The pediatric blind spot: What ocular gene therapy trials still aren’t telling clinicians about children

Gene therapy has moved from proof of concept to clinical reality in inherited retinal diseases (IRDs), with voretigene neparvovec (Luxturna) approved for RPE65-associated disease and a growing pipeline spanning Leber congenital amaurosis (LCA), Leber hereditary optic neuropathy (LHON), achromatopsia, and the retinitis pigmentosa (RP) genes. Yet a structural gap sits underneath the momentum: most IRDs declare themselves in childhood, while the trial evidence is drawn overwhelmingly from mixed or adult-weighted cohorts.

In a review of 16 trials¹, the authors below found that most studies did not stratify outcomes by age, few reported pediatric adverse events separately, and some did not report how many children were enrolled at all—leaving clinicians to counsel families and weigh procedural risk by extrapolating from adult data onto a physiologically and developmentally different patient. That extrapolation is especially consequential in young children, for whom inflammation, cataracts, or prolonged postoperative occlusion can induce amblyopia during the critical period of visual development—an independent, disease-agnostic form of vision loss that adult trials cannot detect by design.

In this Q&A, the review’s authors—Athanasia Sandali, MD, MSc; Anna Nikolaidou, MD, MSc, PhD(c); Theodora Gianni, MD, MSc, FEBO; Zacharenia Tsoukala, MD; Ioannis Tsinopoulos, MD, PhD; and Lampros Lamprogiannis, MD, MSc, PhD, FEBO, FEBO-SP—discuss what the current evidence base does and does not support for the pediatric eye, how vector choice and delivery route shape the safety profile, why sham-eye “controls” in LHON trials are biologically unreliable, and the single reporting change they believe would most improve pediatric evidence as the field moves toward CRISPR-based editing and next-generation capsids.

The Q&A

Most inherited retinal diseases declare themselves in childhood, yet the bulk of the clinical trial evidence comes from mixed or adult-weighted cohorts. Practically speaking, how much of what is “known” about ocular gene therapy applies to the pediatric eye? And where are clinicians extrapolating on thin evidence?

This is, in fact, the most important question we would like our readers to reflect upon. Keep in mind that our review includes trials with any pediatric participants, not only trials conducted exclusively in pediatric populations. Most of the studies we reviewed did not stratify their results by age, and only a minority reported adverse events specifically for children. In contrast, other trials may not even report how many pediatric participants were enrolled—sometimes only giving a mean age with a standard deviation.

The gap we would emphasize most is the interaction between treatment-related ocular complications and visual development, as this is something adult trials cannot address by design: postoperative complications carry the risk of inducing amblyopia during a critical developmental window. In this context, adult tolerability should not be cited as reassurance. Furthermore, small sample sizes, and the resulting inclusion of only one or two children in some studies, mean their results may be more of a case-level observation than cohort data.

Overall, long-term durability, optimal timing of intervention, and outcomes tied to visual development remain areas of real extrapolation—which is precisely why we recommend stratified enrollment and separate pediatric and adult cohorts in future trial design.

Your review pulled 16 trials. What surprised you most about the current pediatric evidence base once you had it all in one table—in terms of maturity, geographic concentration, or the diseases that are and aren’t being addressed?

At the time our review was conducted, the field was still overwhelmingly early-phase, with only five trials having progressed to Phase 3—three of which were the GS010 trials for LHON. Even though LHON is a relatively rare condition, our review found it at the forefront of research for pediatric ocular gene therapy.

Another unexpected finding was that most of the supporting evidence base for Luxturna remained distributed across smaller early-phase studies, despite its FDA approval, which would theoretically warrant larger and more stratified studies.

The geographic concentration was not particularly surprising, as the ClinicalTrials.gov database is predominantly US-based, but it does have real implications for generalizability, given that genetic and phenotypic diversity in IRDs is substantial.

Regarding disease coverage, we found that LCA and LHON together accounted for 10 of the 16 trials, which means some of the more common inherited diseases—such as Stargardt disease or RP—did not have published results at that point in time.

Vector choice and delivery route remain central design decisions. For the subspecialist weighing subretinal versus intravitreal delivery, or AAV serotype selection, what did the pediatric data suggest about the efficacy-versus-safety trade-offs of each approach?

We would refrain from drawing a clean efficacy-versus-safety verdict, since both vector choice and delivery route are influenced by the disease target; for example, subretinal delivery was used predominantly for RPE- and photoreceptor-targeted conditions, while intravitreal delivery was used for conditions where the target cell layer is theoretically more accessible from the vitreous cavity.

Overall, subretinal complications skewed toward surgical and retinal sequelae (retinal tears, subretinal fluid, and maculopathy), while intravitreal complications were most prominently inflammatory or involved the anterior segment. In both cases, pediatric-specific data are limited; we would like to highlight the CNGB3 achromatopsia trial (subretinal delivery), which flagged lenticular opacities in 9 of 12 children. This could pose a risk of amblyopia, as previously discussed, though it is a single-trial finding and should be generalized with caution.

There is no specific evidence on pediatric safety by serotype, but we would emphasize that most adverse events in either delivery route can be attributed to the procedure itself. In any case, it is evident that age-specific procedural risks, and adverse events in general, should be documented far more rigorously than they currently are.

In the LHON trials, treated and sham eyes both showed change, and some studies documented bilateral improvement attributed to interocular vector migration. How should that phenomenon shape how clinicians read efficacy claims—and how controls are designed—in this disease?

As mentioned in our article, at first glance ocular gene therapy is attractive because one eye could theoretically receive the treatment while the contralateral eye serves as a control. The FDA has already advised against this, and through the LHON trials we can discern why: across the GS010/ND4 trials, the sham-treated eye consistently tracked close to—and in some cases even outperformed—the treated eye.

As described in our introduction, unilateral intravitreal AAV injections can produce measurable effects in the fellow eye, likely via vector or transgene product trafficking through astrocyte networks at the optic chiasm. This means the eye that received the “sham” treatment is not a clean biological control but rather contaminated and therefore untrustworthy. Clearly, future gene therapy trials—both in LHON and in other conditions,—need externally controlled designs with natural history cohorts, as it is highly likely that the treatment is not strictly compartmentalized to one eye.

Across the RP trials—MERTK, USH2A, and X-linked RPGR—results were notably mixed, and pediatric participants were often too few to analyze separately. Where do you see the clearest near-term clinical promise?

In the RPGR trial, AAV5-RPGR was found to be safe and well tolerated; most adverse events were procedural and resolved without intervention, and the treatment showed improved retinal sensitivity and functional vision on a vision-guided mobility assessment. Its cohort was also meaningfully larger than those of the other two RP studies, and it was one of the few RP trials in our review with a positive functional outcome.

By contrast, the MERTK and USH2A studies give more cautionary signals: they included fewer patients and showed shorter-term improvements in vision, and, in the case of the USH2A trial, ended in early termination for sponsor reasons unrelated to safety.

The promise shown by the RPGR trial is also reflected in our Table 2 of unpublished/ongoing trials, which indicates a substantial cluster of RPGR-targeted programs, including multiple Phase 2 and Phase 3 studies. The RPGR trials are therefore overall more mature and have a genuinely larger evidence base.

In achromatopsia, higher doses drove functional and quality-of-life gains, but ocular inflammation was the dominant pediatric adverse event. How should clinicians think about dose selection and inflammation risk when the target patients are young children?

The responsible framing is that dedicated, pediatric-specific dose-finding studies need to happen before a specific dose is adopted for children with achromatopsia (or with other diseases, as this applies to all gene therapy treatments discussed in the review).

Extrapolating an adult-optimized dose downward, or assuming that the same dose that shows functional benefit also carries an acceptable inflammation risk for children, is not good practice. In fact, even within the pediatric spectrum, the “perfect” dose may not exist; for example, a 2-year-old and a 16-year-old are immunologically and developmentally different patients, and it is plausible that their tolerance for a given dose, or their risk of dose-related inflammation, is not the same. Thus, stratified enrollment and age-based dosing cohorts are once again imperative in order to draw firm conclusions.

In the meantime, clinicians should probably treat dose selection in young pediatric patients as an open question rather than a settled one and weigh inflammation risk carefully against the functional gains reported in largely adult-driven data.

Your adverse-event analysis separates intravitreal from subretinal profiles and argues that much of the risk is procedure-related rather than construct-related. What does that distinction mean for surgical decision-making and for how clinicians consent pediatric patients?

It is true that the majority of reported adverse events appear to be procedure-related. This suggests that outcomes may actually reflect surgical variables (for example, whether a vitrectomy was performed or whether the lens was manipulated), for which very few details are reported in the included studies. Consequently, it is hard to draw conclusions about how much of a given adverse event profile is attributable to the surgical approach versus the biologic agent.

Future trials should clearly report surgical technique, and clinicians themselves should weigh the surgical risk of subretinal delivery on top of the immunologic risk the vector itself carries. Similarly, the consent conversation should try to separate the intraocular surgery risks from the inherent risks of a specific treatment while also accounting for pediatric-specific stakes such as the risk of amblyopia, which need to be named explicitly rather than folded into a generic surgical-risk disclosure.

You found that very few trials reported safety outcomes specifically for children. What are the real-world consequences of that reporting gap for clinicians trying to assess pediatric risk, and what should sponsors be required to report?

When safety data are not stratified by age, a clinician trying to counsel a family is left extrapolating from adult-dominated safety profiles onto a physiologically different patient. This means making a risk-benefit judgment for a child using pooled statistics that could be masking the risk profile in either direction—better or worse.

Furthermore, as previously mentioned, children face amblyogenic risk that adults do not; thus, an adverse event with negligible functional consequence in an adult (transient inflammation, mild cataract) can have a severe impact in a young child if it is not caught and managed promptly. Without age-stratified reporting, that risk is essentially invisible in the published record. Sponsors should be required to report ages at the individual level, or at minimum as defined subgroups, as well as to stratify adverse events by pediatric versus adult status and to provide details regarding the procedure and the dosing rationale.

Learning effects, test-retest familiarity, and an immature visual system all threaten the validity of functional endpoints in children. In your view, which outcome measures hold up in pediatric cohorts, and which should clinicians interpret with real caution?

Objective outcome measures seem to be the more trustworthy ones. A good example is objective structural imaging (such as OCT-based measures like ganglion cell layer macular volume and nerve fiber layer thickness), which is independent of the child’s cooperation, understanding, and repeated practice. Electrodiagnostic testing (e.g., ffERG, VEPs), performed in accordance with pediatric guidelines and standards, can also provide valuable objective information unrelated to the child’s age and perception.

In contrast, measures such as visual field testing and BCVA, although rooted in clinical practice, are subjective, attention-dependent, and produce a familiarization effect, while patient-reported quality-of-life measures are susceptible to non-visual confounders such as general well-being. As a general rule, objective measurements are always more trustworthy for gauging outcomes than subjective assessments.

You highlight the risk that inflammation, cataract, or prolonged postoperative occlusion could induce amblyopia in children under 10, independent of the primary disease. How should the critical period of visual development reframe the risk-benefit calculus around early intervention?

This point ties together a few things we have already touched on throughout this interview: the amblyopia risk in children under 10, the consent implications of separating surgical from construct-related risk, and the caution around interpreting BCVA in young children all converge on the same underlying idea.

In a young child, any complication carries the added amblyopia stake that simply cannot exist in an adult and thus constitutes an independent form of vision loss on top of the underlying disease. The risk-benefit calculus must therefore be reframed to include the intervention’s potential to trigger a compounding, disease-independent injury.

Although early intervention does carry real appeal (for similar reasons, as it also affects the critical developmental period), it has to be weighed against the vulnerability of the younger visual system. In this aspect, too, age-stratified trial design is essential to capture both this risk and the treatment’s efficacy.

How do you weigh the promise of preventing irreversible photoreceptor loss against exposing an asymptomatic child to procedural and immunologic risk?

In truth, there is no single right answer to this question. On one hand, as noted in our review, gene therapy effectiveness appears to be timeframe- and VA-related, with patients who have better baseline VA showing more significant improvement. Given the progressive nature of these diseases, intervening before photoreceptor degeneration sets in could be impactful for a child’s lifespan.

On the other hand, immunologic and procedural risks are measurable, real costs that a symptomatic child with active, progressive vision loss might reasonably accept, but that a child who is not yet experiencing functional impairment might not.

All in all, this is one of the genuine open ethical questions in the field, and it warrants separate consideration in trial design and consent between symptomatic and non-symptomatic patients.

Your paper calls for age-stratified enrollment, dedicated pediatric cohorts, and adaptive designs. If you were advising the next wave of trials—and looking ahead to CRISPR-based editing and capsid engineering—what single change would most improve the quality of pediatric evidence?

If we had to pick a single change, it would be mandating individual-level, age-stratified outcome and safety reporting as a condition of trial design, built into the protocol from the start. Pediatric participants should be analyzed as a defined, adequately powered cohort rather than an incidental subgroup of an adult-driven trial. This point has compounding value as the field moves toward CRISPR-based editing and next-generation capsids, which are being framed as offering more precise, more durable corrections with potentially reduced immunogenicity.

In fact, since pediatric immune systems and ocular anatomy differ from those of adults, and a one-time edit carries permanence, such technologies would constitute irreversible interventions with even less age-specific safety visibility. Consequently, age-stratified reporting is imperative even—and especially—at this stage, while these earlier-generation vector and antisense trials are still maturing, so that the next generation of gene-editing trials can be designed around real pediatric evidence rather than adult extrapolation.

Is there anything else you would like to add or make sure is discussed?

Thank you for the opportunity to discuss this work. We hope it helps clinicians and families navigate an area of ophthalmology that is evolving quickly, and we look forward to seeing how the field addresses the pediatric evidence gaps we have raised. We dedicate this work to the future children who will benefit from what we have learned so far.

The authors

Athanasia Sandali, MD, MSc

Dr. Athanasia Sandali is a physician and an aspiring ophthalmologist. She earned her medical degree and an MSc in ocular surgery from Aristotle University of Thessaloniki, graduating with distinction. While working towards obtaining her specialty in ophthalmology, she has been actively engaged in various research projects focused on pediatric ophthalmology, including AI-driven diagnostics, gene therapy, and strabismus. Her works have been published in distinguished journals such as Acta Ophthalmologica and the European Journal of Ophthalmology and presented at prestigious congresses such as the Association for Research in Vision and Ophthalmology (ARVO) meeting and the European Association for Vision and Eye Research (EVER) Congress.

Anna Nikolaidou, MD, MSc(Res), MSc, PhD(c)

Dr. Anna Nikolaidou is a physician and PhD candidate in Experimental Medicine at the Institute for Ophthalmic Research, University of Tübingen, where her work focuses on myopia, electrophysiology, and pediatric neuro-ophthalmology. She holds an MD along with two MSc degrees, in medical research methodology and in nanomedicine. Her research spans pediatric ophthalmology, ranging from AI-based diagnostic tools to the ocular effects of autoimmune disease in children, with findings published in prestigious journals including Eye and Pharmaceuticals. She received multiple grants to present at meetings such as the International Society for Clinical Electrophysiology of Vision (ISCEV, 2026) meeting and the Association for Research in Vision and Ophthalmology (ARVO, 2026) meetings.

Theodora Gianni, MD, MSc, FEBO

Dr. Theodora Gianni is a physician and assistant consultant at Agia Sofia Children's Hospital in Athens. She earned her medical degree from the University of Ioannina and holds an MSc from Aristotle University of Thessaloniki, along with the Fellowship of the European Board of Ophthalmology (FEBO). Her research centers on pediatric ophthalmology, including refractive error prevalence in children, ocular manifestations of juvenile systemic lupus erythematosus, and nanomedicine approaches to age-related macular degeneration, with findings published in journals including Eye, Acta Ophthalmologica, and Pharmaceuticals.

Zacharenia Tsoukala, MD, MSc(c)

Dr. Zacharenia Tsoukala is a physician training in pediatrics, with a particular interest in pediatric endocrinology. She earned her medical degree from Aristotle University of Thessaloniki, where she is also pursuing an MSc in Adolescent Medicine and Adolescent Health Care. Her research has addressed diabetes in adolescence as well as ocular gene therapy in pediatric patients, with work presented at conferences including the Panhellenic Congress of Pediatric Endocrinology.

Prof. Ioannis Tsinopoulos, MD, PhD

Dr. Ioannis Tsinopoulos is a professor of ophthalmology at Aristotle University of Thessaloniki, where he practices at the 2nd Department of Ophthalmology, Papageorgiou Hospital, and heads the university's Laboratory of Experimental Ophthalmology. He earned his medical degree from Aristotle University of Thessaloniki and completed his ophthalmology residency and further training in Germany, including an MD thesis at Munich University on phacoemulsification cataract surgery. His research spans cataract and intraocular lens surgery, nanotechnology applications in ophthalmology, and ocular cell biology, with roughly 90 papers in international peer-reviewed journals and a strong presence at both European and international congresses.

Assistant Prof. Lampros Lamprogiannis, MD, PhD, FEBO, FEBO-SP

Dr. Lampros Lamprogiannis is a pediatric ophthalmologist and scientific head of the Pediatric Ophthalmology department at the "Ophthalmica" Institute of Ophthalmology and Microsurgery in Thessaloniki. He was recently elected assistant professor at Aristotle University of Thessaloniki. He trained at Cambridge, King's College Hospital, and Great Ormond Street Hospital for Children, and in 2023 became the first Greek ophthalmologist to hold the European sub-specialty certification in Pediatric Ophthalmology and Strabismus (FEBO-SP). His research spans pediatric ophthalmology, amblyopia and strabismus, and nanotechnology applications such as drug-eluting intraocular lenses, and he has been a frequent presenter at national and European pediatric ophthalmology meetings.

REFERENCE:

1. Sandali A, Nikolaidou A, Gianni T, Tsoukala Z, Tsinopoulos I, Lamprogiannis L. Gene therapy clinical trials for inherited eye diseases: The pediatric perspective. European Journal of Ophthalmology. 2026;0(0). doi:10.1177/11206721261478166