Abstract
Purpose :
The VIEW 1 and VIEW 2 studies led to FDA approval of aflibercept for the treatment of neovascular age-related macular degeneration (nAMD) in 2011. In this electronic health record (EHR) based registry study, we replicated the study design from an observational research perspective and evaluated the outcomes using real-world data (RWD) contained in the American Academy of Ophthalmology IRIS® Registry (Intelligent Research in Sight), the world's largest single-specialty clinical database.
Methods :
The IRIS Registry was queried for eyes older than 50 years with nAMD that were treatment-naive on the index date. Using structured data (ICD coded, CPT codes, etc.) we excluded a prior diagnosis of nAMD, DME, DR, uncontrolled glaucoma, causes for CNV besides nAMD, prior RD, prior uveitis, prior vitrectomy surgery, prior corneal transplantation and those with missing gender/race information. Cohorts emulated the study arms of the VIEW 1 and VIEW 2 trials: cohort 1 received 2 mg of aflibercept every 4 weeks (2q4), cohort 2 received 2 mg of aflibercept every 8 weeks after 3 monthly injections (2q8) and cohort 3 received 0.5 mg ranibizumab every 4 weeks (rq4). We allowed a buffer of +/- 7 days for treatment schedules. By definition, the number of injections in each cohort was 12 (2q4), 7 (2q8) and 12 (rq4).
Results :
The cohorts corresponding to the three arms of the VIEW 1 and VIEW 2 studies included 348 (2q4), 331 (2q8), and 460 (rq4) eyes. We found that 92.5% (2q4), 95.2% (2q8) and 92.6% (rq4) of eyes maintained vision (losing < 15 ETDRS letters), compared to 95.1 to 95.6 (2q4), 95.1 to 95.6% (2q8), and 94.4% (rq4) of eyes in the VIEW studies. We also found that 19.8% (2q4), 28.1% (2q8) and 25% (rq4) of eyes gained 15 or more ETDRS letters and had a mean gain of +3.9 (2q4), +6.9 (2q8) and +5.8 (rq4) ETDRS letters at 12 months. In the VIEW studies, mean BCVA gains were 8.3 to 9.3 letters at 12 months.
Conclusions :
This is the first time that RWD has been used in an attempt to replicate the primary outcomes of a clinical trial in ophthalmology. The proportion of eyes losing <15 across the VIEW 1 and VIEW 2 studies was similar to our registry cohort, indicating proof-of-concept, although the mean gains in letters varied across the studies. Benchmarking real-world endpoints versus randomized controlled trials could be a valuable tool in validating the predictive ability of these endpoints as a measure of effectiveness.
This is a 2021 ARVO Annual Meeting abstract.