Surgical, drug treatments have similar outcomes in patients with vitreous hemorrhage from proliferative diabetic retinopathy

A randomized trial found equivalent improvements in visual acuity at two years with either therapy but identified some differences in outcomes that could help patients and clinicians choose a treatment method.


Surgical and injectable drug treatments were equally effective for proliferative diabetic retinopathy (PDR), a study found.

Researchers compared anti-vascular endothelial growth factor (VEGF) eye injections to removal of blood via vitrectomy surgery and laser photocoagulation in a trial at 39 sites in North America with 205 participants who had vision loss due to vitreous hemorrhage, enrolled from November 2016 to December 2017. The final follow-up visit was completed in January 2020. One eye per participant was randomized to four monthly injections of aflibercept (n=100) or vitrectomy with panretinal photocoagulation (n=105). Both groups could receive aflibercept or vitrectomy during follow-up. The study was supported through a cooperative agreement from the National Eye Institute, the National Institute of Diabetes and Digestive and Kidney Diseases, the NIH, and the U.S. Department of Health and Human Services. Regeneron Pharmaceuticals Inc. provided aflibercept for the study, as well as funds to defray clinical site costs. Results were published Dec. 15, 2020, by JAMA.

The primary outcome was mean visual acuity letter score (range, 0 to 100; higher scores indicate better vision) over 24 weeks. The study was powered to detect a difference of eight letters. Secondary outcomes included mean visual acuity at four weeks and two years.

At baseline, participants had a mean visual acuity letter score of 34.5 (Snellen equivalent, 20/200); 95% of them completed the 24-week visit and 90% (excluding nine deaths) completed the two-year visit. The mean visual acuity letter score over 24 weeks was 59.3 (Snellen equivalent, 20/63) in the aflibercept group and 63.0 (Snellen equivalent, 20/63) in the vitrectomy group (adjusted difference, −5.0; 95% CI, −10.2 to 0.3; P=0.06).

Among 23 secondary outcomes, 15 showed no significant differences between groups. At four weeks, the mean visual acuity letter score was 52.6 (Snellen equivalent, 20/100) in the aflibercept group versus 62.3 (Snellen equivalent, 20/63) in the vitrectomy group (P=0.003). At two years, it was 73.7 (Snellen equivalent, 20/40) versus 71.0 (Snellen equivalent, 20/40), respectively. Over two years, 33 eyes (33%) assigned to aflibercept received vitrectomy and 34 eyes (32%) assigned to vitrectomy received aflibercept.

The authors cautioned that the study may have been underpowered to detect a clinically important benefit in favor of initial vitrectomy with panretinal photocoagulation. They added that although visual outcomes were not significantly different between treatments, other findings from the study could help clinicians guide therapeutic decision, including that vitrectomy offered faster vision restoration, reduced likelihood of recurrent vitreous hemorrhage (15% vs. 49%), and led to greater resolution of neovascularization, while the aflibercept group experienced less frequent center-involved diabetic macular edema and two-thirds avoided vitrectomy. “The decision to initiate treatment using anti-VEGF injections vs vitrectomy with panretinal photocoagulation is influenced by many factors, including anticipated likelihood of patient adherence with follow-up visits, medical comorbidities, access to specialized treatments or medications, and the need or desire to hasten visual recovery, particularly for patients whose fellow eye also does not have good vision,” the authors wrote.

An editorial noted vitrectomy is performed by specially trained surgeons who comprise less than 5% of all ophthalmologists, requires specialized operating facilities, and has related complications. Anti-VEGF injection therapy can be administered by general ophthalmologists and well-trained nurses in a clinic. The costs of anti-VEGF therapy may be unsustainable for many patients, the editorialists noted.

The study suggests that while aflibercept “may be a safe alternative to vitrectomy for the initial management of vitreous hemorrhage from proliferative diabetic retinopathy, vitrectomy with panretinal photocoagulation remains the standard care for this vision-threatening complication for most patients,” the editorialists wrote. “However, both clinical approaches have benefits and risks, and physicians treating patients with vitreous hemorrhage from proliferative diabetic retinopathy can and should personalize the initial therapeutic choice based on individual patient circumstances and influencing factors.”