Findings from a recent study published in Eye tracked the structural evolution and visual decline in poor-responder wet age-related macular degeneration (AMD) patients over time, with the results calling for a shift in how clinicians monitor this population.
Give me some background first..
While anti-vascular endothelial growth factor (VEGF) therapy has changed the game for wet AMD, an estimated 20 to 30% of patients continue to lose vision despite consistent treatment.
Case in point: About one in four wet AMD patients don’t respond well to VEGF injections.
For those poor responders, the standard monitoring metric, central retinal thickness (CRT), may be missing what’s really going on.
Expand on this.
CRT has been the workhorse of wet AMD monitoring for years. It’s objective, easy to pull from optical coherence tomography (OCT), and tracks well with visual outcomes in the big early trials.
- In practice: A bump in CRT usually means the disease is flaring up and it’s time to retreat.
But here’s the issue: Some patients hold steady on CRT, or even see it drop, and still lose vision. That mismatch has pushed researchers to pay closer attention to qualitative structural changes on OCT, including macular atrophy, subretinal fibrosis, and outer retinal integrity markers like the external limiting membrane (ELM) and ellipsoid zone (EZ).
Noted. Now, talk about the study.
A team at the University of Bari in Italy ran a retrospective longitudinal study to characterize what happens structurally in eyes that keep deteriorating despite ongoing anti-VEGF therapy.
The question: Which OCT parameters best predict visual decline at different stages of the disease in this population?
Who was included in the study?
Investigators started with 250 treatment-naive neovascular AMD (nAMD) patients from their retina unit before whittling it down to 70 eyes of 70 patients after applying strict inclusion criteria.
Inclusion criteria included:
- Completed a standard three-injection loading dose of anti-VEGF therapy (ranibizumab 0.5 mg, aflibercept 2 mg, or bevacizumab 1.25 mg)
- Received at least seven injections in the first year
- Lost ≥10 Early Treatment Diabetic Retinopathy Scale (ETDRS) letters from their post-loading dose baseline
Demographics: Average age was 76.8 years, 64.3% female. Mean baseline acuity after loading dose was 70.5 letters (roughly 20/40 Snellen equivalent). Mean follow-up was 38.5 months.
Three timepoints were evaluated: Baseline (4 to 6 weeks post-loading), the visit where 10-letter loss was first documented, and the visit with the worst acuity on record.
Findings?
Macular atrophy went from 7.1% of eyes at baseline to 41.4% at 10-letter loss to 81.4% at worst outcome.
Subretinal fibrosis increased as well: 11.4% at baseline, 25.7% at 10-letter loss, 57.1% at worst outcome.
Outer retinal integrity took a hit across the board:
- External limiting membrane (ELM) integrity: 45.7% → 12.9%
- Ellipsoid zone (EZ) integrity: 42.9% → 10.0%
The kicker: CRT had zero predictive value at any timepoint. It actually followed a paradoxical biphasic pattern, ticking up slightly at 10-letter loss before dropping at worst outcome.
The authors argued that late-stage CRT reduction was probably reflecting tissue loss and neurodegeneration—not a treatment win.
Tell me more.
On multivariate regression, three phases of the disease emerged, each with a different predictive profile:
Phase 1 (baseline): Nothing predicted acuity. No structural parameter reached independent significance, likely because loading therapy had the anatomy looking relatively controlled.
Phase 2 (10-letter loss): Three independent predictors showed up:
- Subretinal fibrosis (β = −0.536)
- Subretinal hyperreflective material (β = −0.350)
- Intraretinal fluid (β = −0.223)
Together, these accounted for 42.8% of the variance in acuity (R² = 0.428).
Phase 3 (worst outcome): Subretinal fibrosis was the only independent predictor left standing (β = −0.469). Once fibrotic scarring took over, it overwhelmed everything else.
Worth noting: Macular atrophy didn’t reach independent significance in any of the models, despite showing up in over 80% of eyes by the end. The authors noted its functional impact was buried under the concurrent fibrotic changes.
Limitations?
As this was a single-center, retrospective study, the usual caveats around causation apply: There was no control group of good responders to compare against.
Other considerations:
- Patients were all on pro re nata (PRN) protocols, but injection frequency, switches between anti-VEGF agents (e.g., ranibizumab, aflibercept, bevacizumab), and retreatment timing weren’t standardized
- The 28% poor-responder rate exceeds the commonly cited 10 to 15% range. The authors chalked that up to their broader definition: capturing anyone with ≥10 letters of loss over extended follow-up, which picks up slower decliners shorter studies miss
- Only spectral-domain OCT was used, without newer tools like OCTA or AI-assisted analysis
Expert opinion?
No outside expert commentary was included in the study.
That said: The authors made a pointed argument that the 10-letter loss timepoint is a critical window, the point where multiple pathological pathways are still active and potentially modifiable before fibrosis takes over for good.
Their position: Monitoring needs to shift away from thickness-based metrics toward qualitative structural assessment centered on fibrotic changes.
Anything else?
One finding that stands out: Intraretinal fluid showed up as an independent predictor of visual decline at the 10-letter loss stage.
- Why that matters: It’s a shift from the traditional view of fluid as just a vascular leak that needs more anti-VEGF. The authors argued it may actually function as a biomarker for ongoing neuroinflammatory damage that contributes to decline on its own.
Also of interest: They flag an open question about whether the treatment regimen itself plays a role. PRN protocols come with gaps between injections, and those gaps may give inflammatory processes room to drive tissue scarring. Whether treat-and-extend regimens or longer-acting agents could slow fibrosis development is something future studies will need to sort out.
Take home.
For clinicians seeing wet AMD patients who keep losing vision on anti-VEGF therapy: CRT alone isn’t cutting it.
Watching for qualitative structural changes on OCT, especially the emergence and progression of subretinal fibrosis, looks like a far better way to track what’s happening in these eyes.
As such: Catching fibrotic changes early, ideally at or before the 10-letter loss mark, may be the best shot at intervening before irreversible scarring sets in.