In the real world, many nAMD patients become 'nonpersistent' with anti-VEGF injections

The treatment of retinal diseases has been dramatically altered by the anti-vascular endothelial growth factor (anti-VEGF) medications such as Eylea (aflibercept) and bevacizumab. But to stay effective, the medications need to be injected as often as every other month after initial monthly loading doses. It has been an open question whether this “treatment burden” of frequent injections would mean that the medications are notably less effective in actual practice than they have shown to be in the more controlled circumstance of randomized clinical trials designed to assess efficacy and safety.

Real-world evidence results presented this morning at the 2026 annual meeting of the American Society of Retina Specialists in Montreal begin to fill in the answer to that question. For the purposes of this study, nonpersistence was defined as a gap of no injection for more than 180 days after the previous injection, and treated eyes, not patients, were counted.

The results presented by Andrew Moshfeghi, M.D., MBA, an associate professor at the USC Gayle and Edward Roski Eye Institute and director of the clinical trials unit for the Department of Ophthalmology at Keck School of Medicine, showed that cumulative incidence of nonpersistence was 38.9% at the end of the first year of scheduled treatment, 52.6% at the second year, 72.9% in the fifth year and 80.1% in the seventh year. Moshfeghi also shared data showing that the mean number of injections fell more than 8.4 the first year to 6.6 the second and then gradually declined to 6.2 in the seventh, and the intervals between injections increased from 59 days the first year to 69 in the seventh. Not unexpectedly, the nonpersistence was associated with a steeper decline in visual acuity in the data that Moshfeghi presented. The visual acuity of the eyes of the patients who were adherent to the scheduled injections improved initially and then gradually decreased to a best-corrected visual acuity (BCVA) — best vision with corrective lenses — that was below their baseline level (-2.1 early treatment of diabetic retinopathy study [ETDRS] letters by year 7). The pattern was the same for the nonpersistent, an initial improvement and then a decline below baseline, but the initial improvement was smaller and the decline below baseline steeper (-8.0 ETDRS letters by year 7).

An analysis of the factors associated with nonpersistence showed that poor visual acuity before treatment, atrophy and delays between treatment and the first injection were associated with a greater likelihood of nonpersistence. Those with the poorest visual acuity to begin (fewer than 35 ETDRS letters) were nearly twice as likely to be nonpersistent than those with the best (70 or greater numbers), according to the data that Moshfeghi presented. Interestingly, the effect of the injections, as measured by a change in visual acuity at follow-up from the previous injection, did not appear to push people toward nonpersistence.

During the question-and-answer period after his and other real-world evidence presentations, Moshfeghi said the practical problems of getting to the facility to get the injections are among the reasons for the “high rates of attrition.”

“A lot of the patients can’t drive, they don’t have good vision, they have to have a medical assistant or a family member come with them — person must take time off from work,” he said.

More durable treatments that don’t have to be administered as often would help with the nonpersistence problem.

“If we can do anything to remove those barriers, like have a longer lasting, more durable treatment that we don’t have to worry, ‘Oh, can the patient come back in 10 weeks versus 12 weeks versus 14 weeks?’ If they can go six months, nine months, a year [between treatments] then we’ll have a lot more faith in the drug still having the intended effect — [and] less onus on the patient to come back so frequently,” Moshfeghi said

Moshfeghi’s data came from an analysis of de-identified electronic health records in the Vestrum Health database. His research focused on treatment-naïve patients with neovascular age-related macular degeneration who started anti-VEGF treatment between 2015 and 2024. With those limitations, the analysis included just over 200,00 eyes from nearly 160,000 patients.