A Systematic Review and Meta-Analyses of Interventional Clinical Trial Studies for Gene Therapies for the Inherited Retinal Degenerations (IRDs)

Abstract

IRDs are one of the leading causes of visual loss in children and young adults. Mutations in over 271 genes lead to retinal dysfunction, degeneration and sight loss. Though no cure exists, gene augmentation therapy has brought hope to the field. This systematic review sought to assess the efficacy of available gene therapy treatments for IRDs. Databases and public resources were searched for randomised controlled trials (RCTs) and non-randomised studies of interventions (NRSIs). Standard methodological procedures were used, including a risk-of-bias assessment. One RCT and five NRSIs were assessed, all for adeno-associated virus two (AAV2)-mediated treatment of RPE-specific 65 kDa (RPE65)-associated LCA (Leber congenital amaurosis). Five outcomes were reported for meta-analyses. Modest improvements in visual acuity, ambulatory navigation/mobility testing or central retinal thickness was observed. There was significant improvement in red and blue light full-field stimulus testing (FST) (red light risk ratio of 1.89, treated v control, p = 0.04; and blue light risk ratio of 2.01, treated v control, p = 0.001). Study design assessment using a ROBIN-I tool (Cochrane Library) showed risk-of-bias judgement to be “low/moderate”, whilst there were “some concerns” for the RCT using a RoB-2 tool (Cochrane Library). Although comparison by meta-analysis is compromised by, amongst other issues, a variable amount of vector delivered in each trial, FST improvements demonstrate a proof-of-principle for treating IRDs with gene therapy.

1. Introduction

The IRDs are a heterogenous group of overwhelmingly monogenic eye conditions that cause premature sight loss [1]. Currently, 271 causal genes have been identified (last updated 8 April 2021 [2]) which have roles in various aspects of photoreceptor and/or retinal pigment epithelium (RPE) function. Dominant, recessive, X-linked, digenic and mitochondrial modes of inheritance have all been described [3,4,5,6,7,8,9,10,11,12,13,14]. IRDs cause progressive retinal degeneration, which results in a variety of progressive symptoms including night blindness (nyctalopia), visual field constriction, central visual loss, dark adaptation problems, photophobia, nystagmus and pupillary abnormalities [14].

There is considerable phenotypic variability between IRDs and historically they have been grouped into several different disease patterns, including retinitis pigmentosa (RP), cone dystrophies, cone-rod dystrophies and Leber congenital amaurosis (LCA) [15]. Further, many attempts have been made at genotype-phenotype correlation. However, in reality, there is both considerable inter-allelic disease overlap and marked intra-allelic disease variability. Thus, the progress made in genetic diagnosis of IRDs has been invaluable.

LCA is a severe congenital or early infant-onset IRD characterised by vision loss, nystagmus, an absence of a normal pupil response and an almost non-recordable ERG [16,17,18,19]. Known genes associated with LCA includes GUCY2D (estimated 10–20% patients), CEP290 (15–20%), CRB (10%), AIPL1 (4–8%) and NMNAT1 (uncertain). Following the original description of the infantile disorder, a subsequent milder form of disease, considered to be on the LCA spectrum, was described that presents in the 6th or 7th year of life and leads to blindness by the age of 30 [20]. Whilst this later-onset disease has been referred to by several different names [21,22,23,24,25], there is considerable overlap with LCA in both genotype and phenotype, with causal genes including RPE65 (5–10%), LRAT (<1%) and RDH12 (4–5%) [18].

Mutations in RPE65 are estimated to account for approximately 5–10% of LCA and approximately 1–2% of retinitis pigmentosa [26]. RPE65 is localized to chromosome 1p31, comprising 14 exons and encoding a 65-Kd protein [23]. RPE65 is a key component of the retinoid visual cycle. It is expressed in the retinal pigment epithelium (RPE) which, together with LRAT, is involved in continuous regeneration of the visual chromophore [24]. LCA-mediated IRDs have a prevalence of between 1 in 33,000 [25] and 1 in 81,000 [27]. In Ireland, there are an estimated ~130 LCA patients [28], while NICE has reported there may be 86 LCA2 patients potentially eligible for EMA-approved gene augmentation therapy [29] in England [30].

Gene augmentation therapy is a novel therapeutic approach for genetic disease that seeks to replace null or loss-of-function protein by expressing wild-type copies of the gene of interest, typically through delivery via a viral capsid [31]. The approach is most applicable to recessive traits, and decades-long efforts have demonstrated the approach efficient at rescuing visual loss in animal models of achromatopsia, X-linked and recessive RP, LCA and Stargardt’ disease, amongst others [32,33].

Subsequently, gene augmentation therapy has begun to be translated in clinical trials. ClinicalTrials.gov have estimated 250 listed studies focused on IRDs, including patient registry studies, natural histories, observational and interventional trials [34]. One of the first genes to be targeted was RPE65-LCA, so chosen because of the relative delay in the development of retinal degeneration despite early-onset visual loss, thus offering a wide treatment window. Gene augmentation therapy (voretigene neparvovec-rzyl/Luxturna [35]) has now received FDA approval in the USA (2017) and EMA approval in Europe (2018) for adult and paediatric disease. This first-in-class treatment gives the field hope that a new class of drugs may arise for IRDs [36,37].

Gene augmentation therapies used to date to treat IRDs are based around similar basic adeno-associated virus (AAV) vectors and their capsids, with a variety of promoters chosen by different research groups and companies. To rigorously determine the efficacies of these new therapies it is critical to assess how results show benefit. This requires that methodology, study design and outcome measures provide a clear and reasonable conclusion for the impact on the patient. To this end, we performed a systematic review and meta-analyses of clinical trials for RP patients undergoing gene therapy.

2. Materials and Methods

2.1. Criteria for Considering Studies for This Review

2.1.1. Types of Studies

Articles eligible for inclusion in this systematic review were interventional clinical trials, either randomized or non-randomized, for gene therapy treatments for IRD patients, published in English searched in the relevant databases from <1946 to 2020 Week 5>.

2.1.2. Types of Participants

All patients who have been diagnosed with IRDs, either non-syndromic or syndromic, were included with no restrictions of age, gender or ethnicity. Clinical trials were excluded from patients with ocular comorbidities, or excluded from patients with complications known to influence visual function. Women who were pregnant or lactating or any participants unwilling to use effective contraception were also ineligible.

2.1.3. Types of Interventions

Studies included any investigational gene therapy interventions for IRDs. There were no comparators available for any approved interventions for IRD patients.

2.2. Types of Outcome Measures

2.2.1. Primary Outcomes

The primary outcome of intervention was a mean change from baseline best corrected visual acuity (BCVA) at one year, as measured by the Early Treatment Diabetic Retinopathy Study (ETDRS) chart and measured by logMAR (standard logarithm of the minimum angle of resolution) [38,39,40,41].

Ambulatory navigation/mobility mazes have been developed by a number of research groups and were included as a primary outcome. However, several methodologies exist. To allow comparison between trials, reporting of these assays used a mean difference, i.e., comparing the proportion of improved performances post-operatively, between groups (a risk ratio [RR]).

2.2.2. Secondary Outcomes

Secondary outcomes included full-field light sensitivity threshold (FST) [42], visual field, visual perception, electroretinogram (ERG), Goldmann visual fields, fundus photography, nystagmus testing, central retinal thickness (as measured by optical coherence tomography (OCT)), pupillary light reflex response (PLR) and diagnostic ophthalmic techniques [38,43,44].

2.2.3. Adverse Events

Adverse events were not searched for vector administration, due to a considerable volume of literature for AAV safety outcomes from several authors [45,46,47,48,49].

2.3. Search Methods for Identification of Studies

2.3.1. Electronic Searches

The literature search used the Cochrane Handbook for Systematic Reviews of Interventions [50], using the Ovid database for MEDLINE and EMBASE.

We used a PICOS strategy to identify a systematic review of interventional clinical trials (the study design) for gene therapies (the intervention) for IRD patients (the population), for the purpose of improving the disorder (the outcomes), given there were no treatments available (the comparison). The PICOS search terms and search strings included 36 words and phrases using MeSH terms and Boolean operators, identified in Table S1 (Supplemental).

Structuring and collection of the relevant studies used the PRISMA checklist process [51]. We used the ROBINS-I risk of bias in non-randomised studies of interventions tool [52,53] and the RoB-2 tool for the Cochrane Collaboration’s process for assessing risk of bias in randomised trials [54].

2.3.2. Searching Other Resources

We searched FDA and clinicaltrials.gov databases, including the Biologics License Application resource (BLA) at FDA.

2.4. Data Collection and Analysis

2.4.1. Selection of Studies

Following database searching, each article was assessed as being definitely relevant, possibly relevant or definitively not relevant. Duplicates were removed and all articles assessed for exclusion and inclusion. All included articles were evaluated for study design and reports and final studies reviewed in depth.

2.4.2. Data Extraction and Management

All relevant data (intervention characteristics, study design, primary and secondary outcomes) were extracted and collected in Excel prior to analysis [55] of all available data with Review Manager (RevMan) 5.4 software [56].

2.4.3. Assessment of Risk of Bias in Included Studies

Selected studies were independently assessed for sources of systematic bias according to the guidelines in the relevant sections for the Cochrane Handbook for the Systematic Reviews of Interventions [50] using ROBINS-I and Rob-2 tools.

2.4.4. Measures of Treatment Effect

Primary and secondary outcome data was assessed in accordance with the methods within each selected study. Number of BCVA letters/logMAR at one year (or more) was used to collect the mean difference (MD), standard deviations [SD] and corresponding 95% confidence intervals (CI), comparing mean change from baseline between groups. Continuous data was additionally recorded for improvements (with 95% CIs) for full-field stimulus testing data (red and blue wavelengths) and retinal thickness. Risk ratios (RRs) (with 95% CIs) for dichotomous outcomes were reported, including the proportion of participants with improved/worsening mobility/ambulatory navigation. A random-effects model in RevMan 5.4 meta-analysis [57] was used for meta-analysis.

2.4.5. Unit of Analysis Issues

For most studies, the unit of analysis was the individual participant (one study eye per participant). Five of six studies used a design with one treated eye compared to an untreated control eye. One study (Russell, 2017 [58]), used a crossover design where both eyes received interventions one year apart.

2.4.6. Missing Data

Missing data was not imputed for the purpose of the analysis while only one study (Russell, 2017 [58]) used both an intent-to-treat (ITT) and a modified intent-to-treat (mITT) model.

2.4.7. Heterogeneity

Heterogeneity was tested between the studies using chi-square analysis with significant heterogeneity (p < 0.05) precluding meta-analysis. An I² value of greater than 50% indicated a substantial statistical heterogeneity.

2.4.8. Assessment of Reporting Biases

ROBIN-I and ROB tools were used to assess risk of bias in the five NSRIs and one RCT respectively. Assessments were made by 2 independent examiners.

3. Results

3.1. Systematic-Review of Search Results

Following the structured search approach (Appendix A.1), 115 peer-reviewed research articles were screened and assessed (Figure 1; Appendix A.2), seven articles were removed due to duplication, leaving one hundred and eight articles to be screened. Eighty-seven records were excluded that did not include relevant information. Twenty-one articles were accessed for eligibility; fifteen articles were excluded: one was not applicable for meta-analysis (a study on gene therapy for choroideremia), five were follow-up studies and nine articles included duplicate data. This left six final articles (Appendix A.3) for review and meta-analysis:

3.2. Outcomes

In total, 23 different assays were reported and analysed across the six studies in Figure 2 and

Table 1. Study selection of six (6) research articles for review and meta-analysis outcomes for LCA2-RPE65 gene therapy.

Studies (n = 6) (Journal)	Trial Identifier (ClinicalTrials.gov)	Study Type/Viral Vector	Titre (vg (a))/Injection Vol.	Population (ITT (b))	Age Range (in Years)	BCVA (c) (logMAR)	Ambulatory Navigation, Low Light Ambient Level (<4 lux) (RR (d)(e))	FST (f) (Red Light), log10(cd.s/m2)	FST (f) (Blue Light), log10(cd.s/m2)	Retinal Thickness (OCT, µm)
Bainbridge et al., 2015(NEJM)	NCT00643747	Phase 1–2, open-label, non-randomized; rAAV 2/2. hRPE65p.hRPE65	1.0 × 1011 to 1.0 × 1012 vg; injection volume of 900 µL to 1 mL	12	6–23 years(Median, 15y; Mean, 14.4y; CI 95%, 3.8)	Mean change BCVA of −0.008 logMAR (treated eyes) vs. −0.063 logMAR (untreated eyes), a difference of 0.06 logMAR (95% CI −0.14, 0.02)	Risk ratio of 5.00 (95% CI 0.27, 93.55)	Data unavailable	Data unavailable	Risk of ratio of 1.50 (95% CI 0.30, 7.43)
Jacobson et al., 2012(Arch Ophthalmol)	NCT00481546	Phase 1, open-label, non-randomized; rAAV2-RPE65	5.96 × 1010 to 17.88 × 1010 vg; injection volume of 150 µL to 450 µL	15	11–30 years(Median, 20y; Mean, 19.6y; CI 95%, 3.1)	Mean change BCVA of −0.12 logMAR (treated eyes) vs. −0.05 logMAR (untreated eyes), a difference of −0.07 logMAR (95% CI −0.18, 0.04)	Risk ratio of 1.18 (95% CI 0.86, 1.61)	Mean change FST of 0.45 log10(cd.s/m2) (treated eyes) vs. −0.02 log10(cd.s/m2) (untreated eyes), a difference of 0.47 log10(cd.s/m2) (95% CI 0.24, 0.70)	Mean change FST of 1.59 log10(cd.s/m2) (treated eyes) vs. −0.05 log10(cd.s/m2) (untreated eyes), a difference of 1.64 log10(cd.s/m2) (95% CI 1.14, 2.14)	Risk of ratio of 1.00 (95% CI 0.31, 3.28)
Le Meur et al., 2018(Mol Ther)	NCT01496040	Phase 1/2, open, non-randomized; AAV2/4.-RPE65-RPE65	1.22 × 1010 to 4.8 × 1010 vg; injection volume of 200 µL to 800 µL	9	9–42 years(Median, 22y; Mean, 24.1y; CI 95%, 7.8)	Mean change BCVA of −0.05 logMAR (treated eyes) vs. −0.02 logMAR (untreated eyes), a difference of −0.03 logMAR (95% CI −0.18, 0.12)	Data unavailable	Data unavailable	Data unavailable	Data unavailable
Russell et al., 2017(Lancet)	NCT00999609	Phase 3, open-labelled, randomised (RCT); AAV2-hRPE65v2	1.5 × 1011 vg; injection volume of 300 µL	31	4–44 years(Median, 11y; Mean, 14.4y; CI 95%, 4.1)	Mean change BCVA of −0.163 logMAR (treated eyes) vs. −0.031 logMAR (untreated eyes), a difference of −0.13 logMAR (95% CI −0.29, 0.03)	Risk ratio of 3.6 (95% CI 1.04, 12.46)	Mean change FST of 1.29 log10(cd.s/m2) (treated eyes) vs. −0.16 log10(cd.s/m2) (untreated eyes), a difference of 1.45 log10(cd.s/m2) (95% CI 0.69, 2.21)	Mean change FST of 1.96 log10(cd.s/m2) (treated eyes) vs. −0.13 log10(cd.s/m2) (untreated eyes), a difference of 2.09 log10(cd.s/m2) (95% CI 0.61, 3.57)	Data unavailable
Testa et al., 2013(Ophthalmology)	NCT00516477	Phase 1, open-label, non-randomized (3-year study); AAV2-hRPE65v2	1.0 × 108 to 5.0 × 108 vg; injection volume of 150 µL to 300 µL	5	11–26 years(Median, 19y; Mean, 19.8y; CI 95%, 7.9)	Mean change BCVA of −0.486 logMAR (treated eyes) vs. −0.264 logMAR (untreated eyes), a difference of −0.22 logMAR (95% CI −0.34, −0.10)	Risk ratio of 1.0 (95% CI 0.71, 1.41)	Data unavailable	Data unavailable	Data unavailable
Weleber et al., 2016(Ophthalmology)	NCT00749957	Phase 1–2, open-label, non-randomized; rAAV2-CB-hRPE65	1.0 × 108 to 5.0 × 108 vg	12	6–39 years(Median, 31y; Mean, 24.8y; CI 95%, 8.5)	Mean change BCVA of −0.025 logMAR (treated eyes) vs. −0.046 logMAR (untreated eyes), a difference of 0.02 logMAR (95% CI −0.06, 0.11)	Data unavailable	Data unavailable	Data unavailable	Data unavailable
Summary meta-analyses		Phase 1, 1/2, 3; AAV2	Range from 1.0 × 108 to 1.0 × 1012 vg; injection vol. 150 µL to 1 mL	Population n = 84	Range 4–44 years	Summary weighted mean difference (MD) of −0.06 logMAR improvement over treated vs. untreated eye (95% CI −0.14, 0.02), p = 0.16	RR improvement of 1.35, over treated vs. untreated eye (95% CI 0.78, 2.2.35), p = 0.29	Summary weighted mean difference (MD) of FST (red) 0.89 log10(cd.s/m2) over treated vs. untreated eye (95% CI −0.06, 1.84), p = 0.07	Summary weighted mean difference (MD) of FST (blue) 1.69 log10(cd.s/m2) over treated vs. untreated eye (95% CI 1.21, 2.16), p = 0.00001	RR improvement of 1.15 (95% CI 0.45, 3.00), p = 0.77

(a) vg—vector genomes; (b) ITT—intention to treat; (c) BCVA—Best corrected visual acuity, (logMAR); (d) RR—risk ratio; (e) 95% CI—95% confidence interval; (f) FST—full-field stimulus testing (red and blue wavelength), log10(cd.s/m²).

(including Figure S1a–c, Appendix A.4 and

Table A1. Summary Trial Inclusion, Exclusion Eligibility and Endpoints.

Study Author		Bainbridge et al., 2015 (NEJM). NCT00643747		Jacobson et al., 2012 (Arch Ophthalmol). NCT00481546		Le Meur et al., 2018 (Mol Ther). NCT01496040		Russell et al., 2017 (Lancet). NCT00999609		Testa et al., 2013 (Ophthalmology). NCT00516477		Weleber et al., 2016 (Ophthalmology). NCT00749957
Principal Investigator		Study Director: Robin R Ali, PhD University College, London		Samuel G. Jacobson, MD, PhD; University of Pennsylvania		Michel WEBER, Professor; CHU Nantes		Albert M Maguire, MD (Children’s Hospital of Philadelphia) and Stephen R Russell, MD (University of Iowa).		Study Director: Clinical Director, Spark Therapeutics		J Timothy Stout, MD, PhD, MBA; Casey Eye Institute, Oregon Health & Science University
Sponsor (Academic/Industry)		University College, London; (Moorfields Eye Hospital NHS Foundation Trust; Targeted Genetics Corporation)		University of Pennsylvania		Nantes University Hospital		Spark Therapeutics		Spark Therapeutics		Applied Genetic Technologies Corporation; (Oregon Health and Science University; University of Massachusetts, Worcester.
Official title		An Open-Label Dose Escalation Study of an Adeno-associated Virus Vector (AAV2/2-hRPE65p-hRPE65) for Gene Therapy of Severe Early-Onset Retinal Degeneration		Phase I Trial of Ocular Subretinal Injection of a Recombinant Adeno-Associated Virus (rAAV2-CBSB-hRPE65) Gene Vector to Patients With Retinal Disease Due to RPE65 Mutations (Clinical Trials of Gene Therapy for Leber Congenital Amaurosis)		Prospective Monocentric Open Label Non Randomized Uncontrolled Phase I/II Clinical Gene Therapy Protocol for the Treatment of Retinal Dystrophy Caused by Defects in RPE65		A Safety and Efficacy Study in Subjects With Leber Congenital Amaurosis (LCA) Using Adeno-Associated Viral Vector to Deliver the Gene for Human RPE65 to the Retinal Pigment Epithelium (RPE) [AAV2-hRPE65v2-301]		A Phase 1 Safety Study in Subjects With Leber Congenital Amaurosis (LCA) Using Adeno-Associated Viral Vector to Deliver the Gene for Human RPE65 Into the Retinal Pigment Epithelium (RPE) [AAV2-hRPE65v2-101]		A Multiple-Site, Phase 1/2, Safety and Efficacy Trial of a Recombinant Adeno-associated Virus Vector Expressing RPE65 (rAAV2-CB-hRPE65) in Patients With Leber Congenital Amaurosis Type 2
Study design		Phase 1–2, open-label, non-randomized;		Phase 1, open-label, non-randomized;		Phase 1/2, open, non-randomized;		Phase 3, open-labelled, randomised (RCT);		Phase 1, open-label, non-randomized (3-year study);		Phase 1–2, open-label, non-randomized;
Treatment		rAAV 2/2. hRPE65p.hRPE65		rAAV2-RPE65		AAV2/4.-RPE65-RPE65		AAV2-hRPE65v2		AAV2-hRPE65v2		rAAV2-CB-hRPE65
Inclusion Criteria:	1	Clinical diagnosis of severe early-onset retinal dystrophy confirmed missense mutation(s) in RPE65	1	RPE65-associated retinal disease (two disease-causing RPE65 mutations);	1	Mutations that code for abnormal RPE65 protein	1	Willingness to adhere to protocol and long-term follow-up as evidenced by written informed consent or parental permission and subject assent (where applicable).	1	Male and female subjects of any ethnic group are eligible for participation in this study, providing they meet the following criteria:	1	Retinal disease consistent with a diagnosis of Leber congenital amaurosis and documented mutations in the RPE65 gene (including null mutations and mutations that code for abnormal RPE65 protein);
			2	Clinical diagnosis of Leber congenital amaurosis (LCA)/early-onset retinal degeneration (EORD) and of severely impaired visual and retinal function, and best corrected visual acuity of 20/40 or worse in the study eye;	2	Presence of characteristic abnormalities in fundus	2	Diagnosis of LCA due to RPE65 mutations; molecular diagnosis is to be performed, or confirmed, by a CLIA-approved laboratory.	2	Must be willing to adhere to protocol and companion protocol for long-term follow-up as evidenced by written informed consent or parental permission and subject assent.	2	At least 6 years of age;
			3	Ability to perform tests of visual and retinal function;	3	Dramatic reduction of both rods ans cones ERG responses	3	Age three years old or older.	3	Adults and children diagnosed with LCA.	3	Good general health without significant physical examination findings or clinically significant abnormal laboratory results;
			4	Visible photoreceptor layer on a standard OCT scan;	4	Low visual acuity <0.32	4	Visual acuity worse than 20/60 (both eyes) and/or visual field less than 20 degrees in any meridian as measured by a III4e isopter or equivalent (both eyes).	4	Molecular diagnosis of LCA due to RPE65 mutations (homozygotes or compound heterozygotes) by a CLIA-approved laboratory.	4	Able to perform tests of visual and retinal function;
			5	Good general health;	5	inform consent signed	5	Sufficient viable retinal cells as determined by non-invasive means, such as optical coherence tomography (OCT) and/or ophthalmoscopy. Must have either: 1) an area of retina within the posterior pole of >100 µm thickness shown on OCT; 2) ≥ 3 disc areas of retina without atrophy or pigmentary degeneration within the posterior pole; or 3) remaining visual field within 30 degrees of fixation as measured by a III4e isopter or equivalent.	5	Age eight years old or older at the time of administration.	5	Visual acuity not better than 20/60 and not worse than hand motion in both the treated eye and the fellow eye;
			6	Ability to comply with research procedures;			6	Subjects must be evaluable on mobility testing (the primary efficacy endpoint) to be eligible for the study. Evaluable is defined as: 1) The ability to perform mobility testing within the luminance range evaluated in the study. Individuals must receive an accuracy score of ≤ 1 during screening mobility testing at 400 lux or less to be eligible; individuals with an accuracy score of > 1 on all screening mobility test runs at 400 lux, or those who refuse to perform mobility testing at screening, will be excluded. 2) The inability to pass mobility testing at 1 lux. Individuals must fail screening mobility testing at 1 lux to be eligible; individuals that pass one or more screening mobility test runs at 1 lux will be excluded.	6	Visual acuity ≤ 20/160 or visual field less than 20 degrees in the eye to be injected.	6	Visible photoreceptor (outer nuclear) layer on a standard optical coherence tomography (OCT) scan;
			7	Specific for Cohorts 1, 2 and 4: 18 years of age and older;							7	Acceptable hematology, clinical chemistry and urine laboratory parameters;
			8	Specific for Cohorts 3 and 5: Between 8 and 17 years of age, inclusive.							8	For females of childbearing potential, a negative pregnancy test at screening and at baseline, and agreement to use effective contraception for 12 months after administration of rAAV2-CB-hRPE65, for sexual activity that could lead to pregnancy;
											9	For males of reproductive potential, agreement to use effective contraception for 12 months after administration of rAAV2-CB-hRPE65, for sexual activity that could lead to pregnancy
Exclusion Criteria:	1	Visual acuity in the study eye better than 6/36 Snellen	1	AAV antibody titers greater than two standard deviations above normal at baseline;	1	Patients with chronic conditions such a haematological, cardiac, renal diseases	1	Unable or unwilling to meet requirements of the study, including receiving bilateral subretinal vector administrations.	1	SUBJECTS WILL NOT BE EXCLUDED BASED ON THEIR GENDER, RACE OR ETHNICITY. Subjects who meet any of the following conditions are excluded from the clinical study: Subjects who meet any of the following conditions are excluded from the clinical study:	1	Pre-existing eye conditions that would preclude the planned surgery or interfere with interpretation of study endpoints or complications of surgery (e.g., glaucoma, corneal or lenticular opacities, or history or retinal detachment);
	2	Hypertension	2	Humoral immune deficiency as evidenced by low tetanus toxoid IgG antibody titers;	2	Patients with, within the past 6 months, a clinically significant cardiac disease or known congestive heart failure, cardiac rhytm and conduction abnormalities	2	Any prior participation in a study in which a gene therapy vector was administered.	2	Unable or unwilling to meet requirements of the study.	2	Presence of epiretinal membrane on OCT;
	3	Diabetes mellitus	3	Pre-existing eye conditions that would preclude the planned surgery or interfere with the interpretation of study endpoints or surgical complications;	3	Patients with pulmonaty dysfunction	3	Participation in a clinical study with an investigational drug in the past six months.	3		3	History of immunodeficiency or other medical conditions that might increase the risk of rAAV2-CB-hRPE65 administration;
	4	Tuberculosis	4	Complicating systemic diseases;	4	Patients with suspected rheumatoid arthritis	4	Use of retinoid compounds or precursors that could potentially interact with the biochemical activity of the RPE65 enzyme; individuals who discontinue use of these compounds for 18 months may become eligible.	4	Participation in a clinical study with an investigational drug in the past six months.	4	Use of anticoagulants or anti-platelet agents within 7 days prior to study agent administration;
	5	Renal impairment	5	Use of anti-platelet agents that may alter coagulation within 7 days prior to study agent administration;	5	Patients with current systemic infection……..	5	Prior intraocular surgery within six months.	5	Pre-existing eye conditions that would preclude the planned surgery or interfere with the interpretation of study endpoints (for example, glaucoma, corneal or lenticular opacities).	5	History of allergy or sensitivity to medications planned for use in the peri-operative period;
	6	Immunocompromise	6	Use of immunosuppressive medications;			6	Known sensitivity to medications planned for use in the peri-operative period.	6	Lack of sufficient viable retinal cells as determined by non-invasive means, such as optical coherence tomography (OCT) and/or ophthalmoscopy. Specifically, if indirect ophthalmoscopy reveals less than 1 disc area of retina which is not involved by complete retinal degeneration (indicated by geographic atrophy, thinning with tapetal sheen, or confluent intraretinal pigment migration), these eyes will be excluded. In addition, in eyes where optical coherence tomography (OCT) scans of sufficient quality can be obtained, areas of retina with thickness measurements less than 100 um, or absence of neural retina, will not be targeted for delivery of AAV2-hRPE65v2.	6	For females of childbearing potential, a positive pregnancy test at screening or baseline (within 2 days before rAAV2-CB-hRPE65 administration);
	7	Osteoporosis	7	Pregnancy or breastfeeding;			7	Pre-existing eye conditions or complicating systemic diseases that would preclude the planned surgery or interfere with the interpretation of study. Complicating systemic diseases would include those in which the disease itself, or the treatment for the disease, can alter ocular function. Examples are malignancies whose treatment could affect central nervous system function (for example: radiation treatment of the orbit; leukemia with CNS/optic nerve involvement). Subjects with diabetes or sickle cell disease would be excluded if they had any manifestation of advanced retinopathy (e.g., macular edema or proliferative changes). Also excluded would be subjects with immunodeficiency (acquired or congenital) as there could be susceptibility to opportunistic infection (such as CMV retinitis).	7	Complicating systemic diseases or clinically significant abnormal baseline laboratory values. Complicating systemic diseases would include those in which the disease itself, or the treatment for the disease, can alter ocular function. Examples are malignancies whose treatment could affect central nervous system function (for example, radiation treatment of the orbit; leukemia with CNS/optic nerve involvement). Also excluded would be subjects with immuno-compromising diseases, as there could be susceptibility to opportunistic infection (such as CMV retinitis). Subjects with diabetes or sickle cell disease would be excluded if they had any manifestation of advanced retinopathy (e.g., macular edema or proliferative changes). Subjects with juvenile rheumatoid arthritis could be excluded due to increased infection risk after surgery due to poor wound healing. Subjects who are positive for hepatitis B, C, and HIV will be excluded.	7	Females who are breast feeding;
	8	Gastric ulceration	8	Individuals (males and females) of childbearing potential who are unwilling to use effective contraception;			8	Individuals of childbearing potential who are pregnant or unwilling to use effective contraception for four months following vector administration.	8	Prior ocular surgery within six months.	8	Use of any investigational agent, or systemic corticosteroids or other immunosuppressive drug(s), within 3 months prior to enrollment;
	9	Severe affective disorder)	9	Any condition that would prevent a subject from completing follow-up examinations during the course of the study;			9	Individuals incapable of performing mobility testing (the primary efficacy endpoint) for reason other than poor vision, including physical or attentional limitations.	9	Known sensitivity to medications planned for use in the peri-operative period.	9	Prior receipt of any AAV gene therapy product;
	10	Pregnancy or lactation	10	Any condition that makes the subject unsuitable for the study;			10	Any other condition that would not allow the potential subject to complete follow-up examinations during the course of the study or, in the opinion of the investigator, makes the potential subject unsuitable for the study.	10	Individuals of childbearing potential who are pregnant or unwilling to use effective contraception for the duration of the study.	10	Any condition which leads the investigator to believe that the participant cannot comply with the protocol requirements or that may place the participant at an unacceptable risk for participation.
			11	Current, or recent participation, in any other research protocol involving investigational agents or therapies;			11	Subjects will not be excluded based on their gender, race, or ethnicity.	11	Any other condition that would not allow the potential subject to complete follow-up examinations during the course of the study and, in the opinion of the investigator, makes the potential subject unsuitable for the study.
			12	Recent receipt of an investigational biologic therapeutic agent.					12	Subjects will be excluded if immunological studies show presence of neutralizing antibodies to AAV2 above 1:1000.
Primary Outcome Measures:	1	intraocular inflammation [Time Frame: at intervals up to 12 months]	1	The primary safety endpoint in this trial is the standard ocular examination. Toxicity will also be assessed by measurement of vision, hematology and serum chemistries, assays for vector genomes, reported subject history of symptoms and adverse events. [Time Frame: 15 years]	1	The drug safety evaluation after administration [Time Frame: After administration of the gene therapy product.The patient will be folloed for the duration of the hospital stay, an average of 7 days]	1	Multi-luminance Mobility Testing (MLMT), Bilateral [Time Frame: One year (change from baseline)]; The MLMT measures changes in functional vision, as assessed by the ability to navigate a course accurately and at a reasonable pace at different levels of environmental illumination. MLMT was assessed using both eyes at 1 or more of 7 levels of illumination, ranging from 400 lux (a brightly lit office) to 1 lux (a moonless summer night). Each light level was assigned a score code ranging from 0 to 6. A higher score indicated that a subject was able to pass the MLMT at a lower light level. A score of −1 was assigned to those who could not pass MLMT at 400 lux. The MLMT of each subject was videotaped and assessed by independent graders. The MLMT score was determined by the lowest light level at which the subject was able to pass the MLMT. The MLMT score change was defined as the difference between the score at Baseline and the score at Year 1. A positive MLMT score change from Baseline to Year 1 visit indicated that the subject was able to complete the MLMT at a lower light level.	1	The primary outcome measures are safety and tolerability. Secondary outcome measure(s) include changes in visual function as measured by subjective, psychophysical tests and by objective, physiologic tests. [Time Frame: Visual function will be measured at designated intervals from baseline visits through 5 years as stated in the protocol.]	1	Number of Participants Experiencing Ocular or Non-ocular Adverse Events [Time Frame: 2 years]
Secondary Outcome Measures:	1	visual function [Time Frame: intervals up to 12 months]	1	Visual function will be quantified prior to and after vector administration in order to determine whether vector administration affects visual function. [Time Frame: 15 years]	2	Biodistribution: Urine sampling and nasal secretion will be collected at several time points after administration of the gene therapy product during all the duration of hospital stay, an average of 7 days.	1	Full-field Light Sensitivity Threshold (FST) Testing: White Light [Time Frame: One year (change from baseline)]; Measures the light sensitivity of the entire visual field by recording the luminance at which a subject reliably reports seeing the dimmest flash.			2	Participants With Changes in Visual Fields [Time Frame: 2 years]; Improvement in the central 30 degree visual field, measured by static perimetry, at one or more time points after treatment, that was greater than the limit of agreement for baseline values.
					3	Different efficacy parameters and immune parameters have to be measured to conclude on the overall amelioration of quality of life of enrolled patients [Time Frame: Between Day −120 and Day −7, Day 5, Day 14, Day 30 Day 60, Day 90, Day 120, Day 180, Day 360]	2	Multi-luminance Mobility Testing (Monocular) [Time Frame: One year (change from baseline)]; The MLMT measures changes in functional vision, as assessed by the ability to navigate a course accurately and at a reasonable pace at different levels of environmental illumination. MLMT was assessed using the first eye at 1 or more of 7 levels of illumination, ranging from 400 lux (a brightly lit office) to 1 lux (a moonless summer night). Each light level was assigned a score code ranging from 0 to 6. A higher score indicated that a subject was able to pass the MLMT at a lower light level. A score of -1 was assigned to those who could not pass MLMT at 400 lux. The MLMT of each subject was videotaped and assessed by independent graders. The MLMT score was determined by the lowest light level at which the subject was able to pass the MLMT. The MLMT score change was defined as the difference between the score at Baseline and the score at Year 1. A positive MLMT score change from Baseline to Year 1 visit indicated that the subject was able to complete the MLMT at a lower light level.			3	Participants With Changes in Best Corrected Visual Acuity [Time Frame: 2 years]; Increase in BCVA of 7 or more letters at Year 2 visit compared to average baseline value
					4	Recording global ERG (electroretinogram)	3	Visual Acuity [Time Frame: One year (change from baseline)]; Measurement of the sharpness of vision, determined by the ability to read letters on a standardized chart from a specified distance.
					5	Patient efficacy questionnaire
					6	Testing of far and near visual acuity, color vision, pupillometry, microperimetry and dark adaptation.

). Safety data was not collected on specific AAV2 vectors, having been examined in other independent studies on interventional clinical studies in the retina [37,59,60,61,62]. Only one assay, visual acuity (VA), was common to all six papers. Of the 23 assays reviewed, only five outcomes were reported for meta-analysis—VA (logMAR), mobility, red light full-field stimulus (FST) testing (log10(cd.s/m²)), blue light full-field stimulus (FST) testing (log10(cd.s/m²)) and central retinal thickness (CRT).

All continuous and dichotomous data was reported. If either continuous and dichotomous data were available, then analysis was used to compare and contrast the models. However, continuous data was preferred from a statistical perspective as some information risked being lost in categorical data.

3.3. Visual Acuity Measured by logMAR

A 0.30 logMAR (3 line) mean post-operative change of VA was accepted as being a “clinically meaningful” improvement [43]. VA results were reported in Figure 3. Overall, outcomes showed a benefit of treatment compared to control eyes, but did not meet statistical significance.

There was a mean [SD] improvement of −0.142[0.181] logMAR letters in treated eyes (n = 73), compared to −0.079[0.103] logMAR letters in untreated (n = 62), showing a difference of logMAR −0.063 (including Table S2 (Supplemental)). RevMan 5.4 analysis showed a statistical difference of −0.06 logMAR (95% CI [−0.14, 0.02], p = 0.16) above. Individually, only one study (Testa, 2013), reported a clinically meaningful improvement in VA, with a mean [SD] improvement of −0.49 [0.04] logMAR letters in the treated eye, compared to a mean [SD] improvement of −0.26 [0.13] logMAR in the untreated eye.

Finally, an analysis of dichotomous data on visual outcomes post treatment (better or worse) was performed. Five studies provided individual patient data to allow this analysis (n = 58 treated eyes; n = 47 untreated eyes). The line of no effect showed an RR of 1.13 (95% CI 0.83, 1.53), indicating an improvement with treatment that did not reached clinical significance (p = 0.44) (Figure S2 (Supplemental)).

3.4. Mobility

Given the disparity between the four different mobility methods used in the studies in terms of size, light intensity, scoring and reporting, no direct comparison was possible. Instead, a meta-analysis of dichotomous data (better/worse post-treatment) was performed, To do so, four sub-groups were defined, according to light intensity used to illuminate the mobility mazes: (a): mobility under a single light of intensity of 4 lux; (b): mobility under a “low” ambient light level (0.2, 0.6, 1, 2 or 4 lux), broadly scotopic light; (c): mobility under a “high” ambient light level (10, 15, 50, 100 or 125 lux), broadly photopic light vision function; (d): mobility under all ambient light levels ranging from 0.2 to 100 lux. Results are summarised in Figure 4 (and in Table S3 (Supplemental)).

Under a light intensity of 4 lux (Figure 4, “Lux 4”), analysis of 4 studies showed an RR of 1.03 (95% CI 0.75, 1.42), indicating an improvement with treatment that did not reach clinical significance (p = 0.84). Under low ambient light (“Low Lux 0.2 to 4”), analysis of 4 studies showed an RR of 1.35 (95% CI 0.78, 2.35), indicating an improvement with treatment that did not reach clinical significance (p = 0.29). Under high ambient light (“High Lux 10 to 100”), analysis of 4 studies showed an RR of 0.42 (95% CI 0.12, 1.50), indicating a worsening with treatment that did not reach clinical significance (p = 0.18). Analysis of all ambient light levels (“All lux levels 0.2 to 100”) of 4 studies showed an RR of 1.15 (95% CI 0.84, 1.58), indicating an improvement with treatment that did not reach clinical significance (p = 0.39).

3.5. Full-Field Stimulus (FST) Testing for Red and Blue Wavelength

Only two studies used FST testing that allowed for a meta-analysis. Both continuous and dichotomous (better/worse) data was analysed (Figure 5a–d).

Under red light FST results, analysis of continuous data, showed a mean difference [MD] of 0.89 log10(cd.s/m²) (95% CI −0.06, 1.84) in treated eyes compared to control, indicating an improvement with treatment that did not reach clinical significance (p = 0.07). Analysis of dichotomous data (better/worse) for red light FST showed a RR of 1.89 (95% CI 1.04, 3.41), indicating an improvement with treatment that reached clinical significance (p = 0.04).

Under blue light FST results, analysis of continuous data, showed a difference of 1.69 log10(cd.s/m²) (95% CI 1.21, 2.16) in treated eyes compared to control, indicating an improvement with treatment that reached clinical significance (p = 0.00001). Analysis of dichotomous data (better/worse) showed a RR of 2.01 (95% CI 1.32, 3.06), indicating an improvement with treatment that reached clinical significance (p = 0.001).

3.6. Central Retinal Thickness (CRT)

Three studies reported CRT outcomes as measured by optical coherence tomography (OCT), but only two included quantitative data that allowed for meta-analysis. Analysis of dichotomous data (thinner/thicker) at 1 year post treatment, showed a RR of 1.15 (95% CI 0.45, 3.00), indicating an increase of CRT with treatment that did not reach clinical significance (p = 0.77) (Figure 6a). Analysis of dichotomous data for a long term timepoint (3 years), showed a RR of 1.29 (95% CI 0.33, 5.10), indicating an increase of CRT with treatment that did not reach clinical significance (p = 0.72) (Figure 6b).

3.7. Risk of Bias Tools within Studies

Cochrane risk-of-bias tools were used to assess study reliability; ROBIN-I methods [52], for non-randomised study designs, and RoB-2 methods [63], for randomised clinical trials. Overall, a risk-of-bias judgement was reported “low/moderate”, with a predicted direction of bias “towards null/unpredictable” for the 5 NRSIs, and a report with “some concerns” and a predicted direction of bias with “favours experimental” for the RCT (

Table 2. Analysis of risk of bias studies for five (5) NRSIs and one (1) RCT.

Study Author & Year	ROBIN-I (Risk of Bias in Non-Randomised Studies of Interventions)		RoB-2 (Risk of Bias in Randomised Studies of Interventions [RCT])
	Risk-of-Bias Judgement	Overall Predicted Direction of Bias	Risk-of-Bias Judgement	Overall Predicted Direction of Bias
Bainbridge et al., 2015	Low/Moderate	Towards null/Unpredictable	N/A	N/A
Jacobson et al., 2012	Low/Moderate	Towards null/Unpredictable	N/A	N/A
Le Meur et al., 2018	Low/Moderate	Towards null/Unpredictable	N/A	N/A
Russell et al., 2017	N/A	N/A	Some concerns	Favours experimental
Testa et al., 2013	Low/Moderate	Towards null/Unpredictable	N/A	N/A
Weleber et al., 2016	Low/Moderate	Towards null/Unpredictable	N/A	N/A

; Appendix A.5,

Table A2. ROBINS-I Risk of Bias in Non-Randomised Studies of Interventions.

	ROBINS-I Risk of Bias in Non-Randomised Studies of Interventions (NRSI)
(1) Yes[Y]; (2) Probably Yes[PY]; (3) Probably no[PN]; (4) No[No]; and (5) No information[NI].
Risk of bias assessment		A		B		C		D		E
Responses underlined in green are potential markers for low risk of bias, and responses in red are potential markers for a risk of bias. Where questions relate only to sign posts to other questions, no formatting is used.		Bainbridge et al., 2015		Jacobson et al., 2012		Le Meur et al., 2018		Testa et al., 2013		Weleber et al., 2016
	Signalling questions	Description	Response options	Description	Response options	Description	Response options	Description	Response options	Description	Response options
Bias due to confounding
	1.1 Is there potential for confounding of the effect of intervention in this study?	4 of the 12 patients used the better eye, contradicting the text suggesting that the poorer eye should be the treated eye, Table 1 (page 1889); the conentration of vector genomes (vg) have different cohorts;	Y	Y; “worse” vs. “better” eyes was logMAR 1.09 vs. 0.96, respectively, and there was no adjustment;	Y	Y; “worse” had an avarage of 31.5 letters while “better”eyes had 41.1 letters and there was no adjustment;	Y	Y; “worse” had a logMAR of 1.47 while “better”eyes had 1.14 logMAR, and there was no adjustment;	Y	Y; “worse” had an avarage of 24.0 letters while “better”eyes had 28.4 letters and there was no adjustment;	Y
	1.2. Was the analysis based on splitting participants’ follow up time according to intervention received?	Not significant time split between control vs. intervention;	N		N		N		N		N
	1.3. Were intervention discontinuations or switches likely to be related to factors that are prognostic for the outcome?		NI		NI		NI		NI		NI
	Questions relating to baseline confounding only
	1.4. Did the authors use an appropriate analysis method that controlled for all the important confounding domains?	There was no information that controlled a confounding domain;	N	See comment on Jacobson, page 21;	N		N		N		N
	1.5. If Y/PY to 1.4: Were confounding domains that were controlled for measured validly and reliably by the variables available in this study?	There was no contolled confounding domains.	N		N		N		N		N
	1.6. Did the authors control for any post-intervention variables that could have been affected by the intervention?	No information;	NI		N		NI		NI		NI
	Questions relating to baseline and time-varying confounding
	1.7. Did the authors use an appropriate analysis method that controlled for all the important confounding domains and for time-varying confounding?	There is no control or adjustment that might impact the outcome between the worse eye vs. the better eye;	N		N		N		N		N
	1.8. If Y/PY to 1.7: Were confounding domains that were controlled for measured validly and reliably by the variables available in this study?	N, see comment above re: better eye of 4 patients were used;	N		N		N		N		N
	Risk of bias judgement	Moderate/(?); confounding occurs due to uncorrected baselines for treated vs. control;	Low/Moderate		Low/Moderate		Low/Moderate		Low/Moderate		Low/Moderate
	Optional: What is the predicted direction of bias due to confounding?	Unpredictable/Favours experimental	Unpredictable		Unpredictable		Unpredictable		Unpredictable		Unpredictable
Bias in selection of participants into the study		Bainbridge et al., 2015		Jacobson et al., 2012		Le Meur et al., 2018		Testa et al., 2013		Weleber et al., 2016
	2.1. Was selection of participants into the study (or into the analysis) based on participant characteristics observed after the start of intervention?		N		N		N		N		N
	2.2. If Y/PY to 2.1: Were the post-intervention variables that influenced selection likely to be associated with intervention?		N		N		N		N		N
	2.3 If Y/PY to 2.2: Were the post-intervention variables that influenced selection likely to be influenced by the outcome or a cause of the outcome?		N		N		N		N		N
	2.4. Do start of follow-up and start of intervention coincide for most participants?		NI		NI		NI		NI		NI
	2.5. If Y/PY to 2.2 and 2.3, or N/PN to 2.4: Were adjustment techniques used that are likely to correct for the presence of selection biases?		NI		N		NI		NI		NI
	Risk of bias judgement	Low	Low		Low		Low		Low		Low
	Optional: What is the predicted direction of bias due to selection of participants into the study?		Towards null		Towards null		Towards null		Towards null		Towards null
Bias in classification of interventions		Bainbridge et al., 2015		Jacobson et al., 2012		Le Meur et al., 2018		Testa et al., 2013		Weleber et al., 2016
	3.1 Were intervention groups clearly defined?		Y		Y		Y		Y		Y
	3.2 Was the information used to define intervention groups recorded at the start of the intervention?		Y		Y		Y		Y		Y
	3.3 Could classification of intervention status have been affected by knowledge of the outcome or risk of the outcome?		PN		PN		PN		PN		PN
	Risk of bias judgement	Low/Moderate	Low		Low		Low		Low		Low
	Optional: What is the predicted direction of bias due to classification of interventions?		Towards null		Towards null		Towards null		Towards null		Towards null
Bias due to deviations from intended interventions		Bainbridge et al., 2015		Jacobson et al., 2012		Le Meur et al., 2018		Testa et al., 2013		Weleber et al., 2016
	If your aim for this study is to assess the effect of assignment to intervention, answer questions 4.1 and 4.2
	4.1. Were there deviations from the intended intervention beyond what would be expected in usual practice?		N		N		N		N		N
	4.2. If Y/PY to 4.1: Were these deviations from intended intervention unbalanced between groups and likely to have affected the outcome?		N		N		N		N		N
	If your aim for this study is to assess the effect of starting and adhering to intervention, answer questions 4.3 to 4.6
	4.3. Were important co-interventions balanced across intervention groups?		NI		NI		NI		NI		NI
	4.4. Was the intervention implemented successfully for most participants?		PY		PY		PY		PY		PY
	4.5. Did study participants adhere to the assigned intervention regimen?		Y		Y		Y		Y		Y
	4.6. If N/PN to 4.3, 4.4 or 4.5: Was an appropriate analysis used to estimate the effect of starting and adhering to the intervention?		Y		Y		Y		Y		Y
	Risk of bias judgement	Low	Low		Low		Low		Low		Low
	Optional: What is the predicted direction of bias due to deviations from the intended interventions?		Towards null		Towards null		Towards null		Towards null		Towards null
Bias due to missing data		Bainbridge et al., 2015		Jacobson et al., 2012		Le Meur et al., 2018		Testa et al., 2013		Weleber et al., 2016
	5.1 Were outcome data available for all, or nearly all, participants?	N/PN; some partial missing data (see mobility data in Figure 2, pg 1892);	Y		Y		Y		Y		Y
	5.2 Were participants excluded due to missing data on intervention status?
			N		N		N		N		N
	5.3 Were participants excluded due to missing data on other variables needed for the analysis?
			NI		NI		NI		NI		NI
	5.4 If PN/N to 5.1, or Y/PY to 5.2 or 5.3: Are the proportion of participants and reasons for missing data similar across interventions?		NA		NA		NA		NA		NA
	5.5 If PN/N to 5.1, or Y/PY to 5.2 or 5.3: Is there evidence that results were robust to the presence of missing data?		NI		NI		NI		NI		NI
	Risk of bias judgement	Low	Low		Low		Low		Low		Low
	Optional: What is the predicted direction of bias due to missing data?		Towards null		Towards null		Towards null		Towards null		Towards null
Bias in measurement of outcomes		Bainbridge et al., 2015		Jacobson et al., 2012		Le Meur et al., 2018		Testa et al., 2013		Weleber et al., 2016
	6.1 Could the outcome measure have been influenced by knowledge of the intervention received?		NI		NI		NI		NI		NI
	6.2 Were outcome assessors aware of the intervention received by study participants?		NI		NI		NI		NI		NI
	6.3 Were the methods of outcome assessment comparable across intervention groups?		Y		Y		Y		Y		Y
	6.4 Were any systematic errors in measurement of the outcome related to intervention received?		PN		PN		PN		PN		PN
	Risk of bias judgement		Low		Low		Low		Low		Low
	Optional: What is the predicted direction of bias due to measurement of outcomes?		Towards null		Towards null		Towards null		Towards null		Towards null
Bias in selection of the reported result		Bainbridge et al., 2015		Jacobson et al., 2012		Le Meur et al., 2018		Testa et al., 2013		Weleber et al., 2016
	Is the reported effect estimate likely to be selected, on the basis of the results, from...
	7.1 … multiple outcome measurements within the outcome domain?	Y; some outcomes were reported to collect some data (e.g., ERG, contrast sensitivity, or FAF) but the outcomes did not show the data; see this main paper of Tuohy & Megaw, Figure 2, showing 23 assays, arranged alphabetically; estimate of 10 of 23 outcomes identified;	NI	See this main paper of Tuohy & Megaw, Figure 2, showing 23 assays, arranged alphabetically; estimate of 8 of 23 outcomes identified;	NI	See this main paper of Tuohy & Megaw, Figure 2, showing 23 assays, arranged alphabetically; estimate of 6 of 23 outcomes identified;	NI	See this main paper of Tuohy & Megaw, Figure 2, showing 23 assays, arranged alphabetically; estimate of 9 of 23 outcomes identified;	NI	See this main paper of Tuohy & Megaw, Figure 2, showing 23 assays, arranged alphabetically; estimate of 5 of 23 outcomes identified;	NI
	7.2 … multiple analyses of the intervention-outcome relationship?	Y; there was no adjusted data	NI		NI		NI		NI		NI
	7.3 … different subgroups?		NI		NI		NI		NI		NI
	Risk of bias judgement	Low/Moderate	Low		Low		Low		Low		Low
	Optional: What is the predicted direction of bias due to selection of the reported result?	Towards null/Unpredictable	Towards null		Towards null		Towards null		Towards null		Towards null
Overall bias		Bainbridge et al., 2015		Jacobson et al., 2012		Le Meur et al., 2018		Testa et al., 2013		Weleber et al., 2016
	Risk of bias judgement		Low		Low		Low		Low		Low
	Optional: What is the overall predicted direction of bias for this outcome?		Towards null		Towards null		Towards null		Towards null		Towards null

. and Appendix A.6,

Table A3. RoB-2—Risk of Bias in Randomised Studies of Interventions.

Cochrane Tool for RoB-2—Risk of Bias in Randomised Studies of Interventions (RoB 2)
(1) Yes[Y]; (2) Probably Yes[PY]; (3) Probably No[PN]; (4) No[No]; and (5) No Information[NI].
Domain 1: Risk of bias arising from the randomization process
Signalling questions	Comments	Response options	Actual responses—Russell et al., 2017
1.1 Was the allocation sequence random?	The study was randomized; the allocation sequence was performed under direction of an independent biostatistician assigned to either intervention or control and the list of patients was created before enrolment.	Y/PY/PN/N/NI	Y
1.2 Was the allocation sequence concealed until participants were enrolled and assigned to interventions?		Y/PY/PN/N/NI	Y
1.3 Did baseline differences between intervention groups suggest a problem with the randomization process?	There were unequal groups (Treated n = 21 vs. Untreated n = 10); in addition, the baselines between Treated and Untreated eyes were unequal; “some concerns”.	Y/PY/PN/N/NI	Y
Risk-of-bias judgement	There is a risk-of-bias due to the small sample and this concern was identified the issue from the authors themselves (see Russell, 2017, page 858).	Low/High/Some concerns	Low/Some concerns
Optional: What is the predicted direction of bias arising from the randomization process?	Favours expirmental; change of LogMAR, ambulatory navigation/mobility and full-field sensitivity improves visual function	NA/Favours experimental/Favours comparator/Towards null/Away from null/Unpredictable	Favours expirmental
Domain 2: Risk of bias due to deviations from the intended interventions (effect of assignment to intervention)
Signalling questions	Comments	Response options	Actual responses—Russell et al., 2017
2.1. Were participants aware of their assigned intervention during the trial?	The study was “open-label” therefore the patients were unblinded.	Y/PY/PN/N/NI	N
2.2. Were carers and people delivering the interventions aware of participants’ assigned intervention during the trial?		Y/PY/PN/N/NI	N
2.3. If Y/PY/NI to 2.1 or 2.2: Were there deviations from the intended intervention that arose because of the trial context?	Deviations may arise on the basis of being unblinded due to the “trial context”, specifcally the absence of unequal baselines) (see interpretation in the Cochrane advice, Rob-2 explanation no. 2.3)	NA/Y/PY/PN/N/NI	PY
2.4 If Y/PY to 2.3: Were these deviations likely to have affected the outcome?	There is “NI”; there was no trial protocol avabilable in the Russell paper, or no IND 13804 document or not included in the BLA 125610 (FDA)	NA/Y/PY/PN/N/NI	NI
2.5. If Y/PY/NI to 2.4: Were these deviations from intended intervention balanced between groups?	See question above, 2.4	NA/Y/PY/PN/N/NI	NI
2.6 Was an appropriate analysis used to estimate the effect of assignment to intervention?		Y/PY/PN/N/NI	Y
2.7 If N/PN/NI to 2.6: Was there potential for a substantial impact (on the result) of the failure to analyse participants in the group to which they were randomized?		NA/Y/PY/PN/N/NI	N
Risk-of-bias judgement		Low/High/Some concerns	Some concerns
Optional: What is the predicted direction of bias due to deviations from intended interventions?		NA/Favours experimental/Favours comparator/Towards null/Away from null/Unpredictable	Favours expirmental
Domain 3: Missing outcome data
Signalling questions	Comments	Response options	Actual responses—Russell et al., 2017
3.1 Were data for this outcome available for all, or nearly all, participants randomized?		Y/PY/PN/N/NI	Y
3.2 If N/PN/NI to 3.1: Is there evidence that the result was not biased by missing outcome data?		NA/Y/PY/PN/N	N
3.3 If N/PN to 3.2: Could missingness in the outcome depend on its true value?		NA/Y/PY/PN/N/NI	N
3.4 If Y/PY/NI to 3.3: Is it likely that missingness in the outcome depended on its true value?		NA/Y/PY/PN/N/NI	N
Risk-of-bias judgement		Low/High/Some concerns	Low
Optional: What is the predicted direction of bias due to missing outcome data?		NA/Favours experimental/Favours comparator/Towards null/Away from null/Unpredictable	Favours expirmental
Domain 4: Risk of bias in measurement of the outcome
Signalling questions	Comments	Response options	Actual responses—Russell et al., 2017
4.1 Was the method of measuring the outcome inappropriate?	The outcome methods for the mobility test (MLMT) had several measurements rolled into one final outcome without sufficient data incuding: (i) speed; (ii) time; (iii) accuracy; (iv) obstacles; (v) time penalties; (vi) lux, and; (vii) the scales of ordinal vs. logarithmic interpretation. Consequently, (a) the full direct data and measurements were not avaiable, and; (b) the final outcome may have different interpretations; see comments from FDA reviewers in the BLA 125610.	Y/PY/PN/N/NI	Y
4.2 Could measurement or ascertainment of the outcome have differed between intervention groups?	As above.	Y/PY/PN/N/NI	PY
4.3 If N/PN/NI to 4.1 and 4.2: Were outcome assessors aware of the intervention received by study participants?	Open-label, therefore there was no blinding.	NA/Y/PY/PN/N/NI	Y
4.4 If Y/PY/NI to 4.3: Could assessment of the outcome have been influenced by knowledge of intervention received?	The study is an open-label, unblinded and approved design; while there may be some unconscious influence, there is no evdience for such, therefore was no influence reported; PN or N.	NA/Y/PY/PN/N/NI	PN/N
4.5 If Y/PY/NI to 4.4: Is it likely that assessment of the outcome was influenced by knowledge of intervention received?		NA/Y/PY/PN/N/NI	PN/N
Risk-of-bias judgement		Low/High/Some concerns	Some concerns
Optional: What is the predicted direction of bias in measurement of the outcome?		NA/Favours experimental/Favours comparator/Towards null/Away from null/Unpredictable	Favours expirmental
Domain 5: Risk of bias in selection of the reported result
Signalling questions	Comments	Response options	Actual responses—Russell et al., 2017
5.1 Were the data that produced this result analysed in accordance with a pre-specified analysis plan that was finalized before unblinded outcome data were available for analysis?	There was no pre-specified analysis plan (a protocol) so there is: NI.	Y/PY/PN/N/NI	NI
Is the numerical result being assessed likely to have been selected, on the basis of the results, from...	The MLMT assay is a novel primary outcome; see question and response above in 4.2	Y/PY/PN/N/NI	NI
5.2 … multiple eligible outcome measurements (e.g., scales, definitions, time points) within the outcome domain?		Y/PY/PN/N/NI	NI
5.3 … multiple eligible analyses of the data?		Y/PY/PN/N/NI
Risk-of-bias judgement		Low/High/Some concerns	Some concerns
Optional: What is the predicted direction of bias due to selection of the reported result?		NA/Favours experimental/Favours comparator/Towards null/Away from null/Unpredictable	Favours experimental
Overall risk of bias
	Comments	Response options	Actual responses—Russell et al., 2017
Risk-of-bias judgement	Further data would be valauble to conclude an overall risk of bias/judgement.	Low/High/Some concerns	Some concerns
Optional: What is the overall predicted direction of bias for this outcome?		NA/Favours experimental/Favours comparator/Towards null/Away from null/Unpredictable	Favours experimental

).

Finally,

Table 3. Summary table of 12 meta-analyses with a total of all treated and untreated eyes showed that three meta-analyses have reported statistical significance within the table below *. Of the 3 of the 12 meta-analyses that reached statistical significance, FST (red light) had an RR improvement of 1.89 (95% CI 1.04, 3.41) p = 0.04; FST (blue light) had a MD improvement of 1.69 (95% CI 1.21, 2.16) p = 0.00001, and finally; FST (blue light) had an RR improvement of 2.01, (95% CI 1.32, 3.06), p = 0.001.

No.	Meta Analyses	Number of Studies for Meta-Analysis	Treated Eyes	Untreated Eyes	Study Author	Continuous (C)/Dichotomous (D)	Formal Result (MD or RR)	95% Confidence Interval	Chi2	I2	Z Efffect	p Value
1	LogMAR visual acuity	6	73	62	Bainbridge, Jacobson, Le Meur, Russell, Testa, Weleber	C	MD −0.06	CI (−0.14, 0.02)	14.39	65%	1.40	0.16
2	LogMAR visual acuity	6	58	47	Bainbridge, Jacobson, Le Meur, Russell, Testa, Weleber	D	RR 1.13	CI (0.83, 1.53)	3.92	0%	0.77	0.44
3	Ambulatory navigation/mobility: Sub-group A (4 lux)	4	42	31	Bainbridge, Jacobson, Russell, Testa,	D	RR 1.03	CI (0.75, 1.42)	2.35	0%	0.21	0.84
4	Ambulatory navigation/mobility: Sub-group B (0.2, 0.6, 1, 2, 4 lux)	4	60	49	Bainbridge, Jacobson, Russell, Testa,	D	RR 1.35	CI (0.78, 2.35)	9.58	69%	1.07	0.29
5	Ambulatory navigation/mobility: Sub-group C (10, 15, 50, 100, 125 lux)	3	38	27	Bainbridge, Jacobson, Russell	D	RR 0.42	CI (0.12, 1.50)	0.89	0%	1.33	0.18
6	Ambulatory navigation/mobility: Sub-group D (0.2–125 lux)	4	66	56	Bainbridge, Russell	D	RR 1.15	CI (0.84, 1.58)	4.30	30%	0.85	0.39
7	FST (red light) measurement of log10(cd.s/m2)	2	32	24	Bainbridge, Russell	C	MD 0.89	CI (−0.6, 1.84)	5.86	83%	1.84	0.07
8	FST (red light) measurement of log10(cd.s/m2)	2	32	24	Bainbridge, Russell	D	RR 1.89	CI (1.04, 3.41)	1.74	43%	2.10	* 0.04
9	FST (blue light) measurement of log10(cd.s/m2)	2	32	24	Bainbridge, Russell	C	MD 1.69	CI (1.21, 2.16)	0.32	0%	6.93	* 0.00001
10	FST (blue light) measurement of log10(cd.s/m2)	2	32	24	Bainbridge, Russell	D	RR 2.01	CI (1.32, 3.06)	0.63	0%	3.23	* 0.001
11	Central retinal thickness (CRT) (1 year)	2	27	27	Bainbridge, Jacobson	D	RR 1.15	CI (0.45, 3.00)	0.16	0%	0.30	0.77
12	Central retinal thickness (CRT) (3 year)	2	21	21	Bainbridge, Jacobson	D	RR 1.29	CI (0.33, 5.10)	0.23	0%	0.36	0.72

provided a summary of 12 meta-analyses reported for each of the outcomes, a PRISMA summary of a structured abstract in Appendix A.7, and a PRISMA checklist in Appendix A.8 (

Table A4. PRISMA List for LCA2 Studies.

Section/Topic	#	Checklist Item	Reported on Page #
TITLE
Title	1	Identify the report as a systematic review, meta-analysis, or both.	Page 1
ABSTRACT
Structured summary	2	Provide a structured summary including, as applicable: background; objectives; data sources; study eligibility criteria, participants, and interventions; study appraisal and synthesis methods; results; limitations; conclusions and implication ns of key findings; systematic review registration number.	Page 53, (Appendix A.7)
BACKGROUND
Rationale	3	Describe the rationale for the review in the context of what is already known.	Page 1–2 (Introduction)
Objectives	4	Provide an explicit statement of questions being addressed with reference to participants, interventions, comparisons, outcomes, and study design (PICOS).	Page 2 (Materials and Methods); page 4 (Results); Table S1 (Supplemental).
METHODS
Protocol and registration	5	Indicate if a review protocol exists, if and where it can be accessed (e.g., Web address), and, if available, provide registration information including registration number.	Informal proposal/protocol assessed and peer-reviewed at the University of Edinburgh.
Eligibility criteria	6	Specify study characteristics (e.g., PICOS, length of follow-up) and report characteristics (e.g., years considered, language, publication status) used as criteria for eligibility, giving rationale.	Table S1 (Supplemental); Appendix A.4.
Information sources	7	Describe all information sources (e.g., databases with dates of coverage, contact with study authors to identify additional studies) in the search and date last searched.	Table S1 (Supplemental); Appendix A.1. page 15.
Search	8	Present full electronic search strategy for at least one database, including any limits used, such that it could be repeated.	Appendix A.1. page 15; Appendix A.2, page 21.
Study selection	9	State the process for selecting studies (i.e., screening, eligibility, included in systematic review, and, if applicable, included in the meta-analysis).	Appendix A.2, page 21; Appendix A.3, page 31.
Data collection process	10	Describe method of data extraction from reports (e.g., piloted forms, independently, in duplicate) and any processes for obtaining and confirming data from investigators.	Page 4 (Materials and Methods).
Data items	11	List and define all variables for which data were sought (e.g., PICOS, funding sources) and any assumptions and simplifications made.	2 NRSI studies were supported by academic funding (Jacobson 2012, Le Meur 2018); 1 NRSI was supported by both academic and industry funding (Bainbridge 2015); 2 NRSI studies were supported by industry funding (Testa 2013, Weleber 2016); 1 RCT was supported by industry funding (Russell 2017).
Risk of bias in individual studies	12	Describe methods used for assessing risk of bias of individual studies (including specification of whether this was done at the study or outcome level), and how this information is to be used in any data synthesis.	Page 4 (Materials and Methods); Appendix A.5 and Appendix A.6.
Summary measures	13	State the principal summary measures (e.g., risk ratio, difference in means).	Page 4 (Materials and Methods).
Synthesis of results	14	Describe the methods of handling data and combining results of studies, if done, including measures of consistency (e.g., I2) for each meta-analysis.	Page 4 (Materials and Methods); Page 8–16.
Risk of bias across studies	15	Specify any assessment of risk of bias that may affect the cumulative evidence (e.g., publication bias, selective reporting within studies).	Page 4 (Materials and Methods); Appendix A.5 and Appendix A.6.
Additional analyses	16	Describe methods of additional analyses (e.g., sensitivity or subgroup analyses, meta-regression), if done, indicating which were pre-specified.	n/a
RESULTS
Study selection	17	Give numbers of studies screened, assessed for eligibility, and included in the review, with reasons for exclusions at each stage, ideally with a flow diagram.	Page 5, Figure 1 (Results)
Study characteristics	18	For each study, present characteristics for which data were extracted (e.g., study size, PICOS, follow-up period) and provide the citations.	Table S1 (Supplemental); Appendix A.2.
Risk of bias within studies	19	Present data on risk of bias of each study and, if available, any outcome level assessment (see item 12).	Appendix A.5 and Appendix A.6.
Results of individual studies	20	For all outcomes considered (benefits or harms), present, for each study: (a) simple summary data for each intervention group (b) effect estimates and confidence intervals, ideally with a forest plot.	Table 3, page 13.
Synthesis of results	21	Present results of each meta-analysis done, including confidence intervals and measures of consistency.	Table 3, page 12.
Risk of bias across studies	22	Present results of any assessment of risk of bias across studies (see Item 15).	Table 2, page 11
DISCUSSION
Summary of evidence	24	Summarize the main findings including the strength of evidence for each main outcome; consider their relevance to key groups (e.g., healthcare providers, users, and policy makers).	Page 15 (Conclusion); Appendix A.7 (Structured summary).
Limitations	25	Discuss limitations at study and outcome level (e.g., risk of bias), and at review-level (e.g., incomplete retrieval of identified research, reporting bias).	Table 2, Page 11; Appendix A.5 and Appendix A.6.
Conclusions	26	Provide a general interpretation of the results in the context of other evidence, and implications for future research.	Page 15 (Conclusion); Table S4a,b.
FUNDING
Funding	27	Describe sources of funding for the systematic review and other support (e.g., supply of data); role of funders for the systematic review.	None.

).

4. Discussion

Inherited retinal dystrophies (IRDs) are a leading cause of visual loss in children and adults of working age. Formerly untreatable, the emergence of gene augmentation therapy represents a real paradigm shift in patient care. We thus performed a systematic review and meta-analysis of interventional clinical trials to assess the efficacy of gene therapies for IRDs, thus delivering useful information for both clinicians and patients. The purpose of this systematic review is also based on a “fair test” [64], grounded in evidence-based medicine [65,66] (and Figure S1). To test such new therapies, it is critical to assess how transparent results show clear benefit for the patient. This requires that methodology, study design and outcome measures have to provide a clear and reasonable conclusion for the impact on the patient. A systematic review and meta-analysis of IRD patient outcomes for gene therapy is critical in order to support the field [67].

A search of peer-reviewed literature found that only gene therapies to treat Leber congenital amaurosis (LCA) met the criteria for addressing the original question (Appendix A.1). LCA is a rare disorder and gene therapy is an expensive treatment, which led to studies with small patient numbers. Further, the particularly severe phenotype of the disease, with low visual acuity from birth, led to difficulties in assessing the effect of treatment.

Of the 6 studies analysed, a significant drawback to the meta-analysis performed here is the variability in vector design and concentration of virus injected sub-retinally. All studies analysed used an AAV2 serotype, with most using an AAV2/2 capsid. However, one study used an AAV2/4 capsid. Further, some studies used a hybrid chicken β-actin promoter with a cytomegalovirus enhancer, whilst some used the human RPE65 promoter. Treatment doses ranged from 10⁸ to 10¹² vg, in volumes from 0.15 to 1.0 mL. This spans a number of logarithmic steps in each dose, potentially compromising the comparison of the results within the 6 studies. Despite all this, in our view, the similarities in products compared in the meta-analyses outweigh the differences. All contain the same recombinant human RPE65 gene, all are packaged in a similar AAV2 vector and all use a similar sub-retinal surgical procedure for delivery. All were used to treat the same trial population (RPE65-LCA2 patients). Further, there were similar criteria for controls and there was considerable overlap in trial duration and endpoints. Finally, we felt the comparison appropriate as pre-clinical work has shown good photoreceptor transduction and expression efficiency. As such, despite the analyses’ obvious limitations, we felt it appropriate in order to increase numbers of this rare disease and thus improve statistical power. Given the differences outlined, it is extremely encouraging to note the significant improvements in full-field stimulus (FST) testing that are seen following meta-analysis.

With gene augmentation in its infancy, it is perhaps unsurprising that there were variabilities in the biomarkers used to determine treatment efficacy. In total, 23 outcome assays were used. Visual acuity was the only outcome used in all six studies analysed. Five studies used Goldmann perimetry, four used ambulatory navigation/mobility and three used electroretinography. A further nineteen assays were used in two or less studies. Many of the assays were not comparable for several reasons. Five studies assessed visual fields using Goldmann perimetry, although different studies presented different isopters with variable follow up time. Further, a lack of quantitative data in some studies meant overall meta-analysis was not possible. Three studies used electroretinography as an assay, but two provided no data. Other assays used in two or more studies were unsuitable for meta-analysis due to a lack of quantitative data or irreconcilable differences in the way data was presented. As the field evolves, it is hoped agreed standards for methods and reporting will be established, allowing for easier meta-analyses of trials.

Visual acuity is the gold standard assay by which retinal disease treatments are assessed. Though our meta-analysis showed only modest improvement with gene therapy (in terms of clinical or statistical significance), the result is perhaps not surprising given the low-vision phenotype of LCA patients. Two studies (Bainbridge and Testa) did not provide raw visual acuity data and instead patient vision was determined from results presented in study graphs. This was undertaken by two independent researchers, with a mean of the two readings being used, but an element of uncertainty remains with the overall result due the unavailability of raw data within the actual papers. It should be noted that an I² value 65% indicates substantial statistical heterogeneity within the VA assays (Figure 3). As such, little weight can be placed upon the outcome of our VA analyses.

The study designs often dictated that the eye with the worse vision was treated, with control eyes having a better baseline vision. Although logMAR vision charts determine a linear improvement in vision with each letter or line gained, if treatment and control arms have different baseline values, bias is introduced and outcomes may be influenced as a result. Without adjustment, it may be unclear what impact arises from the treatment effect, as opposed to the treated eye being worse at baseline. Even adjusted data may not be robust enough to eliminate this confounding factor. Emerging gene therapy trials, where both eyes are treated and compared for one year to a deferred treatment group, should address this issue.

Four studies reported mobility testing as a key outcome. Mobility testing for the MLMT assay (Russell et al.) received criticism by an independent commentator [68] and reviewers in the FDA regarding uneven luminance levels [29]. In addition, we note that the MLMT assay results were indirect. A “passing level” of the assay compared baselines between 1-year timepoints however, the original data for measuring speed, time, accuracy (and further components) for assessment, were not included in the paper or the Biologicals License Application (BLA). Further, the MLMT assay used a logarithmic scale, based on light intensity (lux), which was then subsequently converted to an ordinal scale (ranging from −1 to 6), such that a two-point change in the ordinal scale may have a different interpretation depending on the baseline score (Table S4 (Supplemental)) [68].

Due to disparate methodologies (maze size and design, measurement, quantification and reporting), only analysis of dichotomous data (better/worse post treatment) was possible. This risks overestimating the benefit in certain studies. For example, Russell et al. reasoned that results in their maze required at least 2 levels of improvement on their assessment scale to accept the result as showing therapeutic benefit, whereas we defined even a 1 level gain post-operatively as performing ‘better’.

Some studies had datapoints missing, while quantitative data was missing from others, and required interpretation from results presented in study graphs (Jacobson et al.). Though RevMan analysis of dichotomous data suggested overall improvement in mobility, statistical significance was not reached. At present, there is no better test available for assessing the impact of gene therapy on visual function and so, as the field develops, it would be advantageous if some standardisation of the test could be agreed upon, recently supported by other literature [69,70,71].

The use of full-field stimulus (FST) testing (white, red and blue wavelength) is highly relevant because few research tools can quantify changes in visual perception if sight loss is as severe as it is in an RPE65/LCA2 population. Thus, the FST data carries extra significance. FST results presented in the studies was at times confusing. One study (Russell et al.) alternately presented white light FST results in log10(cd.s/m²) units and −log10(cd.s/m²) units, whilst not commenting on their red and blue light FST results (Russell et al.). A further study only described results in terms of “log10” units, which we interpreted as log10(cd.s/m²) units (Jacobson et al.), thus allowing for meta-analysis. Although the FDA, as part of the Biologics Licence Application (BLA) review for Luxturna [29], stated ‘the direct clinical benefit of FST is not clear’, it is apparent from meta-analysis of these two studies that retinal sensitivity improves with AAV-mediated gene augmentation therapy for RPE65-mediated LCA2. The significance of this cannot be underestimated. It is proof of principle that visual improvement is achievable with this technology and gives us hope that similar benefit could be achieved when other alleles are targeted.

The first attempt to use gene augmentation therapy for retinal disease has led to an FDA and EMA approved product (voretigene neparvovec-rzyl [Luxturna]). Improvements in surgical technique and improved knowledge of treatment technicalities (e.g., virus concentration) could mean subsequent iterations of these therapies show improved efficacy. Further, the more novel outcomes for mobility may drive innovative end-points. Though some concerns were raised by our Cochrane risk of bias analysis, further treatments targeting more common disease-causing genes [37] will mean increased patient numbers in trials and may allow for blinded evaluations, resulting in more robust studies.

As of April 2021, there are > 40 interventional gene therapy trials for IRDs reported at clinicaltrials.gov, from both academic and commercial institutions targeting several different IRD genes [72,73,74,75,76]. This meta-analysis highlights the need for consistency of trial design to allow comparison of gene products, but also shows the potential this technology has for addressing a leading cause of blindness in children and adults of working age.

5. Conclusions

The objective of this work was to conduct a systematic review of interventional clinical trial studies for IRDs and to assess and compare the effectiveness of available gene therapy treatments. Following the search, review and analysis of the relevant studies, the systematic review concluded that a meta-analysis for AAV-RPE65 gene therapy for LCA2 reported a modest improvement for visual acuity, mobility and full-field stimulus testing (FST), with FST improvements reaching statistical significance. In terms of a recommendation to support the IRD patient communities and researchers, we propose that full and open-access data is key. If the field is to be progressed and improved, then objective and transparent results need to be shared in order to improve outcomes, analysis, reporting and interpretation.