Objective-To provide data on the short-term effect of intravitreal bevacizumab for diabetic macular edema (DME).
Design-Randomized phase 2 clinical trial.Participants-121 eyes of 121 subjects (109 eligible for analysis) with DME and Snellen acuity equivalent ranging from 20/32-20/320.Interventions-Random assignment to one of five groups: focal photocoagulation at baseline (N=19, Group A), intravitreal injection of 1.25mg bevacizumab at baseline and 6 weeks (N=22, Group B), intravitreal injection of 2.5mg bevacizumab at baseline and 6 weeks (N=24, Group C), intravitreal injection of 1.25mg bevacizumab at baseline and sham injection at 6 weeks (N=22, Group D), or intravitreal injection of 1.25mg bevacizumab at baseline and 6 weeks with photocoagulation at 3 weeks (N=22, Group E).Main Outcome Measures-Central subfield thickness (CST) on optical coherence tomography and best-corrected visual acuity (VA) were measured at baseline and after 3, 6, 9, 12, 18, and 24 weeks.Results-At baseline, median CST was 411 microns and median Snellen VA equivalent was 20/50. Compared with Group A, Groups B and C had a greater reduction in CST at 3 weeks and about one line better median visual acuity over 12 weeks. There were no meaningful differences between Groups B and C in CST reduction or VA improvement. A CST reduction >11% (the reliability limit) was present at 3 weeks in 36/84 (43%) bevacizumab-treated eyes and in 5/18 (28%) eyes treated with laser alone, and at 6 weeks in 31/84 (37%) and 9/18 (50%) eyes, respectively. Combining focal photocoagulation with bevacizumab resulted in no apparent short-term benefit or adverse outcomes. Endophthalmitis developed in one eye. The following events occurred during the first 24 weeks in subjects treated with bevacizumab without attributing cause to the drug: myocardial infarction (N=2), congestive heart failure (N=1), elevated blood pressure (N=3), and worsened renal function (N=3).Conclusion-These results demonstrate that intravitreal bevacizumab can reduce DME in some eyes, but the study was not designed to determine whether treatment is beneficial. A phase 3 trial would be needed for that purpose.