Post‑injection endophthalmitis (PIE) is an infrequent but potentially devastating complication of intravitreal injections (IVI). The reported incidence per injection ranges from 0.02% to 0.09%, and because patients typically receive multiple injections, the cumulative risk per eye can approach higher values over time. PIE may be infectious or sterile, commonly occurs 3–4 days after injection, and can lead to serious sequelae including retinal detachment and permanent vision loss.
Ocular surface antisepsis—most frequently with povidone–iodine or chlorhexidine—is a routine component of PIE prophylaxis. Beyond antisepsis, substantial variation exists in clinical practice regarding other preventive measures. Surveys of retina specialists have reported wide ranges in the use of pre‑ or post‑injection topical antibiotics, sterile eyelid speculums, sterile drapes, masks during the procedure, sterile gloves, and performing IVIs in operating room environments.
Previous systematic reviews have largely focused on single prophylactic interventions evaluated in isolation. Such pairwise syntheses do not account for the combined strategies commonly used in practice. A network meta‑analysis (NMA) provides a framework to compare multiple prophylactic strategies simultaneously and to estimate relative effects for strategies not directly compared in primary studies.
This protocol outlines a plan to conduct a systematic review with frequentist random‑effects network meta‑analysis to compare the effects of all available prophylactic strategies, alone or in combination, for the prevention of PIE in adults receiving IVI of pharmacologic agents (excluding injections given for treatment of bacterial endophthalmitis).
Specific outcomes to be addressed include:
The protocol also specifies procedures for study selection, data extraction, risk of bias assessment, statistical analysis, and grading the certainty of evidence.
Eligibility criteria will include randomized controlled trials and observational studies that compare any prophylactic strategy for PIE against an alternative strategy, placebo, or no prophylaxis in adults aged 18 years or older who received IVI of any pharmacologic agent. Studies where IVI was performed to treat bacterial endophthalmitis are excluded.
Information sources specified in the protocol include MEDLINE, EMBASE, Cochrane CENTRAL, Web of Science, and ClinicalTrials.gov, searched from inception. Paired reviewers will independently screen titles, abstracts, and full texts; extract data; and assess risk of bias for each prespecified outcome.
When feasible for each outcome, the review will conduct frequentist random‑effects NMAs to estimate relative effects across multiple prophylactic strategies and their combinations. The protocol commits to assessing the certainty of the resulting network estimates using the Grading of Recommendations Assessment, Development and Evaluation (GRADE) approach and to interpreting findings in that context.
The increasing global use of intravitreal therapies—anti‑VEGF agents, corticosteroids, anti‑complement therapies, and others—has made PIE an increasingly relevant safety concern in ophthalmic practice. Given the low absolute incidence per injection but frequent cumulative exposure, identifying effective prophylactic approaches is critical to minimize patient harm.
Surveys of clinical practice reveal marked heterogeneity in prophylactic routines beyond antisepsis, including topical antibiotics, use of sterile eyelid speculums and drapes, provider masking, glove use, and setting of the procedure. This variability underscores the need for a comprehensive comparative synthesis that reflects real‑world combinations of measures.
The protocol adheres to PRISMA‑P reporting standards and is registered on the Open Science Framework. The planned reporting of the completed review will follow PRISMA with the network meta‑analysis extension. Key methodological steps in the protocol include:
The protocol notes that no datasets were generated or analyzed at the protocol stage; relevant data will be made available upon completion of the systematic review.
The authors propose that a systematic review paired with an NMA will provide a comprehensive and current synthesis of the comparative effects of prophylactic strategies for PIE. By explicitly accommodating combinations of measures, the planned NMA aims to reflect the complexity of real‑world practice and to permit estimation of relative effects for strategies lacking direct head‑to‑head evidence.
The protocol recognizes heterogeneity in practice patterns and in published study designs and commits to transparent methodology—registration, adherence to PRISMA guidance, dual review processes, and formal assessment of evidence certainty with GRADE—to enhance trustworthiness of conclusions.
This review protocol is registered on the Open Science Framework: https://doi.org/10.17605/OSF.IO/SJC4U. The authors indicate they will report the completed systematic review in accordance with PRISMA and its network meta‑analysis extension.
The protocol states that the authors received no specific funding for this work. Several authors disclosed multiple consulting relationships, research grants, and advisory roles with industry organizations; the protocol lists these competing interests in detail. The source notes that these declarations do not alter adherence to PLOS ONE policies on data and material sharing.