Secondary Logo

Journal Logo

Clinical Science

Long-Term Maintenance of Corneal Endothelial Cell Density After Corneal Transplantation

Kayukawa, Kanae MD*,†; Kitazawa, Koji MD, PhD*,†; Wakimasu, Koichi MD; Patel, Sanjay V. MD, FRCOphth; Bush, John BA*; Sotozono, Chie MD, PhD*; Kinoshita, Shigeru MD, PhD§,†

Author Information
doi: 10.1097/ICO.0000000000002386


The accelerated decrease of corneal endothelial cell density (ECD) after corneal transplantation is an unresolved and critical problem for corneal surgeons. Reportedly, the decrease of corneal ECD after corneal transplantation is much higher than that after cataract surgery (2.5% per year)1 or in normal healthy subjects (0.6% per year).2 For example, Bourne et al3 reported that the mean ECD continued to decrease 7.8% per year from 3 to 5 years after keratoplasty. Although it is believed that ECD continuously decreases after corneal transplantation, postoperative follow-up examinations sometimes reveal a high corneal ECD in patients many years after undergoing corneal transplantation. Hence, and with this fact taken into consideration, it may be possible that a limited loss of corneal ECD can be attained after surgery in cases undergoing corneal transplantation, as in cataract surgery, if some biological improvements are made. Thus, we investigated this phenomenon in a large cohort of consecutive patients who underwent penetrating keratoplasty (PK).

PK, a transplantation procedure that replaces the full-thickness cornea, continues to be performed worldwide for the treatment of patients afflicted with corneal dystrophies, corneal opacity, severe bullous keratopathy (BK), and corneal shape abnormalities.4,5 Although previous studies have reported the long-term safety and efficacy of PK,6–8 it is widely known that corneal ECD gradually decreases over the long-term postoperative period and can ultimately lead to failure of the transplanted corneal graft. The findings in the Cornea Donor Study (CDS), a multicenter, prospective, double-masked, controlled clinical trial conducted in the United States, demonstrated that corneal ECD gradually diminishes after PK, with a 70% loss of endothelial cells at 5 years postoperative compared with the number of cells before surgery.7,9 It is well known that retaining long-term corneal transparency after surgery depends on how well the transplanted graft maintains a healthy normal ECD. The CDS findings also revealed that younger donor age, female sex, and larger graft diameter were all associated with a higher corneal ECD at 5 years postoperative.7,10–12

Several previous large cohort studies have investigated ECD after corneal transplantation procedures, such as PK,8,11,13,14 Descemet stripping automated endothelial keratoplasty (DSAEK),15,16 and Descemet membrane endothelial keratoplasty.17,18 Those studies reported ECD and endothelial cell loss (ECL) postoperative compared with preoperative for the entire cohort and did not differentiate between grafts with a high or low ECL. It should be noted that the Japanese Ophthalmological Society, a cooperative working group of corneal specialists who developed a grading score for the severity of corneal endothelial damage, has defined a corneal ECD of ≥2000 cells/mm2 as being “normal”.19,20

Thus, in this present study, our primary aim was to obtain a deeper understanding of the factors that might help contribute to maintaining a corneal ECD of ≥2000 cells/mm2 at 5 years after PK.



This retrospective large cohort study was approved by the Kyoto Ethics Review Committee, Kyoto, Japan (Approval #1604). The study protocol was conducted in accordance with the tenets of the Declaration of Helsinki, and the study was registered at UMIN (; R000028643 UMIN000024891).

The data used in this study were that obtained from 1 surgeon at 1 institute to reduce surgery-related bias because of the differences between surgeons, and it should be noted that donor eyes of children (ie, a donor age of <12 years) that may produce an especially high ECD were excluded from the study. Of 872 consecutive eyes that underwent PK at the Baptist Eye Institute, Kyoto, Japan, from 1998 through 2011, 485 eyes that underwent PK by 1 corneal specialist surgeon (S.K.) with internationally shipped donor corneas (ie, with a donor age of >12 years) were involved in this study.

Of the 485 included eyes, 225 were continually followed for 5 years after PK, whereas 260 were excluded beause of the loss of follow-up or the patient follow-up being transferred to another medical institution. Of the 225 eligible cases, graft failure within 5 years after surgery (n = 23 eyes), eyes in which the ECD could not be determined because of trauma (n = 2 eyes), eyes in which corneal infection occurred (n = 2 eyes), and eyes with corneal opacity and/or being borderline before graft failure (n = 14 eyes) were excluded. In addition, 10 Fuchs endothelial corneal dystrophy (FECD) BK eyes were excluded because a reasonable amount of peripheral corneal endothelial cells may exist in FECD BK eyes in comparison with non-FECD BK eyes, thus possibly affecting the statistical analysis (Fig. 1). Thus, the abovementioned exclusion criteria resulted in 174 eyes of 149 patients (63 men and 86 women) with an average age of 65.7 ± 12.7 years (mean ± SD; range: 20–88 years) at the time of surgery being ultimately included in the study. Indications for corneal transplantation included corneal edema (n = 97 eyes, 55.7%), corneal opacity due to infectious keratitis and trauma (n = 51 eyes, 29.3%), corneal dystrophies (n = 19 eyes, 10.9%), and keratoconus (n = 7 eyes, 4.0%). A summary of the patient demographics is shown in Table 1.

Flow chart illustrating the inclusion and exclusion criteria in this retrospective study.
Patient and Donor Demographics and Recipient, Operative, and Donor Characteristics (N = 174)

Surgical Procedure and Postoperative Management

Details of the specific surgical procedure used and the postoperative management administered at the Baptist Eye Institute have previously been reported.21,22 The donor corneas used in this study were obtained from the SightLife Eye Bank (Seattle, WA) and the Rocky Mountain Lions Eye Bank (Aurora, CO). In all cases, except those with keratoconus, the donor graft ranged in size from 0.25 to 0.5 mm larger than the host trephination. In the keratoconus cases, the size of the donor graft was the same as that of the host trephination. The diameter of the donor corneal graft ranged in size from 7.0 to 8.0 mm. In the cases that underwent PK combined with cataract surgery, the cataract extraction was performed by phacoemulsification and aspiration with continuous curvilinear capsulorhexis under open sky. In the cases of aphakia or a ruptured capsule or zonule of Zinn, intraocular lens implantation with suture fixation and anterior vitrectomy were performed when deemed necessary. Cases with elevated intraocular pressure (IOP) after surgery were first treated with an IOP-lowering medication. Glaucoma surgery, including trabeculotomy as a ‘first-choice’ treatment and trabeculectomy in cases with wide-ranging peripheral anterior synechia, was performed only if the elevated IOP could not be medically controlled.

Clinical Evaluation

Data collected in this study included recipient age at the time of surgery, sex, diagnosis, history of diabetes, and whether PK was performed with or without intraocular lens implantation. Donor data included age, history of diabetes, sex, trephine size, preoperative ECD, cause of death, death to preservation time, and death to transplantation time. In all patients, postoperative ECD was measured at 1, 2, 3, 4, and 5 years postoperative by using a noncontact specular microscope (EM-1000 or EM-3000; Tomey Corporation, Nagoya, Japan) and was determined by using a center method with manual adjustment. The preoperative ECD data were obtained by the 2 abovementioned eye banks through specular microscopy after manual identification of the central corneal endothelium in the donor eye. Donor cause of death was classified into 2 groups as follows: 1) acute [eg, heart disease, cerebrovascular disease, or acute respiratory failure (eg, asphyxia)] and 2) chronic (eg, a malignant tumor or chronic liver disease). ECL was expressed as the difference in ECD between preoperative and that at 5 years postoperative as a percentage of preoperative ECD.

Statistical Analysis

Statistical analysis was performed using JMP ver.14 Statistical Software (SAS Institute Inc, Cary, NC). Annual rate of ECL between 1 and 5 years postoperative was calculated based on the linear approximation method by Excel 2019 software (Microsoft corp, Redmond, WA). Multivariate factors analysis adjusting for baseline donor ECD was performed with logistic regression analysis between the cases with and without a normal ECD (ie, greater than or less than 2000 cells/mm2) at 5 years postoperative to investigate the key factors associated with a high corneal ECD after surgery. Odds ratios and 95% confidence intervals were calculated, and a P-value of <0.05 was considered statistically significant.


Among the initial 225 eligible eyes, graft failure occurred in 23 eyes and the rate of graft survival at 5 years postoperative was 89.8% (202 eyes), and there was no allograft rejection in 202 eyes. In 14 eyes, the ECD could not be accurately or definitively determined because of corneal opacity and/or being borderline cases before graft failure. After the exclusion of the eyes with trauma (n = 2) and corneal infection (n = 2), 174 eyes were ultimately included in this study. Of those 174 eyes, 16 (9.2%) retained a central corneal ECD of ≥2000 cells/mm2 at 5 years postoperative. Of those 16 eyes, 6 (37.5%) had a recipient diagnosis of BK. Moreover, the ECDs (mean ± SD) in those 16 eyes before and 5 years after PK were 2837 ± 305 cells/mm2 and 2264 ± 205 cells/mm2, respectively, and the ECL (mean ± SD) at 5 years postoperative was 19.5% ± 9.8%. In the eyes with an ECD of <2000 cells/mm2 at 5 years postoperative, the ECDs (mean ± SD) before and 5 years after PK were 2823 ± 338 cells/mm2 and 823 ± 396 cells/mm2, respectively, and the ECL (mean ± SD) at 5 years postoperative was 70.9% ± 13.2% (Fig. 2). In the 16 eyes with an ECD of >2000 cells/mm2, there was minimal change of corneal ECD after the cell loss associated with the surgery (see Supplemental Figure 1, Supplemental Digital Content 1, The average annual rate of ECL between 1 year and 5 years postoperative in the 16 eyes with a high ECD at 5 years postoperative was 2.3% ± 3.7%, lower than the average annual rate of ECL (15.0% ± 5.0%) in eyes with an ECD of <2000 cells/mm2 at 5 years postoperative (P < 0.001).

Change in ECD over time showing all cases (total cases) and cases with an ECD of >2000 cells/mm2 (ie, an ECD of ≥2000) or <2000 cells/mm2 (ie, an ECD of <2000) at 5 years postoperative. Black dots and lines indicate mean values and SD.

Next, we performed a multivariate analysis adjusting for baseline donor ECD between the cases with an ECD of greater than and less than 2000 cells/mm2 at 5 years postoperative to investigate the key factors associated with a high corneal ECD after surgery. The multivariate analysis findings revealed that the donor-specific and surgery-related factors were not significant factors. Only the recipient diagnosis of BK was found to be significantly associated with an ECD of <2000 cells/mm2 at 5 years postoperative (Table 2). Six eyes (3%) had undergone glaucoma surgery. Of those 6 eyes, 1 (17%) had undergone a trabeculotomy at 3 years after PK, and the ECD in that eye was ≥2000 cells/mm2 at 5 years postoperative. In the other 5 eyes (1 trabeculectomy eye and 4 trabeculotomy eyes), the ECD at 5 years postoperative was <2000 cells/mm2.

Multivariate Factor Analysis Between Eyes With High ECD (ie, Over 2000 Cells/mm2, N = 16) at 5 Years Postoperative and the Other Eyes (N = 158)


The findings of this retrospective study of patients who underwent PK from 1998 thorough 2011 showed that even at 5 years postoperative, the central corneal ECD in 16 patients (9.2%) remained greater than 2000 cells/mm2. Of particular interest was the fact that the average annual rate of ECL in those patients was approximately 2%, which is far lower than the cell loss normally observed after corneal transplantation. In fact, in all of the other treated eyes in this study, there was a significantly higher annual rate of cell loss, even in the absence of allograft rejection.

In this present study, the multivariate analysis findings revealed that none of the donor factors and surgical factors were of any statistical significance. Only the recipient diagnosis of BK was significantly associated with an ECD of <2000 cells/mm2 at 5 years postoperative as several studies have reported.8,13,14,23 In this study, 6 (37.5%) of the 16 eyes with an ECD of ≥2000 cells/mm2 at 5 years postoperative were transplanted for BK, contrary to what one would normally expect with this higher-risk recipient diagnosis. Thus, it is possible that the healthiest, or “ideal”, donor corneas are able to maintain a high ECD over a long-term postoperative period even if the anterior-chamber microenvironment is otherwise not favorable.

We previously reported that the histological analysis findings of corneal endothelium preserved in storage media revealed that dead cells exist on the endothelium of the donor corneas and that those cells are lost after incubation,24 thus implying that the viability of individual donor corneal endothelial cells varies. Endothelial cell viability and function cannot be determined simply by measuring ECD through specular microscopy. Because the preoperative ECD was similar in both groups of eyes in this study, our results suggest that there are potentially unknown biologic factors that influence corneal ECL.

Future investigations of the key factors associated with the extensive longevity of corneal endothelial cells of donor corneas after PK might require tests indicative of endothelial cell function in addition to morphometric measurements by specular microscopy. By identifying the factors, or possible biomarkers, of such ideal corneal endothelial cells, donor corneas could possibly be selected to optimize outcomes. This theory is of high potential value when considering cultivated human corneal endothelial cell injection therapy, a cutting-edge treatment that we recently developed as a novel alternative therapy to traditional corneal transplantation.25 The selection of ideal donor corneal endothelial cells for expansion in vitro could possibly provide cells with good viability and longevity to provide more optimal clinical outcomes, may be even in patients with severe BK. Because cultivated cells from 1 donor can be expanded to treat multiple recipients, such an approach could possibly provide a high clinical value to many recipients with higher-risk diagnoses.

It should be noted that the present study did have several unique characteristics. First, this study involved a large number of participants (n = 174 eyes) who successfully completed a 5-year follow-up course after PK. Second, all surgeries in this study were performed by 1 corneal specialist surgeon at 1 institute. Third, all donor corneas used in this study were obtained from US eye banks. These factors suggest that any clinical bias, such as the surgeon's specific technique and the differences between a domestic and an imported donor cornea, was minimized as much as possible. Thus, we believe that the data in this study regarding the biological change in postoperative ECD after corneal transplantation are significant.

It should also be noted that this present retrospective study did have some limitations. All retrospective studies have some degree of selection bias. It was difficult for elderly patients and patients living in remote areas to continuously visit our institute for long-term follow-up examinations. Similarly, younger patients with keratoconus, that is, those who generally demonstrate good clinical outcomes, often chose to discontinue follow-up.14 In fact, of the 485 consecutive eyes, 260 eyes (53.6%) were not enrolled in this study. However, 225 (46.4%) of the 485 consecutive eyes in this study that underwent PK performed by 1 surgeon were followed up for at least 5 years postoperative. Hence, we consider that the findings in this study provide representative clinical outcomes. In support of this consideration, the postoperative ECD of the excluded 247 eyes was calculated, and there is no significant difference compared with the enrolled population in this study, donor factors including age, sex, preservation time, ECD, and any recipient factors including age, sex, and recipient diagnosis was not much different from the data for the eligible 225 eyes. Another possible limitation is that in all patients, the PK procedure was performed by 1 specific corneal specialist surgeon (S.K.). Thus, a multicenter study would be needed for a more detailed investigation.

The second limitation was that the number of cases with an ECD of ≥2000 cells/mm2 at 5 years postoperative was small (n = 16), thus potentially weakening the power of our statistics and the ability to detect the significant predictive factors. Thus, further studies with more cases are needed to confirm the results in this study.

Third, it should be noted that only PK cases were included in this long-term retrospective study, and it has been reported that there are some differences in the pattern of ECL among eyes that undergo PK, eyes that undergo DSAEK,15 and eyes that undergo Descemet membrane endothelial keratoplasty. However, similar to the results in the CDS,7 Price and associates reported that ECD at 10 years after DSAEK was significantly associated with that at 6 months after surgery,15 thus indicating that for patients undergoing DSAEK, the longevity of the donor-graft corneal endothelial cells might be explained by findings similar to those in this present study.

In conclusion, the findings in this study revealed that some cases of PK, even when performed for a higher-risk indication such as BK, retain a high corneal ECD at 5 years postoperative, contrary to the findings that are typically observed after PK. Hence, our findings suggest that there might be unknown biological factors unrelated to ECD that confer longevity to donor corneas or even biomarkers that predict endothelial cell resiliency. Moreover, the findings in our most recent study indicate that some biological factors in the donor eye corneal endothelial cells may be related to a high ECD at many years postoperative (Data presented at the American Academy of Ophthalmology Annual Meeting, 2019, San Francisco, CA; Manuscript currently in preparation).

Further investigation is needed to better understand the specific factors that affect the longevity of transplanted corneal endothelial cells (ie, the “seeds”) and the anterior-chamber microenvironment (ie, the “soil”) because the selection of ideal donor corneas could possibly drastically improve the prognosis of corneal transplantation in the future.


1. Bourne WM, Nelson LR, Hodge DO. Continued endothelial cell loss ten years after lens implantation. Ophthalmology. 1994;101:1014–1022; discussion 1022-1013.
2. Bourne WM, Nelson LR, Hodge DO. Central corneal endothelial cell changes over a ten-year period. Invest Ophthalmol Vis Sci. 1997;38:779–782.
3. Bourne WM, Hodge DO, Nelson LR. Corneal endothelium five years after transplantation. Am J Ophthalmol. 1994;118:185–196.
4. Tan DT, Dart JK, Holland EJ, et al. Corneal transplantation. Lancet. 2012;379:1749–1761.
5. Pluzsik MT, Seitz B, Flockerzi FA, et al. Changing trends in penetrating keratoplasty indications between 2011 and 2018—Histopathology of 2123 corneal buttons in a single center in Germany. Curr Eye Res. 2020;13:1–6.
6. Thompson RW Jr, Price MO, Bowers PJ, et al. Long-term graft survival after penetrating keratoplasty. Ophthalmology. 2003;110:1396–1402.
7. Sugar A, Gal RL, Kollman C, et al. Factors associated with corneal graft survival in the cornea donor study. JAMA Ophthalmol. 2015;133:246–254.
8. Patel SV, Diehl NN, Hodge DO, et al. Donor risk factors for graft failure in a 20-year study of penetrating keratoplasty. Arch Ophthalmol. 2010;128:418–425.
9. Lass JH, Sugar A, Benetz BA, et al. Endothelial cell density to predict endothelial graft failure after penetrating keratoplasty. Arch Ophthalmol. 2010;128:63–69.
10. Lass JH, Beck RW, Benetz BA, et al. Baseline factors related to endothelial cell loss following penetrating keratoplasty. Arch Ophthalmol. 2011;129:1149–1154.
11. Lass JH, Benetz BA, Gal RL, et al. Donor age and factors related to endothelial cell loss 10 years after penetrating keratoplasty: specular Microscopy Ancillary Study. Ophthalmology. 2013;120:2428–2435.
12. Mannis MJ, Holland EJ, Gal RL, et al. The effect of donor age on penetrating keratoplasty for endothelial disease: graft survival after 10 years in the Cornea Donor Study. Ophthalmology. 2013;120:2419–2427.
13. Tan DT, Janardhanan P, Zhou H, et al. Penetrating keratoplasty in Asian eyes: the Singapore corneal transplant study. Ophthalmology. 2008;115:975–982 e971.
14. Williams KA, Lowe M, Bartlett C, et al. Risk factors for human corneal graft failure within the Australian corneal graft registry. Transplantation. 2008;86:1720–1724.
15. Price MO, Calhoun P, Kollman C, et al. Descemet stripping endothelial keratoplasty: ten-year endothelial cell loss compared with penetrating keratoplasty. Ophthalmology. 2016;123:1421–1427.
16. Ang M, Soh Y, Htoon HM, et al. Five-year graft survival comparing Descemet stripping automated endothelial keratoplasty and penetrating keratoplasty. Ophthalmology. 2016;123:1646–1652.
17. Schlogl A, Tourtas T, Kruse FE, et al. Long-term clinical outcome after Descemet membrane endothelial keratoplasty. Am J Ophthalmol. 2016;169:218–226.
18. Baydoun L, Muller T, Lavy I, et al. Ten-year clinical outcome of the first patient undergoing Descemet membrane endothelial keratoplasty. Cornea. 2017;36:379–381.
19. Kinoshita S, Amano S, Inoue Y, et al. Grading for corneal endothelial damage [in Japanese]. Nippon Ganka Gakkai Zasshi. 2014;118:81–83.
20. Higa A, Sakai H, Sawaguchi S, et al. Corneal endothelial cell density and associated factors in a population-based study in Japan: the Kumejima study. Am J Ophthalmol. 2010;149:794–799.
21. Kitazawa K, Kayukawa K, Wakimasu K, et al. Cystoid macular edema after Descemet's stripping automated endothelial keratoplasty. Ophthalmology. 2017;124:572–573.
22. Kitazawa K, Wakimasu K, Kayukawa K, et al. Moderately long-term safety and efficacy of repeat penetrating keratoplasty. Cornea. 2018;37:1255–1259.
23. Anshu A, Lim LS, Htoon HM, et al. Postoperative risk factors influencing corneal graft survival in the Singapore Corneal Transplant Study. Am J Ophthalmol. 2011;151:442–448 e441.
24. Kitazawa K, Inatomi T, Tanioka H, et al. The existence of dead cells in donor corneal endothelium preserved with storage media. Br J Ophthalmol. 2017;101:1725–1730.
25. Kinoshita S, Koizumi N, Ueno M, et al. Injection of cultured cells with a ROCK inhibitor for bullous keratopathy. N Engl J Med. 2018;378:995–1003.

endothelial cell loss; corneal transplantation; DSAEK; PK; endothelial cell density

Supplemental Digital Content

Copyright © 2020 Wolters Kluwer Health, Inc. All rights reserved.