Figure 1. Endothelial cell density change over time in all subjects with at least 1 gradable postoperative image before either the date of graft failure or the end of 5-year follow-up available (n = 567) (A) and restricted to subjects without graft failure at 5 years (n = 345) (B). Black dots indicate means; horizontal lines in boxes, medians; and bottom and top of boxes, 25th and 75th percentiles, respectively.
Figure 2. Endothelial cell density over time according to donor sex. Black dots indicate means; horizontal lines in boxes, medians; and bottom and top of boxes, 25th and 75th percentiles, respectively.
Lass JH, Beck RW, Benetz BA, Dontchev M, Gal RL, Holland EJ, Kollman C, Mannis MJ, Price F, Raber I, Stark W, Stulting RD, Sugar A, Cornea Donor Study Investigator Group FT. Baseline Factors Related to Endothelial Cell Loss Following Penetrating Keratoplasty. Arch Ophthalmol. 2011;129(9):1149-1154. doi:10.1001/archophthalmol.2011.102
Author Affiliations: Department of Ophthalmology and Visual Sciences, Case Western Reserve University and University Hospitals Eye Institute, Cleveland (Dr Lass and Ms Benetz), and Cincinnati Eye Institute, Cincinnati (Dr Holland), Ohio; Jaeb Center for Health Research, Tampa, Florida (Drs Beck and Kollman and Mss Dontchev and Gal); Department of Ophthalmology and Vision Science, University of California, Davis, Sacramento (Dr Mannis); Price Vision Group, Indianapolis, Indiana (Dr Price); Ophthalmic Partners of Pennsylvania, Bala Cynwyd, Pennsylvania (Dr Raber); Wilmer Eye Institute, The Johns Hopkins University School of Medicine, Baltimore, Maryland (Dr Stark); Emory Eye Center, Emory University, Atlanta, Georgia (Dr Stulting); and W. K. Kellogg Eye Center, University of Michigan, Ann Arbor (Dr Sugar). Dr Stulting is now with the Woolfson Eye Institute, Atlanta.
Group Information: A list of the Cornea Donor Study Investigator Group members was published in Arch Ophthalmol. 2010;128(1):63-69.
Objective To identify baseline (donor, recipient, and operative) factors that affect endothelial cell loss following penetrating keratoplasty for a moderate-risk condition (principally Fuchs dystrophy or pseudophakic or aphakic corneal edema).
Methods In a subset (n = 567) of Cornea Donor Study participants, preoperative and postoperative endothelial cell densities (ECDs) were determined by a central reading center. Multivariate regression analyses were performed to examine which baseline factors correlated with ECD over time.
Results Larger grafts (P < .001), younger donor age (P < .001), and female donor (P = .004) were significantly associated with higher ECD during follow-up. Median endothelial cell loss at 5 years was 68% for grafts larger than 8.0 to 9.0 mm in diameter, 75% for grafts 7.0 mm to smaller than 8.0 mm in diameter, and 74% for grafts 8.0 mm in diameter. Grafts from female donors experienced a 67% cell loss compared with a 72% cell loss among grafts from male donors. Method of tissue retrieval, donor cause of death, history of diabetes, and time from death to preservation or to surgery were not significantly associated with changes in ECD over time.
Conclusions Following penetrating keratoplasty for endothelial dysfunction conditions, larger donor graft size, younger donor age, and female donor were associated with higher ECD over 5 years. These data warrant exploring the possibility that similar associations may exist following endothelial keratoplasty.
Trial Registration clinicaltrials.gov Identifier: NCT00006411
The Cornea Donor Study (CDS) was designed to determine the effect of donor age on penetrating keratoplasty outcomes.1 At 5 years, no significant effect of donor age (up to age 75 years) on graft survival was found.2 Other donor factors (method of retrieval, processing factors, timing of donor cornea use)3 and ABO incompatibility4 likewise had no effect on graft survival. The risk for graft failure was approximately 4-fold higher in eyes with pseudophakic or aphakic corneal edema than in eyes with Fuchs dystrophy, whether pseudophakic or not.5 Prior glaucoma surgery and/or preoperative glaucoma medication use also substantially increased the graft failure rate.5
While donor age had no effect on 5-year graft survival, a slight association between increasing donor age and greater post–penetrating keratoplasty corneal endothelial cell loss was detected in an ancillary study to the CDS, the Specular Microscopy Ancillary Study.6 The Specular Microscopy Ancillary Study also confirmed that there was substantial cell loss in successful grafts at 5 years irrespective of donor age (median loss preoperatively to 5 years, 69% for donors aged <66 years and 75% for donors aged ≥66 years). In this study, we evaluated whether other donor,3,7 recipient,5 and operative5 factors might be associated with endothelial cell loss during the first 5 years after penetrating keratoplasty.
The CDS and the Specular Microscopy Ancillary Study protocols are described in detail in earlier articles.2,3,5- 11 Briefly summarized, subjects eligible for the study were aged between 40 and 80 years and had corneal disease associated with endothelial decompensation and moderate risk of failure (principally Fuchs dystrophy and pseudophakic corneal edema). Eligible corneas were from donors aged 10 to 75 years meeting Eye Bank Association of America standards.3,7,12 Eligibility criteria also included a preoperative baseline eye bank–determined endothelial cell density (ECD) of the central corneal endothelium from 2300 to 3300 cells/mm2. Clinical investigators and subjects were masked to donor age and ECD of the donor cornea. Subject characteristics, including age, were not taken into account when the cornea assignment was made. Study protocols were approved by institutional review boards at each participating site, and written consent was obtained from each subject.
Preoperative care, surgical technique, and postoperative care (including prescription of medications) were provided according to each clinical investigator's customary routine.2,5,11 Study follow-up visits were required at 6 months, 12 months, and thereafter annually for 5 years. Annual follow-up continued unless a regraft was performed.
The eye bank obtained preoperative specular microscopic images of the central donor corneal endothelium. Each clinical site obtained postoperative specular microscopic images of the central corneal endothelium of the graft at the 6-month and annual follow-up visits. The preoperative donor images and postoperative subject images were evaluated for quality and ECD by a central reading center, the Cornea Image Analysis Reading Center (formerly the Specular Microscopy Reading Center) at Case Western Reserve University and University Hospitals Eye Institute, using a previously described variable frame analysis method.6,8- 10 If the independent ECD determinations of 2 readers differed by 5.0% or more, a third determination of ECD was made by an independent adjudicator.8
Analysis was restricted to 567 subjects with at least 1 gradable postoperative image before either the date of graft failure or the end of 5-year follow-up. Subjects with graft failures due to trauma or infection were excluded from the analysis. Images obtained after graft failure were not included in the analysis, and no attempt was made to impute missing image data after graft failure. The analysis was therefore conditional on graft survival.
Baseline donor ECD was determined by the reading center for 376 cases (66%) and by the eye bank for the remaining 191 (34%). The relationships between baseline (donor, recipient, and operative) factors and ECD were explored in multivariate analyses. Cross-sectional regression models were used to evaluate change in ECD from baseline to 5 years, and longitudinal repeated-measures models were used to evaluate ECD change throughout follow-up. All models were fit with the rank-normalized transformation (van der Waerden scores). Final multivariate models were generated through stepwise selection of covariates at a significance level of .01 after adjusting for baseline ECD. The large number of statistical comparisons increases the likelihood of a false-positive, and no attempt was made to control the overall type I error probability in these exploratory analyses. A threshold of P < .01 was used to define statistical significance as a compromise to balance the risks of type I vs type II errors.
All reported P values are 2-sided. Statistical analyses were conducted using SAS version 9.1 statistical software (SAS Institute, Inc, Cary, North Carolina).
The mean (SD) age of the 567 transplant recipients at the time of penetrating keratoplasty was 70 (8) years; 357 (63%) were women and 536 (95%) were white, non-Hispanic individuals. Endothelial failure indications included Fuchs dystrophy in 382 participants (67%), pseudophakic or aphakic corneal edema in 165 (29%), and a variety of other diagnoses in 20 (4%). The mean (SD) donor age at the time of death was 58 (15) years; 191 donors (34%) were women and 539 (95%) were white, non-Hispanic individuals. The primary cause of donor death was cardiovascular disease or stroke (61%), followed by cancer (19%), trauma (9%), respiratory causes (7%), and other causes (4%). Other baseline donor, recipient, and operative characteristics were comparable to previous CDS results.3,5,7,11
The changes in ECD over time are illustrated for all 567 recipients with at least 1 postoperative image (Figure 1A) and for the 345 recipients with an image at 5 years (Figure 1B). The baseline median (interquartile range) ECD was 2698 (2481-2892) cells/mm2, which decreased to 778 (576-1267) cells/mm2 at 5 years. This represented a median (interquartile range) of 70% (55%-77%) loss of cells during the course of the study (Figure 1B).
In the multivariate analysis, younger donor age (P < .001), female donor (P = .004), and larger grafts (P < .001) were significantly associated with higher ECD over the course of follow-up after adjusting for baseline ECD and diagnosis or lens status (Table). The relationship between donor sex and ECD loss over time is illustrated in Figure 2. The association between graft size and ECD loss over time remained significant after adjusting the model for potential effect by site to control for any influence of the individual surgeon's technique on the size of the graft. At 5 years, among 345 eyes that had not experienced graft failure, a smaller decrease in ECD from baseline was associated with larger grafts (P < .001) and younger donor age (P = .009) after adjusting for baseline ECD (Table). When female donor was added to the multivariate model, it demonstrated a weaker association with the 5-year change in ECD (P = .09) than it did in the longitudinal analysis. No other donor, recipient, and operative factors, namely method of tissue retrieval, cause of death, diabetes history, time from death to preservation or to surgery, baseline recipient diagnosis and lens status, prior use of glaucoma medications and/or surgery, and postoperative intraocular pressure, demonstrated significant association with changes in ECD over time (Table).
As we previously reported,6 subjects who received a cornea from a younger donor experienced slightly less long-term endothelial cell loss. This association was most evident with corneas from donors younger than 40 years. Among grafts that survived for 5 years postoperatively, those from the 44 donors younger than 40 years experienced 62% cell loss compared with a 75% loss among the 62 grafts from donors aged between 70 and 76 years (Table). In addition to the reported association with age, the current analyses also identified associations between endothelial cell loss and donor graft size and donor sex.
Graft size was associated with endothelial cell loss after adjusting for baseline ECD and diagnosis or lens status. At 5 years, grafts larger than 8.0 to 9.0 mm in diameter experienced a median 68% cell loss compared with a 75% loss for grafts 7.0 to smaller than 8.0 mm in diameter and a 74% loss for grafts 8.0 mm in diameter. This effect did not appear to be related to the disparity between the graft and recipient bed diameters. Previous investigators have found a similar correlation with larger graft diameters following penetrating keratoplasty and less endothelial cell loss.13,14 However, the advantage of less postoperative cell loss after penetrating keratoplasty with larger donor diameter may be offset by a higher risk of graft rejection.15,16
The observed effect of less postoperative cell loss after penetrating keratoplasty with larger donor diameter may also be applicable in endothelial keratoplasty, as grafts larger than 8.0 mm in diameter are commonly used and the observed rate of cell loss after 6 months postoperatively is slower compared with penetrating keratoplasty.17,18 With endothelial keratoplasty, 26% more cells are transplanted with a 9.0-mm donor graft than an 8.0-mm donor graft and the peripheral cornea may have a higher density of cells than the central cornea, further adding to this benefit.19,20 The postulated beneficial effect of larger endothelial keratoplasty grafts may be greater for those with pseudophakic corneal edema compared with those with Fuchs dystrophy because the peripheral endothelium is more compromised in the former condition.
In the longitudinal multivariate analysis, female donor was significantly associated with higher ECD during the course of follow-up (P = .004) (Table), but this association was not statistically significant at 5 years in those grafts that had not failed (5-year cell loss of 72% vs 67% in grafts from male vs female donors, respectively). Kinoshita made a similar observation in a cohort of 340 penetrating keratoplasties (225 male and 125 female donors) followed up over a 7-year period (Shigeru Kinoshita, MD, PhD, written communication, September 8, 2010), while another study reported that healthy females had a 2.9% greater cell density than males.21 Receptors for both male and female hormones have been shown to be present in the epithelium, keratocytes, and, most relevantly, the endothelium of the human cornea,22 which raises the possibility that these hormones may play a biological role in corneal function. While some studies have reported an effect of these hormones on lacrimal23 and meibomian gland function and corneal thickness,24 none have reported effects on corneal endothelial viability. One study has suggested a differential effect of the sex hormones on cellular apoptosis.25 Further in vitro, ex vivo, and clinical studies are warranted to determine whether hormones, particularly estrogens, play a role in endothelial survival following keratoplasty.
Similar to our finding of no relationship with any donor factors and graft failure,3 the condition of the donor, cause of death, history of diabetes, death to preservation time, death to surgery time, and method of retrieval were unrelated to the degree of endothelial cell loss. In the CDS, although individuals with a diagnosis of pseudophakic or aphakic corneal edema were found to have a 4-fold higher graft failure rate than those with Fuchs dystrophy, independent of lens status,5 recipient diagnosis was not found to be related to cell loss over time. There was also no observed relationship between prior treatment for glaucoma and endothelial cell loss, but these factors also had a highly significant effect on graft failure in the CDS.5 Although this may be related to the small number in the preoperative glaucoma and/or surgery group (n = 69) compared with the group with no medication or surgery (n = 498), Lee and Kim26 reported results for a similarly sized study where those without preoperative glaucoma (n = 53) had substantially less cell loss (mean [SD], 38% [19%]) at 2 years than those with preoperative glaucoma (n = 30; mean [SD], 57% [21%]). The deleterious effect of elevated intraocular pressure remains ill defined, while tube shunts continue to be a major risk for progressive loss of endothelial function with both penetrating27,28 and endothelial18 keratoplasty. This is not surprising considering the marked endothelial cell loss associated with a commonly used shunt, the Ahmed glaucoma valve, without keratoplasty.29
Our findings are strengthened by the prospective study design, large sample size, and determination of the ECD by a central reading center. However, as with all studies of changes of ECD over time following keratoplasty, there are problems inherent in these analyses that cannot be overcome by statistical methods. Outcome data are conditional on surviving grafts; thus, there is the potential for bias due to dropout of eyes from the analysis data set when grafts fail. However, none of these factors reported to be related to endothelial cell loss over time are associated with graft failure over 5 years in the CDS.
In addition to the baseline factors examined in this study, there may be postoperative factors that affected the endothelial cell loss we observed in this cohort. In this analysis, postoperative factors were not examined either because data were not collected (eg, duration of topical corticosteroid use) or because the effect was difficult to determine owing to visit frequency and timing of specular microscopic examination (eg, graft rejection).
The Specular Microscopy Ancillary Study has expanded our understanding of the various baseline factors that have an influence (donor age, graft size, and, surprisingly, female donor) and appear to have no influence (notably death to surgery, cause of death, diabetes, and method of retrieval) on endothelial survival following penetrating keratoplasty for conditions with endothelial dysfunction. However, while endothelial cell loss may be a proxy measure for later graft failure, this relationship is not straightforward. The CDS data showed that an ECD less than 1700 cells/mm2 at 6 months was associated with a graft failure rate of 13% at 5 years, while an ECD of 2500 cells/mm2 or greater at 6 months was associated with a 2% failure rate at 5 years.9 On the other hand, 14% of clear grafts had an ECD less than 500 cells/mm2 at 5 years. Thus, further observation is warranted to determine how the observed decrease in ECD at 5 years affects the graft success rate over a longer period.
Submitted for Publication: January 20, 2011; final revision received March 8, 2011; accepted March 15, 2011.
Correspondence: Jonathan H. Lass, MD, Cornea Donor Study Coordinating Center, Jaeb Center for Health Research, 15310 Amberly Dr, Ste 350, Tampa, FL 33647 (firstname.lastname@example.org).
Author Contributions: Dr Beck, principal investigator for the CDS, had full access to all of the study data and takes responsibility for the integrity of the data and the accuracy of the data analysis.
Financial Disclosure: None reported.
Published Online: May 9, 2011. doi:10.1001/archophthalmol.2011.102
Funding/Support: This work was supported by grants EY12728 and EY12358 from the National Eye Institute and by Eye Bank Association of America, Bausch & Lomb, Inc, Tissue Banks International, Vision Share, Inc, San Diego Eye Bank, The Cornea Society, Katena Products, Inc, ViroMed Laboratories, Inc, Midwest Eye-Banks (Michigan Eye-Bank and Illinois Eye-Bank), Konan Medical Corp, Eye Bank for Sight Restoration, SightLife, Sight Society of Northeastern New York (Lions Eye Bank of Albany), and Lions Eye Bank of Oregon.
Additional Information: The CDS Publications Committee members Christopher R. Croasdale, MD, Mark D. Mifflin, MD, and Joel Sugar, MD, independently reviewed and approved the manuscript for submission.
Additional Contributions: M. Edward Medof, MD, PhD, was consulted on the relationship of sex and endothelial cell loss.
This article was corrected for errors on October 21, 2011.