A Nine-Year Analysis of Practice Patterns, Microbiologic Yield, and Clinical Outcomes in Cases of Presumed Infectious Endophthalmitis

Purpose: To analyze practice patterns, microbiologic yield, and clinical outcomes in cases of presumed infectious endophthalmitis at a single tertiary academic referral center. Design: Retrospective chart review. Participants: Eyes diagnosed with endophthalmitis at the Duke Eye Center between January 1, 2009, and January 1, 2018, that underwent aqueous tap, needle vitreous tap, or mechanical vitreous biopsy with pars plana vitrectomy (PPV) in conjunction with intravitreal antimicrobials. Methods: Charts of 133 eyes of 130 consecutive patients were identified and reviewed for etiology of endophthalmitis, ocular history, interventions, visual outcomes, complication rates, and microbiologic culture results. Main Outcome Measures: Microbiologic yield and corrected visual acuity (VA) at initial presentation and last follow-up (up to 1 year) were the primary outcome measures. Secondary outcome measures included need for subsequent intervention, retinal detachment (RD) rate, and antimicrobial sensitivities. Results: Of the 133 cases of presumed infectious endophthalmitis, 74% were unrelated to cataract surgery. Initial management included intravitreal antimicrobials with aqueous tap in 46% of eyes, needle vitreous tap in 59% of eyes, and mechanical vitreous biopsy with PPV in 14% of eyes. Twenty-seven percent of cases that underwent initial aqueous or needle vitreous tap required subsequent intervention with PPV. Overall, 45% of endophthalmitis cases were culture proven. Needle vitreous tap and mechanical vitreous biopsy with PPV were more likely to yield culture growth compared with aqueous tap (P = 0.042, P = 0.004, respectively). All isolated bacteria were sensitive to the combination of intravitreal vancomycin and ceftazidime. Eyes that underwent initial mechanical vitreous biopsy with PPV had worse VA at presentation (P = 0.024) and at last follow up (P = 0.011) compared with eyes that underwent initial aqueous tap or needle vitreous tap. An RD occurred within 3-12 months in 13% (17/133) of cases, and of these, 41% had an endogenous etiology. Conclusions: Needle vitreous tap and mechanical vitreous biopsy were more effective in yielding positive cultures compared with aqueous tap; however, culture results may not influence initial intravitreal antimicrobial choice. Eyes with presenting VA between counting fingers and hand motion at 1 foot were more likely to undergo initial PPV. (C) 2020 by the American Academy of Ophthalmology