Why delivery matters to families affected by Retinitis Pigmentosa
Gene therapy has created real hope for inherited retinal diseases, including forms of Retinitis Pigmentosa (RP). Yet translating promising science into approved treatments has been difficult. Luxturna® (voretigene neparvovec-rzyl), a treatment for retinal dystrophy caused by variants in RPE65, was the first FDA-approved in vivo gene therapy and remains the only approved retinal gene therapy nine years later.
A 2026 review asks an important question: could the way a gene therapy is delivered into the eye influence whether a clinical trial succeeds? For people with RP and their families, this question matters because a treatment must reach the right retinal cells, cover the areas affected by disease, and do so with an acceptable safety profile. It must also be evaluated using study measures that can reliably detect meaningful change.
Two main ways to deliver retinal gene therapy
Most retinal gene therapy studies use adeno-associated virus (AAV) vectors. These are modified viruses designed to carry a working genetic instruction into cells. A search of ClinicalTrials.gov identified 80 interventional U.S. phase 1–3 studies using AAV vectors. Of these, 45 used subretinal injection (SRI) and 33 used intravitreal injection (IVT).
With subretinal injection, the treatment is placed in the space beneath the retina. This approach can efficiently deliver a gene therapy to photoreceptors—the light-sensing cells affected in many forms of RP—and to the retinal pigment epithelium (RPE), an important support layer for photoreceptors. However, subretinal delivery temporarily detaches part of the retina during surgery. The therapy is therefore largely limited to the area of retina that is detached and exposed to the vector.
Intravitreal injection places treatment into the gel-like vitreous cavity inside the eye. This route offers broader exposure across the retina and avoids surgically detaching the retina. However, intravitreal gene therapy has so far shown limited effectiveness for retinal targets, and higher doses may lead to inflammation.
These differences are not simply technical. They may determine which cells receive the therapy, how much of the diseased retina is treated, and what risks participants may face.
Safety and coverage must be considered together
The review highlights the trade-offs involved in subretinal treatment. Across peer-reviewed publications, 11.6% of eyes receiving subretinal gene therapy experienced serious adverse events. In some studies, all treated participants experienced treatment-related adverse events.
This does not make subretinal delivery unsuitable; it has important strengths, including efficient targeting of photoreceptors and RPE cells. But it emphasizes why the potential benefit of treating a specific retinal region must be weighed against procedure-related risks. This balance may be especially relevant in RP, where degeneration can extend beyond the central retina and progress over time.
Intravitreal delivery may ultimately offer an advantage if researchers can develop vectors capable of reaching the intended retinal cells effectively from the vitreous. The challenge is to expand retinal coverage without creating excessive inflammation or reducing the amount of treatment that reaches the cells most affected by disease.
RPGR-related X-linked RP illustrates the challenge
The review uses X-linked Retinitis Pigmentosa caused by variants in RPGR as a key example. RPGR-related XLRP is an important target for retinal gene therapy research. Of 10 trials identified for this condition, nine used subretinal injection.
These trials commonly used microperimetry and low-light visual acuity as efficacy endpoints. Microperimetry measures retinal sensitivity at specific points, while low-light visual acuity assesses vision under dimmer conditions. Both can be relevant to RP, but selecting the right outcome measure is critical. If an endpoint does not track the changes that are most likely to occur over the duration of a study, a trial may miss an important treatment effect—or may be unable to demonstrate benefit clearly.
Two completed phase 2/3 RPGR gene therapy studies did not meet their primary endpoints. Findings from the XOLARIS XLRP natural history study suggest that other measures may better reflect disease progression in this condition. These include visual field loss, which captures changes in the area a person can see, and central ellipsoid zone width, an imaging measure related to surviving photoreceptor structure in the central retina.
Designing trials around disease biology
The central message is that route of administration, therapeutic mechanism, disease distribution, and trial endpoints need to be planned together. A therapy designed to correct a defect in photoreceptors may require a different delivery strategy from one aimed at another retinal cell type. Likewise, a treatment that reaches only a limited retinal area should be evaluated with measures that can detect change in that area and over an appropriate time frame.
For future RP studies, the goal is not merely to choose between subretinal and intravitreal injection. It is to match the delivery route to the biology of the therapy and the pattern of retinal degeneration. Emerging vectors and delivery approaches that can treat a wider area of retina while reducing surgical injury may improve the prospects for successful trials.
Looking ahead
The active retinal gene therapy pipeline reflects major scientific momentum. At the same time, the gap between numerous clinical trials and a single approved retinal gene therapy shows how demanding development can be. For the RP community, this review offers a practical framework: progress will depend on reaching the right cells, safely treating enough of the retina, and measuring outcomes that reflect how the disease changes. Better alignment of these elements could help turn promising gene therapy programs into treatments that make a meaningful difference for people living with RP.