazards model performed for the whole group of patients (mutations combined) showed that visual recovery was significantly associated only with the 14484/ND6 mutation (P = 0.004), when the therapy was included in the independent variables simply as ‘therapy/not therapy’ (Table 3; Fig. 1). In the second model, including the variables ‘time between disease onset and start of therapy’, ‘average therapy dosage’ and ‘therapy duration’, visual recovery was again significantly associated with the 14484/ND6 mutation (P = 0.009), and also with an earlier start of therapy (P = 0.049). Considering the 11778/ND4 mutation group, the first model showed a significant association between therapy and visual recovery (P = 0.031) (Table 3; Fig. 1), whereas the second model showed that the parameter with the greater effect on visual recovery was an earlier start of therapy (P = 0.005) (Table 3). (Top) Estimates of cumulative (cum) hazard for visual recovery in the whole group, stratified by the presence of 11778/ND4, 3460/ND1 or 14484/ND6 mutation. (Bottom) Estimates of cumulative hazard for visual recovery in 11778/ND4 group stratified by the presence or absence of idebenone therapy. Multivariate prediction analysis (Cox proportional-hazards model) for visual recovery in patients with LHON In both analyses, a backward stepwise method was applied to obtain the more significant model in which all included variables had a P < 0.05. a Included variables: gender, age at onset, type of mutation and therapy/not therapy; outcome: visual recovery. b Included variables: gender, age at onset, type of mutation, time between disease onset and start of therapy, average therapy dosage and therapy duration; outcome: visual recovery. c Earlier start of therapy was associated with visual recovery. Multivariate prediction analysis (Cox proportional-hazards model) for visual recovery in patients with LHON In both analyses, a backward stepwise method was applied to obtain the more significant model in which all included variables had a P < 0.05. a Included variables: gender, age at onset, type of mutation and therapy/not therapy; outcome: visual recovery. b Included variables: gender, age at onset, type of mutation, time between disease onset and start of therapy, average therapy dosage and therapy duration; outcome: visual recovery. c Earlier start of therapy was associated with visual recovery. Concerning the 3460/ND1 and 14484/ND6 mutations, the number of patients available was too low for meaningful statistical analysis (18 and 12, respectively). As expected and confirmed by the Cox analysis, the 14484/ND6 patients had a higher propensity for visual recovery (Carelli et al., 2004; Yu-Wai-Man et al., 2011). Conversely, the 3460/ND1 patients did not, whether treated or untreated (data not shown). In the six patients treated ‘in-between-eyes’ (five carrying the 11778/ND4 and one the 3460/ND1 mutation), the time separating the involvement of the two eyes was significantly longer compared with patients treated after the second eye or the untreated group (Fig. 2). However, all the second eyes eventually became affected and three out of six had 0.05 final visual acuity. In these latter eyes, two had this final visual outcome after visual loss progression and one after recovery from the nadir. Comparison of the interval between asynchronous disease onset in the first and the second eye in idebenone-treated (in-between-eyes), idebenone-treated (IT) (after the second eye) and untreated (NT) patients. The three groups are indicated on the x-axis and on the y-axis is the time lapse in months between the involvement of the first and the second eye. Asterisk indicates P < 0.05 (Kruskall–Wallis: P = 0.027; post hoc Mann–Whitney U tests: idebenone-treated in-between-eyes versus idebenone-treated after the second eye: P = 0.02; idebenone-treated in-between-eyes versus untreated: P = 0.004; idebenone-treated after the second eye versus untreated: P = 0.6). This retrospective evaluation of patients with LHON treated with idebenone shows an increased frequency of recovery of visual acuity, compared with the cohort of untreated patients, which is significant for the 11778/ND4 patients. Remarkably, an early start of therapy was the most predictive factor for visual recovery, mainly driven by the 11778/ND4 patients. The patients who started therapy at the early stage of monocular disease had a significant delay in the involvement of the second eye, which was, however, not spared by the disease. The final visual outcome of these early-treated eyes was not markedly better, though the available data preclude proper extrapolation as to whether they may have diverged from the natural history of the disease. The untreated patients that spontaneously recovered visual acuity tended to have younger age at onset, as previously reported (Carelli et al., 2004; Yu-Wai-Man et al., 2011). Considering all mutations together, there was a trend for earlier onset of visual recovery in treated patients compared with untreated, as also reported by Mashima and colleagues (2000). Furthermore, treated patients recovering vision tended to have longer therapy duration than those that failed to recover. It is also evident that within the treated group, only a proportion of patients responded to treatment. Mutation stratification indicates that the 11778/ND4 patients were the best responders. The present data are not sufficient to verify the treatment response for the 3460/ND1 and 14484/ND6 mutations. The 3460/ND1 patients seem to be poor responders, whereas the 14484/ND6 patients had a high propensity for visual recovery even without treatment. Treatment allocation was neither prospective nor randomized and the limits of this study are those intrinsic to the retrospective evaluation. However, selection bias for treatment was unlikely. The investigators began offering idebenone to all patients and no patient declined treatment once the availability of a treatment option existed. Our study took into consideration a large cohort of untreated patients with different mutations, thus allowing an estimate of the frequency of spontaneous visual recovery, which was similar to the previously published rates (Nikoskelainen et al., 1996; Carelli et al., 2004; Spruijt et al., 2006; Yu-Wai-Man et al., 2011). In the 6-month prospective, randomized, placebo-controlled idebenone trial (RHODOS study), including 85 patients with LHON (55 treated), the primary endpoint (best recovery of visual acuity) failed to show significant differences, whereas significance was found in secondary endpoints and by subgroup analysis (Klopstock et al., 2011). The design of our retrospective study included patients treated within 1 year after disease onset, thus avoiding the period (2–5 years after onset) in which the probability of spontaneous recovery of visual acuity is higher. The RHODOS study, on the contrary, recruited patients up to 5 years after disease onset. The RHODOS study uniformly applied the 900 mg/day idebenone dosage (Klopstock et al., 2011), whereas the patients of our cohort were treated with variable and lower dosage of idebenone (270–675 mg/day), but on average for a longer time. In conclusion, both studies strongly suggest that early and prolonged idebenone treatment in patients with acute LHON may significantly improve the frequency of visual recovery and possibly change the natural history of the disease. Despite the retrospective nature of our study, the results indicate that idebenone administration may be the first effective treatment in LHON. Telethon-Italy (grant #GPP10005). We would like to dedicate, for his invaluable teaching and inspiration, this study to the memory of Prof. Pasquale Montagna who prematurely passed away.