Background: Chronic endothelial cell loss of the graft is common after penetrating keratoplasty. Some kind of subclinical immunological reaction that is not visible at the slitlamp has been suspected as main cause for this phenomenon. Furthermore, migration of graft endothelial cells towards the host cornea has been discussed to add to this loss in special cases. In this study, 3 homogenous patient groups with similar risk of immunological reactions were examined. Main difference between these groups was the potential of graft endothelial cells to migrate towards the host cornea.
Patients and methods: Patients with keratoconus without cataract surgery (group I with little migration potential, n = 273), patients with Fuchs endothelial dystrophy without/with cataract surgery (groups IIa/IIb with moderate migration potential, n = 89/n = 165) and patients with bullous keratopathy after cataract surgery (group III with potentially large migration tendency, n = 188) were included in the study. All patients had a first keratoplasty. Patients with glaucoma or further intraocular procedures after keratoplasty were excluded from the study. Clear graft survival and ratio of grafts without immune reactions were estimated according to Kaplan and Meier. Endothelial analysis concerned only patients without immune reactions and with at least 3 postoperative endothelial cell density values of the graft center (76 patients in I, 18 patients in IIa, 41 patients in IIb and 23 patients in III).
Results: Mean relative loss of endothelial cells per year was 14.0 +/- 19.0 % in group I, 17.0 +/- 19.1 % in group IIa, 20.8 +/- 18.2 % in group IIb and 29.4 +/- 17.6 % in group III (ANOVA, p < 0.01). Five years postoperatively in group I 99 %, in group IIa 98 %, in group IIb 93 % and in group III 69 % of the grafts were centrally clear (log rank test, p < 0.001). In the same period in group I 88 %, in group IIa 86 %, in group IIb 83 % and in group III 81 % of the grafts were free of immune reactions (log rank test, p < 0.05). Reasons for irreversible graft failure were immune reactions (0 in group I; 0 in group IIa; 1 in group IIb; 9 in group III, ), surface disorders (1 in group I; 0 in IIa; 1 in group IIb; 3 in group III) and endothelial failure (0 in group I; 1 in group IIa; 5 in group IIb; 6 in group III) (chi square test, p < 0.01).
Conclusions: In patients with bullous keratopathy endothelial cell loss via migration seems to add significantly to the general chronic loss which is suspected to be immunological. Peripheral migration of endothelial cells, therefore, most probably contributes to limited prognosis of patients with bullous keratopathy in terms of clear graft survival. In consequence, corneal grafts for bullous keratopathy should be as large as immunologically tolerable, and endothelial cell density should be as high as possible in order to counteract this special endothelial loss factor.