Therapeutic keratoplasty for corneal perforation.

To the Editor: We read with interest the article by Gorovoy et al regarding the outcome of 2 patients who had secondary graft failure after Descemet stripping automated endothelial keratoplasty (DSAEK). Although the report documents the successful visual rehabilitation of the patients after late endothelial graft failure, we would like to request the clarification of several points arising from these cases. A note of concern is that both the DSAEK grafts lasted only 1 and 2 years, respectively. Were the cases in the surgeons’ ‘‘learning curve’’ for DSAEK? Since the patients underwent a second allograft, did they receive any oral postoperative immunosuppressive therapy or did the surgeons involved intend to use a prolonged course of topical steroids? Did the removed donor lamella undergo any histological examination to look for activated keratocytes to corroborate the clinical evidence of no interface haze. We have also experienced 2 cases of late endothelial failure after DSAEK. These both occurred after folded forceps insertion before we switched to a glide insertion technique. One of the cases has since undergone graft exchange using a previously described technique. Prefailure vision before late endothelial failure was 20/25. After uncomplicated graft exchange and glide insertion, 2 months post-op the patient’s vision is back to 20/25. We agree with Gorovoy et al who did not note any interface haze or scarring. The removed donor tissue underwent electron microscopic evaluation (Fig. 1). As expected, the endothelial cells were abnormal with increased coefficient of variation of cell area and reduced cell density. However, highmagnification assessment of the keratocytes showed 2 different phenotypes. A mixture of activated and degenerated (Figs. 1C, D) keratocytes were found. This finding is in keeping with cases of late endothelial failure after penetrating keratoplasty where an activated keratocyte response was also found and different to primary graft failure after DSAEK where all the keratocytes had degenerated. In the present case, the majority of activated keratocytes were found in the posterior part of the donor lamella, adjacent to Descemet membrane. The histological evidence corroborates the finding of no interface haze