Adverse Events Halt Gene Therapy Trial for Diabetic Macular Edema

SAN ANTONIO -- Unexpected, severe, vision-threatening adverse events have quashed hopes for a safe and effective gene therapy for diabetic retinal disease.

Three patients with diabetic macular edema (DME) developed hypotony requiring surgical treatment, followed by severe, progressive decline in vision. Additionally, intraocular inflammation (IOI), sometimes severe, occurred in almost all patients treated with ADVM-022, which induces aflibercept protein expression, reported Charles Wykoff, MD, PhD, of the Blanton Eye Institute and Houston Methodist Hospital in Texas, at the American Society of Retina Specialists meeting.

"This is a very, very challenging situation," said Wykoff. "I think some minor inflammation that does not have anatomic or visual consequences is tolerable when you're dealing with the concept of durable treatment and durable production of a pharmaceutical agent inside the eye, but not a situation like this. This is completely unacceptable. If we could have known this data before initiating this trial, it would never have been initiated."

In contrast, encouraging results have come from a trial of the gene therapy in patients with neovascular age-related macular degeneration (AMD), and continued investigation of the therapy will focus on that population, he added. Reasons for the disparate results remain unclear but speculation includes the number and severity of comorbid conditions, particularly vascular disorders, and higher baseline levels of systemic inflammation in the diabetic patients.

ADVM-022 employs a novel biofactory approach to gene therapy, which induces continuous delivery of aflibercept after intravitreal injection. The therapy was designed with the goal of maintaining long-term control of retinal disease without the need for continual injections.

Wykoff reported findings from the randomized, phase II INFINITY trial involving 34 patients with DME. The trial was launched following positive findings from a phase I trial of the gene therapy in patients with AMD. Randomization assigned 25 patients to one of two doses of ADVM-022 (6 or 2 × 10¹¹ viral genomes) and nine to a control group treated with intravitreal injections of aflibercept (Eylea).

The primary endpoint was the time to worsening DME, defined as the need for supplemental aflibercept injections (triggered by a decline in visual acuity ≥5 ETDRS letters or an increase in central subfield thickness [CST] >50 μm). The trial was stopped after the reports of hypotony, all of which occurred in the high-dose group.

A 24-week assessment of the primary endpoint showed a significant prolongation of the time to DME worsening in patients treated with the higher dose of ADVM-022 (P=0.048) and a trend toward prolongation with the lower dose (P=0.210). Improvement in best corrected visual acuity (BCVA) occurred until week 24, followed by a drop off in the high-dose ADVM-022 group because of adverse events. CST improved to a similar degree in all three treatment arms.

Non-ocular adverse events (AEs) occurred in a similar proportion of patients in the three groups, said Wykoff. However, 20 of 25 patients treated with ADVM-022 developed anterior IOI (versus three in the control group) and three had posterior IOI events (none with aflibercept). Additionally, 15 ADVM-o22 patients had iris-related AEs as compared with none in the control group. Most AEs associated with ADVM-022 were mild or moderate in severity, but the higher dose of the gene therapy was associated with more intense IOI.

Session co-moderator Rishi Singh, MD, of the Cleveland Clinic, questioned whether ADVM-022 is a viable therapeutic option, given the risk of severe AEs. "It's incomprehensible. We're concerned when we see 1 or 2% rates of retinal vascular occlusive disease and here we're seeing really high rates of inflammation in the intravitreal studies. What's the future for this therapy?"

Wykoff said studies to date have provided a tremendous amount of data to help understand the risks to move forward with more assurances of safety.

"I think that looking for durable treatments that could be a 'one-and-done' for a substantial portion of patients is incredibly important to move our field forward," he said. "There is a tremendous amount to be learned, and we're trying to turn over every stone to understand why this dose-limiting toxicity developed, how we can prevent it, and how this and every other gene therapy program can move forward."

In contrast to the DME study, 2-year results of the phase I OPTIC trial of ADVM-022 in AMD showed that both doses of the gene therapy provided "robust aflibercept expression and sustained efficacy," said Dante Pieramici, MD, of California Retina Consultants in Bakersfield. The lower dose reduced annualized injection frequency by more than 80% in association with a low rate of ocular AEs. No clinically relevant intraocular pressure events occurred with either dose in patients with neovascular AMD.

The study involved a total of 30 patients in four cohorts. Supplemental intravitreal aflibercept was administered if a patient had ≥10-letter loss in BCVA, an increase in CST >75 μm, or AMD-related vision-threatening hemorrhage. Improvement in BCVA and CST was maintained over time. Pieramici said additional studies of ADVM-022 are warranted in patients with neovascular AMD.