Changing Practice Patterns and Long-term Outcomes of Endothelial Versus Penetrating Keratoplasty: A Prospective Dutch Registry Study




Purpose


To compare graft survival, best-corrected visual acuity (BCVA), endothelial cell density (ECD), and refraction following penetrating keratoplasty (PK) vs endothelial keratoplasty (EK) for Fuchs endothelial dystrophy (FED) and pseudophakic bullous keratopathy (PBK).


Design


Nonrandomized treatment comparison with national registry data.


Methods


All consecutive patients undergoing first keratoplasty for FED and PBK between 1998 and 2014 were analyzed, with a maximum follow-up of 5 years (mean ± SD follow-up 39 ± 20 months, range 0–60 months). Graft survival was analyzed using Kaplan-Meier survival curves and Cox regression analysis. BCVA, ECD, and refractive error were compared using linear mixed models. Main outcome measures were graft survival, BCVA, refraction, and ECD.


Results


A total of 5115 keratoplasties (PK = 2390; EK = 2725) were identified. Two-year graft survival following EK was lower compared with PK (94.5% vs 96.3%, HR = 1.56, P = .001). Five-year survival was comparable for EK and PK (93.4% vs 89.7%, HR = 0.89, P = .261). EK graft survival improved significantly over time while remaining stable for PK. One-year BCVA was better following EK vs PK (0.34 vs 0.47 logMAR, P < .001). Astigmatism was lower 1 year after EK vs PK (−1.69 vs −3.52 D, P < .001). One-year ECD was lower after EK vs PK (1472 vs 1859 cells/mm 2 , P < .001). At 3 years, ECD did not differ between EK and PK.


Conclusions


Long-term graft survival after EK and PK is high and comparable despite lower short-term survival for EK. EK graft survival improved over time, suggesting a learning curve. EK results in better BCVA, lower astigmatism, and similar long-term ECD compared with PK for FED and PBK.


Corneal transplantation remains the only effective treatment for irreversible corneal endothelial dysfunction. For more than a century, penetrating keratoplasty (PK), in which the full-thickness host cornea is replaced with that of a donor, was the gold standard. In the last decade, endothelial keratoplasty (EK) techniques were introduced, allowing selective replacement of diseased corneal layers. Although EK and PK share similar goals, EK is suggested to be safer and allow for faster and more predictable visual recovery. As a result, EK gradually replaced PK as the treatment of choice for endothelial dysfunction. However, high-level evidence from randomized controlled trials (RCT) supporting the superiority of EK is currently lacking.


Recently, the superiority of EK was questioned by 2 national registry studies, which compared the outcomes of EK vs PK in the United Kingdom (UK) and Australia. Both studies reported lower 2-year graft survival rates following EK for the treatment of Fuchs endothelial dystrophy (FED) and pseudophakic bullous keratopathy (PBK). In addition, multiple previous studies suggest that long-term visual acuity following EK may be limited by chronic changes in the recipient cornea, the presence of a donor-recipient interface, and posterior corneal irregularities.


In the Netherlands, corneal transplant surgery is mostly performed in high-volume tertiary centers. Moreover, because Dutch surgeons were actively involved in the development of EK techniques, unique long-term experience has been gained, which may shed light on the relative merits of EK vs PK. In the current study, we report the changing practice patterns in keratoplasty and compare the long-term outcomes of EK and PK in terms of graft survival, visual acuity, refractive error, and endothelial cell density (ECD) for the treatment of FED and PBK based on prospectively collected data from the Netherlands Organ Transplantation Registry over the years 1998–2014.


Methods


Graft Registry


Data for this prospective multicenter registry study was obtained from the Netherlands Organ Transplantation Registry (NOTR), a prospective national database founded by the Netherlands Transplantation Foundation (Nederlandse Transplantatie Stichting [NTS], http://www.transplantatiestichting.nl/over-de-nts ). Informed consent was obtained from patients to participate in the registry and for use of anonymized data for research purposes. Institutional review board approval for data extraction and analysis was provided by the NOTR scientific council. Data were anonymized prior to analysis and the study was performed in accordance with the tenets of the Declaration of Helsinki and Dutch legislation.


Population


The study cohort included all consecutive primary endothelial and penetrating keratoplasties performed in the Netherlands for the treatment of FED or PBK between January 1, 1998 and December 31, 2013. No other inclusion or exclusion criteria were applied. The term EK was used regardless of donor corneal dissection technique, including Descemet stripping automated endothelial keratoplasty (DSAEK) and Descemet stripping endothelial keratoplasty (DSEK).


Data Collection


In the Netherlands, donor corneas are allocated centrally and the allocation process is always registered in NOTR. NTS prospectively collects data on the recipient, donor, eye bank processing, surgical procedure, and follow-up with regular intervals using a standardized electronic data capture system. Data collection continues until graft failure or loss to follow-up ( http://www.transplantatiestichting.nl/professionals/cijfers-voor-professionals-1/transplantatie-follow ). Overall registry data capture during the study period was 89%. Overall lost to follow-up (LTFU) between 1998 and 2014 measured 19% for EK and 31% for PK, P < .001. LTFU after introduction of EK into standard care in 2007 measured 19% for EK and 16% for PK, P = .13.


Outcome Measures


The primary outcome measure of this study was graft survival, with a maximum follow-up duration of 5 years. Grafts were identified as failed when either a reason for graft failure was documented by the treating surgeon or a subsequent record for corneal retransplantation in the same eye was identified in the registry. Graft failure occurring within 3 months of transplantation was defined as primary graft failure (PGF). For failed grafts, interval to graft failure was defined by the duration between transplantation and recorded graft failure. For surviving grafts, censuring was performed at the last recorded visit. All patients with follow-up durations longer than 5 years were censored at the 5-year time point. Secondary outcome measures included best-corrected visual acuity (BCVA), spherical equivalent (SEQ), refractive astigmatism, and ECD. BCVA was defined as Snellen acuity in the grafted eye measured using either spectacles or contact lens. The best visual acuity was noted. Snellen BCVA was converted to logarithm of the minimal angle of resolution (logMAR) for statistical analysis. SEQ was defined as the sum of the spherical value and half the cylindrical value.


Statistical Methods


All statistical procedures were performed using SPSS software version 22.0 (SPSS Inc, Chicago, Illinois, USA). Baseline characteristics were reported as mean ± standard deviation (SD), median, and interquartile range (IQR) or frequencies and percentages, as appropriate. Baseline demographics were compared using a Student t test for numerical variables or χ 2 test for categorical variables. Kaplan-Meier curves and Cox regression analysis were used to assess graft survival time. Transplantation indication was used as stratification variable. Linear mixed models (LMM) were fitted to investigate postoperative BCVA, SEQ, refractive astigmatism, and ECD over time. These dependent variables were reported as estimated marginal means for the combination of transplantation techniques (ie, EK and PK) and follow-up time points, evaluated at the means of the baseline of the dependent. Transplantation techniques, dependent variable at baseline, follow-up time points, and all interactions were included in the model as fixed variables; patient was included as a random variable; follow-up time points were taken as categorical variable. The level of statistical significance was set at .05 for all analyses.




Results


Demographics


Recipient and donor demographics, indication for surgery, surgical procedure, and follow-up duration are given in Table 1 . Demographics did not differ between patients undergoing EK or PK for FED or PBK, respectively. Follow-up duration was shorter for EK than for PK for both indications.



Table 1

Dutch National Registry Study of Endothelial Versus Penetrating Keratoplasty: Recipient and Donor Demographics


































FED EK PK EK vs PK
Number of procedures (n/%) 2441/63% 1449/37%
Recipient age, y (mean ± SD) 71 ± 9 71 ± 10 P = .78
Recipient sex, male (n/%) 955/39% 541/37% P = .27
Donor age, y (mean, ± SD) 67 ± 10 67 ± 10 P = .14
Follow-up, mo (median/IQR) 30/18–48 60/39–60 P < .001


































PBK EK PK EK vs PK
Number of procedures (n/%) 284/24% 941/76%
Recipient age, y (mean ± SD) 72 ± 13 72 ± 12 P = .37
Recipient sex, male (n/%) 121/43% 398/42% P = .93
Donor age, y (mean, SD) 66 ± 10 67 ± 11 P = .08
Follow-up, mo (median/IQR) 24/14–43 48/23–60 P < .001

EK = endothelial keratoplasty; FED = Fuchs endothelial dystrophy; IQR = interquartile range; PBK = pseudophakic bullous keratopathy; PK = penetrating keratoplasty; SD = standard deviation.


Keratoplasty Practice Patterns and Indications


The total number of corneal transplantations performed annually for FED and PBK remained constant between 1998 and 2007. Between 2007 and 2014, the total number of corneal transplantations increased owing to an increase in the number of transplantations performed for FED ( Figure 1 , Top). A coinciding increase in the number of EK procedures and decrease in the number of PK procedures for both indications was observed ( Figure 1 , Bottom). The level of visual limitation at which EK was performed for the treatment of FED decreased significantly between 2006 and 2014 (β = −0.02 logMAR/year, P < .001) while remaining unchanged for PBK (−0.01 logMAR/year, P = .78). The age at which EK was performed did not decrease between 2008 and 2014 for either FED (0.09/year, P = .38) or PBK (−0.46/year, P = .19). Improvement of vision was the main indication for keratoplasty (FED: 95%, PBK: 77%), followed by combined pain reduction and visual improvement (FED: 4.5%, PBK: 16%) and pain reduction only (FED: 0.5%, PBK: 7%).




Figure 1


Practice patterns of corneal transplantation in the Netherlands. Total number of corneal transplantations performed annually for FED (blue) and PBK (green) in the Netherlands (Top). Total number of endothelial and penetrating keratoplasties performed for FED (green and blue, respectively) and PBK (purple and yellow, respectively) in the Netherlands (Bottom). FED = Fuchs endothelial dystrophy; PBK = pseudophakic bullous keratopathy.


Graft Survival


During the first 2 years of follow-up, graft survival was significantly lower following EK compared with PK for FED (95.2% vs 97.4%, hazard ratio [HR] = 1.99 [95% confidence interval (CI): 1.38–2.88], P < .001) and PBK (88.7% vs 94.5%, HR = 2.29 [95% CI: 1.47–3.56], P < .001). The Supplementary Figure (Top left) shows 2-year EK and PK graft survival for FED and PBK (Supplemental Material available at AJO.com ). PGF accounted for 27% and 33% of total EK graft failures and 2% and 3% of total PK graft failures for FED and PBK, respectively. Between 2 and 5 years of follow-up, graft survival was significantly better following EK compared with PK for FED (98.2% vs 95.2, HR = 0.58 [95% CI: 0.36–0.91], P = .019) and PBK (97.2% vs 85.9%, HR = 0.27 [95% CI: 0.10–0.72], P = .009). The Supplementary Figure (Top right) shows 2- to 5-year EK and PK graft survival for FED and PBK. The hazard ratio for EK vs PK decreased over time, dropping below 1 at 15 months of follow-up for FED and PBK. At 5 years follow-up, Cox regression analysis showed no difference in graft survival between EK and PK for FED (94.1% vs 93.3%, HR = 1.23 [95% CI: 0.95–1.60], P = .12) or PBK (87.3% vs 84.1%, HR = 1.18 [95% CI: 0.82–1.71], P = .37). Five-year graft survival was significantly better for FED compared with PBK for EK (94.1% vs 87.3%, HR = 2.29 [95% CI: 1.59–3.30], P < .001) and PK (93.3% vs 84.1%, HR = 2.92 [95% CI: 2.26–3.77], P < .001). Figure 2 (Top left) shows 5-year EK and PK graft survival for FED and PBK. Finally, 5-year graft survival following EK significantly improved over time for both FED (range = 82.0%–96.6%, HR = 0.67 [95% CI: 0.56–0.80], P < .001) and PBK (range = 74.2%–93.4%, HR = 0.71 [95% CI: 0.53–0.95], P = .022) while remaining stable following PK for the treatment of FED (range = 87.1%–92.5%, HR = 1.33 [95% CI: 0.95–1.87], P = .10) and PBK (range = 77.7%–92.6%, HR = 1.01 [95% CI: 0.77–1.32], P = .95).




Figure 2


Endothelial and penetrating keratoplasty graft survival. Five-year graft survival following endothelial keratoplasty (green) and penetrating keratoplasty (blue) for Fuchs endothelial dystrophy (Top left) and pseudophakic bullous keratopathy (Top right). The 2-year cutoff point in graft survival is shown in purple. Evolution of graft survival over time stratified by 2-year intervals for date of surgery following endothelial keratoplasty for Fuchs endothelial dystrophy (Bottom left) and pseudophakic bullous keratopathy (Bottom right).


The evolution of EK graft survival for FED and PBK over time, stratified by 2-year intervals for date of surgery, is shown in Figure 2 (Top right), demonstrating that more recently performed endothelial grafts show better survival.


Best-Corrected Visual Acuity


BCVA following EK and PK for FED and PBK is given in Table 2 . Preoperatively, BCVA was better in patients undergoing EK compared to PK for FED (0.68 ± 0.43 vs. 0.90 ± 0.52 logMAR, P < 0.001) and PBK (1.32 ± 0.65 vs. 1.60 ± 0.62 logMAR, P < 0.001). Correcting for preoperative differences in BCVA, postoperative BCVA was significantly better following EK compared to PK for the treatment of FED until 36 months of follow-up. Thereafter, BCVA remained better in the EK group, but did not reach statistical significance. Correcting for preoperative differences in BCVA, postoperative BCVA was significantly better following EK compared with PK for the treatment of PBK throughout the entire follow-up.



Table 2

Dutch National Registry Study of Endothelial Versus Penetrating Keratoplasty: Best Corrected Visual Acuity
























































FED
Follow-up (months)
Endothelial Keratoplasty Penetrating Keratoplasty EK vs PK
P
N logMAR N logMAR
3 months 2396 0.37 1381 0.53 <.001
6 months 1973 0.32 457 0.47 <.001
12 months 2076 0.29 897 0.39 <.001
24 months 1902 0.27 961 0.34 <.001
36 months 1173 0.26 824 0.35 <.001
60 months 715 0.29 699 0.32 .166
























































PBK
Follow-up
Endothelial Keratoplasty Penetrating Keratoplasty EK vs PK
P
N logMAR N logMAR
3 months 217 0.65 267 0.85 <.001
6 months 210 0.58 516 0.77 <.001
12 months 179 0.55 522 0.74 <.001
24 months 96 0.50 389 0.69 <.001
36 months 48 0.52 270 0.71 <.001
60 months 23 0.51 169 0.70 <.001

EK = endothelial keratoplasty; FED = Fuchs endothelial dystrophy; PBK = pseudophakic bullous keratopathy; PK = penetrating keratoplasty.

Best-corrected visual acuity (logMAR) following EK and PK for FED and PBK, corrected for baseline best-corrected visual acuity (linear mixed-model estimated marginal means).


Refractive Outcomes


Refractive astigmatism and SEQ are given in Table 3 . Correcting for preoperative differences, postoperative SEQ and refractive astigmatism were significantly higher following PK compared with EK throughout the entire follow-up. Refractive shift is shown in Table 4 . Following EK, a significant hyperopic refractive shift measuring 0.34–0.67 diopter was observed throughout the entire follow-up.



Table 3

Dutch National Registry Study of Endothelial Versus Penetrating Keratoplasty: Spherical Equivalent and Refractive Astigmatism
























































Follow-up Endothelial Keratoplasty Penetrating Keratoplasty EK vs PK
P
N SEQ N SEQ
3 months 1956 1.37 532 2.87 <.001
6 months 1849 1.35 1076 2.85 <.001
12 months 1476 1.31 1108 2.81 <.001
24 months 822 1.23 822 2.73 <.001
36 months 505 1.22 647 2.72 <.001
60 months 164 1.15 449 2.70 <.001
























































Follow-up Endothelial Keratoplasty Penetrating Keratoplasty EK vs PK
P
N Astigmatism N Astigmatism
3 months 1956 −1.69 532 −4.19 <.001
6 months 1849 −1.72 1076 −3.68 <.001
12 months 1476 −1.69 1108 −3.52 <.001
24 months 822 −1.77 822 −3.62 <.001
36 months 505 −1.91 647 −3.64 <.001
60 months 164 −2.29 449 −3.74 <.001

EK = endothelial keratoplasty; PK = penetrating keratoplasty; SEQ = spherical equivalent.

Absolute spherical equivalent (SEQ, diopter) and refractive astigmatism (diopter) following EK and PK, corrected for baseline differences (linear mixed-model estimated marginal means).


Table 4

Dutch National Registry Study of Endothelial Versus Penetrating Keratoplasty: Refractive Shift































































Follow-up Endothelial Keratoplasty Penetrating Keratoplasty
N Refractive Shift P N Refractive Shift P
3 months 1956 0.67 <.001 532 −0.63 .001
6 months 1849 0.57 <.001 1076 −0.78 <.001
12 months 1476 0.57 <.001 1108 −1.16 <.001
24 months 822 0.57 <.001 822 −1.89 <.001
36 months 505 0.45 <.001 647 −2.00 <.001
60 months 164 0.34 .026 449 −2.24 <.001

Refractive shift (diopter) following endothelial and penetrating keratoplasty corrected for baseline differences (linear mixed-model estimated marginal means).

Only gold members can continue reading. Log In or Register to continue

Stay updated, free articles. Join our Telegram channel

Jan 5, 2017 | Posted by in OPHTHALMOLOGY | Comments Off on Changing Practice Patterns and Long-term Outcomes of Endothelial Versus Penetrating Keratoplasty: A Prospective Dutch Registry Study

Full access? Get Clinical Tree

Get Clinical Tree app for offline access