Photobiomodulation Shows Vision Gains in Dry AMD at 24 Months

Key Takeaways

• In LIGHTSITE III, most treated eyes gained at least 5 letters by month 21, and smaller proportions achieved gains of 10 or 15 letters.
• No phototoxicity was reported, ocular adverse-event rates were similar between groups, and progression to geographic atrophy was lower with photobiomodulation at month 24 while neovascular AMD occurred in both groups.

In the LIGHTSITE III 24-month readout, multiwavelength photobiomodulation was associated with a +6.2-letter BCVA gain at month 21 in dry AMD. LIGHTSITE III was a randomized, double-masked prospective comparison with sham in eyes with dry AMD.

LIGHTSITE III randomized 148 eyes from 100 subjects with dry AMD to multiwavelength photobiomodulation at 590 nm, 660 nm, and 850 nm or to sham treatment. Each treatment series consisted of nine sessions delivered three times weekly over three to five weeks, repeated every four months for 24 months. That schedule amounted to 54 in-office sessions across two years, a detail the editorial later cited as part of the study burden. Investigators said the prespecified BCVA efficacy endpoint was met at month 21, and the extended regimen formed part of the context for interpreting the results.

At month 21, 61.5% of treated eyes gained at least 5 letters, 23.1% gained at least 10 letters, and 4.4% gained at least 15 letters. Investigators reported no signs of phototoxicity, and at least one ocular-specific adverse event occurred in 38 study eyes from 32 subjects. Those ocular event rates were similar between groups, at 25.5% in sham-treated eyes and 25.8% in photobiomodulation-treated eyes. By month 24, progression to geographic atrophy occurred in 6.8% of photobiomodulation-treated eyes and 24% of sham-treated eyes, while later-stage neovascular AMD occurred in two sham-treated study eyes and seven photobiomodulation-treated study eyes. Three nonstudy eyes also progressed to neovascular AMD.

Several design features limit straightforward interpretation, including the sample size, eye-based analyses with partial bilateral enrollment, and modeling used to address within-subject correlation. Additionally, 73% of subjects were available at month 24, with discontinuations more frequent in the sham group.