The challenge of an adequate outcome in trials for genetic eye disease such as Leber hereditary optic neuropathy

Abstract

With improvements in screening and management for eye diseases such as diabetic retinopathy, inherited eye diseases (IEDs), in particular inherited retinal diseases (IRDs), are becoming the leading cause of blindness for working age adults in developed countries. Hundreds of genes causing IEDs have been discovered. Once confirmed, a genetic diagnosis can lead to a better understanding of the likely natural history, phenotype-genotype correlations and possible eligibility for enrolment in research and clinical trials. Gene-based therapies, notably adeno-associated virus (AAV) gene therapy trials, are underway for many genetic subtypes of IRD. To date, one gene therapy—voretigene neparvovec-rzyl (Luxturna)—has been licenced in the United States, but not yet in Australia. A major challenge for all the trials is the slow (but relentless) progression of most IRDs. In many cases, the test-retest variability of most measures of vision loss is greater than the annual progression of disease. New tests, such as microperimetry, retinal imaging and electrophysiology, have been developed. Even with these new technologies it will still take years to show any meaningful reduction in progression of disease and natural history trials with these technologies are pending. In some cases, restoration of function, such as improved night vision (and navigation in a dim lit environment) in RPE65 patients treated with Luxturna, is a useful indicator of efficacy. But how well Luxturna prevents vision loss over decades remains to be determined. Leber hereditary optic neuropathy (LHON) is an exception to most IEDs, with an acute, dramatic loss of vision in both eyes (from 6/4.8 to 6/120 or loss of 14 lines) in a matter of days or weeks. Despite visual acuity being the most measured parameter in ophthalmic practice, it is analysed infrequently in epidemiology studies. We have minimal data on the population distribution of visual acuity over different ages and very little on the usual changes with aging from childhood to old age. Moreover, the test-retest changes of visual acuity are not wellestablished. The Early Treatment of Diabetic Retinopathy Study helped institute a more standard testing protocol and the “cut-off” for improvement was given as three lines of vision on the ETDRS chart for diabetic maculopathy. This was later used in age-related macular degeneration studies. Visual acuity is not an objective test. The education and experience of the patient and the level of encouragement from the person testing, as well as other subtle changes in factors such as illumination and general well-being of the patient, can influence a result. In addition, patients with central vision loss learn to use their remaining field, identifying eccentric fixation to allow them to see better. Such a dramatic loss of vision and the very high likelihood of vision loss in the second eye meant that early trials of LHON treatment were optimistic in showing a successful result with small numbers of patients. Idebenone showed initial promise with a single case of the 11 778 LHON mutation, but later reports were mixed, leading to a randomized clinical trial that was not really conclusive as to the efficacy of idebenone. A major problem with LHON is the well documented natural history of partial visual recovery without any treatment (particularly with some genetic subtypes [14484mtDNA and 3460mtDNA] and in younger individuals). Even with poorer prognosis and the more common 11 778 mutation, spontaneous recovery is seen. Visual prognosis is better in eyes with less severe reduction of visual acuity 1 year after onset. Visual acuity corresponds with visual field indices in the decline of vision in LHON. Thus an ideal study needs to be a masked clinical trial with treated and untreated patients. The unmasked LHON AAV gene therapy trials have been run in China, Europe and the United States, and unfortunately only mild evidence of slight improvement was found. The studies have compared the treated and untreated eyes, presuming the untreated eye would be an adequate control. However, the untreated eyes have also shown improvement. This leaves us wondering if the improvement is just a placebo effect or bilateral natural history of improvement or alternatively is there a retrochiasmal cross-over of the treatment effect? Does treating an adjacent axon from the contralateral eye benefit axons in the untreated eye? The study by Zhang et al in this journal is a continuation of their earlier trial; they have now treated 53 patients with the 11 778 mutation. The major findings of this study, in which half the patients had lost vision more than 2 years earlier, was that more than half the participants had vision improve more or equal to three lines at 3 months after treatment. However, those who recovered were more likely to have had better vision and visual field parameters (VFI and MD) at the time of treatment—the same indicators for increased likelihood of spontaneous improvement. DOI: 10.1111/ceo.13586