Further studies on the retrocorneal membrane-endothelium relationship.

Partial penetrating keratoplasty may fail for a number of reasons. By improved surgical technique, instrumentation, therapy, and methods of managing donor material some of these reasons are suppressed, thereby increasing the relative importance of the remainder. Hence post-graft membranes (PGMs) have come to occupy a high place in the pathology of corneal grafts (Rycroft, I965; Brown and Kitano, I966; Kurz and D'Amico, I968). Post-graft membrane is the name applied to errant fibrous tissue which often develops after full-thickness keratoplasty from the host/graft junction and forms a sheet of variable dimensions and density behind the graft (Werb, I962). Similar structures which occur after penetration of the posterior corneal layers by mechanical injury other than grafting, and by degenerative processes have been called post-trauma membranes (PTM) and postdisease membranes (PDM) respectively (Sherrard and Rycroft, I 967a). The three aetiological types are grouped under the heading retrocorneal membrane (RCM). Other retrocorneal structures can be classified on anatomical criteria. There are those comprising iris tissue, fibrin, epithelium, cell and blood debris, or a combination of them. The membranes which are under investigation here are those composed of corneal stromal elements. Occasionally they are enhanced by anterior iris synechia but this is to be regarded as a secondary feature. In earlier reports (Sherrard and Rycroft, I967a, b), it was suggested that the corneal endothelium prevents and does not augment PGM growth, and it was mentioned that total PGMs result after the complete removal of the graft endothelium in rabbits. Since that time one hundred further experiments, involving destruction of the corneal endothelium adjacent to retrocorneal wounds in rabbits, have been carried out. Basically the results agree with those of similar investigations by Brown and Kitano (I966), and have not, therefore, been published. However, they greatly strengthen the view that RCMs will develop only in the absence of functional endothelium. Figs I and 2 illustrate PGM formation after partial (Fig. i) and total (Fig. 2) removal of the graft endothelium. The RCM-endothelium relationship has been investigated under more closely controlled conditions. Small fragments of corneal stroma were explanted into the anterior chamber. The explants represented potential RCMs and were either placed directly against healthy endothelium or totally separated from it. They were all of approximately the same size to standardize the mass of potential RCM and all were placed well away from the iris and penetrating corneal wounds to avoid their involvement in iris synechia, foreign body (suture) reaction, vascularization, oedema, epithelialization, and contamination by any PTM issuing from a large wound. Initially only autoexplants were prepared to evade immunological complications. The results demonstrate that the integrity of the endothelium is essential to the prevention of retrocorneal fibrosis of stromal origin.