Author Affiliations: Departments of Ophthalmology (Drs Patel and Bourne) and Health Sciences Research (Ms Diehl and Mr Hodge), Mayo Clinic, Rochester, Minnesota; and Department of Biostatistics, Mayo Clinic, Jacksonville, Florida (Ms Diehl).
To determine donor risk factors for graft failure and late endothelial failure (LEF) 20 years after penetrating keratoplasty (PK).
Prospective, observational study of 500 consecutive PKs performed by one surgeon. After excluding eyes undergoing second grafts, fellow eyes of bilateral cases, and patients who withdrew research authorization, 388 eyes (388 patients) were available for analysis. At specific intervals during 20 years of follow-up after PK, the central corneal endothelium was analyzed by specular microscopy and risk factors for graft failure and LEF were determined.
The 41 grafts examined at 20 years represented 22% of the available clear grafts. Mean (SD) donor endothelial cell loss since the preoperative examination was 74% (9%), with no change in endothelial cell density between 15 and 20 years (P = .93; 36 eyes). Eighty-three grafts were known to have failed, of which 26 were because of LEF. Transplant recipients with a diagnosis of keratoconus had a lower risk of graft failure (P = .03) and LEF (P ≤ .03) compared with those with endothelial dysfunction. Neither the method (McCarey-Kaufman or K-Sol media [n = 275] or organ culture [n = 113]) nor the length of donor preservation was associated with graft failure or LEF. Lower preoperative endothelial cell density (P = .02) and higher endothelial cell loss at 2 months (P = .004) increased the risk of LEF for a recipient diagnosis of endothelial dysfunction but not keratoconus.
Recipient diagnosis is a risk factor for graft failure. The length of corneal preservation is not associated with graft failure or with LEF, but these results should be confirmed in a study in which donor corneas are preserved using current methods and with more complete follow-up.
The number of corneal transplants performed in the United States in 2008 increased by 22% to almost 42 000 compared with 2006.1 Most of the increase can be attributed to the rapid adoption of endothelial keratoplasty (EK) over penetrating keratoplasty (PK),2,3 with 18 000 EK procedures performed in 2008 compared with 6000 in 2006.1 Consequently, the demand for donor corneal tissue has increased, and many surgeons, especially outside the United States, often have to accept tissue with longer preservation times or with lower preoperative endothelial cell densities (ECDs) compared with previously.
In this report, we updated the status of the corneal endothelium at 20 years after PK in our long-term prospective study. We also investigated the relationships between donor preservation time (hereinafter referred to as death to transplant time) and graft failure and between donor ECD and morphometry and late endothelial failure (LEF).4,5
Patients attending the cornea service at Mayo Clinic, Rochester, were enrolled in this prospective observational study. The cohort, which has been described in detail previously,6- 8 included 500 consecutive patients who underwent a PK performed by one of us (W.M.B.) between 1976 and 1986. From the 500 grafts, 36 repeat grafts, 70 grafts in fellow eyes, and 6 patients who withdrew research authorization were excluded from the study, leaving 388 grafts in 388 patients available for analysis (Table 1). At 20 years after PK, 83 grafts were known to have failed and 115 patients were known to have died; of the remaining 190 patients, 41 (22%) were examined at 20 years.
This study adhered to the tenets of the Declaration of Helsinki and was approved by the Mayo Clinic Institutional Review Board. Informed consent was obtained from all patients after explanation of the nature and possible consequences of the study.
Donor corneas were preserved in McCarey-Kaufman medium9 or K-Sol medium at 4°C10 or were placed in organ culture at 34°C11 before transplantation. Death to preservation time was the sum of enucleation time (time from donor death to enucleation) and moist chamber time (time from enucleation to preservation). Death to transplant time was the sum of the death to preservation and storage times (time from the onset of preservation or organ culture to transplantation).
The surgical technique has been reported in detail previously.12 It involved trephination of the donor by a punch from the endothelial side and suturing with a running or an interrupted technique.
The central donor corneal endothelium was photographed by using a specular microscope before preservation or organ culture. Follow-up examinations were scheduled for 2 months and 1, 3, 5, 10, 15, and 20 years after keratoplasty. At each visit, the central corneal endothelium was photographed and the corneal thickness was measured by contact specular microscopy. Patients were examined to detect complications, including graft failure and LEF.
The outlines or apices of at least 50 endothelial cells were digitized from specular micrographs. During the 30 years of analysis for this study, all endothelial images were analyzed by 3 consecutive observers; interobserver and intraobserver variations were assessed by analyzing a training series of images, and both were determined to be satisfactory over the course of the study. We calculated the mean endothelial cell area, mean ECD, coefficient of variation (SD [mean]) of endothelial cell area, and percentage of cells that were hexagonal. Endothelial cell loss was defined as the decrease in ECD between the preoperative and 20-year examinations, expressed as a percentage of the preoperative ECD. The annual rate of endothelial cell loss between 15 and 20 years was calculated by assuming that the endothelial cell loss between these examinations was exponential (first order), using the following equation:
ECD20 = ECD15e−rt,
where the subscripts 15 and 20 indicate the postoperative year, r is the rate coefficient (similar to the annual rate of endothelial cell loss for this interval13), and t is the time in years. During the interval from 15 to 20 years, t = 5 years.
Graft failure was defined as an irreversible loss of graft clarity from any cause, including LEF, which was defined as loss of graft clarity unassociated with a recent graft rejection episode and unresponsive to corticosteroid therapy,6 that is, presumably from endothelial cell attrition.
Differences in the ECD, coefficient of variation of cell area, percentage of hexagonal cells, and corneal thickness at 15 and 20 years after PK were compared using paired t tests if the data were distributed normally or Wilcoxon signed rank tests if the data were not distributed normally. We used Pearson correlation coefficients to evaluate the relationship between normally distributed continuous variables and Spearman correlations for nonnormal variables. Graft failure and LEF estimates were constructed by using the Kaplan-Meier method. Cox proportional hazards models were used to evaluate possible risk factors for graft failure, including donor age and death to transplant time. Preoperative donor endothelial cell variables and early postoperative endothelial cell loss were assessed as possible risk factors for LEF. Risk factors were analyzed for all eyes and according to preservation method and recipient diagnosis; endothelial dysfunction included recipient diagnoses of Fuchs endothelial dystrophy, pseudophakic corneal edema, and aphakic corneal edema. A 2-tailed P ≤ .05 was considered statistically significant. Unless otherwise indicated, data are expressed as mean (SD).
Of the 41 patients examined at 20 years after PK, the preoperative diagnosis was keratoconus in 28 (68%) and endothelial dysfunction in 7 (17%) (Table 1). Seven grafts were known to have failed between 15 and 20 years after PK. One hundred forty-nine patients who were presumed to be alive and without graft failure were not examined at 20 years (Table 2). Of these patients, the preoperative diagnosis was keratoconus in 38 (26%) and endothelial dysfunction in 88 (59%).
At 20 years after PK, mean donor endothelial cell loss since the preoperative examination was 74% (9%) (41 eyes; Table 3). The ECD at 20 years after keratoplasty was 798 (232) cells/mm2, which did not differ from ECD at 15 years (872  cells/mm2; P = .93; the minimum detectable difference was 93 cells/mm2 [α = .05; β = .20; 36 eyes]). The ECD at 20 years ranged from 427 to 1779 cells/mm2; 6 grafts had an ECD of less than 600 cells/mm2, and 1 graft had an ECD of less than 500 cells/mm2. The mean rate of endothelial cell loss between 15 and 20 years after surgery was 0.2% (4.7%) (n = 36).
Thirteen patients attended all follow-up visits through the 20-year period and had no episodes of graft rejection or reoperations that might have affected the corneal endothelium (Table 4); recipient diagnoses were keratoconus (n = 11), Fuchs dystrophy (n = 1), and pseudophakic corneal edema (n = 1). Donor endothelial cell loss since the preoperative examination was 72% (15%). Endothelial cell density at 20 years after keratoplasty was 835 (328) cells/mm2, which did not differ from cell density at 15 years (809  cells/mm2; P = .10; the minimum detectable difference was 176 cells/mm2 [α = .05; β = .20; 13 eyes]). The mean rate of endothelial cell loss from 15 to 20 years after surgery was 0.06% (4.1%) (n = 13).
Donor age correlated with preoperative donor hexagonality (r = −0.44; P < .001; 384 eyes) and with the preoperative donor coefficient of variation of cell area (r = 0.29; P < .001; 387 eyes). The significant relationships remained through 5 years for hexagonality (r = −0.18; P = .01; 182 eyes) and through 3 years for the coefficient of variation of cell area (r = 0.14; P = .03; 226 eyes), but not thereafter.
Corneal thickness did not change between 15 and 20 years after PK (P = .13; the minimum detectable difference was 0.01 mm [α = .05; β = .20; 36 eyes]), but it was thicker at 20 years compared with 2 months after PK (P < .001; 41 eyes) (Table 3). The increase in corneal thickness in the postoperative period was higher for corneas preserved in McCarey-Kaufman medium compared with the corneas preserved in organ culture (P = .04, Student-Newman-Keuls procedure) (Table 5).
Graft failure was known to have occurred in 83 eyes during the 20-year observation period. Late endothelial failure was the leading cause of graft failure, accounting for 26 of the 83 known graft failures (31%) and 20 of 27 graft failures (74%) between 5 and 20 years after PK. The cumulative probability of developing overall graft failure and LEF at 20 years was 30% (95% confidence interval [CI], 23%-36%) and 13% (8%-19%), respectively (Figure).
Cumulative probability of developing overall graft failure and late endothelial failure (LEF) after penetrating keratoplasty. At 20 years, the cumulative probabilities of developing overall graft failure and LEF were 30% and 13%, respectively.
Median donor age was 43 (range, 1-72) years for the overall group, 38 (range, 1-68) years for eyes undergoing transplantation for keratoconus, and 43 (range, 2-72) years for eyes undergoing transplantation for endothelial dysfunction. For all eyes, univariate analyses showed that donor age was associated with overall graft failure (P = .04; a 10-year increment in age increased the risk of failure by 1.14; 95% CI, 1.00-1.28) but not with LEF (P = .07; minimum detectable hazard ratio [HR], 1.39 [α = .05; β = .20]). Donor age was not associated with LEF for eyes undergoing transplantation for keratoconus (P = .29) or for endothelial dysfunction (P = .18).
Median death to transplant time for the corneas stored in McCarey-Kaufman medium was 1.9 days (or 45 hours; range, 1-193 hours; 216 corneas); for the corneas stored in K-Sol medium, 5.6 days (or 134 hours; range, 11-312 hours; 59 corneas); and for the corneas stored in organ culture, 24.2 days (or 580 hours; range, 239-971 hours; 113 corneas). For all 388 eyes, death to transplant time was not associated with graft failure (P = .07) or with LEF (P = .26). When eyes were stratified according to preservation method, death to transplant time was not associated with graft failure overall or with LEF (Table 6). Death to transplant time was not associated with graft failure overall for recipient diagnoses of keratoconus (P = .27), Fuchs dystrophy (P = .32), or endothelial dysfunction (P = .08). Death to transplant time was associated with LEF for a recipient diagnosis of keratoconus (12-hour increment in death to transplant time increased the risk of LEF by 1.06; 95% CI, 1.01-1.12; P = .03) but not Fuchs dystrophy (P = .71) or endothelial dysfunction (P = .99). Stratifying eyes by recipient diagnosis and preservation method resulted in no association between death to transplant time and graft failure overall; a similar subanalysis for LEF was not possible because of the small number of LEF outcomes within each analysis.
For all eyes, death to transplant time did not correlate with the percentage of endothelial cell loss since the preoperative examination at 2 months (r = 0.10; P = .06; 348 eyes), 5 years (r = −0.12; P = .12; 182 eyes), 10 years (r = −0.14; P = .12; 117 eyes), 15 years (r = −0.12; P = .33; 66 eyes), or 20 years (r = 0.17; P = .29; 41 eyes) after PK. Subanalyses of correlations between death to transplant time and endothelial cell loss within each method of preservation were also not statistically significant.
For all eyes, lower preoperative ECD by 500 cells/mm2 increased the risk of LEF by 1.95 (P = .006; 95% CI, 1.22-3.13), and lower preoperative hexagonality by 10% increased the risk of LEF by 1.47 (P = .03; 95% CI, 1.04-2.07). Preoperative coefficient of variation was not associated with increased risk of LEF (P = .15). For all eyes at 2 months after surgery, lower ECD by 500 cells/mm2 increased the risk of LEF by 1.65 (P < .001; 95% CI, 1.25-2.18). The ECD at 2 months after surgery did not predict LEF in eyes undergoing transplantation for keratoconus (P = .90) but did predict LEF in eyes undergoing transplantation for Fuchs dystrophy (P = .005; lower ECD by 500 cells/mm2 increased the risk by 2.16; 95% CI, 1.26-3.72) (Table 7). Postoperatively, coefficient of variation and hexagonality at any interval through 3 years were not associated with subsequent LEF (P > .13).
In univariate analyses, a recipient diagnosis of Fuchs dystrophy increased the risk of LEF by 5.68 compared with a recipient diagnosis of keratoconus (P = .009; 95% CI, 1.55-20.82), and a recipient diagnosis of pseudophakic or aphakic corneal edema increased the risk of LEF by 4.73 compared with a recipient diagnosis of keratoconus (P = .03; 1.14-19.6). The risk of LEF associated with a recipient diagnosis of Fuchs dystrophy did not differ from that for a recipient diagnosis of pseudophakic or aphakic corneal edema. Intraoperative use of sodium hyaluronate, 1%, as a viscoelastic agent was not associated with graft failure overall (P = .26) or with LEF (P = .14).
In a multivariate analysis that included donor preservation method and time, donor and recipient ages, preoperative donor ECD, and recipient diagnosis, only recipient diagnosis was a risk factor for overall graft failure; there were too few outcomes to perform a multivariate analysis for LEF. A recipient diagnosis of keratoconus had a lower risk of overall graft failure than did a diagnosis of Fuchs dystrophy (HR, 0.41; 95% CI, 0.18-0.91; P = .03) or pseudophakic corneal edema (0.37; 0.16-0.89; P = .03). The risk of graft failure between eyes undergoing transplantation for Fuchs dystrophy and for pseudophakic corneal edema did not differ (HR, 0.91; 95% CI, 0.45-1.85; P = .80).
Mean endothelial cell loss between 15 and 20 years after PK was 0.2% per year and was similar to the rate of 0.6% per year of endothelial cell loss in normal corneas.14 Individual grafts varied widely from the mean during the second decade after PK as indicated by the large standard deviations of endothelial cell loss. Despite our small number of surviving grafts at 20 years after PK, mean ECD and morphometry remained relatively stable between 15 and 20 years and were accompanied by stable corneal thickness. The stability in ECD and cell loss should be interpreted with caution because grafts that failed were excluded from subsequent examinations, which might have lowered the apparent rate of cell loss.
Older donors slightly increased the risk of overall graft failure in a univariate analysis but were not significant in a multivariate analysis after adjusting for recipient diagnosis. Donor age was not a risk factor for LEF, for which the smallest detectable HR with our statistical power was 1.39; by restricting the same analysis to recipient diagnoses of endothelial decompensation, which have a moderate risk of failure compared with keratoconus, we still found no relationship between donor age and LEF, although our statistical power was reduced further. At 5 years after PK, the Cornea Donor Study (CDS), which included recipient diagnoses of endothelial decompensation only, found no relationship between donor age and overall graft failure.15 Although these early results of the CDS suggest that the donor pool can be expanded by using older donors, the study has been extended for 10-year follow-up to determine longer-term outcomes. Indeed, the significant univariate relationship between donor age and graft failure at 20 years in the present smaller study was evident at 10 years after surgery7 but not at 5 years (P = .41),6 indicating the need for longer-term analysis.
Increasing the death to transplant time is an alternative method of expanding the donor pool, which might become necessary as the demand for donor tissue increases. Our data showed no evidence that longer death to transplant times were associated with graft failure overall or with LEF. Optisol-GS (Bausch & Lomb Surgical, Irvine, California), which currently is the preferred preservation medium in the United States, was not available when the patients in this study received their transplants. Optisol-GS contains dextran and chondroitin sulfate,16 whereas McCarey-Kaufman medium contained dextran but not chondroitin sulfate, and K-Sol medium contained chondroitin sulfate but not dextran.10 Given these differences, we do not expect that the results of our study would have been different if Optisol-GS had been available. In a retrospective study with 3-year follow-up, Wagoner and Gonnah17 found no relationship between prolonged preservation time (all >7 days) in Optisol-GS and graft failure, and, similarly, the Singapore Corneal Transplant Study found no relationship between preservation time in Optisol-GS and graft failure.18 The findings in our study provide additional evidence that increased death to transplant time has no effect on graft failure, which is important information for surgeons who import donor tissue or accept donor tissue with extended preservation times. Data from the Minnesota Lions Eye Bank, St Paul, show that the death to transplant time increased from 4.5 days in 2006 to 5.4 days in the first half of 2009 and that the increase coincided with the availability and demand for precut donor tissue for EK (Raylene Streed, BA, Technical Director, personal communication). Our data provide some reassurance that it is acceptable to use donor tissue with extended preservation times, and indeed, Optisol-GS was approved by the US Food and Drug Administration for preservation through 14 days.
More than 100 corneas in our study were preserved by using organ culture, which is the primary method of corneal preservation in Europe. We found no association between the duration of organ culture, which ranged from 10 to 40 days, and graft failure, LEF, or endothelial cell loss at any examination through 20 years. We did not examine endothelial cell loss during the preservation period, but, in a randomized paired-donor study, Thuret et al19 found that cell loss was higher with 21 to 34 days of organ culture compared with 5 to 12 days, although cell loss after PK did not differ during the first year. With short-term follow-up, Frueh and Böhnke20 suggested that extended donor tissue culture was appropriate for elective surgery; our long-term data support their conclusion and indicate that organ culture can be an effective method of expanding the donor pool.
A previous study by our group found that lower preoperative ECD and increased endothelial cell loss at 2 months were risk factors for LEF and speculated that improvements in corneal preservation might render endothelial cells less susceptible to surgical trauma.5 The same relationships remained significant through 20 years for all eyes, but, in a subanalysis according to recipient diagnosis, the relationship was only significant for endothelial dysfunction and not for keratoconus. Possible explanations are the larger number and better quality of recipient endothelial cells in the eyes with keratoconus compared with those with endothelial dysfunction or that our study lacked power to detect a relationship for the small keratoconus group. Data from the CDS will confirm or refute our findings with respect to endothelial dysfunction. At 20 years after PK, we also found that decreased preoperative hexagonality (ie, increased pleomorphism) was associated with a higher risk of LEF. Increased pleomorphism and polymegethism are apparent in corneas of contact lens wearers21,22 and in diabetic patients23 and have been suggested to represent stress to the endothelium.24 Similarly, it is conceivable that increased pleomorphism in the preop erative donor cornea is an indicator of endothelial cell stress during preservation and might help determine which donors are susceptible to higher cell loss during surgery. We also found significant associations between donor age and endothelial morphologic characteristics, suggesting that endothelial cell stress might be more apparent in older donors. In future prospective studies, changes in morphometry should be reported during the preservation period and postoperatively to determine their value. The length of donor preservation was not associated with higher early endothelial cell loss.
Central graft thickness gradually increased during the observation period, but the increase in thickness was higher for corneas preserved in McCarey-Kaufman medium compared with those preserved by organ culture. In a paired donor study, Rijneveld et al25 also found that corneas preserved in McCarey-Kaufman medium became thicker than the fellow corneas preserved by organ culture at 14 years after PK. They found no difference in ECD, suggesting that the difference in corneal thickness was not attributable to endothelial function. Corneas in organ culture are more edematous and typically are preserved longer than corneas in cold storage, and both of these factors have been associated with increased proteoglycan loss during preservation.26 This proteoglycan loss during preservation might explain the relationship between thinner grafts in the early postoperative period and longer preservation times27 and might also result in reduced swelling pressure and therefore thinner corneas in the longer term. Regardless of the etiology, the difference in corneal thickness in the small study by Rijneveld et al25 was not associated with a difference in visual acuity.
During the 30-year study period, loss to follow-up caused by patient death, patient unavailability, and graft failure has increased such that only 22% of the presumed surviving grafts were available for examination at 20 years. Our data should therefore be interpreted with caution because the loss to follow-up may have introduced bias and, as a result, the rate of endothelial cell loss and risk of graft failure might in fact be higher than reported. In addition, keratoconus accounted for 68% of the clear grafts examined at 20 years after PK, whereas only 21% of the original 388 grafts were performed for keratoconus. Surviving grafts in the second decade of this study were biased toward recipients with a diagnosis of keratoconus because of their younger age, lower risk of graft failure, and easier ability to return for follow-up examinations. Despite these inevitable limitations after 2 decades of follow-up, our prospective data provide valuable insight into endothelial cell survival after keratoplasty and the risk of graft failure. Whether our data will be directly applicable to endothelial cell loss and LEF after EK, which has rapidly become the procedure of choice for treating endothelial dysfunction, remains to be determined.3 Endothelial cell loss during the first 6 months after EK is much higher than that after PK,28 and our data would therefore predict higher rates of LEF after EK than after PK.5 Nevertheless, endothelial cell loss between 6 months and 3 years after EK might in fact be significantly lower than that after PK,29 which might mitigate our prediction if the lower rate of cell loss is sustained over the long term. Differences in graft size and wound configuration between EK and PK might explain the lower cell loss after 6 months.
In summary, we have examined this cohort of patients for more than 20 years after PK, and lower preoperative donor ECD and higher early endothelial cell loss, but not increasing donor age, increased the risk of LEF after PK. Corneal preservation time did not affect endothelial cell loss or predict graft failure. Although the CDS will definitively answer the question about donor age and graft failure, the effect of preservation time on graft failure warrants a randomized, prospective study to confirm our findings and support extended preservation as a method of expanding the donor pool.
Correspondence: Sanjay V. Patel, MD, Department of Ophthalmology, Mayo Clinic, 200 First St SW, Rochester, MN 55905 (firstname.lastname@example.org).
Submitted for Publication: July 6, 2009; final revision received October 7, 2009; accepted October 15, 2009.
Financial Disclosure: None reported.
Funding/Support: This study was supported in part by grant EY02037 from the National Institutes of Health; an unrestricted grant from Research to Prevent Blindness, Inc (Department of Ophthalmology, Mayo Clinic, Rochester); an Olga Keith Wiess Special Scholar grant from Research to Prevent Blindness, Inc (Dr Patel); and the Mayo Foundation.
Previous Presentations: This study was presented in part at the XVII International Congress of Eye Research; October 30, 2006; Buenos Aires, Argentina, and at the Federated Societies Scientific Session; November 10, 2006; Las Vegas, Nevada.
Thank you for submitting a comment on this article. It will be reviewed by JAMA Ophthalmology editors. You will be notified when your comment has been published. Comments should not exceed 500 words of text and 10 references.
Do not submit personal medical questions or information that could identify a specific patient, questions about a particular case, or general inquiries to an author. Only content that has not been published, posted, or submitted elsewhere should be submitted. By submitting this Comment, you and any coauthors transfer copyright to the journal if your Comment is posted.
* = Required Field
Disclosure of Any Conflicts of Interest*
Indicate all relevant conflicts of interest of each author below, including all relevant financial interests, activities, and relationships within the past 3 years including, but not limited to, employment, affiliation, grants or funding, consultancies, honoraria or payment, speakers’ bureaus, stock ownership or options, expert testimony, royalties, donation of medical equipment, or patents planned, pending, or issued. If all authors have none, check "No potential conflicts or relevant financial interests" in the box below. Please also indicate any funding received in support of this work. The information will be posted with your response.
Some tools below are only available to our subscribers or users with an online account.
Download citation file:
Web of Science® Times Cited: 19
Customize your page view by dragging & repositioning the boxes below.
Enter your username and email address. We'll send you a link to reset your password.
Enter your username and email address. We'll send instructions on how to reset your password to the email address we have on record.
Athens and Shibboleth are access management services that provide single sign-on to protected resources. They replace the multiple user names and passwords necessary to access subscription-based content with a single user name and password that can be entered once per session. It operates independently of a user's location or IP address. If your institution uses Athens or Shibboleth authentication, please contact your site administrator to receive your user name and password.