ABSTRACT.Purpose: To compare the frequency of rejection and graft failure after Descemet's stripping automated endothelial keratoplasty (DSAEK) and penetrating keratoplasty (PK) for secondary endothelial failure. Methods: Seventy-eight eyes undergoing DSAEK from 2006 to 2013 and 80 eyes undergoing PK from 1998 to 2013 were included. Indications were endothelial failure due to pseudophakic bullous keratopathy (82%), previous eye trauma (8%), uveitis (4%) or other causes (6%). Rejection episodes and graft failure were recorded, together with preoperative risk factors (glaucoma, neovascularization and active inflammation). Kaplan-Meier survival curves with up to 4 years follow-up were used to compare DSAEK with PK, as well as low-risk with high-risk eyes. Results: During follow-up, 15% of the PK-treated eyes and 4% of the DSAEKtreated eyes experienced rejection. Graft failure occurred in 8% of the PKtreated eyes and in 7% of the DSAEK-treated eyes. No significant differences were found in the rate of rejection (p = 0.11), graft failure (p = 0.29) or rejection-related graft failure (p = 0.30) between DSAEK and PK in low-risk eyes. In DSAEK, high-risk eyes (mainly with glaucoma) experienced significantly more rejection episodes (p = 0.01), graft failures (p = 0.04) and rejectionrelated graft failures (p = 0.04) compared with low-risk eyes. No differences were seen between high-risk and low-risk PK-treated eyes (p > 0.40). Conclusion: In the present, relatively small study, no differences in rejection rate or graft failure were found between DSAEK and PK in low-risk eyes. However, glaucoma seems to increase the risk of graft complications after DSAEK for secondary endothelial failure. Long-term studies of larger samples sizes are needed to confirm this.