Screening donor eyes for keratoconus (KC) and previous PRK or LASIK treatment is becoming more and more important as significantly more than 1 million LASIK procedures are performed per year on a world wide basis, with the number still rapidly growing. Eyes that have undergone these procedures are clearly unsuitable for transplantation, and the clinical changes caused by these procedures are barely detectable by slit-lamp examination even immediately after surgery, if at all. The major requirement for a sufficiently reliable screening by computer-assisted video-keratography (CAVK) is a corneal surface of sufficient quality for obtaining reproducible surface maps.
In this eye-bank study, we evaluated and now propose the use of a score to reflect epithelial surface integrity. In this study, we also tested the usefulness of the newly released CAVK system Keratron Scout (Optikon 2000, Rome, Italy), for donor-eye screening.
MATERIALS AND METHODS
Forty eyes from 20 donors were enucleated with great care for the complete preservation of the epithelium within 24 hours postmortem (which is the usual time frame in our eye bank). The 12 o'clock position on each eye was marked by suturing the rectus superior muscle. In no case was epithelium removed because all corneas were suitable for transplantation. Fifteen donors were male, five female. The mean age of the donor was 61.7 years, with an age range of 7–88 years. The mean time from death of the donor to enucleation was 15.4 hours (± 5.6 h). The corneal surface and stroma were examined with a hand-held slit lamp, and epithelial defects and other corneal pathology were noted in detail. The low intraocular pressure regularly found in postmortem eyes was raised to within a normal range by injecting sterile, balanced salt solution (BSS) into the vitreous cavity. An eye pressure of approximately 22 mm Hg was aimed for and carefully checked with applanation tonometry (Fig. 1), with great care taken to not induce epithelial irregularities or defects. Balanced salt solution was used to keep the corneas moist before corneal topography was performed. The whole procedure was performed under sterile conditions to maintain asepsis and suitability for transplantation.
Screening of all eyes was performed within 30 minutes after enucleation and immediately after raising the intraocular pressure. An “OR trolley” model of the Placido-based corneal topography system Keratron Scout (Optikon 2000), which allows measurements in horizontal orientation, was used to assess corneal topography (Fig. 2, 3).
The Placido image consists of 28 rings; 7168 points on the corneal surface are measured, and more than 75,000 analyzed. Image capture is triggered by an infrared (IR) beam that intercepts the corneal vertex. Corneal geometry reconstruction is based on an arc-step algorithm, producing “true curvature” color maps with high accuracy. 1 Eight pictures per eye were taken with the Keratron Scout (Optikon 2000) in all cases. The measurements were centered on the pupil. The repeatability was tested with the “repeatability check,” an integrated feature of the Keratron Scout (Optikon 2000) software (Fig. 4). Images that were significantly out of range were discarded, if proposed by the program. One image of the remaining series of each eye was randomly selected and used for analysis. According to topographic regularity (and taking into consideration the slit-lamp findings), corneas were classified by using a score that grades epithelial surface regularity (Table 1). Corneas with a completely regular surface and with no “disturbances” in their topography maps within the 7-mm zone were classified as belonging to “group 0” (Fig. 5A). Corneas with an irregularity in just one quadrant were classified as “group 1” (Fig. 5B). Corneas with irregularities in two quadrants were considered “group 2” (Fig. 5C). Those showing irregularities in three quadrants were “group 3” (Fig. 5D), and corneas with a completely irregular topography or an irregularity in the corneal center were classified as “group 4” (Fig. 5E). Two observers (J.S. and J.R.) classified each topography map separately and independently. In cases of disagreement between these observers, final scoring was achieved by a joint review of the topography map.
Topography maps were then categorized as “normal,” “myopic laser surgery suspect” (displaying central flattening with associated midperipheral steepening), “hyperopic laser surgery suspect,” or “keratoconus suspect” (the latter two displaying excessive central corneal steepening).
Central corneal thickness was measured by ultrasonic pachymetry (DGH 2000, DGH Technology Inc., Exton, PA, U.S.A.). Measurement was performed 10 times per eye, and the mean value was taken for analysis.
To detect significant differences for the comparison of the five score groups, the ANOVA model was applied with a type I error rate of 5%. 2 It is well known that multiple comparisons demand the use of a type I error rate adjustment to protect against an increase in the overall type I error rate. 2 The p value adjustment was done with the Bonferroni-method for each table to end up with a reasonable overall type I error rate of 5%. 3 Hence the result of one t test was considered as statistically significant, if and only if the corresponding p value was less than 0.05/3. 3 To find a relation among different pairs of variables (age, postmortem time, pachymetry), the corresponding correlation coefficients were computed.
The 15 corneas displaying a regular epithelial surface were considered to belong to group 0. Fourteen corneas with just minor irregularities were categorized as belonging to group 1. Four corneas were classified into group 2, three corneas into group 3, and four corneas into group 4 (Fig. 6). Of the latter, in one pair of donor eyes (61-year-old male donor, enucleation 14 hours postmortem), analysis of corneal topography was impossible because of bullae in the epithelium; in another eye the optic nerve was cut too short making it impossible to “normalize” the intraocular pressure. The corneal topography abnormalities correlated with the epithelial defects noted by the slit lamp. There was no statistically significant correlation between the epithelial regularity score and age of the donor or time from death to enucleation. Mean pachymetry was 732 ± 76 μm. No statistically significant correlation could be found between central corneal thickness and donor age or time from death to enucleation.
Condition of corneal surface allowed reliable screening with CAVK in 36 eyes (corneas of group 0, 1, 2, 3). The mean simulated K-reading of all eyes was 43.64 (±1.54) D, and the mean astigmatism was 1.26 (±0.91) D. The mean simulated K reading was 44.3 (±1.47) D in group 0, 43.15 (±1.57) D in group 1, 42.29 (±0.81) D in group 2, and 43.99 (± 1.14) D in group 3. Mean astigmatism was 0.89 (±0.61) D in group 0, 1.18 (±0.81) D in group 1, 2.54 (±1.16) D in group 2, and 1.94 (±0.95) D in group 3. Analysis of topography maps showed neither excessive corneal steepening typical for KC nor central flattening with associated midperipheral steepening, which is typically found after myopic laser surgery.
Keratoconus, a chronic noninflammatory corneal thinning disorder leading to scarring and progressive thinning, has a reported incidence between 54.5 and 230 per 100,000 in the general population. 4 Although late stages of the disease process certainly may be detected by the established routines of an eye bank, early cases of KC may not easily be identified. Several authors have proposed an undetected KC in the donor cornea as an explanation for a “recurrence” in the graft. 5–8
With the increasing number of refractive procedures worldwide, there is growing concern that refractively altered corneas will make their way into eye banks. Because the morphologic changes after a LASIK or PRK procedure in most cases are barely detectable by slit-lamp examination in the living eye and probably even less so in the postmortem cornea, there is an substantial risk of including such a cornea in a transplantation schedule. 9
Todd et al. tried to project the effect of refractive procedures on the future corneal donor pool. They calculated that by 2020 almost 50 million Americans might have undergone laser vision correction, which would at that time exclude one in four potential donors, considering the demographic overlap between donors and refractive patients. 10
Several authors have reported the use of various CAVK-systems for donor-eye screening. 11–14 This is the first report on the use of the Keratron Scout (Optikon 2000) in an eye bank. The main advantage of this system is its high degree of accuracy resulting from its arc-step algorithm and the high number of measured and analyzed points. 1 The possibility of performing measurements in the horizontal orientation makes screening of an enucleated eye much easier, because no globe-fixing devices are needed. In addition, the scout topographer is obtainable as a battery-operated hand-held model, which makes it feasible to perform screening for abnormal shapes in the morgue before any surgical step is taken, with the exception of pressurization of the globe. This is an important innovation because many eye banks do not enucleate the whole globe but prefer in situ excision of the corneoscleral disc.
The most important criterion for a reliable screening by CAVK is the quality of the epithelium. Because corneal transplantation is preferably performed with an intact epithelial layer, screening methods should leave the corneal surface intact, and avoiding abrasion of the epithelium.
Several factors can adversely affect the postmortem quality of the epithelium, causing an irregular corneal surface, e.g. lagophthalmus leading to a localized dryness of the surface, or postmortem edematous changes in the epithelium increasing the risk of mechanical damage. Enucleation of the donor eye should be performed with great care so as not to induce scratches or warpage on the epithelium.
We found it critically important to constantly moisturize the corneal epithelium with fluid before topography was performed to achieve a “smooth” surface. The lubricant should reduce the “roughness” of the postmortem epithelial surface; however, its viscosity should not preclude the detection of specific details of the topography. We found balanced salt solution appropriate for this purpose.
With the proposed “epithelium regularity score,” a cut-off point has to be set to distinguish between safely usable corneas and those with an irregular corneal surface that would lead to unreliable screening. We believe that safe screening seems to be possible in eyes with a “score 2,” whereas in corneas with irregularities in three quadrants, each case must be decided on an individual basis. Topography maps of “score 4” corneas are clearly not useful for these screening efforts.
As the eye pressure regularly and rapidly is diminished to low levels after death, the globe loses its shape with the cornea collapsing. It is therefore mandatory to reestablish a near-normal eye pressure for these screening purposes when using a placido-based topography system. This can be achieved by injecting balanced salt solution into the vitreous cavity under sterile conditions to maintain suitability for transplantation. Furthermore, tonometry should not result in damage of the epithelial surface. Systems that use a piezoelectric pressure transducer for intraocular pressure measurement would be ideal but usually are not available to a standard eye bank. The careful use of a hand-held sterilized surgical applanation tonometer seems to represent an acceptable alternative.
Mean central corneal thickness in our sample was 732 μm, which is similar to the pachymetry results of other studies. 11 Because of the irregularity of the swelling, one cannot deduce preoperative corneal thickness from postoperative pachymetry findings. Therefore postenucleation central corneal pachymetry findings alone cannot be the only criteria used to discriminate “normal” from “refractive” or “keratoconic” corneas with their characteristic central corneal thinning over time. The relatively low numbers in each group may account for the inability to find a statistically significant correlation between pachymetry results and the time from death to enucleation in this study. Several authors analyzed the influence of corneal swelling on corneal topography, and their results must be taken into consideration to assess the validity of postmortem topography. This is quite important, because false-positive and false-negative screening results should be minimized.
Ousley et al. 15 and Simon et al. 16 found corneal steepening in corneal topography of de-epithelialized eye-bank eyes after dehydration with dextran, but changes were clinically relatively small (<1D). Maloney found flattening of cadaver corneas of 0.02 D for every 10 μm of thickening. 17 Because the epithelium was removed before corneal topography was evaluated in the three studies mentioned, the results could not be completely compared with those in our study, in which the epithelial layer was left in place. Other authors found no significant topographic changes after experimentally induced corneal edema in living subjects. 18
Screening donor corneas for KC and refractive corneal surgery with CAVK would add an additional safety feature to the existing screening procedure including slit-lamp examination, blood screening, review of donor case histories, microbiologic studies, and evaluation of endothelial cell density. This system easily could be installed in any eye bank and could be handled by technicians. We have shown that the Keratron Scout (Optikon 2000) could become a useful tool for this screening purpose, but additional studies and tests are needed to determine the postmortem sensitivity and specificity in detecting those eyes that already have undergone ablative refractive procedures.
1. Mattioli R, Tripoli NK. Corneal geometry reconstruction with the Keratron Videokeratographer. Optom Vis Sci 1997; 74: 881–94.
2. Neter J, Wassermann W, Kutner M. Applied linear statistical models: Regression, analysis of variance and experimental design
, 3rd ed New York, NY: Irwin Series in Statistics, 1990.
3. Miller J. Simultaneous statistical interference
, 2nd ed. New York: Springer Verlag, 1981.
4. Rabinowitz YS. Keratoconus
. Surv Ophthalmol 1998; 42: 297–319.
5. Rubinfeld RS, Traboulsi EI, Arentsen JJ, et al. Keratoconus
after penetrating keratoplasty. Ophthalmic Surg 1990; 21: 420–2.
6. Eiferman RA. Recurrence of keratoconus
[letter]. Br J Ophthalmol 1984; 68: 289–90.
7. Tuft SJ, Gregory W. Long-term refraction and keratometry after penetrating keratoplasty for keratoconus
1995; 14: 64–7.
8. Belmont SC, Muller JW, Draga A, et al. Keratoconus
in a donor cornea
[letter]. Refract Corneal Surg 1994; 10: 658.
9. Mannis MJ, McDonough G, Howard K, et al. Screening donor eyes that have undergone PRK
1997; 16: 683–5.
10. Todd M, Nickel MS, Timothy B, et al. Projected effect of LASIK
on the future corneal donor pool [abstract]. Cornea
2000; 19: 257.
11. Terry MA, Ousley PJ. New screening methods for donor eye-bank eyes. Cornea
1999; 18: 430–6.
12. Terry MA, Ousley PJ, Rich LF, et al. Evaluation of prior photorefractive keratectomy in donor tissue. Cornea
1999; 18: 353–8.
13. Lim-Bon-Siong R, Williams JM, Samapunphong S, et al. Screening of myopic photorefractive keratectomy in eye bank eyes by computerized videokeratography. Arch Ophthalmol 1998; 116: 617–23.
14. Schelonka LP, Ogawa GSH. Corneal topography
of human cadaver eyes donated months after radial and astigmatic keratotomy. Cornea
1997; 16: 689–92.
15. Ousley PJ, Terry MA. Hydration effects on corneal topography
. Arch Ophthalmol 1996; 114: 181–5.
16. Simon G, Small RH, Ren Q, et al. Effect of corneal hydration on Goldmann applanation tonometry and corneal topography
. Refract Corneal Surg 1993; 9: 110–7.
17. Maloney RK. Effect of corneal hydration and intraocular pressure on keratometric power after experimental radial keratotomy. Ophthalmology 1990; 97: 927–33.
18. Rom ME, Keller WB, Meyer CJ, et al. Relationship between corneal edema and topography
. CLAO J 1995; 21: 191–4.