Boston Type 1 Keratoprosthesis versus Repeat Donor Keratoplasty for Corneal Graft Failure

Boston Type 1 Keratoprosthesis versus Repeat Donor Keratoplasty for Corneal Graft Failure A Systematic Review and Meta-Analysis Sumayya Ahmad, MD,1 Priya M. Mathews, MD, MPH,1 Kristina Lindsley, MS,2 Majed Alkharashi, MD,5 Frank S. Hwang, MD,4 Sueko M. Ng, MHS,2 Anthony J. Aldave, MD,3 Esen Karamursel Akpek, MD1 Purpose: To compare repeat penetrating keratoplasty (PK) with Boston type I keratoprosthesis (KPro) implantation for full-thickness donor corneal graft failure. Design: Previous donor graft failure is a common indication for both PK and KPro implantation. Selection of the surgical procedure is entirely dependent on the surgeon because there are no studies available for guidance. Therefore, a systematic review was undertaken to examine vision, device retention, graft clarity, and postoperative glaucoma and infection outcomes after repeat PK versus KPro implantation. Methods: Articles with data regarding repeat PK published between 1990 and 2014 were identified in PubMed, EMBASE, the Latin American and Caribbean Health Sciences Literature Database, and the Cochrane Central Register of Controlled Trials and were reviewed. Results were compared with a retrospective review of consecutive, nonrandomized, longitudinal case series of KPro implantations performed at 5 tertiary care centers in the United States. Visual acuity at 2 years was the primary outcome measure. The proportion of clear grafts in the repeat PK group, device retention in the KPro group, and the development of postoperative glaucoma and infection were secondary outcome measures. Results: The search strategy identified 17 128 articles in the PK analysis. After screening, 26 studies (21 case series and 5 cohort studies) were included in the review. Pooled analysis of the 26 unique studies demonstrated a 42% (95% confidence interval [CI], 30%e56%) likelihood of maintaining 20/200 or better at 2 years after repeat PK, compared with an 80% (95% CI, 68%e88%) probability with KPro implantation. The probability of maintaining a clear graft at 5 years was 47% (95% CI, 40%e54%) after repeat PK, whereas the probability of retention of the KPro at 5 years was 75% (95% CI, 64%e84%). The rate of progression of glaucoma at 3 years was 25% (95% CI, 10%e44%) after repeat PK and 30% in the KPro cohort. Conclusions: These results demonstrate favorable outcomes of KPro surgery for donor corneal graft failure with a greater likelihood of maintaining visual improvement without higher risk of postoperative glaucoma compared with repeat donor PK. Ophthalmology 2015;-:1e13 ª 2015 by the American Academy of Ophthalmology. Supplemental material is available at www.aaojournal.org.

Corneal transplantation is one of the most commonly performed transplantation procedures in the developed world. In 2014 in the United States, 46 173 corneal transplantation surgeries were performed.1 Of these, 19 294 were penetrating keratoplasty (PK) procedures and the rest were anterior or endothelial lamellar grafts. Repeat grafts (n ¼ 4399) accounted for the second most common indication for PK, after keratoconus. Despite the high rate of success of primary PK for indications such as keratoconus, corneal scarring, or endothelial failure, repeat PK is a well-known risk factor for graft failure.2e6 The most common cause of graft failure is endothelial  2015 by the American Academy of Ophthalmology Published by Elsevier Inc.

rejection.7 For patients with a high risk of graft failure, another option is available. The Boston type 1 keratoprosthesis (KPro) has gained popularity in the last decade because of favorable outcomes reported by the most recent published series.8e13 There are no published guidelines as to when a KPro may benefit patients with donor graft failure compared with repeat PK. It is also unknown how many repeat PKs can be performed with reasonable expectation of improvement in vision. A recent Cochrane review demonstrated an absence of studies directly comparing the outcomes of these 2 procedures.14 The purpose of this review was to evaluate

http://dx.doi.org/10.1016/j.ophtha.2015.09.028 ISSN 0161-6420/15

1

Ophthalmology Volume -, Number -, Month 2015 systematically the clinical outcomes of repeat PK versus KPro implantation based on available published studies.

Methods Search Strategy We searched the Cochrane Central Register of Controlled Trials, which contains the Cochrane Eyes and Vision Group Trials Register, Excerpta Medica dataBASE (EMBASE), the Latin American and Caribbean Health Sciences Literature Database, and PubMed in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analysis consensus statement.15 We conducted a highly sensitive search to identify studies examining keratoprostheses and human donor corneas for patients with graft failure. The electronic searches were performed on August 13, 2014, by one author (K.L.). A detailed description of the search strategy for the various databases appears in the Appendix (available at www.aaojournal.org). The clinical trial registries search as well as a prior report did not identify any ongoing clinical trials comparing these 2 outcomes.14

Inclusion and Exclusion Criteria Three pairs of authors (K.L. and M.A., S.A. and P.M.M., and S.M.N. and F.S.H.) assessed the search results independently. Authors classified each record based on title and abstract as (1) definitely relevant, (2) possibly relevant, or (3) definitely not relevant. Studies assessed as (1) or (2) were assessed for eligibility independently by 2 authors (S.A., P.M.M.). Discrepancies were resolved by consensus. No study was classified as unclear after review of the full text. We planned to include nonrandomized control trials, prospective and retrospective cohort studies, and interventional case series that were published between 1990 and 2014. Studies that reported on fewer than 20 patients or cases were excluded. We grouped multiple reports from the same study when authors only reported updated information (e.g., results from longer follow-up, data for a subset of patients) on the same group of patients. We included studies of participants 18 years of age and older with corneal opacity who had 1 or more failed PKs. Penetrating keratoplasty was defined as a full-thickness optical donor corneal transplantation performed with the intention of improving the vision. Anterior or deep anterior lamellar keratoplasty or endothelial keratoplasty cases were excluded. We did not exclude studies that also included participants younger than 18 years of age unless the results were exclusively or mostly in pediatric patients. We also excluded studies that examined patients exclusively with ocular surface disease (OSD) because of the significant comorbidity in these eyes. However, studies that included patients with OSD as part of their larger grouped results were not excluded. When multiple publications from the same study population were available, we checked for duplicate analysis and included only the most recent publication. A total of 19 unique studies were identified within the keratoprosthesis analysis.6,8e13,16e47 There were 14 studies reporting on the KPro, which consisted of 13 case series and 1 large multicenter report. The Boston Type 1 KPro Study Group report consisted of 19 surgeons at 18 medical centers who performed surgery between 2003 and 2008. Several of these surgeons published separate reports of outcomes from their own patients, making it difficult to assess the cohorts independently. Although visual acuity (VA) and retention rates were reported in many of these studies, few studies reported outcomes at a specified time point and there were scarce data beyond 1 year. In addition, many of these cohorts did not

2

report outcomes separately for patients with a history of previously failed PK. Most importantly, most previous studies included the same eye twice in analyses even if prior KPro implantations were performed in the same eye. Considering the scarcity of primary outcomes reported at the time point of interest, we decided to use the results from a long-term multicenter cohort reporting on the Boston type 1 KPro. We believed that this would allow a more accurate comparison of outcomes for patients with a history of graft failure.

Data Extraction and Quality Assessment Two pairs of authors (S.M.N. and M.A., and K.L. and F.S.H.) extracted data independently onto data extraction forms adapted from Cochrane Eyes and Vision Group forms. Study characteristics extracted for each relevant study included methods, participants, interventions, outcomes, and funding sources. We resolved discrepancies by consensus or by conferring with a third author (S.N.).

Statistical Analysis The primary outcome for this review was the proportion of patients maintaining VA of 20/200 or better at 2 or more years after repeat PK surgery. Secondary outcomes included the proportion of patients achieving VA of 20/40 or better, progression of glaucoma, and rate of infectious keratitis. The proportion of graft failures at years 1, 2, and 5 after repeat PK surgery also was assessed. Glaucoma progression was defined as elevated intraocular pressure (IOP) requiring initiation of drops or additional therapy, but actual definition varied in each report. We planned to report the measure of effect as a relative risk with 95% confidence intervals (CIs) for comparative studies, or proportions for noncomparative case series. We planned to report dichotomous secondary outcomes in the same manner. We used the I2 statistic to examine statistical heterogeneity, with I2 values larger than 20% indicating substantial statistical heterogeneity. We also assessed clinical heterogeneity based on the characteristics of participants in the included studies, such as age, number of prior transplants, and underlying comorbidities (e.g., retinal detachment, glaucoma, and OSD). We expected substantial heterogeneity because of wide variation in the patient population (i.e., indication for surgery, preoperative VA, comorbid ocular conditions) and considered different ways to assess the outcomes of interest. Therefore, we planned to use the random effects model to account for the high interstudy variation. We intended to perform the analyses separately for cases with a history of only 1 prior PK and more than 2 prior PKs. Three pairs of review authors (S.M.N. and M.A., S.A. and P.M.M., and S.M.N. and F.S.H.) assessed independently the sources of systematic bias in studies according to the methods described in chapter 13 of the Cochrane Handbook for Systematic Reviews of Interventions.48 We resolved discrepancies through discussion. The meta-analysis was performed using the meta package from R statistical software (http://www.inside-r.org/ packages/cran/meta).49

Methods for Boston Type 1 Keratoprosthesis Analysis This was a retrospective multicenter review of eyes that underwent KPro implantation by 1 of 5 high-volume KPro surgeons at 5 tertiary care referral centers in the United States (Department of Ophthalmology, Albany Medical Center, Albany, New York; Flaum Eye Institute, Rochester, New York; Jules Stein Eye Institute, Los Angeles, California; Wills Eye Institute, Philadelphia Pennsylvania; and Wilmer Eye Institute, Baltimore, Maryland). All

Ahmad et al



KPro versus Repeat PK for Graft Failure

adult patients who underwent KPro surgery from January 2003 through December 2006 were considered. January 2003 was used as a cutoff to include only the eyes that received the fenestrated back plate device with lower complication rates (in contrast to the outdated solid back plate). A detailed description of the participants, data gathered, surgical technique, and follow-up care can be found elsewhere.50 Additional consecutive patients who met the study criteria at 1 center (Wilmer Eye Institute, Baltimore, MD) from January 2007 through July 2011 also were included. The study was reviewed and approved by the institutional review board at each site in accordance with the Declaration of Helsinki and complied with Health Insurance Portability and Accountability Act.

Data Collection and Analysis Information from each eye was collected retrospectively between May 2011 and April 2012 and entered in a uniform Microsoft Excel spreadsheet (Microsoft Corp, Redmond, WA) at each site. Data collection was completed as of November 2014 to allow for maximum follow-up. De-identified data then were reviewed by 2 of the authors (S.A. and P.M.M.) for completeness and consistency. Patients younger than 18 years at the time of surgery without at least 1 postoperative follow-up visit or without at least 1 prior PK were excluded from the analyses. Ocular surface disease was defined as the presence of severe keratoconjunctivitis sicca, cicatrizing conjunctivitis from chemical or thermal trauma, or autoimmune causes such as mucous membrane pemphigoid, StevenseJohnsons syndrome, or atopic disease. Snellen acuity was recorded as a primary outcome and converted to logarithm of the minimum angle of resolution (logMAR) units for the purposes of the analysis. Counting fingers, hand movements or light perception, and no light perception VA were converted to 1.8 logMAR, 2.3 logMAR, and 2.6 logMAR, respectively.51 The data were entered in Microsoft Excel spreadsheets, and further statistical analysis was performed using Stata software version 13.1 (Stata Corp, College Station, TX). Patient outcomes were estimated using Kaplan-Meier survival curves for all patients. Success was defined in 2 ways: (1) maintenance of VA 20/200 or better at various time points after the surgery or (2) retention for KPro and clear graft for repeat PK.

Results We identified 17 128 titles by the electronic searches as of August 14, 2014 (Fig 1). After removing 6546 duplicates, we identified 10 582 titles that potentially were eligible. After removing 2132 articles that were published before 1990, we evaluated 8450 studies by abstract review. Based on this assessment, we excluded 8062 studies that were not eligible because they did not meet inclusion criteria or did not specifically report on patients with regrafts (n ¼ 7899), were written in a foreign language or were conference abstracts (n ¼ 163), or both. We did not identify any relevant randomized control trials or clinical control trials comparing artificial versus donor corneas in patients undergoing repeat corneal transplantation. We retrieved 388 articles for full text review and further excluded studies that did not include patients with a history of failed PK, had fewer than 20 repeat PK patients, or did not report outcomes for repeat PK patients separately. We identified 45 reports of 26 nonrandomized studies evaluating outcomes for patients undergoing repeat PK after a history of failed corneal transplantation.3,5,52e76 The 26 studies (21 case series3,5,52,54,56,57,59,60,62e66,68e76 and 5 cohort studies53,55,58,61,67) included in this review enrolled 29 160 patients with 29 855 eyes. Of these patients, 5599 patients had undergone repeat PKs (Fig 1).

Figure 1. Flowchart describing the selection of studies for penetrating keratoplasty analyses.

The meta-analysis for primary outcome included 10 unique studies (3344 patients). Many studies did not specify the number of prior PKs each patient received. There were insufficient data available for a meta-analysis of the results of repeat PKs performed in patients who had 2 or more previous graft failures. Only 3 studies provided sufficient data for a VA outcome (20/40 vision) for repeat PK after only 1 previous graft failure. The meta-analysis for secondary outcomes included 12 unique studies. For the repeat PK after only 1 previous graft failure analysis, 10 studies provided sufficient data for a secondary outcomes analysis (1-, 2-, and 5-year survival and glaucoma progression).

Repeat Penetrating Keratoplasty Data Visual Acuity. Nine studies reported the proportion of patients who achieved a VA of 20/200 or better after repeat PK, with most having follow-up for at least 2 years after the surgery (range, 6e52 months; mean, 37 months).3,52,59,62,64,65,71,72,74 The likelihood of attaining 20/200 vision in patients followed up for at least 2 years after surgery was 42% (95% CI, 30%e56%; I2 ¼ 96.3%; Fig 2). Ten studies reported on the proportion of patients achieving a VA of 20/40 or better at least 2 years after repeat PK surgery (range, 6e52 months; mean, 35 months). In the meta-analysis, the random effects pooled estimate of achieving 20/40 VA at least 2 years after repeat PK surgery was 16% (95% CI, 11%e23%; I2 ¼ 89.6%; Fig 3).3,52,59,62,64,65,69,71,72,74 For the first repeat PK analysis, 3 studies reported on the proportion of patients achieving 20/40 vision at 2 years.59,64,69 The random effects pooled estimate of achieving 20/40 VA at least 2 years after the first repeat PK surgery was 16% (95% CI, 10%e22%; I2 ¼ 0%), which was identical to the all-repeat PK analysis. Graft Failure. The estimated proportion of patients with failure (reported by 17 studies) was 21% (95% CI, 15%e28%;

3

Ophthalmology Volume -, Number -, Month 2015

Figure 2. Table and graph showing the random effects analysis evaluating the proportion of patients with repeat grafts who achieved 20/200 or better vision at 2 years. CI ¼ confidence interval; W ¼ weight.

I2 ¼ 94.6%) at 1 year (Fig 4), 33% (95% CI, 25%e41%; I2 ¼ 89.8%) at 2 years (reported by 12 studies; Fig 5), and 53% (95% CI, 46%e60%; I2 ¼ 93.4%) at 5 years (reported by 16 studies; Fig 6).3,5,52e64,67e71,73e75,77,78 For patients with a single prior graft, the proportion of patients with repeat PK failure at 1 year was 27% (95% CI, 15%e41%; I2 ¼ 94.5%), the proportion at 2 years was 40% (95% CI, 28%e52%; I2 ¼ 39.7%), and the proportion at 5 years was 51% (95% CI, 34%e68%; I2 ¼ 92.4%).3,5,58e61,64,69,70,75 Complications. Eight studies reported the rate of glaucoma after repeat PK surgery.6,59,60,63,65,69,70,72 The proportion of patients having glaucoma after repeat PK was 25% (95% CI, 10%e 44%; I2 ¼ 95.4%; Fig 7). The average follow-up for these studies was 31 months (range, 9e72 months). For patients with a single prior graft, the proportion of patients having glaucoma at 2 years after repeat PK was 38% (95% CI, 21%e56%; I2 ¼ 86.5%). Three studies reported the infectious keratitis rate after repeat PK surgery, and their mean follow-up was 47 months (range,

25e72 months).52,59,72 The proportion of patients having infectious keratitis after repeat PK was 18% (95% CI, 9%e30%; I2 ¼ 83.4%; Fig 8). Because of the lack of reporting, it was unclear for most studies whether the cases were enrolled consecutively, whether they were assessed at similar time points, and when they were lost to followup. Additionally, there was a reporting bias in that many studies did not provide outcome data that may have been collected.

Keratoprosthesis Data A total of 104 eyes of 98 patients with a prior history of graft failure underwent KPro surgery for the first time between January 2003 and December 2006 at the above-mentioned sites. Of these, 31 had a history of only 1 previously failed PK. The mean followup time was 44.9 months (range, 0.03e105 months), with more than half of the eyes (53%) having more than 4 years of follow-up. Demographic characteristics are listed in Table 1.

Figure 3. Table and graph showing the random effects analysis evaluating the proportion of patients with repeat grafts who achieved 20/40 or better vision at 2 years. CI ¼ confidence interval; W ¼ weight.

4

Ahmad et al



KPro versus Repeat PK for Graft Failure

Figure 4. Table and graph showing the random effects analysis evaluating the proportion of patients with repeat grafts who have graft failure at 1 year. CI ¼ confidence interval; W ¼ weight.

Visual Acuity. Seventy-six eyes (76/104 [73%]) achieved 20/ 200 VA after the device was placed. At year 1, 17% of patients maintained 20/40 or better vision, and at years 2 and 5, 19.6% and 20%, respectively, maintained it. At year 1, 62% of eyes had a VA of 20/200 or better, and at years 2 and 5, 57.1% retained that level of vision, although VA data were not available for all patients

during each time point (Table 2). The probability of maintaining 20/200 or better VA was 90% (95% CI, 78%e94%) at 1 year, 80% (95% CI, 68%e88%) at 2 years, and 60% (95% CI, 45%e 71%) at 5 years using survival analysis models (Fig 9). For patients with a history of only 1 previously failed PK, the probability of maintaining 20/200 or better vision was 88%

Figure 5. Table and graph showing the random effects analysis evaluating the proportion of patients with repeat grafts who have graft failure at 2 years. CI ¼ confidence interval; W ¼ weight.

5

Ophthalmology Volume -, Number -, Month 2015

Figure 6. Table and graph showing the random effects analysis evaluating the proportion of patients with repeat grafts who have graft failure at 5 years. CI ¼ confidence interval; W ¼ weight.

(95% CI, 59%e97%) at 1 year, 74% (95% CI, 45%e89%) at 2 years, and 56% (CI, 25%e78%) at 5 years using survival analysis models. Retention. The explantation rate was 26% (27/104) at a mean follow-up of 50 months. The probability of explantation of the KPro was 1.1% (95% CI, 1%e7%) at 1 year, 5.5% (95% CI, 2%e 13%) at 2 years, and 25% (95% CI, 16%e36%) at 5 years using survival analysis models (Fig 10). For patients with a history of only 1 previously failed PK, the probability of explantation of the KPro was 3.7% (95% CI, 0%e23.5%) at 1 year, 7.8% (95% CI, 2%e28%) at 2 years, and 34% (95% CI, 17%e58%) at 5 years. Of the 4 underlying causes, OSD and infection had a

higher association with removal of the KPro (P ¼ 0.023 and hazard ratio of 2.53, and P ¼ 0.07 and hazard ratio of 2.35 for OSD and infection, respectively). Congenital corneal abnormalities were not associated with a higher risk of failure, and Fuchs dystrophy or bullous keratopathy were associated with a higher probability of retention (P ¼ 0.05; hazard ratio, 0.35). Complications. At year 2, 28.8% of patients experienced an elevation in IOP and 7.9% required glaucoma surgery. The rates were similar for patients with a prior history of only 1 failed PK (29% and 9.7%, respectively). At year 5, 30.7% of patients experienced an elevation in IOP and 13.5% required glaucoma surgery. For patients with a prior history of only 1 failed PK, 32.2%

Figure 7. Table and graph showing the random effects analysis evaluating the proportion of patients with repeat grafts in whom glaucoma develops. CI ¼ confidence interval; W ¼ weight.

6

Ahmad et al

KPro versus Repeat PK for Graft Failure



Figure 8. Table and graph showing the random effects analysis evaluating the proportion of patients with repeat grafts in whom infectious keratitis develops. CI ¼ confidence interval; W ¼ weight.

experienced an elevation in IOP and 9.7% required surgery at year 5. The rate of infectious keratitis and endophthalmitis at 5 years was 2.9% and 10.3%, respectively, for all patients and 6.5% and 14.3%, respectively, for patients with a history of only 1 previously failed PK (Table 3).

Discussion In our pooled results, we found a 42% probability of achieving 20/200 or better VA at 2 years with a repeat PK compared with an 80% probability at 2 years with a KPro implantation. We also found a 16% probability of achieving 20/40 or better VA Table 1. Baseline Characteristics of Study Participants Who Received a Boston Type 1 Keratoprosthesis for Graft Failure All Patients (n [ 104) Demographics Age at time of surgery, yrs 63.3 (21e90) (range) Female gender (%) 59 (56.7) Visual acuity, logMAR 1.92 (0.40e2.30) (range) Length of follow-up, mos 44.9 (0.03e105) (range) No. of prior transplants (%) 1 31 (29.8) 2 44 (42.3) 3þ 29 (27.9) Initial corneal diagnosis, no. (%) Ocular surface disease 16 (15.4) Infectious cause 13 (12.5) Congenital corneal 17 (16.3) abnormalities Fuchs dystrophy/bullous 36 (34.7) keratopathy/keratoconus Unknown/other 22 (21.5) Glaucoma status, no. (%) Preoperative glaucoma 68 (66) Taking glaucoma 50 (49.5) medications Previous glaucoma surgery 34 (32.6) Concurrent glaucoma 23 (22) surgery

Patients with 1 Prior Graft (n [ 31) 69.2 (23e90) 18 (58.1) 1.83 (0.54e2.3) 42.9 (0.03e83.9)

31 (100)

5 (16.1) 8 (25) 3 (9.7) 120 (32) 5 (16) 21 (67) 16 (51.6) 22 (71) 6 (19)

logMAR ¼ logarithm of the minimum angle of resolution.

at 2 years with a repeat PK and 19.6% with a KPro implantation. If we compare these results at the average follow-up for the PK studies (37 months) with those of KPro implantation, the visual prognosis remains higher for KPro implantation than for PK (72% for VA 20/200 with KPro implantation at 37 months). For patients who underwent a repeat PK after only 1 failed PK, achieving 20/40 or better VA at 2 years also was comparable with patients who had history of 2 or more previously failed PKs or a KPro implantation (16%, 17.6%, and 19.6%, respectively). There was significant variability noted within the repeat PK results. This was likely because of the diversity of underlying diagnoses, mean follow-up time, and the clinical setting in which the surgery was performed. The Australian Corneal Graft Registry reported better visual outcomes but a low rate of OSD (0.03%), with most patients having an underlying diagnosis of keratoconus and Fuchs dystrophy.71 In contrast, a King Khaled Eye Specialist Hospital study and a Turkish study had higher rates of OSD and unknown diagnoses, resulting in worse visual outcomes.52,64,72 These findings are consistent with other studies suggesting that the underlying initial diagnosis can serve as a predictor for regraft success.4 Furthermore, the King Khaled Eye Specialist Hospital study had a longer follow-up time (44 months) and the Turkish report pooled together second and third regrafts in their visual outcomes analysesdall features that could result in lower success rates. The patients in our KPro cohort are comparable with those of these pooled studies, with a wide range of underlying indications for the device (27% with OSD and infectious causes). Previous large case series also have demonstrated encouraging visual outcomes for patients receiving a KPro, which have demonstrated a range between 56.5% and 100% maintaining 20/200 or better VA (range of follow-up, 8.5 monthse3 years).8,9,11,13,36 The Boston KPro Study Group found rates of 56.4% with 20/200 or better VA and 22.6% with 20/40 or better VA at a mean follow-up of 14.5 months, which are consistent with our own results (Table 2).8 Inconsistencies of data reporting at various time points for the case series confound our outcomes, making survival analysis a more reliable indicator of vision potential. Survival analysis models from other studies have demonstrated a rate of 20/200 vision between 75% and 80% at 2-year follow up for KPro implantation, which is consistent with our own findings.10,36

7

Ophthalmology Volume -, Number -, Month 2015 Table 2. Visual Acuity Distributions in Patients Who Received a Boston Type 1 Keratoprosthesis for Graft Failure No. of Years after Surgery All Patients (%)

Patients with 1 Prior Graft (%)

Visual Acuity

1 (n ¼ 69)*

2 (n ¼ 56)*

3 (n ¼ 47)*

4 (n ¼ 41)*

5 (n ¼ 35)*

1 (n ¼ 18)*

2 (n ¼ 17)*

3 (n ¼ 12)*

4 (n ¼ 11)*

5 (n ¼ 7)*

20/40 <20/40e20/200 <20/200e20/400 CFeHM LP NLP

17.4 44.9 7.2 23.2 4.4 2.9

19.6 37.5 3.6 23.2 10.7 5.4

27.7 29.8 6.4 14.9 10.6 10.6

26.8 29.3 9.8 14.6 7.3 12.2

20.0 37.1 20.0 5.7 11.5 5.7

16.7 22.2 11.1 33.3 11.1 5.6

17.6 29.4 5.9 29.4 11.9 5.8

8.3 33.3 0 33.3 16.7 8.3

18.2 27.2 9.1 9.1 18.2 18.2

0 57.1 28.6 0 14.3 0

CF ¼ counting fingers; HM ¼ hand movements; LP ¼ light perception; NLP ¼ no light perception. *Number of patients at each time point with recorded visual acuity data

The retention rate in the KPro cohort was 98.9% at 1 year, 94% at 2 years, and 75% at 5 years using survival analysis models. The survival rate (clear graft) for the PK group was 79% at 1 year, 67% at 2 years, and 47% at 5 years. For patients with a history of only 1 prior graft, the survival rate (clear graft) for repeat PK was 73% at 1 year, 60% at 2 years, and 49% at 5 years, whereas the retention with KPro was 96%, 92%, and 66% at years 1, 2, and 5, respectively. Failure also was correlated with underlying diagnosis and healthcare setting. Three of the studies with high rates of 1-year PK failure were reports from a developing country (India) with increased OSD and variability in follow-up care and quality of donor tissue.5,57,69 The Singapore Eye Study also had a diverse population with both rural and urban patients and similar results. In contrast, the Price Vision Group had much better outcomes at 5 years for regrafts.67,68,71 The reason for this is unclear, but may be secondary to their healthier patient population, uniformity of care compared with multicenter studies, and a center effect in which centers that perform a high number of grafts report better outcomes. This center effect also may be secondary to more diligence in obtaining best-corrected VA in tertiary care centers, where there are specialty services, such as rigid

gas-permeable contact lens fitting performed by experienced optometrists. The KPro retention rates we report are similar to the multicenter Boston KPro Study Group findings. Further, a recent American Academy of Ophthalmology report noted an average retention of 88% with follow-up time between 8.5 and 21 months.8,9,11,79 Studies suggest that each subsequent regraft results in a higher likelihood of graft failure and worse visual prognosis.65,74,80 The average number of prior transplants in the KPro cohort was 2.1, which is higher than most regrafts reported in our PK analyses, in which several studies included patients who only had 1 prior regraft. When examining the regrafting success rate for the third or higher PK, the results are dismal. Beckingsale et al75 reported a 60% and 44% survival for second and third regrafts at 5 years, respectively, which has been echoed by others.58 The Wills Eye Hospital group reported a higher survival rate than most (78% for the third grafts and 73% at 2 years), which may be a reflection of their tendency to graft eyes they believe may have a higher chance of success.74 We also attempted to compare outcomes for patients with a history of a single previously failed PK. Interestingly, the repeat PK analyses demonstrated slightly

Figure 9. Graph showing the probability of maintaining 20/200 or better vision with a Boston type 1 keratoprosthesis after graft failure.

Figure 10. Graph showing the probability of retaining a Boston type 1 keratoprosthesis after graft failure.

8

Ahmad et al

KPro versus Repeat PK for Graft Failure



Table 3. Cumulative Incidence of Postoperative Complications in Patients Who Received a Boston Type 1 Keratoprosthesis for Graft Failure No. of Years after Surgery All Patients (%)

Patients with 1 Prior Graft (%)

Complication

1

2

3

4

5

1

2

3

4

5

KPro removal Retroprosthetic membrane* Persistent epithelial defect* Retinal detachment Elevated IOP Glaucoma requiring additional surgery Corneal infiltrate Endophthalmitis* Sterile keratolysis (melt)*

6.7 33.0 3.1 4.8 25 6.7 1.9 2.1 3.1

16.3 44.8 6.9 8.6 28.8 7.9 2.9 4.6 5.7

17.3 52.3 6.9 11.5 28.8 11.5 2.9 4.6 5.8

22.1 56.7 8.6 12.5 29.8 11.5 2.9 8.6 8.6

26.0 62.3 9.1 14.4 30.7 13.5 2.9 10.3 9.1

6.4 30 0 3.2 22.6 6.5 3.2 3.3 0

9.6 44.8 3.4 16.1 29.0 9.7 6.5 3.4 0

22.6 58.3 4.2 16.1 29.0 9.7 6.5 8.3 0

32.2 66.7 9.5 16.1 32.2 9.7 6.5 14.3 9.5

32.2 66.7 9.5 16.1 32.2 9.7 6.5 14.3 9.5

IOP ¼ intraocular pressure; KPro ¼ Boston type I keratoprosthesis. *Number at risk for these complications is the number of patients who retained the KPro at each time point.

high rates of graft failure at years 1 and 2 compared with the multiple graft analysis. This may be secondary to the few studies reporting such results and because several of these studies were from developing countries. At year 5, the rates of failure were almost identical for the first repeat PK analysis (51% vs. 53%). Unsurprisingly, the KPro implantations performed after only 1 failed donor PK had worse outcomes than those with a history of multiple failed grafts. This is likely because of selection bias. Typically, the eyes that were deemed higher risk for a traditional PK are offered a KPro. Indeed, 71% of these eyes had a history of prior glaucoma surgery, as compared with 33% for patients with 1 or more PKs, and a higher rate of KPro explantation at 5 years compared with the all-graft analysis (34% vs. 25%). Our results show a 25% incidence of glaucoma requiring intervention at 5 years after repeat PK, as compared with 30.7% by 5 years with KPro implantation. The PK studies had a wide range of definitions of glaucoma. The Toronto Western Hospital group defined it as elevated IOP, whereas the Wills Eye Hospital study and the Turkish study described it only within the context of failure or of vision loss secondary to glaucoma.65,70,72 The wide range of follow-up and variability in defining the disease confounds our pooled analysis. Regardless, our estimate is consistent with others’ findings; 2 other studies found a rate of 35% to 38% at 6 to 7.5 years after regrafting, respectively, when glaucoma was defined as elevation in IOP.59,81 It is well known that both repeat PK and KPro implantation are associated with high rates of glaucoma. This may be the result of a closed-angle mechanism in repeat PK, in which progressive synechiae have been documented, and also may be confounded by steroid use.82e84 There are further challenges in patients with KPro implantation. Because obtaining an accurate IOP often is not possible, monitoring the disease remains a challenge. A recent report demonstrated that the risk of glaucomatous vision loss (decrease in VA by 3 lines or more) is 27% at 3 years with KPro.85 However, this study also had a higher rate of primary KPro implantations, and the authors now place

tube shunts simultaneously with the KPro in all patients with a prior history of glaucoma, which is a practice that many centers also have adopted.86,87 Of note, there is potential morbidity associated with tube shunts, including postoperative tube exposure, which may result in delayedonset endophthalmitis. The patient population in our multicenter cohort had a high percentage of preoperative glaucoma (66%) and glaucoma surgery (33%), and roughly one quarter had simultaneous glaucoma surgery, highlighting the conservative measures taken by the providers. Finally, the results of the glaucoma analysis also may be confounded by the differences in IOP monitoring and detection of glaucoma progression by the various providers, which may be particularly difficult in the early postoperative period. The rate of infectious keratitis was 18% in the pooled PK results, whereas the rate of infectious keratitis in our KPro cohort demonstrated a 2.9% cumulative incidence at 5 years. A Moorfields study reported the highest rate of infectious keratitis after regraft of 30% at 6 years, representing their higher number of patients with ocular surface and herpetic disease.59 However, the progression to endophthalmitis is not clear from the studies included. A recent pooled analysis of KPro implantation found that the incidence of endophthalmitis was 5.4% of all patients who were reported in the literature between 2001 to 2011 with a wide range of follow-up and is consistent with our findings.88 This study does not address the question of under what circumstances and clinical grounds a corneal surgeon should go forward with a repeat donor corneal transplantation versus considering whether KPro implantation is appropriate in a patient already with a failed donor graft. However, our results show that over the intermediate-term postoperative follow-up period, the patients receiving KPro are more likely to achieve improvement in their VA and to retain this improvement than patients undergoing a repeat PK. In distinct contrast to common perceptions, the postoperative complication profiles were similar in each group.

9

Ophthalmology Volume -, Number -, Month 2015 The strengths of the meta-analysis are the broad range of studies included and diverse patient populations represented. Strengths of the retrospective cohort include its long follow-up time and multicenter feature. Limitations of this meta-analysis include potential biases in the original studies, methodologic issues, and different strategies for adjusting for confounders. Limitations of the KPro analysis include the comparison of a pooled analysis with a retrospective cohort as well as a survival bias. The patient population of our KPro cohort received care at a tertiary care setting, and there may be potential bias toward having patients with more complex eyes. Likewise, patients with significant pathologic features, including intraocular lens explantation and anterior or pars plana lensectomy (which commonly were performed concurrently with the KPro), could affect complication rates in the KPro cohort. Furthermore, it is difficult to compare retention with failure in these inherently diverse patient populations because they often have complex underlying pathologic features. We also acknowledge the difficulty in comparing proportions of failure in a metaanalysis and comparing the results with a retrospective case series. However, the scarcity of outcomes reported at the time points of interest and in patients with a history of graft failure convinced us to use this dataset rather than a pooled comparison. The diversity of underlying diagnoses for the KPro cohort is analogous to the pooled analysis. We believe that inclusion of OSD in both cohorts was important, particularly because of its significant burden on patients undergoing both interventions. The PK studies included a range of underlying pathologic characteristics, and most did not report separately on patients with OSD. Exclusion of all patients with OSD would have precluded our meta-analysis. However, overall we believe that our outcomes are consistent with publications from other groups reporting on both KPro implantation and repeat PK outcomes. Although KPro surgery was once considered a last-resort procedure owing to its side effect profile of glaucoma and infection, new designs and better postoperative management have led to more encouraging outcomes.50,89e91 Furthermore, the cost effectiveness of KPro implantation also may parallel that of PK surgery.18 We suspect that misconceptions of outcomes of the KPro implantation, stemming from the initial design, may be resulting in underuse of the device. Indeed, the Massachusetts Eye and Ear Infirmary reported only 600 grafts performed for KPro implantation in 2013 (Gelfand L, personal communication, 2013), as compared with 4261 for repeat PKs.1 Many providers, even within tertiary care settings, often are not comfortable managing KPro procedures and their complications. There also may be a need for outreach to and education for anterior segment surgeons who wish to use the device. That being said, there may be some evidence that the center effect could play an important role for patients with high-risk grafts and repeat PKs. Our results suggest that outcomes may be better for those centers that provide care to such patients on a regular basis. In conclusion, in selected patients who have experienced 1 or more failed full-thickness donor corneal transplantations, KPro implantation may be a superior alternative to provide visual rehabilitation with similar postoperative

10

risks to PK surgery with regard to serious complications. However, it must be emphasized that although our results indicate a similar complication profile and superior visual outcomes for KPro implantation, as soon as a KPro fails and requires explantation, the patient may have irreparable severe vision loss. The significant increase in explantation over time, particularly over the 3- to 5-year follow-up, is particularly significant. Although the graft clarity is a reasonable measure of success for PK at 5 years, loss of graft clarity does not inevitably imply permanent loss of vision or the inability to regraft or even convert to KPro implantation. This becomes particularly relevant in younger patients, for whom the KPro may not be the best option. Finally, newer techniques such as Descemet’s stripping endothelial keratoplasty also may play a role in selected patients with a previously failed PK. The preoperative patient profile predicting success with each of these procedures needs to be determined in prospective, randomized studies with long postoperative follow-up. Acknowledgments. The authors thank Esen K. Akpek, MD (Wilmer Eye Institute, Baltimore, MD); Anthony Aldave, MD (Jules Stein Eye Institute, Los Angeles, CA); James V. Aquavella, MD (Flaum Eye Institute, Rochester, NY); Michael Belin, MD (University of Arizona Department of Ophthalmology, Tucson, AZ); and Sadeer Hannush, MD, (Wills Eye Institute, Philadelphia, PA) for performing the KPro implantation surgeries; Janek Klawe for assisting in the outcomes analysis; and Divya Srikumaran for gathering the KPro data.

References 1. Eye Bank Association of America. Eye Banking Statistical Report. Washington, D.C.: Eye Bank Association of America; 2014. 2. Patel SV, Hodge DO, Bourne WM. Corneal endothelium and postoperative outcomes 15 years after penetrating keratoplasty. Am J Ophthalmol 2005;139:311–9. 3. Bersudsky V, Blum-Hareuveni T, Rehany U, Rumelt S. The profile of repeated corneal transplantation. Ophthalmology 2001;108:461–9. 4. Claesson M, Armitage WJ. Clinical outcome of repeat penetrating keratoplasty. Cornea 2013;32:1026–30. 5. Dandona L, Naduvilath TJ, Janarthanan M, et al. Survival analysis and visual outcome in a large series of corneal transplants in India. Br J Ophthalmol 1997;81: 726–31. 6. Harissi-Dagher M, Dohlman CH. The Boston Keratoprosthesis in severe ocular trauma. Can J Ophthalmol 2008;43:165–9. 7. Khodadoust A. The allograft rejection reaction: the leading cause of late failure of clinical corneal grafts. New York: Elsevier; 1973. 8. Zerbe BL, Belin MW, Ciolino JB. Results from the multicenter Boston Type 1 Keratoprosthesis Study. Ophthalmology 2006;113:1779.e1–e7. 9. Aldave AJ, Kamal KM, Vo RC, Yu F. The Boston type I keratoprosthesis: improving outcomes and expanding indications. Ophthalmology 2009;116:640–51. 10. Chew HF, Ayres BD, Hammersmith KM, et al. Boston keratoprosthesis outcomes and complications. Cornea 2009;28: 989–96. 11. Ciolino JB, Belin MW, Todani A, et al. Retention of the Boston keratoprosthesis type 1: multicenter study results. Ophthalmology 2013;120:1195–200.

Ahmad et al



KPro versus Repeat PK for Graft Failure

12. Greiner MA, Li JY, Mannis MJ. Longer-term vision outcomes and complications with the Boston type 1 keratoprosthesis at the University of California, Davis. Ophthalmology 2011;118: 1543–50. 13. Bradley JC, Hernandez EG, Schwab IR, Mannis MJ. Boston type 1 keratoprosthesis: the University of California Davis experience. Cornea 2009;28:321–7. 14. Akpek EK, Alkharashi M, Hwang FS, et al. Artificial corneas versus donor corneas for repeat corneal transplants. Cochrane Database Syst Rev 2014;11:CD009561. 15. Liberati A, Altman DG, Tetzlaff J, et al. The PRISMA statement for reporting systematic reviews and meta-analyses of studies that evaluate health care interventions: explanation and elaboration. Ann Intern Med 2009;151:W65–94. 16. Akpek EK, Harissi-Dagher M, Petrarca R, et al. Outcomes of Boston keratoprosthesis in aniridia: a retrospective multicenter study. Am J Ophthalmol 2007;144:227–31. 17. Aldave AJ, Sangwan VS, Basu S, et al. International results with the Boston type I keratoprosthesis. Ophthalmology 2012;119:1530–8. 18. Ament JD, Stryjewski TP, Ciolino JB, et al. Cost-effectiveness of the Boston keratoprosthesis. Am J Ophthalmol 2010;149: 221–228.e2. 19. Aquavella JV, Qian Y, McCormick GJ, Palakuru JR. Keratoprosthesis: current techniques. Cornea 2006;25:656–62. 20. Crawford GJ, Hicks CR, Lou X, et al. The Chirila keratoprosthesis: phase I human clinical trial. Ophthalmology 2002;109:883–9. 21. Dunlap K, Chak G, Aquavella JV, et al. Short-term visual outcomes of Boston type 1 keratoprosthesis implantation. Ophthalmology 2010;117:687–92. 22. Ghaffariyeh A, Honarpisheh N, Karkhaneh A, et al. FyodorovZuev keratoprosthesis implantation: long-term results in patients with multiple failed corneal grafts. Graefes Arch Clin Exp Ophthalmol 2011;249:93–101. 23. Guell JL, Arcos E, Gris O, et al. Outcomes with the Boston type 1 keratoprosthesis at Instituto de Microcirugia Ocular IMO. Saudi J Ophthalmol 2011;25:281–4. 24. Harissi-Dagher M, Khan BF, Schaumberg DA, Dohlman CH. Importance of nutrition to corneal grafts when used as a carrier of the Boston keratoprosthesis. Cornea 2007;26:564–8. 25. Hicks CR, Crawford GJ, Lou X, et al. Corneal replacement using a synthetic hydrogel cornea, AlphaCor: device, preliminary outcomes and complications. Eye (Lond) 2003;17: 385–92. 26. Hicks CR, Crawford GJ, Tan DT, et al. Outcomes of implantation of an artificial cornea, AlphaCor: effects of prior ocular herpes simplex infection. Cornea 2002;21: 685–90. 27. Hicks CR, Macvie O, Crawford GJ, Constable IJ. A risk score as part of an evidence-based approach to the selection of corneal replacement surgery. Cornea 2005;24:523–30. 28. Hille K, Hille A, Ruprecht KW. Medium term results in keratoprostheses with biocompatible and biological haptic. Graefes Arch Clin Exp Ophthalmol 2006;244:696–704. 29. Jiraskova N, Rozsival P, Burova M, Kalfertova M. AlphaCor artificial cornea: clinical outcome. Eye (Lond) 2011;25: 1138–46. 30. Kamyar R, Weizer JS, de Paula FH, et al. Glaucoma associated with Boston type I keratoprosthesis. Cornea 2012;31:134–9. 31. Khan BF, Harissi-Dagher M, Khan DM, Dohlman CH. Advances in Boston keratoprosthesis: enhancing retention and prevention of infection and inflammation. Int Ophthalmol Clin 2007;47:61–71.

32. Nouri M, Terada H, Alfonso EC, et al. Endophthalmitis after keratoprosthesis: incidence, bacterial causes, and risk factors. Arch Ophthalmol 2001;119:484–9. 33. Patel AP, Wu EI, Ritterband DC, Seedor JA. Boston type 1 keratoprosthesis: the New York Eye and Ear experience. Eye (Lond) 2012;26:418–25. 34. Rudnisky CJ, Belin MW, Todani A, et al. Risk factors for the development of retroprosthetic membranes with Boston keratoprosthesis type 1: multicenter study results. Ophthalmology 2012;119:951–5. 35. Sayegh RR, Ang LP, Foster CS, Dohlman CH. The Boston keratoprosthesis in Stevens-Johnson syndrome. Am J Ophthalmol 2008;145:438–44. 36. Yaghouti F, Nouri M, Abad JC, et al. Keratoprosthesis: preoperative prognostic categories. Cornea 2001;20:19–23. 37. Goldman DR, Hubschman JP, Aldave AJ, et al. Postoperative posterior segment complications in eyes treated with the Boston type I keratoprosthesis. Retina 2013;33:532–41. 38. Kim MJ, Yu F, Aldave AJ. Microbial keratitis after Boston type I keratoprosthesis implantation: incidence, organisms, risk factors, and outcomes. Ophthalmology 2013;120: 2209–16. 39. Sejpal K, Yu F, Aldave AJ. The Boston keratoprosthesis in the management of corneal limbal stem cell deficiency. Cornea 2011;30:1187–94. 40. Aquavella JV, Qian Y, McCormick GJ, Palakuru JR. Keratoprosthesis: the Dohlman-Doane device. Am J Ophthalmol 2005;140:1032–8. 41. Hicks CR, Crawford GJ, Dart JK, et al. AlphaCor: clinical outcomes. Cornea 2006;25:1034–42. 42. Verdejo-Gomez L, Pelaez N, Gris O, Guell JL. The Boston type I keratoprosthesis: an assessment of its efficacy and safety. Ophthalmic Surg Lasers Imaging 2011;42:446–52. 43. Trichet E, Carles G, Matonti F, et al. [Alphacor keratoprosthesis: device, surgical technique and clinical outcomes]. J Fr Ophtalmol 2013;36:393–401. 44. Shihadeh WA, Mohidat HM. Outcomes of the Boston keratoprosthesis in Jordan. Middle East Afr J Ophthalmol 2012;19: 97–100. 45. Shapiro BL, Cortes DE, Chin EK, et al. High-resolution spectral domain anterior segment optical coherence tomography in type 1 Boston keratoprosthesis. Cornea 2013;32:951–5. 46. Robert MC, Biernacki K, Harissi-Dagher M. Boston keratoprosthesis type 1 surgery: use of frozen versus fresh corneal donor carriers. Cornea 2012;31:339–45. 47. Koller B, Neuhann T, Neuhann I. [Results with the Boston keratoprosthesis]. Ophthalmologe 2012;109:454–61. 48. Reeves B. Including non-randomized studies. In: Higgins JPT, Green S, eds. Cochrane Handbook for Systematic Reviews of Interventions, version 510 (updated March 2011). The Cochrane Collaboration; 2011. Available at: www.cochranehandbook.org. 49. Schwarzer G. Meta-analysis with R. Vienna, Austria: The R Foundation for Statistical Computing; 2010:1–65. 50. Srikumaran D, Munoz B, Aldave AJ, et al. Long-term outcomes of Boston type 1 keratoprosthesis implantation. Ophthalmology 2014;121:2159–64. 51. Schulze-Bonsel K, Feltgen N, Burau H, et al. Visual acuities “hand motion” and “counting fingers” can be quantified with the Freiburg visual acuity test. Invest Ophthalmol Vis Sci 2006;47:1236–40. 52. Al-Mezaine H, Wagoner MD. Repeat penetrating keratoplasty: indications, graft survival, and visual outcome. Br J Ophthalmol 2006;90:324–7.

11

Ophthalmology Volume -, Number -, Month 2015 53. Borderie VM, Boelle PY, Touzeau O, et al. Predicted longterm outcome of corneal transplantation. Ophthalmology 2009;116:2354–60. 54. De Cock R. Penetrating keratoplasty in the West Bank and Gaza. Eye (Lond) 1994;8(Pt 1):29–34. 55. Fasolo A, Capuzzo C, Fornea M, et al. Risk factors for graft failure after penetrating keratoplasty: 5-year follow-up from the corneal transplant epidemiological study. Cornea 2011;30: 1328–35. 56. Hayashi K, Kondo H, Maeno A, Takimoto M. Long-term changes in corneal endothelial cell density after repeat penetrating keratoplasty in eyes with endothelial decompensation. Cornea 2013;32:1019–25. 57. Joshi SA, Jagdale SS, More PD, Deshpande M. Outcome of optical penetrating keratoplasties at a tertiary care eye institute in Western India. Indian J Ophthalmol 2012;60:15–21. 58. Kelly TL, Coster DJ, Williams KA. Repeat penetrating corneal transplantation in patients with keratoconus. Ophthalmology 2011;118:1538–42. 59. Kirkness CM, Ezra E, Rice NS, Steele AD. The success and survival of repeat corneal grafts. Eye (Lond) 1990;4(Pt 1): 58–64. 60. Parmar P, Salman A, Kalavathy CM, et al. Outcome analysis of cataract surgery following therapeutic keratoplasty. Cornea 2005;24:123–9. 61. Patel HY, Ormonde S, Brookes NH, et al. The New Zealand National Eye Bank: survival and visual outcome 1 year after penetrating keratoplasty. Cornea 2011;30:760–4. 62. Patel NP, Kim T, Rapuano CJ, et al. Indications for and outcomes of repeat penetrating keratoplasty, 1989e1995. Ophthalmology 2000;107:719–24. 63. Rahman I, Carley F, Hillarby C, et al. Penetrating keratoplasty: indications, outcomes, and complications. Eye (Lond) 2009;23:1288–94. 64. Randleman JB, Song CD, Palay DA. Indications for and outcomes of penetrating keratoplasty performed by resident surgeons. Am J Ophthalmol 2003;136:68–75. 65. Rapuano CJ, Cohen EJ, Brady SE, et al. Indications for and outcomes of repeat penetrating keratoplasty. Am J Ophthalmol 1990;109:689–95. 66. Sellami D, Abid S, Bouaouaja G, et al. Epidemiology and risk factors for corneal graft rejection. Transplant Proc 2007;39: 2609–11. 67. Tan DT, Janardhanan P, Zhou H, et al. Penetrating keratoplasty in Asian eyes: the Singapore Corneal Transplant Study. Ophthalmology 2008;115:975–982.e1. 68. Thompson RW Jr, Price MO, Bowers PJ, Price FW Jr. Longterm graft survival after penetrating keratoplasty. Ophthalmology 2003;110:1396–402. 69. Vanathi M, Sharma N, Sinha R, et al. Indications and outcome of repeat penetrating keratoplasty in India. BMC Ophthalmol 2005;5:26. 70. Weisbrod DJ, Sit M, Naor J, Slomovic AR. Outcomes of repeat penetrating keratoplasty and risk factors for graft failure. Cornea 2003;22:429–34. 71. Williams KA, Lowe M, Bartlett C, et al. Risk factors for human corneal graft failure within the Australian corneal graft registry. Transplantation 2008;86:1720–4.

72. Yalniz-Akkaya Z, Burcu Nurozler A, Yildiz E, et al. Repeat penetrating keratoplasty: indications and prognosis, 1995e2005. Eur J Ophthalmol 2009;19:362–8. 73. Yamagami S, Suzuki Y, Tsuru T. Risk factors for graft failure in penetrating keratoplasty. Acta Ophthalmol Scand 1996;74: 584–8. 74. Yildiz EH, Hoskins E, Fram N, et al. Third or greater penetrating keratoplasties: indications, survival, and visual outcomes. Cornea 2010;29:254–9. 75. Beckingsale P, Mavrikakis I, Al-Yousuf N, et al. Penetrating keratoplasty: outcomes from a corneal unit compared to national data. Br J Ophthalmol 2006;90:728–31. 76. Inoue K, Amano S, Oshika T, et al. A 10-year review of penetrating keratoplasty. Jpn J Ophthalmol 2000;44:139–45. 77. Inoue K, Amano S, Oshika T, Tsuru T. Risk factors for corneal graft failure and rejection in penetrating keratoplasty. Acta Ophthalmol Scand 2001;79:251–5. 78. Yamagami S, Suzuki Y, Ohya T, et al. [Multivariate analysis of risk factors of rejection in penetrating keratoplasty]. Nihon Ganka Gakkai Zasshi 1994;98:1097–100. 79. Lee WB, Shtein RM, Kaufman SC, et al. Boston Keratoprosthesis: outcomes and complications. A report by the American Academy of Ophthalmology. Ophthalmology 2015;122:1504–11. 80. Maguire MG, Stark WJ, Gottsch JD, et al. Risk factors for corneal graft failure and rejection in the collaborative corneal transplantation studies. Collaborative Corneal Transplantation Studies Research Group. Ophthalmology 1994;101:1536–47. 81. Rumelt S, Bersudsky V, Blum-Hareuveni T, Rehany U. Preexisting and postoperative glaucoma in repeated corneal transplantation. Cornea 2002;21:759–65. 82. Goldberg DB, Schanzlin DJ, Brown SI. Incidence of increased intraocular pressure after keratoplasty. Am J Ophthalmol 1981;92:372–7. 83. Polack FM. Glaucoma in keratoplasty. Cornea 1988;7:67–70. 84. Simmons RB, Stern RA, Teekhasaenee C, Kenyon KR. Elevated intraocular pressure following penetrating keratoplasty. Trans Am Ophthalmol Soc 1989;87:79–91. discussion 91e3. 85. Ahmad S, Akpek EK, Gehlbach PL, et al. Predictors of visual outcomes following Boston type 1 keratoprosthesis implantation. Am J Ophthalmol 2014;159:739–47. 86. Crnej A, Paschalis EI, Salvador-Culla B, et al. Glaucoma progression and role of glaucoma surgery in patients with Boston keratoprosthesis. Cornea 2014;33:349–54. 87. Huh ES, Aref AA, Vajaranant TS, et al. Outcomes of pars plana glaucoma drainage implant in Boston type 1 keratoprosthesis surgery. J Glaucoma 2014;23:e39–44. 88. Robert MC, Moussally K, Harissi-Dagher M. Review of endophthalmitis following Boston keratoprosthesis type 1. Br J Ophthalmol 2012;96:776–80. 89. Ilhan-Sarac O, Akpek EK. Current concepts and techniques in keratoprosthesis. Curr Opin Ophthalmol 2005;16:246–50. 90. Traish AS, Chodosh J. Expanding application of the Boston type I keratoprosthesis due to advances in design and improved post-operative therapeutic strategies. Semin Ophthalmol 2010;25:239–43. 91. HarissieDagher M. Boston keratoprosthesis: expanding the boundaries. Digit J Ophthalmol 2011;17:53–4.

Footnotes and Financial Disclosures Originally received: June 27, 2015. Final revision: September 10, 2015. Accepted: September 18, 2015. Available online: ---.

12

1

The Wilmer Eye Institute, The Johns Hopkins University School of Medicine, Baltimore, Maryland.

2

Manuscript no. 2015-1082.

Department of Epidemiology, Johns Hopkins Bloomberg School of Public Health, Baltimore, Maryland.

Ahmad et al



KPro versus Repeat PK for Graft Failure

3

Cornea and Refractive Surgery Services, The Jules Stein Eye Institute, Los Angeles, California.

Data collection: Ahmad, Mathews, Ng, Lindsley, Alkharashi, Hwang, Akpek

4

Obtained funding: none Overall responsibility: Ahmad, Mathews, Lindsley, Alkharashi, Hwang, Ng, Aldave, Akpek

Cornea and External Disease, The Kresge Eye Institute, Wayne State University, Detroit, Michigan.

5

Consultant, King Saud University, Riyadh, Saudi Arabia.

Financial Disclosure(s): The author(s) have no proprietary or commercial interest in any materials discussed in this article. Supported by National Eye Institute, National Institutes of Health, Bethesda, Maryland (grant no: 1 U01 EY020522 [K.L., S.M.N.]). Author Contributions: Conception and design: Ahmad, Mathews, Lindsley, Ng, Akpek Analysis and interpretation: Ahmad, Mathews, Ng, Lindsley, Aldave

Abbreviations and Acronyms: CI ¼ confidence interval; IOP ¼ intraocular pressure; KPro ¼ Boston type I keratoprosthesis; logMAR ¼ logarithm of the minimum angle of resolution; OSD ¼ ocular surface disease; PK ¼ penetrating keratoplasty. Correspondence: Esen Karamursel Akpek, MD, Ocular Surface Diseases Clinic, The Wilmer Eye Institute, Johns Hopkins University School of Medicine, 600 North Wolfe Street, Maumenee #317, Baltimore, MD 21287e9238 E-mail: [email protected].

13