Stereotactic radiotherapy for neovascular age-related macular degeneration: year 3 and 4 results of a randomised, double-masked, sham-controlled, recall, post-primary device trial (STAR)

Objectives:
At the 2-year primary outcome, the STAR trial showed that stereotactic radiotherapy (SRT) significantly reduces anti-vascular endothelial growth factor (VEGF) injection frequency for neovascular age-related macular degeneration (AMD), with non-inferior vision. This preplanned recall aimed to assess results beyond the primary outcome.

Design:
Randomised, double-masked, sham-controlled, recall, post-primary device trial.

Setting:
30 UK National Health Service Hospitals.

Participants:
411 participants aged at least 50 years with chronic, pre-treated, active AMD.

Intervention and follow-up:
Participants received one-off 16-Gray or sham SRT delivered using a robotically-controlled device. After 2 years of monthly study visits, participants reverted to routine care, with anti-VEGF drug selection and dosing intervals based on local practice, but with masking maintained, and repeat data collection at year 3 and 4 study visits.

Main outcome measures:
The primary efficacy outcome at year 4 was the number of anti-VEGF injections, tested for superiority (fewer injections). The main secondary outcome was visual acuity, tested for non-inferiority (five-letter margin). Safety outcomes included adverse events, serious adverse events and microvascular abnormalities. The same analyses were undertaken at years 2, 3 and 4. A within-trial costing analysis was undertaken for participants with 4 years’ follow-up.

Results:
Of 411 participants (204 [58%] female), 274 were allocated to SRT and 137 to sham. The year 4 intent-to-treat efficacy analysis included 222 (81%) SRT, and 106 (77%) sham SRT participants respectively. The SRT group received a mean (standard deviation) of 19.1 (10.9) injections over 4 years versus 21.6 (11.3) with sham, an adjusted reduction of 3.2 injections (95% confidence interval of difference: -5.7 to -0.7). The SRT group averaged 8.4 injections in years 3-4, versus 8.3 with sham. The SRT group final visual acuity change was 8.3 letters worse than sham (95% confidence interval of difference: –12.7 to –4.0). Adverse event rates were similar between groups, but reading centre-detected microvascular abnormalities occurred in 126 SRT-treated eyes (58%, 126/218) and 16 (16%, 16/102) sham-treated eyes.

Conclusions:
Although the overall reduction in intravitreal therapy was maintained out to year 4, the inferior vision in SRT-treated eyes effectively reverses the conclusions of the year 2 analysis, and no longer supports the use of SRT to treat nAMD. Long-term, standard care, masked, post-primary recall studies may provide useful clinical insight.

Trial registration: ClinicalTrials.gov (NCT02243878).