Purchase this article with an account.
G. Thuret, N. Delesalle, J. Dubus, L. Fleury, L. Mouillot, P. Gain; Determination of Corneal Endothelial Cell Density in French Eye Banks: Second National Review. Invest. Ophthalmol. Vis. Sci. 2010;51(13):1168.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
In European eye banks, 15 to 20% of corneas are discarded for endothelial cell density (ECD) < 2000 cells/mm². Given the importance of having a precise, robust and reproducible ECD counting method, the Afssaps, organized during 2008-09 a national review of the quality of ECD determination in the 18 french eye banks.
The study reproduced a first review driven in 2003 by St-Etienne (Transplantation2004;78:1299). Five test corneas consisting of flat mounted alizarin red stained corneal endothelium (ECD: 825, 1185, 2047, 2262, and 2952 cells/mm2, by 2 calibrated methods) were successively sent to the eye banks. All the technicians counted using their routine method(s). A questionnaire gathered data about the methods of corneas preparation, counting (manual or by image analysis), and microscope calibration.
Only 5 banks (28%) used image analysis to validate their grafts (manual counting for the others). Corneas preparation and counting methods were not standardized. Test slides: 430 counts were done by 70 eye banks technicians. Forty-two % deviated by more than 10% from the expected ECD. Among them, 128 were over-estimated (max +88% with 1550 given for a real ECD of 825 cells/mm2) and 52 were under-estimated (min -31%, with 812 given for a real ECD of 1185). Two banks constantly over-estimated (in the mean +32% and +43%, no calibration and/or material problem) but the 16 other banks were in average within +/-13% from expected ECDs. For manual counts only, a statistically significant inter-banks difference was observed. ECD by image analysis were more reproducible and closer to the expected values than manual ones. Compared to 2003, reliability of ECD determination globally improved.
ECD measurement appears acceptable in all but two French eye Banks. Methods for corneal preparation and counting strategies should be standardized. Image analysis systems prove more reliable (precise and low intra and inter rater variability) than manual methods. Recommendations will be edited to improve endothelial quality assessment of corneal grafts.
This PDF is available to Subscribers Only