BACKGROUND Although Descemet-stripping automated endothelial keratoplasty has replaced penetrating keratoplasty for primary treatment of endothelial disorders, many patients have already undergone penetrating keratoplasty. It is unclear when repeat penetrating keratoplasty is necessary or when endothelial keratoplasty may restore clarity to a failed graft. METHODS Retrospective case series of patients undergoing Descemet-stripping automated endothelial keratoplasty after penetrating keratoplasty by three surgeons at an academic tertiary care centre. METHODS Eight patients with Descemet-stripping automated endothelial keratoplasty after penetrating keratoplasty from 2006 to 2009. METHODS Microkeratome-prepared Descemet-stripping automated endothelial keratoplasty donor tissue was used. In seven cases, the penetrating keratoplasty bed was neither stripped nor scraped, and in one, scraping only was performed. METHODS Preoperative and 6-month postoperative best-corrected visual acuities in logMAR (logarithm of the minimum angle of resolution). RESULTS The average pre-Descemet-stripping automated endothelial keratoplasty best-corrected visual acuity was 1.375, and the average best-corrected visual acuity 6months postoperatively was logMAR 1.0, a 2.5-fold improvement in the minimum angle of resolution (P=0.22). Seven of the eight patients showed an improvement in best-corrected visual acuity, and one patient had failure of Descemet-stripping automated endothelial keratoplasty and required penetrating keratoplasty. Five had a postoperative event: one had a gap that resolved spontaneously, three required rebubblings (injections of air only without otherwise repositioning the graft), and one experienced graft failure. CONCLUSIONS Descemet-stripping automated endothelial keratoplasty can successfully rescue a prior penetrating keratoplasty, even with a fairly high detachment rate. Given these favourable visual outcomes, further study of this promising strategy is justified.