Abstract
Purpose:
To determine whether the reasons for keratoplasty have changed between 10 years in a tertiary care setting, with special attention to the rate of glaucoma drainage devices as a reason for keratoplasty.
Methods:
Patients aged >=18 who underwent keratoplasty at Mayo Clinic, Rochester, MN from 2005-2006 and 2015-2016 were studied. All reasons for keratoplasty performed in the study time period are assessed, including patients who previously had a glaucoma drainage device implanted in the same eye. After analyzing the reasons for keratoplasty, we assess whether the reasons for keratoplasty have changed between 2005-2006 and 2015-2016 in association with the increasing placement of glaucoma drainage devices.
Results:
The number of keratoplasty procedures performed in the two time periods increased by 62% from 163 (2005-2006) to 264 (2015-2016), while glaucoma drainage device placement increased by 164% from 80 GDD (2005-2006) to 211 GDD (2015-2016). While the performance of keratoplasty increased between the two points in time, the frequency of each cause for keratoplasty did not change significantly. The majority of keratoplasties were performed due to corneal disease, and glaucoma drainage devices made up a small portion of reasons for keratoplasty (2005-2006, 4.29%; 2015-2016, 5.68%).
Conclusions:
The frequency of glaucoma drainage devices as a reason for keratoplasty has not changed significantly between 10 years in this tertiary care setting. Patients with glaucoma drainage devices who later required keratoplasty had associated features including multiple surgical procedures and co-morbid infection, pseudophakic bullous keratopathy, Fuchs dystrophy, PXE, uveitis, and congenital glaucoma.
Keywords: glaucoma drainage device, glaucoma, corneal graft, corneal transplant, keratoplasty
Précis:
Over 10 years in a tertiary care setting, glaucoma drainage devices have not increased as a reason for keratoplasty.
Introduction:
Corneal decompensation is the most common long-term complication of glaucoma drainage devices (GDD), with evidence of corneal edema reported in 13% (45/345) of eyes 5 years after GDD surgery. [1, 2] There is evidence of a significant short term and long term decrease in corneal endothelial cell density following GDD implantation. [3–5] The patient population undergoing GDD implantation share other features that put them at risk of reduced corneal endothelial cell density including advanced age, multiple ocular surgeries, chronic inflammation, exfoliative glaucoma, and high intraocular pressure (IOP), which may further increase their risk for corneal decompensation. [6–8] Other mechanisms specific to GDD present an ongoing risk for corneal damage and decompensation, including hypotony, shallow anterior chamber, tube position, tube-cornea touch, and nutritional steal. [8–12] Medicare data has identified a 27% incidence of corneal edema 6 years after GDD versus 9% following trabeculectomy. [13] Despite these risks, use of GDD to treat glaucoma has increased while the use of trabeculectomy decreased. [14, 15] Theoretically, there may be a corresponding increase in the number of keratoplasties performed due to GDD. There is limited information on whether GDD is a main cause of corneal decompensation leading to keratoplasty in a tertiary care setting. To better understand whether increased use of GDD has influenced corneal decompensation to a significant extent, this study assesses the number of patients with GDD implants undergoing keratoplasty at two-time points. The purpose of this study is to determine whether the rate of keratoplasty due to GDD has changed over 10 years in a tertiary care setting. The total number of GDD and keratoplasties performed were quantified to define the growth of each procedure over 10 years while the number of surgeons performing each procedure in the two time periods was noted.
Methods:
The study was approved by the institutional review board of Mayo Clinic. The study complied with the Health Insurance Portability and Accessibility Act and adhered to the tenets of the Declaration of Helsinki. Informed consent was waived and the data were deidentified.
Patients:
The charts of all patients 18 years and older who underwent keratoplasty including penetrating keratoplasty (PK) and endothelial replacement (EK) procedures at our academic tertiary care center between 2005-2006 (n1=163) and 2015-2016 (n2=264) were reviewed. Patients were identified using PK and EK procedure codes to search the institution’s clinical data repository for procedures performed in the given time periods. For each patient, the reason for initial keratoplasty was evaluated. If the keratoplasty performed in the specified time periods was not the initial corneal transplant, the medical record was examined to identify the date and reason for the first keratoplasty. Any eye that had a GDD prior to keratoplasty was categorized as secondary to GDD in order to avoid underestimating the contribution of GDD to corneal decompensation. Patients who received a GDD after keratoplasty were included in the total number of GDD performed, but their keratoplasty was attributed to reasons other than secondary to GDD. Patients who received a corneal transplant with keratoprosthesis implantation were excluded.
Evaluation of Cause for Corneal Transplantation:
Because of the complexity underlying reasons for corneal graft failure, this study focuses on reasons for initial keratoplasty only. The reason for each patient’s first keratoplasty was determined by review of the entire medical record. The reasons for keratoplasty were categorized as follows: Fuchs dystrophy, keratoconus, infection (either active infection or scar due to prior infection), trauma, GDD, pseudophakic bullous keratopathy, multiple reasons including GDD, multiple reasons (not GDD), and other. Additional data extracted from the medical record were patient demographics (age, gender, and race), previous GDD procedure in the involved eye, and the total number of previous intraocular surgeries for the involved eye. If the patient had corneal transplant in both eyes during the study period, each involved eye was counted as a unique keratoplasty.
Statistical Analysis:
Patient demographics and clinical data were compared between two different study periods using paired t-tests for continuous data or Fisher exact tests for dichotomous data. Frequency and reasons for corneal transplant were compared between the two different study periods using Fisher exact tests. For all data p<0.05 was used to determine statistical significance unless otherwise noted.
Results:
163 patients underwent keratoplasty (PK and EK procedures) at the academic tertiary care center from 2005-2006 and 264 from 2015-2016. This was the first keratoplasty for 88% (144/163) of patients from 2005-2006 and 82% (217/264) from 2015-2016. For the other 19 patients in 2005-2006, initial keratoplasties were identified between 1976-2003 with 8 done at Mayo Clinic Rochester, MN and 11 done at an outside institution. For the 47 patients from 2015-2016, initial keratoplasties were identified between 1966-2016 with 29 out of 47 initial keratoplasties performed at Mayo Clinic Rochester, MN and 18 keratoplasties performed at an outside institution.
Patient Demographics
Patient demographics including age, gender, and race were statistically similar across the two different time periods (Table 1). There was a significant difference in the mean time from first GDD to initial keratoplasty between the two groups (26 months in 2005-2006 and 89.3 months in 2015-2016 (p=0.0136)) (Table 2).
Table 1: Patient Demographics for 427 Patients Enrolled.
Demographic factors were similar for patients undergoing keratoplasty in the two time periods.
| 2005-2006 (n=163) | 2015-2016 (n=264) | P value | |
|---|---|---|---|
| Mean Age in Years | 68 | 67.2 | 0.64 |
| Gender (%) | 0.54 | ||
| Female | 61% | 58% | |
| Male | 39% | 42% | |
| Race (% (n)) | 0.10 | ||
| White | 92.64% (151) | 91.29% (241) | |
| Black/African American/African | 1.23% (2) | 3.03% (8) |
Table 2: Time to Keratoplasty and Multiple Procedures.
GDD: glaucoma drainage device
Months from GDD to keratoplasty reported as mean (range) for n=7 (2005-2006) and n=15 patients (2015-2016). Comparison using the Wilcoxon Rank Sum test revealed a statistically significant increase in mean number of months from GDD to keratoplasty. The number of prior surgeries in the affected eye for all patients is reported as mean (range) for n=163 (2005-2006) and n=264 (2015-2016). The number of prior surgeries in the affected eye for patients with multiple reasons for keratoplasty including GDD is reported as mean (range) for n=7 (2005-2006) and 14 patients (2015-2016).
| 2005-2006 | 2015-2016 | p value | |
|---|---|---|---|
| Months from GDD to keratoplasty | 26 (4, 44) | 89.3 (6, 224) | 0.0136* |
| Number previous surgeries in eye, all patients | 0.68 (0, 6) | 0.85 (0, 9) | 0.12 |
| Number previous surgeries in eye, multiple reasons including GDD | 2.71 (1, 6) | 4 (2, 9) | 0.14 |
Reasons for keratoplasty
The reasons for keratoplasty included: Fuchs dystrophy, multiple reasons (not including patients with GDD), keratoconus, PBK, multiple reasons including GDD, infection, other, trauma, and GDD only (Table 3). Among those categorized as multiple reasons (not GDD), reasons included infection, pseudophakic bullous keratopathy, Fuchs, PXE, multiple procedures, uveitis, and congenital glaucoma. In each time period, the reasons for keratoplasty were similar between the 2005-2006 and 2015-2016 time periods (Figure 1). The frequency of each reason for keratoplasty shows no significant statistical change between the two points in time.
Table 3: Reasons for Keratoplasty.
GDD: glaucoma drainage device
Paired t-tests showed no statistically significant difference for any of the reasons for keratoplasty between the two-time points.
| Reason | 2005-2006 (n, %) | 2015-2016 (n, %) |
|---|---|---|
| Fuchs | 79 (48.5%) | 135 (51.1%) |
| Infection | 16 (9.8%) | 30 (11.4%) |
| Active infection | 5 (3.1%) | 17 (6.46%) |
| Scar due to prior infection | 11 (6.7%) | 13 (4.94%) |
| Keratoconus | 15 (9.2%) | 28 (10.6%) |
| Multiple reasons (not GDD) | 15 (9.2%) | 23 (8.7%) |
| Pseudophakic bullous keratopathy | 17 (10.4%) | 17 (6.4%) |
| Multiple reasons including GDD | 7 (4.3%) | 14 (5.3%) |
| Other | 6 (3.7%) | 9 (3.4%) |
| Trauma | 8 (4.9%) | 7 (2.7%) |
| GDD only | 0 | 1 (0.4%) |
| TOTAL | 163 | 264 |
Figure 1:
GDD: glaucoma drainage device
PBK: Pseudophakic bullous keratopathy
Reasons accounting for corneal graft by year.
In patients with multiple reasons for keratoplasty including GDD, GDD played the following role in each time period studied. From 2005-2006 (n1=163), keratoplasty was performed in 7 (4.3%) patients who had GDD in the same eye. All 7 of the patients received EK procedures. Five had a Baervedlt implant, 1 an Ahmed, and 1 a different type of tube shunt. The mean time from GDD placement to keratoplasty was 26 months with a range of 4 to 44 months. Of these 7, 1 GDD was malpositioned (too close to the cornea on exam or repositioned surgically at the time of the keratoplasty). Other contributing factors in the eye included infection, PBK, uveitis, trauma, and multiple procedures.
In 2015-2016 (n2=264), keratoplasty was performed in 14 (5.3%) patients with GDD. 5 patients underwent EK and 12 patients underwent PK. Of these 14 patients, 4 had corneal decompensation attributed to GDD and 2 other GDD were noted to be malpositioned. Five had a Baerveldt implant, 5 an Ahmed, and 4 were other tube shunts. The mean time from GDD placement to keratoplasty was 89.3 months with a range of 6 to 224 months. Other contributing factors in this group included endophthalmitis, angle-closure, PBK, hypotony, congenital glaucoma, and multiple procedures. Including the 1 patient whose keratoplasty was indicated exclusively due to GDD, a total of 5.68% (15/264) of keratoplasties from 2015-2016 may be attributed to GDD. The frequency of GDD as a reason for keratoplasty did not statistically differ between the two time periods (Table 4).
Table 4: Number of corneal grafts attributable to GDD.
GDD: glaucoma drainage device
The number of keratoplasty procedures attributable to GDD did not differ statistically between the two time periods.
| Not caused by GDD | Caused by GDD | p=0.39 | |
|---|---|---|---|
| 2005-2006 | 156 | 7 | |
| 2015-2016 | 249 | 15 |
Growth rate of keratoplasty procedures and GDD implantation over 10 years
In the 2005-2006 time period, 163 keratoplasties (PK and EK) were performed by 4 surgeons while 80 GDD were implanted by 3 surgeons at Mayo Clinic, Rochester, MN. From 2015-2016, 264 keratoplasties were performed by 3 surgeons and 211 GDD were implanted by 3 surgeons at Mayo Clinic, Rochester, MN (Table 5). The proportion of EK vs. PK in each time range differed significantly with more EK performed in 2015-2016 compared to the 2005-2006 time range (p<0.0001) (Table 5). There was a 62% increase in the number of keratoplasties performed while there was a 164% increase in GDD implantation over the 10 year period, where percent increase is calculated as 100*(n2-n1)/n1. Of the 80 GDD performed from 2005-2006, 9 eyes (11.3%) went on to undergo EK or PK (median time from GDD to keratoplasty was 1735 days with a range of 42 to 3956 days). Of the 211 GDD implanted from 2015-2016, 5 eyes (2.4%) underwent EK or PK (median value 468 days, range of 171 to 581 days from GDD to keratoplasty).
Table 5: Percent increase in GDD and corneal graft procedures over 10 years.
GDD: glaucoma drainage device
Percent increase is calculated as 100*(n2−n1)/n1. 3 GDD surgeons practiced in 2005-2006 and 3 GDD surgeons practiced in 2015-2016, with 1 of those being the same individual across both time periods. There were 4 surgeons performing corneal grafts in 2005-2006, and 3 from 2015-2016, with those 3 individuals being the same across both time periods. The proportion of PK versus EK corneal graft procedures was significantly different between the 2 time periods.
| 2005-2006 | 2015-2016 | Percent Increase | |
|---|---|---|---|
| # GDD procedures | 80 | 211 | 164% |
| # GDD surgeons (#same as previous time point) | 3 | 3 (1) | |
| # Corneal grafts | 163 | 264 | 62% |
| # PK | 128 | 81 | p<0.0001* |
| #EK | 35 | 183 | |
| # Corneal graft surgeons (#same as previous time point) | 4 | 3 (3) |
Discussion:
While previous studies report the occurrence of keratoplasty as an outcome measure following GDD, this study investigates the overall reasons for keratoplasty at two-time points over 10 years to determine if GDD is increasing as a cause for keratoplasty. Although the number of keratoplasties performed increased over a ten year period by 62% from 163 to 264, the frequency of each cause for keratoplasty did not change significantly. GDD made up only a small portion of reasons for keratoplasty (2005-2006, 4.29%; 2015-2016, 5.68%). In this population, the increase in keratoplasties is not due to an increase in number of surgeons. The number of surgeons performing EK and PK procedures actually decreased from 4 to 3, with those 3 surgeons being the same individuals in both time points. One potential reason for the increase in procedures despite the retirement of one surgeon from the practice is increased surgeon comfort with the technique leading to more frequently offering and performing the procedures.
This study identified Fuchs dystrophy as the leading indication for keratoplasty, accounting for approximately half of the keratoplasties performed in this setting. The second most common was multiple reasons (not GDD) and keratoconus followed by PBK, accounting for 6.4% to 10.6% of keratoplasties performed. The frequency of patients with multiple reasons (not GDD) reflects the complexity of patients seen in clinic who undergo keratoplasty and suggests the importance of multiple previous eye surgeries as a factor leading to corneal decompensation. Active infection or scar due to prior infection accounted for 3.1% to 6.7% of keratoplasties, respectively. Other causes and trauma were reasons for less than 4.9% of keratoplasties each.
GDD has empirically been thought to be a common indication for keratoplasty.[16, 17] In line with national trends, GDD placement increased over a ten year time period between 2005 and 2016 at the institution. The number of surgeons performing GDD remained the same, suggesting growth of the department did not contribute to the increase in GDD procedures. Even with the observed increase in both GDD implant (164%) and keratoplasty (62%) over the last 10 years in this tertiary care setting, the frequency of GDD as a cause for keratoplasty has not significantly increased and has remained relatively low in comparison to other indications for keratoplasty.
In this study, factors associating GDD and corneal decompensation include multiple surgical procedures and co-existing diseases such as infection, PBK, Fuchs dystrophy, PXE, uveitis, and congenital glaucoma. There are many proposed mechanisms for the relationship between GDD and corneal decompensation. For example, positioning too close to the cornea can lead to tube-cornea touch inducing endothelial damage. Alternatively, some patients may develop a shallow anterior chamber (hypotony) following GDD implantation, which is also thought to contribute to corneal endothelial cell loss.[8–10] In addition, glaucoma patients who require GDD often share other known risk factors for corneal endothelial cell loss, including advanced age, multiple ocular surgeries, chronic inflammation, and high intraocular pressure (IOP).[6–8] New evidence suggests changes to the metabolic profile of aqueous following GDD placement further contributes to corneal decompensation. [12]
The minimal role of GDD as a cause for keratoplasty over ten years despite a significant increase in performance of both procedures suggests that surgeons may be improving their technique of placing GDD over time. Additionally, time from first GDD implant to keratoplasty increased from an average of 26 to 89.9 months. This change was statistically significant (p=0.0136) and may indicate improvement in technique of GDD placement. An important factor for the increased number of keratoplasties is likely the lower threshold for surgery with evolving endothelial replacement procedures such as Descemet stripping and Descemet membrane endothelial keratoplasty. In this study, lower threshold for surgery does not appear to represent an increase in GDD associated keratoplasty.
Limitations & Future directions:
This study retrospectively reviewed cohorts of patients undergoing keratoplasty at two points in time. Due to the scope and cross-sectional nature of the study, statistical analysis was limited to comparing the frequencies between two points in time. This study design had the advantage of a much longer time scale and broader analysis than previous prospective studies in the field and provided insight to the variety of conditions which lead patients to cornea clinic for first keratoplasty. The conclusion that GDD is not increasing as a reason for keratoplasty is limited to a single tertiary care facility. The increase in each type of procedure is likely multifactorial and influenced by a variety of factors that were not measured here, including increasing referrals, number of providing surgeons, and surgeon comfort with procedures. This was not a population-based study, so the confounding factors of changing referral patterns and lack of continuity in follow-up may impact the reliability of the conclusions. The study predominantly included GDD placed by Mayo Clinic surgeons, but 3 of the 23patients with GDD undergoing keratoplasty had their GDD placed by referring physicians (one in 2005-2006 and two in 2015-2016).
Future investigations should examine more patients longitudinally and examine the incidence of corneal decompensation over time rather than between two-time points. Such longitudinal analysis would strengthen the conclusions drawn here. Additionally, the overlapping risk factors for advanced glaucoma predisposing to GDD and keratoplasty should be analyzed to determine their interaction and aid in early identification of patients at risk for going on to keratoplasty. A regression analysis of GDD patients with corneal decompensation to identify the most important risk factors for corneal decompensation would be useful but demand a larger cohort. From another perspective, it would be valuable to understand whether the dramatic change in PK and EK procedures have influenced outcomes in terms of glaucoma and re-grafting. Understanding influence on patients’ perceived quality of life, especially those that receive multiple procedures, is another important measure of success. Along with data on risk factors for corneal decompensation, this would represent a powerful clinical tool for physician-patient decision making.
Acknowledgments:
Supported in part by Mayo Foundation, Rochester, MN. Catherine Knier was supported by the National Institute of General Medical Sciences (T32 GM 65841) and thanks to the Mayo Clinic MSTP for fostering an outstanding environment for physician-scientist training.
Supported in part by National Institutes of Health research grant(s) and Mayo Foundation, Rochester, MN.
Footnotes
The author(s) have no proprietary or commercial interest in any materials discussed in this article.
Disclosure: The authors declare no conflict of interest.
References:
- 1.Christakis PG, Kalenak JW, Tsai JC, et al. The Ahmed Versus Baerveldt Study: Five-Year Treatment Outcomes. Ophthalmology 2016;123(10):2093–102 doi: 10.1016/j.ophtha.2016.06.035 [DOI] [PubMed] [Google Scholar]
- 2.Gedde SJ, Herndon LW, Brandt JD, et al. Postoperative complications in the Tube Versus Trabeculectomy (TVT) study during five years of follow-up. Am J Ophthalmol 2012;153(5):804–14 e1 doi: 10.1016/j.ajo.2011.10.024 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Casini G, Loiudice P, Pellegrini M, et al. Trabeculectomy Versus EX-PRESS Shunt Versus Ahmed Valve Implant: Short-term Effects on Corneal Endothelial Cells. Am J Ophthalmol 2015;160(6):1185–90.e1 doi: 10.1016/j.ajo.2015.08.022 [DOI] [PubMed] [Google Scholar]
- 4.Kim KN, Lee SB, Lee YH, Lee JJ, Lim HB, Kim C-S. Changes in corneal endothelial cell density and the cumulative risk of corneal decompensation after Ahmed glaucoma valve implantation. Br J Ophthalmol 2016;100(7):933–38 doi: 10.1136/bjophthalmol-2015-306894 [DOI] [PubMed] [Google Scholar]
- 5.Topouzis F, Coleman AL, Choplin N, et al. Follow-up of the original cohort with the Ahmed glaucoma valve implant. Am J Ophthalmol 1999;128(2):198–204 [DOI] [PubMed] [Google Scholar]
- 6.Buys YM, Chipman ML, Zack B, Rootman DS, Slomovic AR, Trope GE. Prospective randomized comparison of one- versus two-site Phacotrabeculectomy two-year results. Ophthalmology 2008;115(7):1130–33 e1 doi: 10.1016/j.ophtha.2007.09.007 [DOI] [PubMed] [Google Scholar]
- 7.Gagnon MM, Boisjoly HM, Brunette I, Charest M, Amyot M. Corneal endothelial cell density in glaucoma. Cornea 1997;16(3):314–8 [PubMed] [Google Scholar]
- 8.Lee E-K, Yun Y-J, Lee J-E, Yim J-H, Kim C-S. Changes in corneal endothelial cells after Ahmed glaucoma valve implantation: 2-year follow-up. Am J Ophthalmol 2009;148(3):361–7 doi: 10.1016/j.ajo.2009.04.016 [DOI] [PubMed] [Google Scholar]
- 9.Koo EB, Hou J, Keenan JD, Stamper RL, Jeng BH, Han Y. Effects of Glaucoma Tube Surgery on Corneal Endothelial Cells: A Review. Eye Contact Lens 2016;42(4):221–4 doi: 10.1097/ICL.0000000000000171 [DOI] [PubMed] [Google Scholar]
- 10.Seo JW, Lee JY, Nam DH, Lee DY. Comparison of the changes in corneal endothelial cells after pars plana and anterior chamber Ahmed valve implant. J Ophthalmol 2015;2015:486832 doi: 10.1155/2015/486832 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Lee RK, Fantes F. Surgical management of patients with combined glaucoma and corneal transplant surgery. Curr Opin Ophthalmol 2003;14(2):95–9 [DOI] [PubMed] [Google Scholar]
- 12.Williamson BK, Hawkey NM, Blake DA, et al. The Effects of Glaucoma Drainage Devices on Oxygen Tension, Glycolytic Metabolites, and Metabolomics Profile of Aqueous Humor in the Rabbit. Transl Vis Sci Technol 2018;7(1):14 doi: 10.1167/tvst.7.1.14 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Stein JD, Ruiz D Jr., Belsky D, Lee PP, Sloan FA. Longitudinal rates of postoperative adverse outcomes after glaucoma surgery among medicare beneficiaries 1994 to 2005. Ophthalmology 2008;115(7):1109–16 e7 doi: 10.1016/j.ophtha.2008.03.033 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Arora KS, Robin AL, Corcoran KJ, Corcoran SL, Ramulu PY. Use of Various Glaucoma Surgeries and Procedures in Medicare Beneficiaries from 1994 to 2012. Ophthalmology 2015;122(8):1615–24 doi: 10.1016/j.ophtha.2015.04.015 [DOI] [PubMed] [Google Scholar]
- 15.Vinod K, Gedde SJ, Feuer WJ, et al. Practice Preferences for Glaucoma Surgery: A Survey of the American Glaucoma Society. J Glaucoma 2017;26(8):687–93 doi: 10.1097/IJG.0000000000000720 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Duarte MC, Herndon LW, Gupta PK, Afshari NA. DSEK in eyes with double glaucoma tubes. Ophthalmology 2008;115(8):1435, 35.e1 doi: 10.1016/j.ophtha.2008.03.003 [DOI] [PubMed] [Google Scholar]
- 17.Nahum Y, Mimouni M, Busin M. Risk Factors Predicting the Need for Graft Exchange After Descemet Stripping Automated Endothelial Keratoplasty. Cornea 2015;34(8):876–9 doi: 10.1097/ICO.0000000000000455 [DOI] [PubMed] [Google Scholar]