Drilling Deeper for Treatment Choices in Diabetic Macular Edema

Abstract

Afterdecadesofusing focal/grid laserphotocoagulation to reduce the riskofvision losswithdiabeticmacularedema(DME), thecurrent therapyhasshifted to intravitreal injectionsofanti– vascular endothelial growth factor (VEGF) agents, which can result in gains in visual acuity (VA). Investigators from theDiabetic Retinopathy Clinical ResearchNetwork1 report in this issueof JAMAOphthalmology the findings fromthe secondary analyses of data previously reported in a comparative trial of 3 anti-VEGFagents—aflibercept, bevacizumab, and ranibizumab—for the treatment of DME.2 The current report examines the1-yearoutcomesof subgroups thatwerebothprespecified and post hoc.1 The comparative trial randomized 660 participants with DME into 3 groups: participants received 0.05-mL intravitreal injections of 2 mg of aflibercept, 1.25 mg of bevacizumab, or 0.3 mg of ranibizumab.2 All 3 drugs resulted in substantial improvement, on average, in VA at 1 year, with mean improvement occurring by 1 month. Prespecified subgroup analyses were based on baseline VA, worse VA (Snellen equivalent of 20/50 or worse), or better VA (Snellen equivalent of 20/32 or 20/40). All 3 agents were considered to have similar effects on VA for participants whose baseline VA was 20/30 or 20/40. However, for participants whose baseline VA was 20/50 or worse, improvement in VA was greatest with aflibercept while the effects of ranibizumab and bevacizumab on VA were similar. In addition, the participants who received aflibercept also required less supplemental focal/grid laser photocoagulation than did the patients who received ranibizumab or bevacizumab. In thecurrent report, additional exploratoryanalyseswere conducted, again stratified by baseline VA, to evaluate the 1-year VA outcomes in greater detail.1 In addition to themean VA gain at 1 year, the proportion of participants experiencing visual gain or loss of 10, 15, or more letters were previously published.2 The additional data presented included an unusualdetailed listingofVAoutcomesat 1 year, includinganalyses of the proportion of participants achieving VA of 20/20 at 1 year in each of the treatment groups. Post hoc analyseswere alsoconductedbasedonstratificationbybaseline retinal thickness as measured by optical coherence tomography in addition tobaselineVA.Central subfield thicknesswasdivided into thicker (>400μm)or thinner (250-399μm)categories.Thescatter plot in eFigure 1 in the current report1 showedonly amodest correlationof retinal thicknesswithVA, similar to the scatterplotpublished inastudypreviously reportedbytheDiabetic Retinopathy Clinical Research Network.3 Participants with thickenedor thinned retinamayhaveawide rangeofVAs.The current analyses showed separate interactions of baseline VA with treatment and baseline central subfield thickness with treatment, but when both 2-way interactions were included in the analyses, only baseline VA remained statistically significant.1 It is not surprising that, regardless of central subfield thickness, baseline VA remains the most important factor in determining the visual outcomes at 1 year for the treatment, given themodest correlationwithVAof retinal thickness as measured by optical coherence tomography. Despite these findings in the subgroup analyses, the investigators expressed concern about the efficacy of bevacizumab.1Asnoted intheabstract, for thesubgroupofeyes with better initial VA and thicker central subfield thickness, “some VA outcomesmay beworse in the bevacizumab group than in theafliberceptandranibizumabgroups.”1Do thesenew data persuade us that there are clinically important differences among the 3 anti-VEGF agents, especially for bevacizumab? I concurwith the investigators’ ownstatements about thesedata: “Given theexploratorynatureof theseanalysesand the small sample size within subgroups, caution is suggested when using the data to guide treatment considerations for patients.”1 These data cannot provide support to further differentiate the effects of the 3 treatments. In fact, these data fully support the investigators’ initial conclusion that, for persons with DME presenting with better initial VA, all 3 antiVEGF agents were appropriate while the greatest visual gain was achieved with aflibercept treatment in participants presenting with the poorest VA. The proportion of persons in the general population presenting with DME with visual impairment of 20/50 or worse is difficult to ascertain because clinical trials have limitations on the degree of VA for patient eligibility. Population-based studiesdonothavesufficientnumbersofpersonsaffectedwith DME to provide reliable statistics (R. Klein, MD, written communication, August 13, 2015). However, the Early Treatment Diabetic Retinopathy Study,4 which was a clinical trial conducted in the 1980s that acceptedparticipantswithDMEwith a variety of levels of VA and with no lower limits on VA, suggested that 75%of patients presenting to a general clinicwith DMEhave a VA of 20/40 or better. One could assume that this proportionwouldbemarkedly increased in thepresent, as the medical care of glycemic andbloodpressure control andmanagement of dyslipidemia for persons with diabetes mellitus have vastly improved over the decades. Regardless of the precise number of personswith varying degrees of baseline VA seeking treatment for DME, the issues surrounding compounding pharmacies for bevacizumab, or types of health insurance coverage, bevacizumab remains an important player in this trio of anti-VEGF therapies. Because of itsmarkedly reduced cost andbecause of the findings from thiscomparative trial,bevacizumabstill canbea first-line treatRelated article page 127 1-Year Efficacy of Anti-VEGF Agents in Diabetic Macular Edema Original Investigation Research