Why Treatment Intervals Diverge Across Retinal Anti-VEGF Indications

Key Takeaways:

• Extended intervals are earned by documented response, not assumed from the label. In one 96-week trial, roughly half of patients on the 16-week arm qualified for intervals of 20 weeks or longer, but only after a year of therapy.
• Macular edema following retinal vein occlusion extends less readily than neovascular (wet) age-related macular degeneration (AMD), and branch occlusions extend better than central ones. Set the cadence by occlusion type.
• In retinopathy of prematurity the surveillance clock runs longer than the treatment clock. Anti-vascular endothelial growth factor (anti-VEGF) treatment moves the earliest permissible stopping point, and full vascularization remains the only reliable one.

One drug class spans retinal indications from the neonatal intensive care unit to the geriatric clinic, and the interval question has no single answer. Labeled maintenance dosing across these agents ranges from every 4 weeks to every 20 weeks, with the longest interval available only in neovascular AMD and diabetic macular edema (DME), only after a year of successful response.¹ Durability is a property of the disease, not the molecule.

Durability earned, not assumed

In neovascular AMD, the 96-week PULSAR trial randomized 1,009 treated patients, after 3 initial monthly doses, to aflibercept 8 mg at 12- or 16-week intervals or aflibercept 2 mg every 8 weeks. Among week-96 completers on the 16-week arm, 78 percent had last assigned intervals of at least 16 weeks, 53 percent qualified for at least 20 weeks, and 31 percent qualified for 24 weeks, with a mean of 8.2 injections against 12.8 in the 2 mg arm. Those are intervals patients qualified for; because the study ran 96 weeks, the longest anyone could complete was 20 weeks. Gains from baseline of roughly 5.5 letters were maintained, meeting noninferiority against a 4-letter margin.²

Labeling stops short of the trial's longest assigned interval: the 20-week option applies in neovascular AMD and DME only, while diabetic retinopathy is labeled every 8 to 12 weeks and retinal vein occlusion every 8 weeks.¹ Guidance declines to name a preferred regimen, holding that no consensus exists on ideal intervals.³

Vein occlusion extends differently

In the 72-week BALATON and COMINO trials, all patients received faricimab in a modified treat-and-extend period, with results reported by prior randomization. At week 68, the proportion on intervals of 12 weeks or longer was 64.¹ percent of the prior faricimab stratum and 56.9 percent of the prior aflibercept stratum in branch occlusion, against 45.5 and 50.1 percent for those same 2 strata in central and hemiretinal occlusion.⁴ Those intervals outrun the label, which specifies monthly dosing with no extension schedule; an April 2026 revision removed the previous 6-month duration limit but did not add interval extension.⁵

Longer horizons agree. Roughly half of patients with branch occlusion and 56 to 75 percent with central occlusion still require therapy beyond 5 years.⁶ In a series followed a mean of 54 months, the maximal recurrence-free interval widened only from 6.4 to 8.5 weeks, and 27 patients left the cohort through loss to follow-up or death.⁷

The prematurity surveillance clock

Here the question inverts. Screening may end at 45 weeks postmenstrual age absent type 1 disease, but where anti-VEGF caused regression the floor moves to at least 65 weeks. The guidance then concedes its own limit: neither initial regression nor the 45-week threshold can be relied on, and full retinal vascularization "is the only criterion listed above that can be relied on as a valid conclusion point" – yet it is not always achieved with these agents alone, leaving prolonged observation and clinical judgment in its place.⁸

That incomplete vascularization is persistent avascular retina, named in the third edition of the International Classification of Retinopathy of Prematurity and documented by location and extent.⁹ No formal management guidelines exist.¹⁰ Recurrence risk is zone-dependent: in a meta-analysis whose overall pooled comparison was not significant, recurrence after anti-VEGF was lower than laser in zone I and higher in zone II, with events from 7 to more than 50 weeks.¹¹

Where extension becomes undertreatment

Registry data show what happens when intervals lengthen without a plan. In neovascular AMD, mean acuity rose 3.0 letters at year 1 and fell to a net loss of 4.6 letters by year 6, and 38.8 percent discontinued, though only 6,878 of the 160,423 eyes with acuity data contributed at year 6.12 Loss to follow-up affected roughly 1 of 9 patients.³

Monitoring does not relax as intervals widen. Increases in intraocular pressure have been seen within 60 minutes of injection, and labeling instructs that pressure and optic nerve head perfusion be monitored. Intraocular inflammation, including retinal vasculitis, appears in class warnings.^1,5 Continuous port delivery of ranibizumab carries a boxed warning for an endophthalmitis rate up to 3 times that of monthly injections.¹³ A documented response earns the interval. The visit is what confirms it still holds.