Background: Assessment of corneal tissue quality before graft is mainly based upon the determination of endothelial cell density (ECD) by eye banks. These cells are responsible for corneal transparency, and ECD correlates with graft survival. In France and often elsewhere in Europe, ECD is measured using a "naked-eye" procedure under a light microscope. To measure objectively the reliability of ECD determination in France, we developed four test corneas with a known ECD.
Methods: The test corneas consisted of 1 mm2 of human corneal endothelium with stained cell borders. The 64 technicians of the 21 French eye banks counted according to the protocol applied in their respective centers.
Results: More than half of the 256 counts (152, 59%) deviated by more than 10% from actual ECD. Of the counts, 85 (33%) were over-estimated, and 67 (26%) were under-estimated. Deviation ranged between 42% under-estimation and 82% over-estimation. Eight banks (38%) constantly over-estimated, and nine (43%) under-estimated ECD. Half of the inter-technician gaps within an eye bank were more than 10%, with a maximum of 51%.
Conclusions: This audit highlights the unacceptable lack of reliability of manual ECD assessment in French eye banks. This surely indicates the delivery of poor quality corneas for graft in certain centers and wastage in others. We urgently advocate normalization of French counting methods. This may require upgrading to a computer-aided method.