In 1990, when a four-year-old child received the first successful gene therapy for a rare immunodeficiency, gene therapies were heralded as the future of pharma research and development (R&D). Their development has since been marred by safety concerns, and only in the past decade have there been positive late-stage trials and even FDA approvals. Heading into 2023, the gene therapy field still has a long way to go to fulfill its promised potential.

Experts say the biggest challenge to designing successful gene therapy trials is selecting the right primary endpoint. In some cases, overly ambitious outcome measures can limit gene therapies that help patients even if they are not functional cures. In other cases, sponsors must strike a delicate balance between objective biomarker endpoints and subjective patient assessments prone, which are to more variability.

As tensions over gene therapy endpoints come to the fore, the field as a whole has seen unprecedented recent growth. The number of gene therapy trial initiations as a percentage of total drug trials is on pace to reach a new peak in 2023, according to GlobalData’s Clinical Trials Database. GlobalData is the parent company of Clinical Trials Arena.

Looking ahead, the future of gene therapy trials could even entail developing new primary endpoints altogether. “It’s up to pharmaceutical investigator groups, the FDA, and the EMA to come together to come up with new endpoint measures,” says Dr. Byron Lam, a gene therapy investigator and neuro-ophthalmologist at the University of Miami. “We need to determine a new way forward.”

Managing high expectations

Though gene therapies can provide functional cures for certain diseases, many are too heterogeneous for a one-size-fits-all approach to be feasible, Lam explains. In many cases, choosing the right endpoint comes down to effectively managing expectations.

The tension over endpoints for gene therapies for inherited retinal diseases (IRDs) highlights this central challenge. In 2017, Roche’s Luxturna (voretigene neparvovec) became the first and only FDA-approved gene therapy for treating an IRD. Lam says Luxturna causes “remarkable vision improvement,’ but only for a unique type of IRD called Leber congenital amaurosis (LCA). LCA is easier to rescue than most other IRDs because patients are very young with most of their retinal structures still intact, he explains.

Meanwhile, other IRDs entail older patients and more diverse underlying causes, warranting different expectations and trial endpoints. Luxturna’s Phase III primary endpoint was a mobility test with a large range, which Lam says is less sensitive for mild-to-moderate IRDs.

Other IRD endpoint possibilities include the well-known best corrected visual acuity (BCVA), where patients read from a set of increasingly smaller letters across the room. However, the EMA and FDA have different thresholds for clinically meaningful benefit, and BCVA is less accurate for IRDs that maintain central vision for longer periods, Lam notes. Ultimately, the onus is on investigator groups to develop new endpoints for tougher-to-treat IRDs that are achievable but still confer clinically meaningful benefit, he notes.

Defining a successful gene therapy

Even for indications where there is an agreed-upon primary endpoint, the metrics of success are often under debate. In the rare blood disorder hemophilia A, experts agree that increasing factor VIII activity is an important primary endpoint, but they disagree on by how much, explains Radek Kaczmarek, PhD, who has researched gene therapies at Indiana University. Some clinicians want complete normalization, which would entail 50% to 150% factor VIII activity, while other experts say improving patients to moderate disease, or 5% to 40% activity, is still a success.

Kaczmarek says there is no way to tell if a patient with hemophilia A receiving gene therapy will have a complete response, a partial response, or no response at all. As a result, endpoints should manage expectations and acknowledge that partial responses generate some clinically meaningful improvement, he notes.

To complicate matters, the most common gene therapy delivery approach—adeno-associated virus (AAV) vectors—can only be administered once. Most patients develop antibodies against AAV vectors after a single gene therapy that preclude them from future gene therapies with the same approach, Kaczmarek notes.

Still, given the known success of many hemophilia patients who received gene therapies, the hemophilia community often expects an absolute cure for all patients, Kaczmarek says. When designing gene therapy clinical trials, he says communicating reasonable expectations is key.“Despite their many shortcomings, gene therapies are clearly the best to date in terms of providing patients with long-term hemostatic correction,” he notes.

Choosing the right patients

Given the inherent risk of gene therapies—including overreactive immune responses and even mortality—many are reserved for patients with late-stage disease. However, patients with the most advanced disease are often the least likely to benefit from gene therapies, explains Dr. Seppo Ylä-Herttuala of the University of Eastern Finland. The result is a tough ethical dilemma for trial sponsors to navigate.

Gene therapy trials for coronary heart disease, where Ylä-Herttuala works, mostly target patients with severe disease, corresponding to New York Heart Association (NYHA) Class 4. These patients are usually older than 70, so their bodies cannot regenerate anymore due to normal aging processes, Ylä-Herttuala explains. Meanwhile, younger patients with less severe heart disease, including NYHA classes II and III, would likely be more responsive to gene therapy interventions, he notes.

Overall, gene therapy trials are generally much smaller than trials for other therapies in the same treatment area. This can also make it harder for gene therapies to demonstrate statistical significance along their primary endpoints with a relatively small sample size.

Phase III gene therapy trials for cardiovascular diseases enrolled an average of 139 patients over the past decade, compared to 732 patients for Phase III cardiovascular trials in general. Gene therapy studies also had more difficulty enrolling as many subjects as planned, according to GlobalData’s database.

To navigate through smaller sample sizes and potentially suboptimal patient selection in earlier studies, Ylä-Herttuala says sponsors should use a biomarker-based approach. For Phase II trials, he recommends surrogate endpoints that can generate data within six months. For coronary heart disease, this could be absolute blood flow in the myocardium.

However, for subsequent Phase III trials, sponsors will still need to use “hard” endpoints to gain approval, Ylä-Herttuala notes. In heart disease, this normally entails major adverse cardiac events (MACE) or mortality.

Patient-reported outcomes for gene therapies

Patients, clinicians, and regulators have recently been pushing for more patient-reported outcomes (PROs) across clinical trials. However, for many smaller gene therapy studies, PROs pose an added risk due to their variability, explains Dr. Peter Marinkovich, a dermatologist at Stanford University who has researched gene therapies for epidermolysis bullosa.

Larger clinical trials can overcome the variability of PROs through larger sample sizes, Marinkovich explains. However, smaller gene therapies should be cognizant of the added risk of PROs when designing a primary endpoint plan, he adds.

For Phase III trials in epidermolysis bullosa, regulators give sponsors two primary endpoint options: 50% wound clearance after six months with PROs of pain, or 100% wound clearance after six months without PROs. “It’s a matter of how confident you are in wound healing,” Marinkovich says.

Looking into the future, experts agree biomarkers could help better determine which patients will respond to different treatments and endpoints. This could allow more sponsors to pursue narrow enrollment and choose objective endpoints that are reproducible in smaller trials.

Nevertheless, experts say biomarker development, like gene therapies more broadly, has yet to reach its potential. “What biomarkers predict how well a patient will respond to gene therapy?” asks Ylä-Herttuala. “That’s the million-dollar question.”


  • When designing gene therapy trials, sponsors should balance high expectations with achievable endpoints that identify patient improvement even without a full cure.
  • Trial sponsors should consider enrolling patients earlier in their disease progression, though this can prove challenging given the inherent risk of gene therapies.
  • Although PROs are gaining traction as primary endpoints in drug development, gene therapy trial sponsors should note their added risk for smaller studies.

Editorial content is independently produced and follows the highest standards of journalistic integrity. Topic sponsors are not involved in the creation of editorial content.