Early Versus Delayed Vitrectomy for Nondiabetic Vitreous Hemorrhage

Abstract

Purpose: To analyze the surgical and visual outcomes in patients who underwent early versus delayed vitrectomy for dense nondiabetic, nonvascular vitreous hemorrhage from undetermined cause. Methods: A retrospective, consecutive series evaluating all patients with a preoperative diagnosis of nondiabetic vitreous hemorrhage treated surgically at The Retina Institute over the past 5 years. Vitreous hemorrhage had to be described as dense, with first-order vessels not visualized on funduscopic examination. Exclusion criteria included a diagnosis of diabetes mellitus, retinal vascular disease, any prior intraocular surgery other than cataract extraction, or any previously diagnosed ocular conditions. A total of 275 patients and 286 eyes were evaluated, with 52 patients and 52 eyes meeting inclusion criteria. Surgical treatment was with standard pars plana vitrectomy. Results: Of the 52 eyes that met inclusion criteria, 32 (61.5%) were found to have at least 1 retinal break. Sixteen eyes (30.1%) had a rhegmatogenous retinal detachment. Mean time from diagnosis of the vitreous hemorrhage to surgical intervention was 15 days (range, 0-78 days). Early vitrectomy was defined as surgical intervention within 10 days of diagnosis, while late vitrectomy was defined as surgical intervention following 10 days. Of eyes that were found to have a retinal detachment, 75% were phakic and 25% were pseudophakic. Final postoperative visual acuity in the early surgery cohort was 20/66, while in the late surgery cohort was 20/89, which did not achieve statistical significance ( P = .67). Conclusion: A significant percentage of patients with retinal detachment were phakic at the time of vitreous hemorrhage (75%) as compared to pseudophakic. Data from the largest series to date indicate that in phakic patients with dense hemorrhage, the risk of retinal detachment is high and should prompt consideration for more urgent surgical intervention.