Abstract
Purpose :
The CHAMP Phase 3 trial (NCT03350620) evaluated the effectiveness of NVK002, a novel, preservative-free formulation of low-dose atropine in slowing myopia progression in children over 3 years. In a pre-specified analysis, NVK002 0.01%, was statistically significant compared to placebo in increasing the proportion of responders (<0.50D myopia progression over 36 months) and slowing the progression of myopia (in spherical equivalent refraction, SER). During the study, adherence, COVID and confounding treatments all may have affected the efficacy outcomes. Subjects who discontinued treatment and switched to orthokeratology, multifocal contact lens or compounded atropine may have affected the treatment effect size versus placebo. We tested the hypothesis that excluding confounding data due to prohibited therapies from efficacy analysis increases the treatment effect size of active treatments compared to placebo.
Methods :
Subjects in the US and Europe aged 3 to ≤17 years with an SER between −0.50 D to −6.00 D, were randomized to receive once-daily placebo, NVK002 0.01%, or 0.02% eye drops for 36 months. The post hoc analyses of responder proportion and SER progression excluded post-last dose data of subjects who discontinued NVK002, remained in the study and switched to prohibited therapies. These analyses were performed on a modified intent-to-treat (mITT) population (subjects aged 6 to 10 years at baseline).
Results :
Of 489 subjects in the mITT set, 23 (9 placebo, 6 0.01%, 8 0.02%) discontinued study treatment and switched to prohibited therapies. Post hoc analysis, excluding the post-last dose data of these 23 subjects, showed increased treatment effect sizes compared to the original analyses for the proportion of responders (0.02%/placebo: Odds ratio=2.5, p=0.16 [post hoc]; OR=1.8, p=0.37 [original]) and the LSM Difference from placebo in SER (0.02% - placebo: LSMD=0.15D, p=0.013 [post hoc]; LSMD=0.099 D, p=0.101 [original]). Small increases in treatment effect sizes were also observed for NVK002 0.01% after exclusion of the confounding data.
Conclusions :
Confounding treatments can have a major effect on efficacy outcomes in long term myopia trials. For CHAMP, the exclusion of 23 patients data who switched to confounding therapies but continued in the study largely explained the observed performance of NVK002 0.02% and increased the observed treatment effect of NVK002.
This abstract was presented at the 2023 ARVO Annual Meeting, held in New Orleans, LA, April 23-27, 2023.