Professional Documents
Culture Documents
• PURPOSE: To assess the relationship of demograph- IPL thickness (P = .045), superotemporal macular GC-
ics, clinical characteristics and structural optical coher- IPL thickness (P = .047), and inferotemporal macular
ence tomography (OCT) findings to long-term visual out- GC-IPL thickness (P = .030).
comes in patients with Leber hereditary optic neuropathy • CONCLUSION: We identified OCT biomarkers associ-
(LHON) treated with idebenone. ated with long-term (ie, 2-year) visual outcomes in pa-
• DESIGN: Retrospective, interventional, noncompara- tients with LHON treated with idebenone therapy in the
tive clinical cohort study. first year after disease onset. Thinning of the GC-IPL in
• METHODS: In this study, a total of 17 participants (34 the superior and temporal parafoveal regions was associ-
eyes) with LHON treated with idebenone therapy within ated with worse long-term visual outcomes in these pa-
1 year after disease onset and 2 years (24 months) of reg- tients. (Am J Ophthalmol 2023;247: 35–41. © 2022
ular follow-ups were retrospectively enrolled. At base- Elsevier Inc. All rights reserved.)
line, structural OCT volume scans of the macula and
optic nerve were reviewed to measure metrics reflecting
L
neuronal loss (ie, macular ganglion cell and inner plexi- eber hereditary optic neuropathy (LHON) is
form layer [GC-IPL] and peripapillary retinal nerve fiber one of the most common mitochondrial diseases
layer [RNFL] thicknesses). Stepwise multiple regression and it is caused by mitochondrial DNA (mtDNA)
analyses were computed to assess associations between point mutations involving complex I of the oxidative
final best-corrected visual acuity (BCVA) at 2 years and phosphorylation chain.1–4 The 3 most common mutations
change in BCVA from baseline at 2 years as dependent eventually resulting in LHON are at nucleotide positions
variables with demographics, clinical characteristics, and 11778/ND4, 3460/ND1, and 14484/ND6. Typically, pa-
OCT metrics at baseline (visit before the initiation of tients with LHON may experience a painless, subacute, and
treatment). simultaneous or sequential loss of central vision, this often
• RESULTS: The BCVA was 1.6±0.8 logMAR (Snellen occurring early in life.2 , 4–6
VA of ∼20/800) at baseline (visit before the initia- LHON is noteworthy in its tissue selectivity, as this dis-
tion of treatment) and 1.0±0.7 logMAR (Snellen VA ease is mainly limited to injuring retinal ganglion cells and
of 20/200) at the 2-year follow-up visit (P < .0001). their axons converging in the optic nerve.1–4 In the early
Mean±SD change in BCVA from baseline at 2 years phases of LHON, the smaller-caliber fibers of the papil-
was –51.9%±35.9%. In multivariable analysis, the lomacular bundle are predominantly damaged.7–9 Succes-
strongest associations with final BCVA were with base- sively, the fiber decline may advance and eventually involve
line BCVA (P = .012), superior macular GC-IPL thick- the whole optic nerve.10
ness (P = .044), superotemporal macular GC-IPL thick- A lot of evidence has suggested that idebenone may be
ness (P = .010), and inferotemporal macular GC-IPL an efficacious and safe therapy in patients with LHON.11
thickness (P = .015). Similarly, the strongest associa- Idebenone is a synthetic analogue of coenzyme Q10 with a
tions with delta BCVA were with superior macular GC- shorter and less lipophilic tail, the latter characteristic al-
lowing an easier penetration of mitochondrial membranes.
Accepted for publication November 4, 2022. Furthermore, the activation of idebenone takes place in the
From the Vita-Salute San Raffaele University, Milan (E.B., A.B., cytoplasm and is thus dependent on mitochondrial func-
M.L.C., C.B., M.B., G.L., F.B., P.B.); IRCCS San Raffaele Scientific In- tion. Therefore, the transformation into the reduced form
stitute, Milan (E.B., A.B., M.L.C., C.B., M.B., G.L., F.B., P.B.); and De-
partment of Ophthalmology, University of Turin (M.R.), Italy may allow idebenone to shuttle the electrons directly to
Inquiries to Enrico Borrelli, Department of Ophthalmology, University complex III, bypassing the complex I dysfunction that char-
Vita-Salute San Raffaele, Milan, Italy.; e-mail: borrelli.enrico@yahoo.com acterizes LHON.12
4 Enrico Borrelli and Alessandro Berni contributed equally to the work
presented here and should therefore be regarded as equivalent authors.
• STATISTICAL ANALYSIS: Statistical calculations were At baseline, the average peripapillary RNFL thickness
performed using Statistical Package for Social Sciences was 94.6±32.3 μm, and the average macular GC-IPL thick-
(version 20.0; SPSS Inc). ness was 57.3±11.5 μm. Table 3 summarizes the other struc-
To detect departures from normal distribution, a Shapiro- tural parameters in our study cohort of LHON subjects.
Wilk test was performed for all variables. Means and SDs
were computed for all quantitative variables. Visual acuities • REGRESSION ANALYSIS: Results of linear regressions are
and CV metrics (ie, mean defect) at baseline and 2-year summarized in Table 4.
follow-up visits were compared by conducting paired sam- In multivariable analysis, the strongest associations with
ples t test or related-samples Wilcoxon signed rank test. final BCVA were with baseline BCVA (P = .012), supe-
Baseline clinical and imaging variables were explored rior macular GC-IPL thickness (P = .044), superotemporal
for associations with final BCVA at 2 years and change macular GC-IPL thickness (P = .010), and inferotemporal
in BCVA from baseline at 2 years. Stepwise multiple re- macular GC-IPL thickness (P = .015).
gression analyses with the above-mentioned outcomes were Similarly, the strongest associations with delta BCVA
computed, with entry selection criterion set at P <.10 and were with superior macular GC-IPL thickness (P = .045),
stay criterion set at P <.05 for variable selection. Measure- superotemporal macular GC-IPL thickness (P = .047), and
ments of bilateral eyes were nested within subject. inferotemporal macular GC-IPL thickness (P = .030).
RESULTS DISCUSSION
• CHARACTERISTICS OF PATIENTS INCLUDED IN THE In this longitudinal investigation, we explored baseline pre-
ANALYSIS: Clinical characteristics of enrolled patients are dictors of long-term visual outcomes in LHON patients
presented in Table 1. Of the 17 patients with LHON who treated with idebenone within the first year after disease on-
were enrolled (15 males), 34 eyes were included in the anal- set. Overall, we reported novel OCT biomarkers (ie, GC-
ysis. Mean age was 32.3±17.8 years (range 13-59 years). IPL thickness) for visual improvement at 24 months af-
Three of 17 patients were younger than 15 years (ie, child- ter the initiation of therapy. Notably, results of the present
hood LHON).24 , 25 study demonstrated that this association is significant in
The BCVA was 1.6±0.8 logMAR (Snellen VA of the superior and temporal parafoveal sectors of the macula,
∼20/800) at baseline (visit before the initiation of treat- whereas the nasal and inferior GC-IPL thicknesses seem to
ment) and 1.0±0.7 logMAR (Snellen VA of 20/200) be poorly associated with visual improvement. Among the
at the 2-year follow-up visit (P < .0001) (Table 2). demographic and clinical factors, age, sex, and time elapsed
Mean ± SD change in BCVA from baseline at 2 years between symptoms’ onset and initiation of therapy did not
was –51.9%±35.9%. The visual field mean defect was – significantly impact the 2-year visual outcome.
18.2±11.6 at baseline and –15.2±9.9 at the 2-year follow- As asserted above, a lot of evidence has proposed that
up visit (P < .0001) (Table 2). Symptomatic disease dura- idebenone represents an effective and safe treatment in pa-
tion was 5.3±4.9 months before the baseline visit. tients with LHON.11 , 14 Accordingly, our study cohort of
BCVA = best corrected visual acuity (logMAR), LHON = Leber hereditary optic neu-
ropathy.
Quantitative values are expressed in mean ± SD.
LHON patients treated with idebenone within the first year imaging surrogate of optic nerve fibers’ atrophy, a preserved
after disease onset experienced a significant 2-year improve- RNFL thickness indeed may be the result of a combination
ment in visual acuity and visual field. of fibers’ swelling and atrophy, rather than an indicator of
The introduction of OCT has expanded the characteri- true preservation of optic nerve fibers. Therefore, the peri-
zation of macular and optic nerve structural modifications papillary RNFL thickness seems to not appropriately reflect
occurring in LHON subjects. In detail, unaffected patients optic nerve atrophy in the earliest stages of LHON disease.
in the asymptomatic or presymptomatic stages appear to be Accordingly, our analysis revealed that peripapillary RNFL
characterized by a thickening of the peripapillary RNFL in metrics were not associated with better visual outcomes in
the inferior and temporal regions.10 Conversely, structural patients with LHON treated with idebenone therapy.
OCT during the acute phase demonstrated the presence of Conversely, a previous report has suggested that the sta-
inferior and temporal peripapillary RNFL and nasal macular tus of the GC-IPL in the macular region may better re-
GC-IPL thinning—these regions reflecting the topograph- flect the amount of preserved ganglion cells and optic nerve
ical distribution of the papillomacular bundle.18 , 26 fibers in patients with LHON.28 Accordingly, the present
Also, in the latter phase, structural OCT usually de- study highlights the distinctive relationship between inner
tects a swelling (ie, thickening) of the superior and nasal retina damage and final visual acuity in LHON patients
peripapillary RNFL.18 , 26 , 27 The RNFL swelling is likely treated with idebenone: considering final visual acuity as
to depend on the compensatory increase of mitochondrial the dependent variable, we observed a direct relationship
biogenesis and/or axonal stasis along the fibers. Following with variables defining a loss of ganglion cells (ie, macu-
the acute stage, a diffuse thinning of the macular GC-IPL lar GC-IPL thickness), even after accounting for confound-
and peripapillary RNFL may be detected in patients with ing factors, such as age, gene mutation, and symptomatic
LHON.10 , 27 , 28 disease duration before initiation of treatment, which are
Previous studies have suggested that a cross-sectional as- known to modify the visual outcome in these patients.27
sessment of peripapillary RNFL thickness may fail to pro- Similarly, the change in visual acuity between baseline and
vide OCT biomarkers associated with visual outcomes in 2-year follow-up visits was significantly associated with vari-
patients with LHON.28 Although the RNFL thinning is an ables defining a damage of the inner retina.
β (95% CI) P Value β (95% CI) P Value β (95% CI) P Value β (95% CI) P Value
BCVA = best-corrected visual acuity, GC-IPL = ganglion cell and inner plexiform layer, LHON = Leber hereditary optic neuropathy,
RNFL = retinal nerve fiber layer.
Previous preclinical evidence has suggested that sessed the protective effect of idebenone in a mouse model
idebenone may have a protective role against retinal of LHON, whereby mitochondrial complex I dysfunction
ganglion cell loss occurring in LHON, and more im- was caused by exposure to rotenone.
portantly, this treatment may restore function of spared In this model, idebenone proved to protect against the
ganglion cells. In particular, Heitz and associates27 have as- loss of retinal ganglion cells. Also, consistent with this pro-
Funding/Support: This study received no funding. Financial Disclosures: The authors indicate no financial support or conflicts of interest. All authors
attest that they meet the current ICMJE criteria for authorship.
Acknowledgment: We thank Leonardo Caporali, IRCCS Istituto delle Scienze Neurologiche di Bologna, Italy, for providing us the LHON gene mutation
analysis.