Amniotic Membrane Transplantation for Corneal Perforation: A Systematic Review and Meta-analysis with a Proposed Size-based Treatment Algorithm

Corneal perforation is a sight- and globe-threatening emergency, and amniotic membrane transplantation (AMT) is widely used for tectonic reconstruction. However, the influence of perforation size on outcomes has not been synthesized within a unified framework. In accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses statement, we systematically reviewed studies reporting the clinical outcomes of AMT for corneal perforation, descemetocele, impending perforation, or severe corneal melting and performed a random-effects proportion meta-analysis (logit transformation, DerSimonian– Laird) of anatomical success. The primary analysis was restricted to endpoint-harmonized early anatomical closure of perforations ≤ 3 mm, whereas a broader sensitivity analysis additionally incorporated studies reporting first-procedure or recurrence-free outcomes. Of the 279 records identified, 16 full-text reports were assessed for eligibility, and 14 studies were included (7 in the quantitative synthesis). In the primary analysis (5 studies, 131 eyes, all ≤ 3 mm), the pooled anatomical success rate was 94.4% (95% CI 76.1–98.9%; I² = 62.7%), whereas in the extended sensitivity analysis (7 studies, 177 eyes), it was 89.8% (95% CI 74.9–96.3%; I² = 59.0%). Reported success was consistently high in perforations ≤ 3 mm, frequently approaching 100%, whereas quantitative data on AMT without major structural adjuncts for perforations > 3 mm were essentially absent. Narrative synthesis demonstrated a progression from conventional multilayer AMT to fibrin glue–assisted, Swiss-roll, pleats-fold, and umbilical cord patch–assisted reconstruction for larger or more complex defects. AMT is an effective globe-preserving intervention for small corneal perforations (≤ 3 mm), whereas current evidence supporting AMT alone for larger perforations is insufficient. We propose a provisional, hypothesis-generating, size-based treatment framework whose ≤ 3 mm component is supported by the available evidence and whose > 3 mm component and finer size thresholds require prospective validation.