Outcomes and risk factors associated with endophthalmitis after intravitreal injection of anti-vascular endothelial growth factor agents.

OBJECTIVE To describe outcomes of and risk factors for endophthalmitis after intravitreal anti-vascular endothelial growth factor (VEGF) injection. DESIGN Single-center, consecutive, case series and retrospective case-control study. PARTICIPANTS Between January 1, 2009, and May 31, 2010, 16 vitreoretinal surgeons administered a total of 27 736 injections. During this period, 23 cases of presumed infectious endophthalmitis occurred. Each surgeon used his own preferred injection technique. INTERVENTION Vitreous or aqueous tap, or both, with intravitreal antibiotic injection and subsequent topical antibiotic and steroid drops. MAIN OUTCOME MEASURES Visual acuity, bladed lid speculum use, conjunctival displacement, hemisphere of injection, bevacizumab versus ranibizumab, and infectious organism. RESULTS Seven of 23 cases had positive culture results; 3 grew coagulase-negative Staphylococcus. All cases had pain and vitritis on average 3.4 days (range, 1-6 days) after injection, with no difference between culture-positive and culture-negative groups. Eighteen (78%) of 23 cases had a hypopyon. Fifteen of 23 cases returned to baseline vision (±2 lines) within 3 months. Neither lid speculum use (0.10% vs. 0.066% in the no-use group; P = 0.27), conjunctival displacement (0.11% vs. 0.076% in the no-displacement group; P = 0.43), hemisphere of injection (0.11% superior vs. 0.079% inferior; P = 0.56), or bevacizumab versus ranibizumab (0.11% vs. 0.066%; P = 0.21) affected risk. Analysis of only culture-positive results yielded similar results. There was no statistically significant difference between the proportion of culture-negative cases after bevacizumab injection (83%) versus ranibizumab injection (55%; P = 0.13). CONCLUSIONS Most patients in whom presumed infectious endophthalmitis develop after anti-VEGF injection regained baseline vision after treatment. Bladed lid speculum use, conjunctival displacement, hemisphere of injection, and type of anti-VEGF agent did not affect risk. No difference in culture-negative endophthalmitis rates was detected after bevacizumab versus ranibizumab injection. Neither the presence of pain, vitritis, decreased vision, hypopyon, nor the interval between injection and development of symptoms differentiate culture-positive from culture-negative cases. Because a subgroup of patients had poor outcomes, a low threshold for vitreous tap with intravitreal antibiotic injection may be warranted. FINANCIAL DISCLOSURE(S) The author(s) have no proprietary or commercial interest in any materials discussed in this article.