Year-2 Protocol T data paints different scenario from year-1

The Diabetic Retinopathy Clinical Research Network’s (DRCR.net) Protocol T-the first head-to-head-to-head comparison among aflibercept (Eylea, Renegeron Pharmaceuticals), ranibizumab (Lucentis, Genentech), and bevacizumab (Avastin, Genentech)-found in its first-year results that all three agents improved vision and reduced edema effectively.

However, the DRCR.net found differences in a pre-planned subanalysis of baseline vision, showing better outcomes with aflibercept than the other two anti-vascular endothelial growth factor (anti-VEGF) agents in patients with baseline vision worse than 20/50. That difference was depleted by Year 2.

"There were 660 eyes randomized, and we had great retention,” said John A. Wells, III, MD, “More than 95% of patients completed the primary outcome visit, and over 90% the final two-year visit.”

Year 2 results

There were no differences between the 3 groups in baseline characteristics that included vision, optical coherence tomography (OCT), central subfield thickness, and history of prior laser or anti-VEGF therapy, said Dr. Wells, who is in private practice in Columbia, SC.

As in other studies, “the number of injections required in the second year decreased by about 40% relative to the first year,” Dr. Wells said. “Overall, all three groups were given a median of 15 to 16 injections over the course of the 2 years of the study.”

In the first and in the second year, eyes in the aflibercept group were less likely to receive laser treatment than eyes in the other two groups. Compared to the first year, in the second year there was about a 50% reduction in the amount of laser applied in all three groups.

Overall, eyes in the aflibercept group were the least likely to receive laser, whereas eyes in the bevacizumab group were the most likely to receive laser at anytime over the 2 years.

The mean change in visual acuity at 2 years for the full cohort showed “the difference between aflibercept and ranibizumab had disappeared and was no longer significant, while aflibercept remains superior to bevacizumab,” Dr. Wells said.

Planned subgroup analysis

But the comparison has “limited utility” because of the close correlation to baseline vision, he said. For Protocol T, the DRCR.net had predetermined to stratify outcomes into “better baseline vision”-those with 20/32 to 20/40 on the Early Treatment of Diabetic Retinopathy Study chart-and worse baseline vision, people with 20/50 or worse on the ETDRS chart.

At year 2, the better baseline vision group sustained their initial gains of about 8 letters,” Dr. Wells said. “An area-under-the-curve (AUC) analysis of this better baseline vision subgroup also confirmed that there were no differences between the 3 arms at the 2-year endpoint.”

However, at 2 years in the worse baseline group, “the difference between aflibercept and ranibizumab had disappeared and was no longer significant, whereas aflibercept remains superior to bevacizumab,” Dr. Wells said, adding the AUC analysis found the average daily change in vision was greater for aflibercept over 2 years than the other 2 agents.

Vision & anatomic outcomes

By Year 2, “about 50% to 60% of eyes gained 3 or more lines of vision, which calls into question the clinical relevance of the difference in the treatments at the 2-year endpoint. However, a mean AUC analysis of eyes gaining 3 or more lines over 2 years showed some superiority of aflibercept,” Dr. Wells said.

Anatomically, bevacizumab reduces the edema less effectively than the other two agents, he pointed out. Continuing with the split between the better baseline and worse baseline groups, Dr. Wells said the reduction was almost 50% less with bevacizumab than with the other 2 agents at years 1 and 2 in eyes with better baseline vision. But in the eyes with worse baseline vision, bevacizumab was inferior only to aflibercept at Year 2.

“Eyes treated with bevacizumab were less likely to achieve the normal OCT central subfield thickness of less than 250 µm in both the better baseline vision subgroup and in the worst baseline vision subgroup,” he said.

Focal grid laser during the course of the study had little effect, Dr. Wells added.

“When we looked at the effect on OCT thickness over the course of the 2 years, in the better baseline vision subgroup eyes that received no focal laser at all during showed no additional reduction in OCT thickness in the second year,” Dr. Wells explained. “Similarly, though, eyes that received laser between 6 months and 2 years in this better baseline vision subgroup also showed no additional reduction in OCT thickness in the second year.”

In the worse baseline vision subgroup, eyes that never received the laser showed no additional reduction in OCT thickness in the second year. If the eyes had undergone laser, though, those in the bevacizumab arm showed additional reductions in OCT edema in Year 2.

Safety outcomes were similar across all three groups during Years 1 and 2, with the exception of pre-specified Antiplatelet Trialists’ Collaboration (APTC) adverse events, where 12% of the patients in the ranibizumab group experienced an event compared to 5% and 8% in the aflibercept and bevacizumab groups, respectively.

Conclusions

Vision gains from baseline were sustained through 2 years, and the differences that favored aflibercept over ranibizumab in the worse baseline vision group had dissipated by Year 2, he said.

“The implication of the increased rate of APTC events with ranibizumab found in this study is uncertain and inconsistent with many prior trials and warrants further study,” Dr. Wells said. Any post-hoc analyses “should be viewed with caution” but “can provide insights that might be useful as clinicians apply these study results to their practices.”

John A. Wells, MD

E: jackwells@palmettoretina.com

Dr. Wells is an investigator for Genentech and Regeneron. This article was adapted from a presentation Dr. Wells delivered at Retina Subspecialty Day prior to the 2016 American Academy of Ophthalmology meeting.