[Prognosis in repeat keratoplasty: per indication analysis in a large monocentric cohort]

Data from national graft registries suggests a limited prognosis in the group of all repeat keratoplasties. This group might rather consist mostly of cases already classified as high risk in the first place, however. Thus, repeat keratoplasties in normal risk indications might have a substantially better prognosis. In particular, any specific increase of immune reactions following repeat keratoplasty as compared to first keratoplasties is unknown at present. Furthermore, the contribution of chronic endothelial cell loss to graft failure following repeat keratoplasty has not yet been investigated. The aim of this analysis on a large retrospective cohort is to estimate the incidence of graft rejections and failures als well as chronic endothelial cell loss following repeat keratoplasty.

All keratoplasties that had been performed from 1986 to 2006 at the University Eye Hospital Düsseldorf were selected for the 4 most relevant indications. A total of 2073 first keratoplasties and 236 repeat keratoplasties for keratoconus, Fuchs endothelial dystrophy, bullous keratopathy and herpetic corneal scars were included in the analysis. For each indication, the incidence of graft rejections and failures was compared between first and repeat keratoplasty using the Kaplan-Meier method and log-rank test. Endothelial cell loss was quantified by means of exponential regression analysis in the cases without immune reactions. Data of 65 repeat keratoplasties and 532 initial keratoplasties were available for analysis in a generalised linear model.

After five years, the risk of graft rejections was not increased in repeat keratoplasties for keratoconus as compared to initial keratoplasties. In Fuchs endothelial dystrophy (p = 0.03) and herpetic corneal scars (n. s.) an increase of 11 % was observed. In bullous keratopathy the increase was as high as 24 % (p < or = 0.01). Irreversible graft failure increased by 4 % in keratoconus, 11 % in Fuchs endothelial dystrophy, 7 % in herpetic eye disease and by 31 % in bullous keratopathy. Only the latter difference reached statistical significance (p < or = 0.01), however. In this group, chronic endothelial cell loss was accelerated to a great extent as well: annual cell loss averaged 43 % in repeat keratoplasties whereas this rate was as low as 20 % for the initial surgery (p = 0.01). In the other indications, acceleration of cell loss in repeat keratoplasties as compared to the first transplantation did not exceed 12 % (p = 0.03).

The prognosis of repeat keratoplasties for keratoconus does not seem to fall short of that of initial keratoplasty. A slight deterioration can be assumed in Fuchs endothelial and herpetic eye disease dystrophy. Repeat keratoplasties in initial bullous keratopathy, however, seem to have a substantially poorer prognosis with respect to immune reactions, graft survival and chronic endothelial cell loss.