Potential bias of preoperative intravitreal anti-VEGF injection for complications of proliferative diabetic retinopathy

PURPOSE: Intravitreal anti-VEGF injection (IVI) is administered before vitrectomy to assist management of proliferative diabetic retinopathy (PDR)-related complications. In the clinical setting, retinal surgeons determine the use of preoperative IVI based on individual criteria. In this study, we investigated factors related to the potential bias of retinal surgeons in using IVI prior to vitrectomy for PDR-related complications, and evaluated the real-world outcomes of surgeon-determined preoperative IVI. METHODS: Medical records of 409 eyes of 409 patients who underwent 25-gauge vitrectomy for PDR complications at seven Japanese centers (22 surgeons) were retrospectively reviewed. Ocular factors, demographic and general clinical factors, surgical procedures, and postoperative complications were compared between IVI group (patients who received preoperative IVI; 87 eyes, 21.3%) and non-IVI group (patients who did not receive preoperative IVI; 322 eyes, 78.7%). In addition, baseline HbA1c in IVI group and non-IVI group was compared between eyes with and without postoperative complications. RESULTS: At baseline, IVI group was younger (P<0.001), had shorter duration of diabetes treatment (P = 0.045), and higher frequencies of neovascular glaucoma [NVG] (P<0.001) and tractional retinal detachment [TRD] (P<0.001) compared to non-IVI group. Although IVI group had higher frequencies of intraoperative retinal break and tamponade procedure, there were no significant differences in postoperative complications and additional treatments between two groups. Baseline HbA1c levels were also not correlated with postoperative complications of VH, NVG, and RD both in IVI group and non-IVI group. Logistic regression analysis identified age (P<0.001, odds ratio [OR] 0.95), presence of NVG (P<0.001, OR 20.2), and presence of TRD (P = 0.0014, OR 2.44) as preoperative factors in favor of IVI. CONCLUSIONS: In this multicenter real-world clinical study, younger age and presence of NVG and TRD were identified as potential biases in using IVI before vitrectomy for PDR complications. Eyes that received preoperative IVI had more intraoperative retinal breaks requiring tamponade than eyes not receiving IVI, but postoperative outcome was not different between the two groups.