A pilot study of viscoelastic agent to prevent recurrent vitreous hemorrhage after vitrectomy for proliferative diabetic retinopathy

BACKGROUND: To evaluate the possibilty of preventing recurrent vitreous hemorrhage (RVH) after vitrectomy in proliferative diabetic retinopathy (PDR) patients with unabsorbed vitreous hemorrhage (VH) by intravitreal injection of viscoelastic agent (VA) at the end of the surgery and compared its effect with triamcinolone acetonide (TA). METHODS: This was a pilot prospective, observational study. PDR patients with VH who underwent vitrectomy were assigned to 3 groups according to the tamponade applicated at the end of the surgery, including VA group (intravitreally injected 1 ml VA if the retina was prone to bleed during the operation), TA group (intravitreally injected 2 mg TA when there was much exudates), or balanced salt solution (BSS) group (no tamponade). Then postoperative follow-up was performed routinely until 6 months after surgery. The primary outcome was the incidence of RVH, secondary outcome were the best-corrected visual acuity (BCVA) and introcular pressure (IOP). Cataract formation and other complication were also assessed. RESULTS: A total of 68 eyes, from 68 patients, were included. 18,18,32 eyes were enrolled in the VA group, TA group and BSS group, respectively. The integral incidence of RVH after vitrectomy was 5.6%, 5.6% and 12.5% respectively (P = 0.602). There was no early RVH in VA or TA group, whereas 3 early RVHs were identified in BSS group, however there was no significant difference (P = 0.171). Every group had one late RVH case. In all groups, final BCVA showed significant improvement compared to baseline. BCVA at any postoperative visit showed no significant differences among 3 groups. Mean IOP was higher 1 week after surgery in VA group compared with the other groups; however, in other times the differences were not significant. No cataract formation and other complication was noted in 3 groups. CONCLUSION: Intravitreal injection of VA or TA at the end of vitrectomy for PDR patients with unabsorbed VH tend to reduce the incidence of early RVH after vitrectomy similarly. As VA was preferred to applicate in the eyes that were prone to bleed, intravitreal injection of VA at the end of vitrectomy might be a promising method for preventing RVH in PDR patients.