Gas Tamponade for the Prevention of Postoperative Vitreous Hemorrhaging After Diabetic Vitrectomy: A Randomized Clinical Trial

December 2022

Written by:  Jordan Deaner, MD
Mid Atlantic Retina / Wills Eye Hospital

Philadelphia, PA

Rush RB, Velazquez JC, Rosales CR, Rush SW. Gas Tamponade for the Prevention of Postoperative Vitreous Hemorrhaging After Diabetic Vitrectomy: A Randomized Clinical Trial. Am J Ophthalmol. 2022;242:173-180. doi:10.1016/j.ajo.2022.06.015

Disclosures: EyePoint – Consultant, Alimera – Consultant

Vitreous hemorrhage (VH) is a frequent post-operative complication in patients undergoing pars plana vitrectomy (PPV) for indications associated with proliferative diabetic retinopathy (PDR). Post-operative vitreous hemorrhage can be visually debilitating and may even require a second surgery to recover vision. Properative anti-VEGF therapy has helped to lower the risk of post-operative VH, reducing the incidence to 13 to 25% of cases. The authors of this study compared the incidence of post-operative vitreous hemorrhage after PPV for non-clearing VH secondary to proliferative diabetic retinopathy between sulfur hexafluoride (SF6) tamponade and basic saline solution (BSS).


This study randomized 106 eyes of patients with non-clearing VH secondary to PDR without vitreoretinal adhesion to PPV with either 20 to 30% SF6 gas tamponade or BSS vitreous substitute. All eyes received intravitreal ziv-aflibercept preoperatively, 1 to 10 days prior to PPV. All eyes that were phakic underwent combination cataract extraction with intraocular lens implantation during the same surgical session. A 23-gauge or 25-gauge PPV was performed, along with meticulous diathermy and membrane peeling at the operating surgeon’s discretion. Laser photocoagulation was performed in all 4 quadrants out to the vitreous base if not already present.  Eyes were randomized to the study group intraoperatively after vitrectomy was completed to limit surgeon technique bias. Primary outcome measure was incidence of post-operative VH between groups over the 6-month follow-up period. Secondary outcomes included unplanned PPV for post-operative VH and best corrected visual acuity (BCVA) at 6-month follow-up.


There was no statistically significant difference between baseline demographic or clinical characteristics. Vitrectomy surgery time was significantly longer in the SF6 group compared to the BSS group (47.0 vs 38.1 minutes, P=0.02). The incidence of post-operative VH over the 6-month follow-up period was 11.1% (6 of 54 eyes) in the SF6 group and 33.3% (14 of 42 eyes) in the BSS group (P=0.01). The incidence of unplanned PPV secondary to post-operative VH was 3.7% (2 of 54 eyes) in the SF6 group and 14.2% (6 of 42 eyes) in the BSS group (P=0.06). At month 6, BCVA was significantly improved from baseline in both groups from 1.60 to 0.67 logMAR in the SF6 group (P<0.01) and from 1.59 to 0.75 logMAR in the BSS group (P<0.01). There was not a statistically significant difference in the BCVA improvement from baseline to 6-month follow-up between the two cohorts (P=0.58).


Overall, this trial showed a significantly lower incidence of post-operative VH when SF6 was used in comparison to BSS in PDR patients undergoing PPV for non-clearing VH. The authors provide a few hypotheses for why gas tamponade may decrease the risk of post-operative VH, including mechanical tamponade on fragile blood vessels, promotion of antifibrinolytic substances near the areas of neovascularization and hemorrhage, and condensed procoagulants in the more limited fluid phase.