Efficacy and Clinical Predictors of Topical NSAIDs as Bridging Therapy for Diabetic Macular Edema: A Systematic Review and Meta-Analysis
In the authors' words
Introduction & Objectives: Diabetic Macular Edema (DME) is a primary cause of vision loss. While anti-VEGF injections are the gold standard, prolonged referral delays in primary care settings often lead to irreversible photoreceptor damage. This study aims to evaluate the efficacy of topical Non-Steroidal Anti-Inflammatory Drugs (NSAIDs) as a temporary bridging therapy and to develop a new clinical scoring system to guide primary care physicians in early intervention. Methods: A systematic review was conducted to identify clinical predictors of responses to topical NSAIDs in DME patients, followed by a meta-analysis to calculate pooled Odds Ratios (OR) for each predictor. These values underwent natural log transformation (beta = ln(OR)) to derive a point-based clinical scoring system. Result: Twelve studies were analyzed. Predictor variables and assigned points derived from pooled ORs included: glycemic control (HbA1c < 8.0%; OR 2.41, 9 points), symptom duration (< 6 months; OR 2.11, 7 points), baseline visual acuity (> 6/12; OR 1.96, 7 points), and retinopathy stage (mild-to-moderate NPDR; OR 1.77, 5 points). The proposed 28-point scoring system categorizes patients into three tiers: High Probability (21-28 points) for immediate NSAID initiation; Moderate Probability (12-20 points) for combined systemic optimization; and Low Probability (0-11 points) for prioritized emergency referral. Evidence confirms that potent topical NSAIDs effectively stabilize macular thickness and prevent exudative progression during the referral window. Conclusion: This study concludes that the OR-derived scoring system effectively empowers primary care physicians to manage DME during referral windows. Topical NSAIDs serve as a safe, low-risk bridging strategy to mitigate permanent vision loss and reduce the burden of blindness in resource-limited settings.
Appeared: Friday, November 6. medRxiv. Preprint, not yet peer-reviewed.