Repeated photodynamic treatments as a risk factor for choroidal newvessels ingrowth

Purpose: Photodynamic therapy with Verteporfin (PDT) has been proven to be effective in reducing the visual loss due to age-related macular degeneration (AMD) complicated by choroidal newvessels (CNV). In order to achieve permanent closure of the CNV, multiple treatments are normally required. Even though clinical trials have shown that repeated PDT sessions do not damage visual function, some changes on the retinal structures have been suspected. We report here a series of 23 patients treated with multiple PDT sessions who developed recurrent CNV in a round area perfectly corresponding to the PDT diode-laser spot. Methods: Fluorescein angiographies of 349 AMD patients (age 71±9 ys, 196 women) who had had at least 2 PDT treatments for subfoveal CNV were retrospectively analyzed. The persistent/recurrent CNV requiring additional PDT were compared with the angiography at baseline and with the diameter of the PDT diode-laser spot. Results: Out of 349 AMD patients who received multiple treatments, 23 subjects (23 eyes) (6.6%) showed recurrences in a round area surrounding the CNV, resembling a circle perfectly corresponding to the spot of the diode laser which had been applied. The mean diameter of the recurrent CNV was 5.2 mm (SD=1.2). At baseline, before any PDT, the CNV could be classified as "predominantly classic" in all patients. All the recurrences appeared as "classic" CNV. Of these 23 eyes, 18 had had 2 previous PDT, 3 previous PDT in 5 eyes. The round area of recurrent CNV was then retrospectively analyzed in the angiograms at baseline and it included (i) the "classic" neovascular lesion, (ii) a surrounding area with late "oozing" interpreted as the "occult" component and (iii) normal retina. After PDT, the recurrent CNV developed in the last two areas, while the original neovascular membrane seemed to be collapsed and not leaking. Conclusions: In selected patients with AMD PDT is able to prevent severe visual loss by collapsing CNV, thus reducing or stopping subretinal exudation. The mechanism of selective action of PDT on the CNV has only been deducted from animal experiments and not yet demonstrated in humans. Different clinical studies have showed effects of PDT on normal choriocapillaris and retinal pigment epithelium included in the treatment spot. We have reported here a series of patients where recurrent "classic" CNV seemed to develop in the retinal area exposed to the PDT. The CNV grew from "occult" lesions or within normal retina. A possible explanation for this paradoxical effect is that PDT could interfere with the mechanisms which inhibit choroidal angiogenesis.