Diagram demonstrates total donor eyes used for statistical analysis following exclusion for those missing data points. Further classification was completed based on diabetes history and severity.
A, Endothelial cell distribution and mean endothelial cell count for eyes from donors without diabetes (N = 19 396 eyes). B, Endothelial cell distribution and mean endothelial cell count for eyes from donors with diabetes (N = 8552 eyes).
Margo JA, Munir WM, Brown CH, Hoover CK. Association Between Endothelial Cell Density and Transplant Suitability of Corneal Tissue With Type 1 and Type 2 Diabetes. JAMA Ophthalmol. Published online December 29, 2016. doi:10.1001/jamaophthalmol.2016.5095
What is the effect of type 1 or 2 diabetes on endothelial cell density, corneal transplant suitability, and technician-induced cell damage on eye bank donor characteristics?
This database study found that 8552 of 27 948 eyes (30.6%) during the 3-year study period were from donors with diabetes. No correlation was identified between the presence of diabetes or severe diabetes and poor transplant suitability, reduced endothelial cell counts, or technician-induced tissue damage.
Regardless of disease severity, the presence of diabetes in corneal donors was not shown to reduce endothelial cell density or limit tissue suitability for transplantation.
The rate of types 1 and 2 diabetes in the United States is increasing. The effect of diabetes on corneal donor tissue is unknown.
To determine the association between endothelial cell density and suitability for transplantation in cornea donors with type 1 or 2 diabetes and determine the effect of diabetes on technician-induced endothelial damage during cornea donor tissue processing.
Design, Setting, and Participants
Donor information was obtained from the SightLife Eye Bank for donors from June 1, 2012, to June 30, 2015. The presence of diabetes was determined based on donor medical history. Severe diabetes was classified based on the presence of comorbidities of diabetes. The donor data set contained information on 34 497 donated eyes during the 3-year period, including donor demographics, time from death to refrigeration and preservation of the cornea, endothelial cell count, lens status, medical and surgical history, and suitability for transplantation.
Main Outcomes and Measures
Endothelial cell density, suitability for transplantation based on tissue analysis, and technician-induced endothelial damage.
Among 14 532 donors (mean [SD] age, 58.6 [13.4] years; 8516 men and 6016 women), the mean (SD) endothelial cell count was 2732 (437) cells/mm2. Type 1 or 2 diabetes was listed in the medical history for 8552 of 27 948 donor eyes (30.6%); 5242 eyes (18.8%) were from patients with severe diabetes. After adjusting for age, race/ethnicity, sex, lens status, time from death to refrigeration, and time from death to preservation, the presence of diabetes (adjusted odds ratio, 0.79; 95% CI, 0.51-1.22; P = .28) and severe diabetes (adjusted odds ratio, 95% CI, 0.86; 95% CI, 0.54-1.39; P = .54) were not associated with poor transplant suitability based on results of tissue examination. Donors with diabetes (mean [SD] cell count difference, 9.0 [6.7] cells/mm2; 95% CI, –4.1 to 22.2; P = .18) and severe diabetes (mean [SD] cell count difference, 7.7 [8.1] cells/mm2; 95% CI, –8.1 to 23.6; P = .34) did not exhibit lower cell counts. Technician-induced endothelial damage occurred in 59 corneas (0.2%) but was not associated with the presence of diabetes (adjusted odds ratio, 1.23; 95% CI, 0.66-2.32; P = .52).
Conclusions and Relevance
These data suggest that cornea donors have a high frequency of diabetes. However, this analysis was not able to show that the presence of diabetes was associated with technician-induced endothelial damage, reduced transplant suitability, or reductions in endothelial cell counts. Eye banks may need to collect medical history in a more robust manner. Additional studies may be valuable to determine the effect on long-term transplant outcomes of diabetes in cornea donors.
Type 1 and type 2 diabetes are increasingly common systemic diseases: the incidence of diabetes in the United States continues to rise each year. According to the 2014 National Diabetes Statistics Report, 21 million people, 9.3% of the total US population, are currently diagnosed with diabetes.1 In those 65 years or older, the prevalence of diabetes is 25.9%. Altogether, 4.2 million people had a diagnosis of diabetic retinopathy between 2005 and 2008. Saaddine and colleagues2 reported that the number of people with diabetic retinopathy in the United States is forecasted to triple, totaling more than 16 million by 2050, indicating that the severity and burden of disease are also increasing. It is therefore important to recognize that diabetes is also increasingly prevalent in the population of corneal transplant donors and that a higher percentage of transplant recipients are receiving tissue from donors with diabetes.
It is well documented in the literature that both morphologic and functional corneal changes occur in patients with diabetes.3- 8 This finding is evidenced clinically by the high prevalence of ocular surface disease, persistent or nonhealing epithelial defects, reduced tear secretion, reduced corneal sensitivity, and diabetic keratopathy, which is found in up to 50% of patients with diabetes.3- 5 Rosenberg and colleagues6 studied 44 eyes and found that patients with diabetes had reduced epithelial thickness, increased overall corneal thickness, and a reduction in nerve fiber bundles, suggesting a mechanism for diabetic epitheliopathy. Corneal endothelial changes also have been studied in small populations. Lee and colleagues7 studied 300 eyes and found that the presence of diabetes was associated with reduced endothelial cellular density and hexagonal shape and increased variation in cell size, particularly in patients who had had diabetes for more than 10 years. Sudhir and colleagues8 studied a larger group of 1191 patients with diabetes. On multivariable analysis, they found a reduced cell count of 66 cells/mm2 in patients with diabetes compared with control participants but did not find evidence of increased corneal thickness or variability in cell size.
Endothelial dysfunction also has been found to be more prevalent after intraocular surgery among patients with diabetes compared with controls without diabetes, suggesting a possible cellular susceptibility to trauma.9,10 Previous studies have likewise shown that the prevalence of diabetes in transplanted corneal tissue is lower than the prevalence in the general population, which could support the fact that corneas from donors with diabetes are more likely to have poor tissue integrity and fail eye bank screening.11 Since studies performed in small populations suggest that diabetes may reduce endothelial cell counts, change cellular morphology, and affect postsurgical outcomes, we set out to further examine this association by using a large cohort of eyes from an eye bank donor database to determine how diabetes affects corneal endothelium and the quality of transplanted tissue.
The purpose of this study was to determine the association between endothelial cell density and the presence of type 1 or 2 diabetes in cornea donors and to establish if diabetes or severity of disease independently affects suitability for transplantation. Secondarily, we sought to determine if technician-induced endothelial damage was more common in donors with diabetes. We hypothesized that corneas from donors with diabetes may experience physiological changes, making them more susceptible to trauma during preparation. Although previous studies have reviewed outcomes among transplanted tissue, our study examined all donor tissue rather than only high-quality, preselected tissue for transplantation.
The SightLife Eye Bank (Seattle, Washington) provided a deidentified data set containing information on all donated eyes from June 1, 2012, to June 30, 2015. Variables listed included age, sex, and race/ethnicity of the donor; date and time of death; time from death to refrigeration and time from death to preservation; endothelial cell count; lens status, approved tissue outcome; reason tissue was not suitable for transplantation, if applicable; and medical and surgical history of the donor. In addition, if a tissue did not pass screening, the reason for failure was categorized by the eye bank into epithelial, stromal, or endothelial problems. For endothelial screening, tissue failed for the following 4 reasons: endothelial cell dropout, defined as extensive cell loss seen on specular microscopy; endothelial cell count less than 2000 cells/mm2; poor cell appearance on specular microscopy, such as polymegathism and or pleomorphism; and other nonspecified problems with the endothelium. Technician-induced endothelial damage was categorized separately. The approved tissue outcome (transplantable or nontransplantable) was categorized separately from overall suitability for transplantation, which was based on both tissue evaluation and medical history of the donor.
Medical history was sorted to first identify the variation in recording the history of diabetes. Keywords used to identify diabetes and associated comorbidities were searched in the entire donor data set to stratify donors based on the presence and severity of diabetes (Box). Documentation of the medical history was dependent on the eye bank staff member evaluating the donor and donor’s medical history before harvesting tissue. The presence or absence of diabetes was determined based on the donor’s medical history. A donor was classified as having had severe diabetes based on the presence in the medical history of any of a set of terms, such as diabetic ketoacidosis, diabetic retinopathy, or insulin-dependent diabetes mellitus (Box), or if they had a diagnosis of diabetes plus at least 1 of a set of associated comorbidities, such as hemodialysis, chronic renal failure, or peripheral neuropathy (Box).
DM type 2
Gastroparesis secondary to IDDM
With retinopathy and neuropathy
Type II with retinopathy
DM with neuropathy
With neuropathy and retinopathy
With peripheral neuropathy
With related neuropathy
With retinopathy and peripheral neuropathy
With peripheral neuropathy, retinopathy, or PVD
Diabetic nephropathy and neuropathy
IDDM with polyneuropathy
NIDDM with neuropathy and retinopathy
NIDDM with peripheral neuropathy and nephropathy
CKD secondary to diabetic nephropathy
Retinopathy in both eyes
Nephropathy and neuropathy
Chronic renal failure
Abbreviations: CKD, chronic kidney disease; CRF, chronic renal failure; DKA, diabetic ketoacidosis; DM, diabetes mellitus; ESRD, end-stage renal disease; ESRF, end-stage renal failure; HD, hemodialysis; IDDM, insulin-dependent diabetes mellitus; NIDDM, non–insulin-dependent diabetes mellitus; PD, peritoneal dialysis; PVD, peripheral vascular disease.
a Donors were classified as having nonsevere diabetes if they only had any diagnosis of diabetes.
b Donors were classified as having severe diabetes if they had any of the listed conditions in the severe diabetes category or had diabetes plus any associated comorbidity.
Donors whose records were missing endothelial cell counts, lens status, time from death to refrigeration and/or preservation, documentation of tissue outcome, or medical and surgical history were excluded from analysis. Technician-induced endothelial damage was defined by the eye bank as damage to the endothelium that occurred during tissue processing that resulted in the inability to categorize the tissue as suitable for transplantation. The University of Maryland’s Institutional Review Board evaluated the study, and exemption was granted for nonhuman participants research.
In statistical analysis, tests of association between dichotomous outcomes, including poor suitability for transplantation, endothelial damage, and diabetes status (yes vs no), were adjusted for covariates using logistic regression, including age, race/ethnicity, sex, time from death to preservation, time from death to refrigeration, and lens status. Endothelial cell counts were approximately normally distributed and analyzed with linear regression. Repeated-measures analysis was used to account for within-individual correlation between the left and right eyes coming from the same donor.
The deidentified eye bank data set contained information on 34 497 donated eyes. After excluding eyes that were missing medical or surgical history, tissue examination and lens status, the data set contained data on 30 931 donated eyes, which was used to examine correlates of technician-induced endothelial damage. After excluding donors with missing endothelial cell count, the data set contained 27 948 donor eyes from 14 532 donors. This data set was the primary data set and was used for all analyses except for those examining correlates of endothelial damage (Figure 1).
There were more eyes from males (8516 of 14 532 [58.6%]) than females (6016 of 14 532 [41.4%]) in the eye bank data set. Donor eyes with pseudophakia represented 9.8% (n = 2741) of the entire donor population. Mean (SD) endothelial cell count was 2732 (437) cells/mm2 and mean (SD) donor age was 58.6 (13.4) years. Type 1 or 2 diabetes was listed in the medical history for 8552 donor eyes (30.6%). A total of 5242 eyes (18.8%) were from donors with severe diabetes, comprising 61.3% of eyes from donors with diabetes (Figure 1).
Baseline characteristics for donors with and without diabetes showed that the mean (SD) age of donors with diabetes (62.3 [10.1] years) was significantly higher than the mean age of those without diabetes (56.9 [14.3] years; P = .001) (Table). There was no significant difference between groups in sex (without diabetes: male, 5880 of 10 070 [58.4%]; with diabetes: male, 2636 of 4462 [59.1%]) or mean (SD) time from death to preservation (without diabetes, 10.6 [5.5] hours; with diabetes, 10.7 [5.4] hours). Mean (SD) time from death to refrigeration was slightly longer and statistically significant in the group without diabetes vs the group with diabetes (3.6 [2.7] vs 3.5 [2.4] hours; P < .001). There was a significantly higher proportion of eyes with pseudophakia in the group with diabetes than in the group without diabetes (1432 of 8552 [16.7%] vs 1309 of 19 396 [6.7%]; P < .001).
Failed endothelial screening was defined by the eye bank as any of the following: endothelial cell drop out (extensive cell loss on specular microscopy), endothelial cell count less than 2000 cells/mm2, poor appearance of the endothelium on specular microscopy, and other nonspecified problems with the endothelium. Failed endothelial screening accounted for 144 of donor eyes (0.5%). In the group with diabetes and the group without diabetes, there were 49 of 8552 donor eyes (0.6%) and 95 of 19 396 donor eyes (0.5%), respectively, with poor suitability for transplantation based on results of endothelial examination alone.
After adjusting for age, race/ethnicity, sex, time from death to refrigeration, time from death to preservation, and lens status, the presence of diabetes (adjusted odds ratio, 0.79; 95% CI, 0.51-1.22; P = .28) was not associated with poor suitability for transplantation based on results of endothelial tissue examination. Furthermore, severe diabetes was also not associated with poor suitability for transplantation (adjusted odds ratio, 0.86; 95% CI, 0.54-1.39; P = .54).
After adjusting again for patient demographics and tissue preservation characteristics, donors with diabetes did not exhibit lower endothelial cell counts (mean [SD] cell count difference, 9.0 [6.7] cells/mm2; 95% CI, –4.1 to 22.2; P = .18) (Figure 2). On subgroup analysis after stratification for severity, donors with severe diabetes also did not exhibit lower endothelial cell counts (mean [SD] cell count difference, 7.7 [8.1] cells/mm2; 95% CI; –8.1 to 23.6; P = .34).
Technician-induced endothelial damage led to failed suitability for transplantation in 59 (0.2%) corneas. In the 59 eyes that failed transplant screening owing to technician-induced endothelial damage, this failure was not associated with the presence of diabetes (adjusted odds ratio, 1.23; 95% CI, 0.66-2.32; P = .52). Because no donor eyes with severe diabetes had technician-induced endothelial damage, a subgroup analysis in these eyes was unable to be performed.
When eyes with phakia from donors with diabetes were stratified separately, they were not more likely to have lower endothelial cell counts than eyes with phakia from donors without diabetes (mean [SD] cell count difference, 6.0 [6.9] cells/mm2; 95% CI, –7.6 to 19.6; P = .39). Likewise, eyes with pseudophakia from donors with diabetes did not have lower endothelial cell counts compared with eyes with pseudophakia from donors without diabetes (mean [SD] cell count difference, 15.4 [22.2] cells/mm2; 95% CI, –28.1 to 58.8; P = .49). Finally, when stratified to groups with phakia and pseudophakia, tissue from donors with diabetes was not associated with poor suitability for transplantation based on results of endothelial examination (adjusted odds ratio, 0.55; 95% CI, 0.24-1.06; P = .07 and adjusted odds ratio, 0.94; 95% CI, 0.49-1.81; P = .86, respectively) when compared with those without diabetes.
Following corneal donor procurement, rigorous screening and analysis of donor tissue is performed to determine if the tissue is suitable for transplantation. In 2013, poor tissue characteristics and quality control problems excluded more than 32 000 corneas from transplantation. The most common reason a cornea failed transplant screening was endothelial abnormalities.12 Endothelial cell health and survival determines the longevity of a corneal graft following transplantation and certain donor characteristics may contribute to failure of a corneal graft. In addition, in previous studies, multiple host factors have been shown to contribute to graft failure, including the use of topical medications for glaucoma, the presence of ocular surface disease, and the presence of diabetes.13
In addition to host factors, transplanted tissues from patients with certain medical conditions could be damaged on the molecular or subclinical level such that the tissue initially appears to be of high quality or passes screening but has an increased propensity to fail, especially when coupled with certain host risk factors that also contribute to graft failure or rejection. Currently, it is unknown if tissue from donors with diabetes is more or less likely to be suitable for transplantation.
However, studies have examined the effect of diabetes in donors on transplant outcomes. Lass and colleagues14 studied the long-term effect of diabetes on time to graft failure and endothelial cell count in 1090 patients. They found no association between diabetes in donors and penetrating keratoplasty at 10 years. They also found no difference in baseline or 10-year endothelial cell density after adjusting for covariates. However, there was variability in documenting the presence of diabetes in donors, which could have affected the study. Vislisel and colleagues11 compared corneal graft survival using tissues from donors with and without diabetes in 183 patients undergoing Descemet stripping automated endothelial keratoplasty and penetrating keratoplasty. They also found similar rates of graft failure in tissue coming from donors with and without diabetes in all patients undergoing keratoplasty during a 72-month follow-up period. However, the small sample size limited the ability to examine the effect of diabetes in donors on different types of keratoplasty. The Cornea Donor Study also examined the association between donor factors and 5-year graft survival in 1090 patients.15 The study examined tissue characteristics, donor demographics, and history of diabetes in the corneal donor. Graft failure rates were not affected by any donor characteristic, including history of diabetes. They did not specifically study endothelial cell counts.
To our knowledge, this is the first study to use a large data set of corneal donors and examine the effect of diabetes in donors and disease severity on suitability for transplantation, endothelial cell density, and technician-induced damage. Our results show that the presence of diabetes was not associated with poor suitability for transplantation based on results of endothelial tissue examination, was not associated with technician-induced endothelial damage and, overall, did not lead to reduced endothelial cell density. By stratifying donors into categories of severe and nonsevere diabetes, we were able to show that regardless of the severity of diabetes, donors did not exhibit reduced endothelial cell counts. Given that more donors with pseudophakia had diabetes, we controlled for lens status in the multivariable analysis and analyzed lens status separately to determine if eyes with phakia or pseudophakia from donors with diabetes were more likely to fail endothelial tissue examination or have reduced cell counts. Neither eyes with pseudophakia or phakia from donors with diabetes had statistically significantly reduced cell counts or were more likely to fail transplant screening.
Greiner and colleagues16 studied more than 350 corneas for Descemet membrane endothelial keratoplasty preparation. They found no difference in characteristics of donor tissue between people with diabetes and those without diabetes; however, they did report increased failure of tissue preparation for Descemet membrane endothelial keratoplasty in donor tissue coming from patients with diabetes (odds ratio, 9.20; 95% CI, 2.89-29.32; P = .001) and recommended caution in using tissue from donors with diabetes for preparation of grafts for Descemet membrane endothelial keratoplasty. Similarly, Vianna and colleagues17 studied risk factors for failure of eye bank tissue preparation for Descemet membrane endothelial keratoplasty and found that the rate of failure of tissue preparation was 5.2%. Donors with diabetes were 3-fold more common in the failure group and longer duration of diabetes was associated with higher rates of failure of tissue preparation. We found no association between failure of tissue preparation and diabetes; however, we did not stratify the type of tissue preparation and thus our sample likely included a large number of full-thickness graft tissue. With penetrating keratoplasty tissue being less complex to prepare, the effect of diabetes on specific types of tissue preparation may have been overlooked.
The advantage of our study was that it was a large retrospective study with 34 497 eyes. We were able to determine the prevalence of diabetes in corneal donors and its effect on suitability for transplantation. By eliminating host factors in this study, the tests of the association of diabetes with graft characteristics may be stronger than studies that examine graft outcomes, as graft failure is usually multifactorial and more difficult to analyze in regard to single causative agents. Our study directly examines the effect of diabetes on tissue rather than multifaceted tissue transplant outcomes. As highlighted in the Table, we found differences between donors with and without diabetes including older age, nonwhite race/ethnicity, and a higher percentage of eyes with pseudophakia from donors with diabetes. Time from death to refrigeration was also shorter in the group with diabetes; however, despite statistical significance, the difference was only 0.1-hour (6 minutes) and likely bears no clinical significance. Furthermore, these factors were all controlled for in multivariable analysis.
Our study contrasts with previous studies, which describe reduced endothelial cell density in patients with diabetes. We attempted to examine donors on a larger scale while also controlling for factors that may reduce endothelial cell counts, such as age and lens status; however, we were unable to group donors by duration of diabetes, which other smaller studies were able to achieve. Our results align with the previously mentioned graft outcome studies11,14,15 showing that patients who receive corneal tissue from donors with diabetes do not have poorer graft outcomes. Our findings highlight that it remains unclear what specific amount of cellular loss is clinically significant enough to affect overall graft outcomes.
Limitations of our study include the fact that the data set was from 1 eye bank. In addition, the validity of the data recording relies on staff to correctly review and document the donor’s medical record in the data set. As in similar studies mentioned previously, we also found significant variability in the way that diabetes was listed in the data set, which likely represents the variability in how physicians document diabetes within the medical record. There were 2 ways that a donor’s diabetes could be classified as severe. If a donor had a condition directly associated with diabetes listed in the medical history, such as diabetes with neuropathy or retinopathy or diabetic ketoacidosis, they were placed into the severe diabetes category. Donors with a diagnosis of diabetes and a separately listed comorbidity, such as renal failure or neuropathy, were also assumed to have severe diabetes. This assumption may have classified more donors into the severe diabetes group and may contribute to the high prevalence of severe diabetes in this study cohort.
In addition, the way in which endothelial cell density is recorded is also inherently flawed as it is based on a mean of measurements by specular microscopy and may not represent the overall accurate endothelial cell count. It is also important to keep in mind that measuring transplant suitability based on results of endothelial examination alone may not be a sensitive enough measure of overall endothelial health.
Although eye banks cannot control host factors, they do possess the ability to screen for conditions that may increase transplant failure or rejection risk and they are vital in maintaining the quality control of transplanted tissue. It is, therefore, important to continually study the effect of chronic disease on the suitability of corneal donor transplant and donor outcomes. Information about donor medical history may need to be collected in a more robust and regimented manner to include other disease-specific details, such as type, severity, control, and treatment. This study helps further explain the association between diabetes and its effect on corneal tissue and highlights the importance of donor medical history and its association with the quality of transplanted tissue.
Accepted for Publication: November 5, 2016.
Corresponding Author: Wuqaas M. Munir, MD, Department of Ophthalmology and Visual Sciences, University of Maryland School of Medicine, 419 W Redwood St, Ste 470, Baltimore, MD 21201 (email@example.com).
Published Online: December 29, 2016. doi:10.1001/jamaophthalmol.2016.5095
Author Contributions: Drs Margo and Munir had full access to all the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.
Study concept and design: Margo, Munir, Brown.
Acquisition, analysis, or interpretation of data: All authors.
Drafting of the manuscript: Margo, Munir.
Critical revision of the manuscript for important intellectual content: All authors.
Statistical analysis: Brown.
Obtained funding: Margo, Munir.
Administrative, technical, or material support: Margo, Munir, Hoover.
Study supervision: Munir.
Conflict of Interest Disclosures: All authors have completed and submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest and none were reported.
Funding/Support: The Eye Bank Association of America provided funding support for statistical analysis and to obtain the eye bank data set through the Richard Lindstrom Research Grant.
Role of the Funder/Sponsor: The funding source had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.