Eyeworld

73 EW RETINA July 2016 This rate in Protocol T seems to be faster with 2 mg aflibercept than with 0.3 mg ranibizumab or 1.25 mg bevacizumab among patients with entry level ETDRS vision of 20/50 or worse, at least during the first year of the study, using the treatment protocol followed in Protocol T in patients who have not received anti-VEGF therapy for 1 year prior to initiation of therapy. Dr. Zarbin said he prefers to start all patients on ranibizumab because it has a shorter systemic half-life, which may or may not factor into systemic side effects. But he certain- ly uses the other anti-VEGFs as well. "In a clinical setting, if we're unhappy with an initial response, there's nothing wrong with trying a different approach," Dr. Zarbin said. "The biggest difference for me with the year 2 results is that I'm not as rigid about treatment as I was after the publication of the year 1 results." EW References 1. Wells JA, et al. Aflibercept, bevacizumab, or ranibizumab for diabetic macular edema: two- year results from a comparative effectiveness randomized clinical trial. Ophthalmology. 2016;123:1351–1359. 2. Wells JA, et al. Aflibercept, bevacizumab, or ranibizumab for diabetic macular edema. N Engl J Med. 2015;372:1193–203. 3. Elman MJ, et al. Intravitreal ranibizumab for diabetic macular edema with prompt versus deferred laser treatment: 5-year randomized trial results. Ophthalmology. 2015;122:375– 81. 4. Kaiser PK. Prospective evaluation of visual acuity assessment: a comparison of snellen versus ETDRS charts in clinical practice (An AOS Thesis). Trans Am Ophthalmol Soc. 2009;107:311–24. 5. Avery RL, et al. Systemic safety of pro- longed monthly anti-vascular endothelial growth factor therapy for diabetic macular edema: a systematic review and meta-analy- sis. JAMA Ophthalmol. 2015;134:21–9. Editors' note: Dr. Wells has financial interests with Genentech (South San Francisco) and Regeneron (Tarrytown, New York). Dr. Zarbin has financial interests with Genentech. Contact information Wells: jackwells@palmettoretina.com Zarbin: zarbin@njms.rutgers.edu Vision outcomes No clinician likes to leave persistent edema month-to-month, Dr. Wells said—and both ranibizumab and aflibercept reduced central subfield thickness by similar amounts in years 1 and 2, and both reduced DME considerably more than bevacizumab (65% and 71% of eyes achieved a normal CST of 250 mi- crons or less at the 2-year endpoint, respectively, compared to 41% with bevacizumab). In the better base- line vision subgroup, bevacizumab reduced the edema about 50% less than the other 2 drugs across the 2 years. But in the worse seeing eyes, while inferior at 1 year, bevacizumab caught up with the other 2 in the second year, possibly because of more laser in that group. "Ranibizumab is not as cheap as bevacizumab, but it's cheaper than aflibercept. My first choice for better seeing eyes is ranibizumab 0.3 mg," he said, but added he "would defi- nitely give" aflibercept to a patient with worse-seeing vision. "We know from other stud- ies with longer follow-up that the number of injections continues to decrease over time—the DRCR's Protocol I showed patients received 9 injections on average in the first year, but by year 5 there were no in- jections needed," Dr. Wells said, and the edema remained consistent. 3 Current industry-sponsored studies are investigating at what threshold fluid can be tolerated provided vision remains stable, Dr. Wells said. There is also a study enrolling eyes with very good vision (20/20 or 20/25), as there is current- ly no study data to indicate how to treat these eyes. "I don't always treat eyes with good vision. In fact, I probably don't treat them nearly as much. I don't do nearly as much macular laser now as I used to and only use steroids in pseudophakic eyes," Dr. Wells said. Perhaps most telling for clini- cians, however, is that patients were unmasked at the end of year 1 and only 1 patient switched treatment regimens. This information and the year 2 data seem to provide a convergence of evidence that any efficacy differences among the 3 treatments may not be clinically im- portant with the possible exception of the rate of visual improvement. But "in the second year, the differences almost disappeared," Dr. Wells said. "By year 2, it was 58% for aflibercept, 55% for ranibizumab, and 52% for bevacizumab in the worse baseline vision eyes, and 20% for aflibercept, 19% for ranibizumab, and 17% for bevacizumab in the better baseline vision eyes." So the clinical advantage of aflibercept over the other 2 agents was diminished by the 2-year endpoint. Protocol T confirmed the rapid decline in necessary number of in- jections to maintain improvement. In the first year, subjects received an average of 10 injections, which decreased to 5 in year 2. "The reason probably is the dis- ease-modifying effect of the drug," Dr. Zarbin said. "It's a hypothesis, but I think the reason people need fewer injections in year 2 is because in many cases, their retinopathy is regressing. In some studies, the ret- inopathy regression seems to begin after the third injection, so it can happen pretty early." Both the VIVID/VISTA and RISE/ RIDE studies found that anti-VEGF injections "doesn't simply make the edema go away, they also induce regression of the severity of diabetic retinopathy," Dr. Zarbin said. Safety signals Patients may be confused about the systemic safety of the drugs, as the Antiplatelet Trialists' Collaboration (APTC) reported events seemed to be higher in the bevacizumab and ranibizumab groups, but the Proto- col T study authors attributed it to chance. "If you look at some of the most general studies where ranibi- zumab and aflibercept were both used, there's no signal that we could find that made us think that there's something to this," Dr. Wells said. "It's not something that's going to alter my treatment plans." Conversely, some studies have shown a "very clear safety signal" with aflibercept and ranibizumab, 5 but those were trial subjects receiv- ing monthly injections at the high- est doses possible, Dr. Zarbin said. "Would you see that same safety sig- nal if people were being dosed PRN or treat and extend? Maybe not." not have received any anti-VEGF therapy within the previous year. There's "a perfectly good reason to do that in the context of this trial, but the fact is most of our patients don't meet that enrollment criterion," Dr. Zarbin said. Another detail regarding patient selection is that most physicians in clini- cal practice measure visual acuity with Snellen charts and not with ETDRS charts. Visual acuity with ETDRS charts tends to be 1–2 lines better than what is measured using Snellen charts, and the degree of disparity increases as the patient's visual acuity worsens. 4 Thus, ETDRS vision of 20/50 is likely to be in the 20/60–20/70 range for office Snellen visual acuity. Dr. Wells said in his practice, "I apply the results of the study to the Snellen vision we measure in our office. I generally treat as long as they are improving; if they stop im- proving, I add laser if they have had at least 6 injections. After an added laser, if they don't improve further, I withhold their treatment and see what happens. Or if they have bad vision and still have an edema, I'll think about adding steroids." Clinical messages Protocol T evaluated changes in large groups of patients, and it's a difficult task to individualize outcomes to the patient in front of the clinician, Dr. Wells said. Talking about an 18-letter vision gain, for example, as found with aflibercept in worse-seeing eyes, is over prom- ising for the individual patient. Dr. Wells stressed this was a mean improvement in a large group of eyes, and numerous patients would, therefore, fall well below the thresh- old. A more clinically meaningful measure is the proportion of eyes gaining 3 or more lines of vision, and in the first year, there was a clear clinical benefit of aflibercept over the other 2 agents: 67% of eyes treated with aflibercept versus 50% of ranibizumab-treated eyes and 41% of bevacizumab-treated eyes achieved that level of vision im- provement. Expressed another way, aflibercept-treated eyes were 63% more likely to gain 3 or more lines than eyes treated with bevacizum- ab, and 34% more likely than eyes treated with ranibizumab.

• Cover