Abstract:AIM: To evaluate the efficacy and safety of autologous serum enriched with topical insulin compared with topical insulin alone and autologous serum alone in patients with persistent corneal epithelial defects (PEDs) unresponsive to standard therapy. METHODS: This study included 45 eyes of 45 patients with PEDs persisting for more than 14d despite conventional treatment. Patients were randomized into 3 equal groups (all, n=15): autologous serum enriched with insulin 1 IU/mL (Group A, 8 males), autologous serum 20% (Group B, 7 males), and insulin 1 IU/mL (Group C, 9 males). Treatments were administered six times daily until complete epithelial closure. The primary outcome was time to full epithelial healing. Secondary outcomes included best-corrected visual acuity (BCVA), epithelial thickness on anterior segment optical coherence tomography (AS-OCT), recurrence, and tolerance. RESULTS: The mean ages in Groups A, B, and C were 58.4±12.1y, 59.7±11.4y, and 57.9±13.0y, respectively (P=0.92). Complete epithelial healing was 100%, 86.7%, and 93.3% of eyes in Groups A, B, and C, respectively (P=0.76). Mean healing time among healed eyes (n=15, 13, and 14) differed significantly across groups [one-way analysis of variance (ANOVA), P=0.007]; post-hoc Tukey analysis showed that Group A (9.5±2.8d) healed significantly faster than Group B (14.3±4.5d; P=0.005), with a non-significant difference versus Group C (11.8±3.9d; P=0.24). BCVA improved in all groups, with the greatest gain in Group A (0.89 to 0.52 logMAR). Optical coherence tomography (OCT) demonstrated earlier epithelial thickening and homogeneity in Group A. Recurrence within 1mo among healed eyes was low (6.7%–15.4%) with no significant difference among the 3 groups. All treatments were well tolerated without adverse events. CONCLUSION: In this pilot study, insulin-enriched autologous serum is associated with significantly faster epithelial healing than autologous serum alone. A numerically shorter healing time is observed compared with topical insulin alone, but this difference is not statistically significant. The combination is well tolerated over the study period. Given the small sample size, these preliminary findings should be interpreted with caution and require confirmation in larger, multicenter, adequately powered randomized controlled trials.