Objective To analyze the risk factors for persistent corneal epithelial defects (PCED) after pars plana vitrectomy (PPV) in patients with proliferative diabetic retinopathy (PDR). Methods A total of 201 PDR patients (201 eyes) who received PPV were enrolled in this retrospective study. There were 86 males (86 eyes) and 115 females (115 eyes). The patients aged from 30 to 81 years, with the mean age of (57.94±9.65) years. Among them, 159 patients were ≥50 years of age, and 42 patients were ＜50 years of age. There were 36 patients with HbA1c ＜7.0%, 165 patients with HbA1c ≥7.0%. There were 93 right eyes and 108 left eyes. There were 93 right eyes and 108 left eyes. The diabetic retinopathy stages were as follows: stage Ⅳ in 24 eyes, stage Ⅴ in 78 eyes and stage Ⅵ in 99 eyes. The operation time was ranged from 1 to 4 hours, with an average of 2 hours. Among the 201 eyes, corneal epidermis were scraped on 25 eyes; 70 eyes were combined with cataract surgery; a laser photocoagulation count ＜1000 points was performed in 78 eyes, and ＞1000 points were performed in 123 eyes. Sixty-one eyes involved intravitreal silicone oil tamponade, 18 eyes involved intravitreal tamponade with C3F8, and 122 eyes were not involved with intraocular tamponade. Postoperative persistent intraocular hypertension was defined as an intraocular pressure (IOP) ≥21 mmHg (1 mmHg=0.133 kPa) after PPV with necessary treatment using IOP-lowering medications for ≥2 weeks. The diagnostic criteria for corneal epithelial defects were taken from the Expert Consensus on Clinical Diagnosis and Treatment of Corneal Epithelial Defect in China (2016). The corneal epithelial defect was diagnosed as PCED if it was treated with common methods such as a lacrimal substitute or corneal contact lens, but showed no improvement and no signs of healing for ≥2 weeks. The incidence of PCED after eye surgery was recorded and its related risk factors were analyzed. A multivariate logistic regression was used to analyze the risk factors for PCED, which were expressed as a odds ratio (OR) and a 95% confidence interval (CI). Results Of 201 eyes, 16 eyes (7.96%) presented with PCED after surgery and 185 eyes (92.04%) with no PCED. There was no significant difference in the age, sex, and eyes between the patients with or without PCED (χ2=6.548, 0.927, 0.044; P=0.011, 0.336, 0.833). A multivariate logistic regression showed that intraoperative epithelial debridement (OR=13.239, 95%CI 2.999−58.442, P=0.001), intraoperative treatment in combination with cataract surgery (OR=7.448, 95%CI 1.975−28.091, P=0.003), intravitreal tamponade with C3F8 (OR=11.344, 95%CI 2.169−59.324, P=0.004), and postoperative persistent intraocular hypertension (OR=10.462, 95%CI 2.464−44.414, P=0.001) were risk factors for PCED after PPV. Conclusion Intraoperative epithelial debridement, intraoperative treatment in combination with cataract surgery, intravitreal tamponade with C3F8, and postoperative persistent intraocular hypertension are risk factors for PCED in patients with PDR after PPV.