AIM:To examine which anesthesia general or local is more effective for penetrating keratoplasty (PKP).METHODS: Patients with indications for PKP (n=141) were enrolled in a prospective study and randomly divided into general anesthesia group (group A, 70 eyes) and local anesthesia group (group B, 71 eyes). Patients received optical PKP (group A1, 30 eyes; group B1, 30 eyes) or therapeutic PKP (group A2, 40 eyes; group B2, 41 eyes). Measurement of anterior chamber treatment time (T) for PKP patients and the ratio (R) of the area of the pupils to that of recipient graft region. T and R values, as well as perioperative and postoperative complications, were compared between groups A and B using t-test or χ² test.RESULTS: Patients were followed for 2wk after PKP. T was (13.45±8.64)min for group A and (7.36±5.24)min for group B, a statistically significant difference (P<0.001). The R value for group A was stable during the operation, while for PKP patients in group B the value initially increased then gradually decreased to normal after suturing. In group B, extrusion of intraocular contents occurred in 5 eyes, and iridal prolapse occured in 11 cases; no perioperative complications occurred in group A. Relapse rate for fungal keratitis was 13.04% in group B and 0% in group A.CONCLUSION: Under general anesthesia, pupils remaine stable during PKP and perioperative complications are averted. General anesthesia gives more time to treat pathological changes in the anterior chamber and treatment success rate is higher.