DSAEK after previous failed PK yields low graft dislocation rate

Cornea. 2011;30(4):414-418.

Issue: May 25, 2011

Graft dislocation of Descemet stripping automated endothelial keratoplasty after previous failed penetrating keratoplasty proved comparable to that of primary DSAEK, a study found.

"Our series represents the largest to date, multicenter, retrospective chart review of DSAEK in eyes with prior failed [PK]," the study authors said.

The retrospective study included 97 eyes of 90 patients who underwent DSAEK after previously failed PK. Average patient age was 73 years (range: 47 to 96 years).

Results showed that graft dislocation occurred in 30 eyes (31%); 29 eyes required re-bubbling and one graft reattached spontaneously. Ninety-eight percent of grafts remained attached for the entire follow-up period. Two eyes (2%) required repeat DSAEK after primary graft failure.

Endothelial grafts dislocated in 10 of 15 eyes with glaucoma drainage devices (67%) and 20 of 82 eyes without drainage devices (24%); the difference was statistically significant (P = .009).

The graft dislocation rate was 24% in eyes with grafts larger than the host PK graft, 18% in eyes with grafts equal to the host and 42% in eyes with grafts smaller than the host. Between-group differences were statistically insignificant.

Dislocation occurred in five of 21 eyes (24%) with sutures remaining after PK and in 22 of 76 eyes (29%) from which sutures had been removed.

Except in 13 eyes with visual worse than 20/400 and severe ocular comorbidities, Snellen visual acuity improved an average of four lines, the authors reported.

While the published DSEK dislocation rates do reach into the 30% range, most DSEK surgeons’ dislocation rates are currently in the 10% to 20% range. Additionally, surgeons performing DSEK on previous PKs are very likely to be experienced DSEK surgeons. Take home message: I believe the dislocation rate of 31% after PK is higher than expected after DSEK in non-PK eyes. Surgeons and patients should know this.

Initially it was taught that the DSEK must be larger than the PK. Then it was taught that the DSEK must be smaller than the PK. This article showed a trend to fewer dislocations when the DSEK was larger than the PK, although it did not reach statistical significance. It does not appear to me as if multivariate analysis was performed to remove the confounding variable of tube shunts. It is possible one size group had many more tube shunts than another, which would skew the results. Take home message: We do not know what size DSEK is best after PK.

– Christopher J. Rapuano, MD
Chief, Cornea Service, Wills Eye Institute, Philadelphia
Disclosure: Dr. Rapuano has no relevant financial relationships.