ImportanceThere are reported benefits from vitrectomy for diabetic macular edema (DME); however, data precede anti–vascular endothelial growth therapy (VEGF) therapy, supporting a need to assess the current role of vitrectomy.ObjectiveTo determine rates of recruitment and efficacy outcomes of vitrectomy plus internal limiting membrane (ILM) peeling adjunctive to treat-and-extend (T&E) anti-VEGF injections for diabetic macular edema (DME).Design, Setting, and ParticipantsThis was a single-masked, multicenter randomized clinical trial at 21 sites in the United Kingdom from June 2018 to January 2021, evaluating single eyes of treatment-naive patients with symptomatic vision loss from DME for less than 1 year. Inclusion criteria were best-corrected visual acuity (BCVA) Early Treatment Diabetic Retinopathy Study letter score greater than 35 (approximate Snellen equivalent, 20/200 or better) and central subfield thickness (CST) greater than 350 μm after 3 monthly intravitreal injections of ranibizumab or aflibercept. Data analysis was performed in July 2023.InterventionsPatients were randomized 1:1 into vitrectomy plus standard care or standard care alone and further stratified into groups with vs without vitreomacular interface abnormality. Both groups received a T&E anti-VEGF injection regimen with aflibercept, 2 mg, or ranibizumab, 0.5 mg. The vitrectomy group additionally underwent pars plana vitrectomy with epiretinal membrane or ILM peel within 1 month of randomization.Main Outcomes and MeasuresRate of recruitment and distance BCVA. Secondary outcome measures were CST, change in BCVA and CST, number of injections, rate of completed follow-up, and withdrawal rate.ResultsOver 32 months, 47 of a planned 100 patients were enrolled; 42 (89%; mean [SD] age, 63 [11] years; 26 [62%] male) completed 12-month follow-up visits. Baseline characteristics appeared comparable between the control (n = 23; mean [SD] age, 66 [10] years) and vitrectomy (n = 24; mean [SD] age, 62 [12] years) groups. No difference in 12-month BCVA was noted between groups, with a 12-month median (IQR) BCVA letter score of 73 (65-77) letters (Snellen equivalent, 20/40) in the control group vs 77 (67-81) letters (Snellen equivalent, 20/32) in the vitrectomy group (difference, 4 letters; 95% CI, −8 to 2; P = .24). There was no difference in BCVA change from baseline (median [IQR], −1 [−3 to 2] letters for the control group vs −2 [−8 to 2] letters for the vitrectomy group; difference, 1 letter; 95% CI, −5 to 7; P = .85). No difference was found in CST changes (median [IQR], −94 [−122 to 9] μm for the control group vs −32 [−48 to 25] μm for the vitrectomy group; difference, 62 μm; 95% CI, −110 to 11; P = .11).Conclusions and RelevanceEnrollment goals could not be attained. However, with 47 participants, evidence did not support a clinical benefit of vitrectomy plus ILM peeling as an adjunct to a T&E regimen of anti-VEGF therapy for DME.Trial Registrationisrctn.org Identifier: ISRCTN59902040