Purpose
To evaluate systematic differences in ECD measurements between non-contact specular microscopy, and sucrose assisted light microscopy, in organ cultured human donor corneas in the Ghent University Hospital Cornea Bank.
Methods
Measurements obtained between 1997 and 2011 in 1016 corneas from 551 donors were analyzed. Only donor corneas with both specular microscopy EDCs and sucrose-assisted light microscopy ECDs were included. In 193 corneas, Topcon SP1000 specular microscopy ECDs were compared to sucrose ECDs. In 167 corneas. Topcon SP2000P (Topcon Corp, Tokyo, Japan) ECDs were compared to sucrose ECDs. A manual counting technique using calibrated graticules on printed photographs was used for SP1000 and Sucrose ECDs, and a center counting method on the instrument itself was used for SP2000P ECDs. Bland-Altman plots and estimates were used for analysis, based on a linear mixed model analysis allowing for paired corneas.
Results
Results are shown in figures 1 and 2. SP1000 ECDs had a mean difference of 254.5 cells/mm2 (lower) with sucrose ECDs, the limits of agreement (horizontal black lines) were + 274.3 and - 783.2. SP2000P ECDs had a mean difference of 71.2 cells/mm2 (lower) with sucrose ECDs, with limits of agreement: + 805.2 and -947.6. Unequal SD’s of two paired measurements cause a correlation between the sum and the mean of those two measurements (Pitman’s test). Hence the negative slope in the SP1000-sucrose ECD difference vs. mean regression line , and positive slope in the SP2000-sucrose ECD difference vs. mean regression line (sloped red lines)
Conclusions
Substantial systematic differences and huge individual differences exist between ECDs obtained with different measurement methods. Specular microscopy in donor eyes resulted in only a minority of cases in usable ECDs, and because of this and larger endothelial cell counting sample sizes, sucrose ECDs remain the gold standard in ECD determination in organ cultured donor corneas. Erroneous magnification calibration in SP1000 was shown to result in a larger systematic difference with sucrose ECD. The small systematic error between SP2000P ECDs and sucrose ECDs indicate that comparison of donor ECDs to in-vivo ECDs may be justifiable, on the condition that instruments are calibrated correctly. Huge individual measurement differences may occur.
Keywords: 481 cornea: endothelium •
741 transplantation