- Email: [email protected]
Graft Reepithelialization after Penetrating Keratoplasty Using Organ-Cultured Donor Tissue
Graft Reepithelialization after Penetrating Keratoplasty Using Organ-Cultured Donor Tissue Vincent M. Borderie, MD, PhD, Olivier Touzeau, MD, Tristan Bourcier, MD, PhD, Cécile Allouch, MD, Laurent Laroche, MD Objective: To analyze graft reepithelialization after penetrating keratoplasty using organ-cultured donor tissue. Design: Retrospective observational cohort study. Participants: One thousand one hundred thirty-eight consecutive penetrating keratoplasties (1003 patients) performed between 1992 and 2004 were analyzed, and 1003 were included (1 graft per patient was included). Methods: Slit-lamp examination after fluorescein staining. Main Outcome Measure: Graft reepithelialization time was recorded. Results: The average graft reepithelialization time was 4.6⫾13.2 days (range, 1–210 days; median, 2.5 days). Complete corneal epithelial healing was obtained in 1 day in 28.5% of patients, in 3 days in 65.8%, in 7 days in 93.6%, and in 14 days in 97.0%. Postoperative chronic epithelial defects occurred in 3.0% of patients. In univariate analysis, death-to-storage time, storage time, deswelling time, rejection risk, trephination size, suture method, use of dexamethasone ointment during surgery, use of sodium hyaluronate during surgery, and use of high postoperative steroid regimen significantly influenced the graft reepithelialization time (P⬍0.05). In multiple regression, only the rejection risk ( ⫽ 0.07; P ⫽ 0.045), use of sodium hyaluronate at the end of surgery ( ⫽ ⫺0.17; P ⫽ 0.001), and steroid regimen ( ⫽ 0.09; P ⫽ 0.022) significantly influenced the graft reepithelialization time. Conclusions: The use of sodium hyaluronate at the end of surgery to coat the ocular surface shortened graft reepithelialization time after penetrating keratoplasty using organ-cultured donor tissue. Graft reepithelialization time was longer in high-risk recipients, and it increased with high dexamethasone eyedrops regimen. Ophthalmology 2006;113:2181–2186 © 2006 by the American Academy of Ophthalmology.
After keratoplasty, the donor corneal epithelium is replaced by the recipient epithelium. This leads to graft reepithelialization, which usually takes no longer than 1 week after transplantation.1 Rapid reepithelialization of the graft is essential for recovery of corneal transparency, for recovery of the epithelial barrier function, and for healing. Conversely, a chronic epithelial defect is a risk factor for inflammation, infection, corneal scarring, corneal thinning, perforation, and graft failure. The objectives of the present study were to analyze graft reepithelialization after penetrating keratoplasty using organ-cultured donor tissue and to evaluate predictive factors for graft reepithelialization time. Originally received: October 27, 2005. Accepted: June 12, 2006. Manuscript no. 2005-1026. From the Department of Ophthalmology, Centre Hospitalier National d’ Ophtalmologie des XV–XX, Paris VI University, Paris, France. Supported by the Paris VI University (Université Pierre et Marie Curie). The authors have no proprietary, commercial, or financial interests in any of the products described in the article. Correspondence and reprint requests to Vincent M. Borderie, MD, PhD, Service V, Centre Hospitalier National d’Ophtalmologie des XV–XX, 28 rue de Charenton, 75012 Paris, France. E-mail: [email protected]. © 2006 by the American Academy of Ophthalmology Published by Elsevier Inc.
Patients and Methods Study Design We studied 1138 consecutive penetrating keratoplasty procedures using organ-cultured donor tissue carried out in 1003 patients between December, 1992, and December, 2004. One hundred thirty-five of 1138 (12%) patients received 2 grafts in the same or contralateral eye during the study period. Because these were not independent observations, we analyzed our data including only the first graft performed during the study period for these 135 patients. Therefore, 1003 penetrating keratoplasty procedures performed in 1003 patients were included in the present study. In accordance with French law, an institutional review board/ethics committee approval was not required for this study because no modifications to French standards of treatment and follow-up were made.
Donor Corneas Donor corneas were stored as previously described.2 They featured an average donor age of 69.7⫾14.0 years (mean⫾standard deviation) and an average death-to-storage time of 29.1⫾13.0 hours. They were stored in organ culture medium for an average time of 19.8⫾4.4 days, followed by incubation in deswelling medium for an average time of 2.0⫾0.6 days. The donor cause of death was cancer in 47.6%, cardiovascular diseases in 28.5%, respiratory diseases in 9.8%, alcoholism in 6.4%, infectious diseases in 5.3%, ISSN 0161-6420/06/$–see front matter doi:10.1016/j.ophtha.2006.06.020
2181
Ophthalmology Volume 113, Number 12, December 2006 Table 1. Patient Characteristics Characteristic Preoperative diagnosis Keratoconus Bullous keratopathy Fuchs’ dystrophy Other corneal dystrophies Corneal scar, corneal ulcer, interstitial keratitis Trauma Regraft Postoperative lens status Phakic Posterior chamber intraocular lens Anterior chamber intraocular lens Aphakic Preoperative intraocular pressure ⱕ20 mmHg and no history of glaucoma ⬎20 mmHg and/or history of glaucoma Recipient rejection status Low risk High risk Previous glaucoma surgery None One or more procedure Use of antiglaucoma drugs None One or more drug
nⴱ
%
214 429 101 27 110 42 80
21.4 42.8 10.0 2.7 11.0 4.2 7.9
329 284 301 89
32.8 28.3 30.0 8.9
783 220
78.1 21.9
770 233
76.8 23.2
All transplants were performed at a single institution by 1 of 11 surgeons. Two surgeons (VMB and LL) carried out 90% of the transplants. The surgical procedure has been reported elsewhere.3 All donor buttons were punched from the posterior corneal surface by using the Hanna device (Moria, Antony, France), and the donor corneal epithelium was not removed. The average recipient trephination size was 7.93⫾0.21 mm. The average donor trephination size was 8.22⫾0.22 mm. For graft suturing, a 16- to 24-bite running suture was used in 19.7% of patients, 24 interrupted sutures were used in 12.2% of patients, and a combination of 8 interrupted sutures and a 16-bite running suture was used in 68.1% of patients.
930 73
92.7 7.3
Medical Treatment
799 204
79.7 20.3
1003 patients (1 graft per patient was included).
and miscellaneous diseases in the remaining 2.4%. Donor corneas were harvested by enucleation between December, 1992, and January, 1997, whereas in situ excision was used between February, 1997, and December, 2004, after French law was modified. They were incubated in 50 ml or 100 ml of organ culture medium at 31° C for 14 to 35 days. This organ culture period was followed by a 1- to 4-day deswelling period. The organ culture medium consisted of Iscove modified Dulbecco medium (Eurobio, Les Ulis, France), fetal calf serum, ␣-thioglycerol, -2-mercaptoethanol, penicillin, streptomycin, and amphotericin. The deswelling medium was made of the same medium supplemented with Dextran T40 (Eurobio). At the end of the organ culture period, the donor corneal endothelium was assessed by means of light microscopy after trypan blue staining, and the donor corneal epithelium was evaluated macroscopically for the absence of ulcer. ABO blood group–matched tissue was used in 68.1% of grafts, Rhesus D blood group–matched tissue was used in 91.7% of grafts, and sex (HY)-matched tissue was used in 72.4% of grafts.
Recipients The mean age of the patients was 59.0⫾21.9 years. Patient characteristics are shown in Table 1. High-risk recipients were defined as having a vascularized cornea (2 or more quadrants of corneal vascularization) or a history of irreversible corneal allograft rejection in the operated eye. Patients with complete limbal stem cell deficiency, severe dry eye (i.e., keratoconjunctivitis sicca with filamentary keratitis, severe punctuate keratitis leading to stromal involvement, or chronic geographic ulcers), severe lid abnormalities (i.e., lid trauma, entropion, or ectropion that cause repercussions on the cornea), or complete corneal anesthesia were not included in this study because they were not considered as candidates for keratoplasty before surgical treatment of the condition (if available). No patients with severe dry eye or complete corneal anesthesia were grafted during the study period. Conversely, 13
2182
patients with limbal stem cell deficiency underwent limbal transplantation (8 autografts and 5 allografts) followed by penetrating keratoplasty and were included in the study. Similarly, 15 patients with severe lid abnormalities underwent lid surgery followed by penetrating keratoplasty and were included in the study.
Surgical Procedures
At the end of surgery, all patients received dexamethasone and neomycin eyedrops. In addition, in 48.2% of eyes, the ocular surface was coated (the ocular surface was covered by sodium hyaluronate before patching the eye) with 1% sodium hyaluronate (Healon; Pharmacia, Saint Quentin en Yvelines, France); in 33.6%, it was coated with a dexamethasone and oxytetracycline ointment (Sterdex; Ciba Vision Ophthalmics, Toulouse, France); in 1.9%, it was coated with both sodium hyaluronate and dexamethasone ointment; and the remaining 16.3% of eyes received nothing other than eyedrops at the end of the surgical procedure. No randomization was carried out for using sodium hyaluronate or dexamethasone ointment. The decision to use sodium hyaluronate or dexamethasone ointment was made with no precise guidelines. Before December, 1992, the routine treatment was dexamethasone ointment coating, which was used in all keratoplasty patients at the end of surgery. The use of sodium hyaluronate was started at the beginning of the study, and it was more frequent at the end of the study than at the beginning. Patients who were considered at risk of postoperative epithelial keratopathy were more likely to receive sodium hyaluronate. Eyes were patched from the end of surgery until the first postoperative day. During hospitalization, 95.1% of patients received dexamethasone (with neomycin) eyedrops and all patients received norfloxacin eyedrops. For statistical analysis, the steroid regimen (Table 2) was considered as a quantitative variable ranging from 0 (i.e., no steroid eyedrops) to 16 (i.e., eyedrops every hour starting at 7 AM and until 10 PM). The regimen was chosen according to recipient rejection risk and preoperative diagnosis. Table 2. Postoperative Dexamethasone Regimen Regimen
n
%
None 3 times daily 4 times daily 5 times daily 6 times daily 8 times daily 10 times daily 16 times daily
51 274 292 87 134 78 11 76
5.1 27.3 29.1 8.7 13.3 7.8 1.1 7.6
1003 patients (1 graft per patient was included).
Borderie et al 䡠 Graft Reepithelialization after Penetrating Keratoplasty Dexamethasone and neomycin eyedrops were maintained for at least 12 months (with progressive decrease in steroid regimen), except in patients with dexamethasone-induced increase in intraocular pressure. In these patients, these eyedrops were replaced by fluorometholone eyedrops.
Table 4. Influence of Qualitative Variables on Graft Reepithelialization Time in Univariate Analysis (Wilcoxon Test and Kruskal-Wallis Analysis of Variance; only significant variables are shown) Variable
Main Outcome Measure Corneal epithelial healing was assessed daily by slit-lamp examination after fluorescein staining. Patients were hospitalized up to the time of graft reepithelialization, which was the main outcome measure. However, patients with postoperative chronic epithelial defects (i.e., persistent epithelial defects after the second postoperative week) were followed up as outpatients. They were treated with various methods (i.e., sodium hyaluronate eyedrops, vitamin A ointment, bandage soft contact lens, lateral tarsorrhaphy, or amniotic membrane transplantation), but eyes were not patched.
Statistical Analysis The observed distribution of graft reepithelialization time was compared with the corresponding theoretical normal distribution by means of the chi-square test. Because significant differences were found between the observed distribution and the theoretical distribution (data not shown), nonparametric tests were used for the subsequent statistical analysis. A univariate analysis was performed first with variables that could influence graft reepithelialization. The influence of quantitative variables on graft reepithelialization time was analyzed with the Spearman regression test. The influence of qualitative variables on graft reepithelialization time was analyzed with the Wilcoxon test and the Kruskal-Wallis analysis of variance. We then performed a multivariate analysis by multiple regression, including variables that were significant at a univariate level (P⬍0.05). We studied both donor variables (i.e., donor age, donor cause of death, death-to-storage time, storage time, deswelling time, ABO compatibility, rhesus D compatibility, HY compatibility), recipient variables (i.e., recipient age, preoperative diagnosis, lens status, recipient rejection status, previous glaucoma, previous glaucoma surgery, previous use of antiglaucoma drugs), surgical variables (i.e., recipient trephination size, suture method, combined procedures), and medical treatments (i.e., use sodium hyaluronate at the end of surgery, use of dexamethasone ointment at the end of surgery, and postoperative steroid regimen). Statistical analysis was carried out using Statistica software version 6.1 (StatSoft France, Maisons-Alfort, France).
Results The average graft reepithelialization time was 4.6⫾13.2 days (range, 1–210; median, 2.5). Complete corneal epithelial healing was obtained in 1 day in 28.5% of patients, in 3 days in 65.8%, in
Table 3. Correlation between Quantitative Variables and Graft Reepithelialization Time in Univariate Analysis (Spearman Regression; only significant variables are shown) Variable
rs
P Value
Death-to-storage time Storage time Deswelling time Recipient trephination size Dexamethasone regimen
0.07 0.11 0.11 0.16 ⫺0.31
⬍0.05 ⬍0.05 ⬍0.05 ⬍0.05 ⬍0.05
Rejection risk Low risk High risk Suture method Running suture Interrupted sutures Mixed suturing Use of dexamethasone ointment Yes No Use of sodium hyaluronate Yes No
Mean Reepithelialization Time (days)
P Value
3.7 7.5
⬍0.05
4.8 6.6 4.2
⬍0.001
5.6 4.0
⬍0.001
3.2 6.0
⬍0.001
7 days in 93.6%, and in 14 days in 97.0%. Postoperative chronic epithelial defects occurred in 3.0% of patients. They healed within 3 to 30 weeks after surgery. In univariate analysis, death-to-storage time, storage time, deswelling time, rejection risk, trephination size, suture method, use of dexamethasone ointment, use of sodium hyaluronate, and steroid regimen all significantly influenced the graft reepithelialization time (P⬍0.05; Tables 3, 4). In multiple regression, only the steroid regimen ( ⫽ 0.09, P ⫽ 0.022; Fig 1), use of sodium hyaluronate ( ⫽ ⫺0.17, P ⫽ 0.001), and rejection risk ( ⫽ 0.07, P ⫽ 0.045) significantly influenced the graft reepithelialization time. Graft epithelial healing time was shorter in patients who received sodium hyaluronate at the end of surgery, and it was longer in high-risk patients and in patients treated with frequent steroid regimen. In the 148 low-risk patients who received sodium hyaluronate at the end of surgery and who were treated after surgery with dexamethasone eyedrops no more than 4 times daily, complete corneal epithelial healing was obtained in 1 day in 58.1%, in 3 days in 88.5%, and in 7 days in 97.3%. Among 30 patients who had chronic epithelial defects, 17 (57%) were high-risk recipients, 10 (33%) received sodium hyaluronate at the end of surgery, and 14 (47%) were treated with dexamethasone eyedrops after surgery more than 4 times daily. Rejection risk and use of sodium hyaluronate seemed to be statistically independent (2 ⫽ 2.7, P ⫽ 0.10), whereas patients who received sodium hyaluronate, as well as high-risk patients, were treated significantly more frequently with a steroid regimen (P⬍0.01, Wilcoxon test).
Discussion In this series of 1003 penetrating keratoplasties using organcultured donor tissue, complete corneal epithelial healing was obtained during the first postoperative week in 93.6% of patients, which is similar to what was observed by Lindstrom et al7 (i.e., intact epithelium or only small defects by 7 days after surgery in 42 of 47 grafts; 90%) and in other studies.4 – 6,8 In a series of 102 grafts, Lass et al4 reported persistent epithelial defects beyond 1 week in 11 eyes (11%), which corresponds to a 89% rate of epithelial healing at 7 days. In another study published by Lass et al,6 this early postoperative complication was observed in 2 of 42
2183
Ophthalmology Volume 113, Number 12, December 2006
Figure 1. Bar graph showing graft reepithelialization time according to the steroid regimen (multiple regression:  ⫽ 0.09, P ⫽ 0.022). SEM ⫽ standard error of the mean.
eyes (5%). Very similar results were published by Lass et al8 in 1 further study including 210 grafts.8 In this latter series, 90.5% of eyes had intact epithelia by 1 week. This was reported in 88% of postkeratoplasty patients in the study by Naor et al.9 In the present study, intact graft epithelium was found in 28.5% of patients on day 1 and in 65.8% on day 3. In a series of 32 grafts, Bourne et al10 reported a mean rate of intact epithelium on day 1 of 64.5%. In the study by Kim et al,11 13 of 40 transplanted corneas (32%) had no epithelial defect on the first postoperative day. More recently, in a series of 91 penetrating keratoplasties, Machado et al12 reported that 65% of the patients had epithelial defects on the first postoperative day. This feature was observed in 26 of 84 patients (69%) in the study published by Chou et al.13 In the present study, the median and mean graft reepithelialization time were 2.5 days and 4.6 days, respectively, which is slightly longer than what was reported by Meyer and Bobb14 (i.e., a 2-day median epithelial healing time; n ⫽ 66) and by Sugar et al15 (i.e., a 2.5-day mean epithelial healing time; n ⫽ 39). In the present study, organ culture was used for donor tissue storage. Organ-cultured corneas feature a decrease in the number of epithelial cell layers (down to 2 cell layers), with no macroscopic epithelial ulcers at the end of storage.16 –18 However, fluorescein staining was not used to assess precisely the donor epithelial status before and after storage in the present study. Organ culture media are devoid of chondroitin sulfate, which is present in many cold storage media. Thus, comparing our results with those reported previously with 4° C-stored corneas is only indirect. Our study includes a large number of patients, allowing us to investigate risk factors in a multivariate analysis. We found 3 significant factors related to graft reepithelialization time that were statistically independent of each other. Furthermore, multiple regression permitted adjustment for the effect of the significant factors. Conversely, most studies in
2184
the literature include a lower number of patients (e.g., no more than 100 patients), which may allow the effect of a particular factor to be demonstrated, but makes adjustment of the significant factor on other factors more difficult to achieve. In univariate analysis, several factors had a significant influence on graft reepithelialization time. Three donor factors (death-to-storage time, storage time, and deswelling time) were significant, indicating presence of a relationship between donor death-to-surgery time and reepithelialization time (the longer the time between donor death and surgery, the longer the graft epithelial healing time). Donor cause of death (which could influence the state of epithelium before storage) had no significant influence on graft reepithelialization. In a series of 84 penetrating keratoplasties, Chou et al13 reported a significant association of longer death-toenucleation times with increased risk of epithelial defects. In a series of 91 penetrating keratoplasties, Machado et al12 reported that patients with macroepithelial defects in the first postoperative month received older donor tissue with longer preservation-to-surgery time. Kim et al11 reported that the degree of epithelial defect on day 1 after penetrating keratoplasty was associated significantly with the interval from storage to surgery. In a series of 66 penetrating keratoplasties, Meyer and Bobb14 demonstrated that postoperative epithelial healing time was prolonged significantly when donor corneal epithelium was missing at keratoplasty. It is well established that corneal tissue storage induces epithelial damage, which increases with storage time.17,19,20 This was confirmed in a large series of 1101 donor corneas (Cornea Donor Study). In this study, fewer epithelial abnormalities were observed as death-to-preservation time decreased.21 This has been demonstrated after hypothermic storage and after organ culture. In corneas stored at 4° C, preservation induces time-dependent epithelial cell apoptosis.22,23 It is likely that a donor corneal transplant with intact epithelium will facilitate postoperative epithelial healing as
Borderie et al 䡠 Graft Reepithelialization after Penetrating Keratoplasty compared with a graft with impaired epithelium.23 However, this effect of donor epithelium condition does not seem to be strong, because donor factors in our study that were significant in univariate analysis were no longer significant in multivariate analysis, showing that the influence of recipient and therapeutic factors was more important. Several medical treatments seemed to have a significant influence on graft reepithelialization in the present study. The use of sodium hyaluronate at the end of surgery to coat the ocular surface seemed to permit significant decrease in graft reepithelialization time both in univariate and multivariate analysis. This factor had the strongest effect in multiple regression ( ⫽ ⫺0.17). Coating the ocular surface with sodium hyaluronate at the end of surgery is considered to be an advisable intraoperative prevention of postoperative epithelial defects.24 However, its clinical effect was not demonstrated. Sodium hyaluronate 0.1% provides better protection of the chick corneal epithelium against dryness than hydroxyethylcellulose 0.1% or phosphate-buffered saline.25 It was demonstrated previously that intraoperative hydration of the ocular surface with sodium hyaluronate increases graft epithelial healing 1 week after surgery.26 However, this way of using sodium hyaluronate during surgery induces difficulties in graft suturing. We used sodium hyaluronate only at the end of surgery, which had no effect on the suturing. The postoperative dexamethasone regimen also had a significant influence on graft reepithelialization in our study, both in univariate and multivariate analysis. The use of dexamethasone ointment at the end of surgery to coat the ocular surface also was associated with a negative effect on graft reepithelialization, but significance was reached only in univariate analysis and not in multivariate analysis. Diclofenac sodium eyedrops, a nonsteroidal antiinflammatory drug, have been shown to be toxic to the corneal epithelium after keratoplasty.27 Sugar et al15 reported a trend, although not significant, toward increase in corneal epithelial healing time in postkeratoplasty patients treated with topical prednisolone acetate as compared with untreated postkeratoplasty patients. In human corneal epithelial cell culture, dexamethasone increases cell proliferation and it induces cell apoptosis. This results in decreased cell proliferation for high concentrations of dexamethasone.28 The negative effect of a frequent use regimen of dexamethasone on graft reepithelialization cannot be considered as an argument not to use corticosteroids after keratoplasty because they may be necessary to prevent rejection. Furthermore, frequent use regimens are useful in patients at high risk of rejection.29 In the present study, 95% of patients received neomycin eyedrops after surgery. Aminoglycosides were demonstrated to be toxic to corneal epithelial cells in vitro. A 30or 60-minute exposure of cultured rabbit corneal epithelial cells to any of 4 aminoglycosides (tobramycin, amikacin, neomycin, and gentamicin) inhibits cell growth in a non– dose-dependent fashion.30 In vivo neomycin was shown to retard normal epithelial healing of deepithelialized corneas.31 It may be possible that neomycin induced a delay in graft reepithelialization in our patients. However, because most patients received neomycin, it is not possible to demonstrate this effect statistically. Among other pharmaceutical agents,
human epidermal growth factor has been evaluated for corneal epithelial healing after keratoplasty. No significant acceleration of corneal reepithelialization could be demonstrated with the use of recombinant human epidermal growth factor.32 Among surgical factors, larger trephination sizes and use of interrupted sutures were associated with longer reepithelialization times. This could be explained by larger area of transplant to be reepithelialized after keratoplasty and, more importantly, surgical damage associated with interrupted sutures. However, these factors no longer were significant in multivariate analysis. One recipient factor (i.e., rejection risk) had a significant influence on graft reepithelialization time both in univariate and multivariate analysis. Grafts of high-risk patients were associated with longer reepithelialization times. These patients featured stromal vascularization or previous graft rejection in the same eye. The latter corneal condition often is associated with corneal vascularization. It seems likely that the donor corneal condition associated with longer graft reepithelialization is the result of impaired limbal stem cell function. However, except for patients with severe limbal deficiency who are not candidates for keratoplasty without associated limbal cell transplantation, it is currently not possible to assess limbal function reliably before keratoplasty. High-risk patients may have a certain level of limbal function impairment, because corneal vascularization is part of the limbal deficiency syndrome. Recipient factors rarely were shown to influence graft reepithelialization. A study by Mannis et al33 including 120 keratoplasties showed that surface keratopathy was associated with older recipients. Abnormalities of the tear film or the lids and corneal hypoesthesia also may influence graft reepithelialization.25 These may be more frequent in high-risk recipients than in low-risk recipients. Patients with complete limbal stem cell deficiency, severe dry eye, severe lid abnormalities, or complete corneal anesthesia were not included in our study. However, tear film break-up time, Schirmer test, corneal sensation, and lid condition were not recorded in all our patients before transplantation. For this reason, these measures could not be included in the statistical analysis. Penetrating keratoplasty performed during an active stage of herpetic disease also is a strong risk factor for postoperative persistent epithelial defect.34 In the present study, all patients with herpetic keratitis were grafted when inflammation was controlled. In conclusion, the use of sodium hyaluronate at the end of surgery to coat the ocular surface may be of value because it decreased graft reepithelialization time after penetrating keratoplasty in the present nonrandomized, retrospective study using organ-cultured tissue. A randomized trial including both organ-cultured and 4° C-stored corneas is needed to draw firm conclusions. Graft reepithelialization time was longer in high-risk recipients and it increased with high-frequency dexamethasone eyedrops regimen. Corticosteroids should be used at the lowest regimen, permitting full control of the host immune response to transplantation.
2185
Ophthalmology Volume 113, Number 12, December 2006
References 1. Hori J, Streilein JW. Dynamics of donor cell persistence and recipient cell replacement in orthotopic corneal allografts in mice. Invest Ophthalmol Vis Sci 2001;42:1820 – 8. 2. Borderie VM, Scheer S, Touzeau O, et al. Donor organ cultured corneal tissue selection before penetrating keratoplasty. Br J Ophthalmol 1998;82:382– 8. 3. Borderie VM, Touzeau O, Allouch C, et al. The results of successful penetrating keratoplasty using donor organcultured corneal tissue. Transplantation 1999;67:1433– 8. 4. Lass JH, Reinhart WJ, Bruner WE, et al. Comparison of corneal storage in K-Sol and chondroitin sulfate corneal storage medium in human corneal transplantation. Ophthalmology 1989;96:688 –97. 5. Lass JH, Reinhart WJ, Skelnik DL, et al. An in vitro and clinical comparison of corneal storage with chondroitin sulfate corneal storage medium with and without dextran. Ophthalmology 1990;97:96 –103. 6. Lass JH, Bourne WM, Musch DC, et al. A randomized, prospective, double-masked clinical trial of Optisol vs DexSol corneal storage media. Arch Ophthalmol 1992;110:1404 – 8. 7. Lindstrom RL, Kaufman HE, Skelnik DL, et al. Optisol corneal storage medium. Am J Ophthalmol 1992;114:345–56. 8. Lass JH, Musch DC, Gordon JF, Laing RA, Corneal Preservation Study Group. Epidermal growth factor and insulin in corneal preservation: results of a multi-center trial. Ophthalmology 1994;101:352–9. 9. Naor J, Slomovic AR, Chipman M, Rootman DS. A randomized, double-masked clinical trial of Optisol-GS vs Chen Medium for human corneal storage. Arch Ophthalmol 2002; 120:1280 –5. 10. Bourne WM, Nelson LR, Maguire LJ, et al. Comparison of Chen Medium and Optisol-GS for human corneal preservation at 4 degrees C: results of transplantation. Cornea 2001;20: 683– 6. 11. Kim T, Palay DA, Lynn M. Donor factors associated with epithelial defects after penetrating keratoplasty. Cornea 1996; 15:451– 6. 12. Machado RA, Mannis MJ, Mandel HA, et al. The relationship between first postoperative day epithelial status and eventual health of the ocular surface in penetrating keratoplasty. Cornea 2002;21:574 –7. 13. Chou L, Cohen EJ, Laibson PR, Rapuano CJ. Factors associated with epithelial defects after penetrating keratoplasty. Ophthalmic Surg 1994;25:700 –3. 14. Meyer RF, Bobb KC. Corneal epithelium in penetrating keratoplasty. Am J Ophthalmol 1980;90:142–7. 15. Sugar A, Bokosky JE, Meyer RF. A randomized trial of topical corticosteroids in epithelial healing after keratoplasty. Cornea 1984;3:268 –71. 16. Borderie VM, Kantelip BM, Delbosc BY, et al. Morphology, histology and ultrastructure of human C31 organ-cultured corneas. Cornea 1995;14:300 –10.
2186
17. Van Horn DL, Doughman DJ, Harris JE, et al. The ultrastructure of human organ-cultured cornea. II. Stroma and epithelium. Arch Ophthalmol 1975;93:275–7. 18. Borderie VM, Baudrimont M, Lopez M, et al. Evaluation of the deswelling period in dextran-containing medium after corneal organ culture. Cornea 1997;16:215–23. 19. Means TL, Geroski DH, L’Hernault N, et al. The corneal epithelium after Optisol-GS storage. Cornea 1996;15:599 – 605. 20. Leibowitz HM, Moore TE. Keratoplasty. In: Leibowitz HM, Waring GO III, eds. Corneal Disorders: Clinical Diagnosis and Management. 2nd ed. Philadelphia, PA: W.B. Saunders; 1998:842– 69. 21. Cornea Donor Study Group. Baseline donor characteristics in the Cornea Donor Study. Cornea 2005;24:389 –96. 22. Komuro A, Hodge DO, Gores GJ, Bourne WM. Cell death during corneal storage at 4 degrees C. Invest Ophthalmol Vis Sci 1999;40:2827–32. 23. Nelson LR, Hodge DO, Bourne WM. In vitro comparison of Chen medium and Optisol-GS medium for human corneal storage. Cornea 2000;19:782–7. 24. Nelson JD. Epithelial problems. In: Brightbill FS, ed. Corneal Surgery: Theory, Technique, and Tissue. 2nd ed. St Louis, MO: Mosby; 1993:235– 46. 25. Wysenbeek YS, Loya N, Ben Sira I, et al. The effect of sodium hyaluronate on the corneal epithelium: an ultrastructural study. Invest Ophthalmol Vis Sci 1988;29:194 –9. 26. Reed DB, Mannis MJ, Hills JF, Johnson CA. Corneal epithelial healing after penetrating keratoplasty using topical Healon versus balanced salt solution. Ophthalmic Surg 1987;18:525– 8. 27. Shimazaki J, Saito H, Yang HY, et al. Persistent epithelial defect following penetrating keratoplasty: an adverse effect of diclofenac eyedrops. Cornea 1995;14:623–7. 28. Bourcier T, Forgez P, Borderie V, et al. Regulation of human corneal epithelial cell proliferation and apoptosis by dexamethasone. Invest Ophthalmol Vis Sci 2000;41:4133– 41. 29. Collaborative Corneal Transplantation Studies Research Group. The Collaborative Corneal Transplantation Studies (CCTS): effectiveness of histocompatibility matching in high-risk corneal transplantation. Arch Ophthalmol 1992;110:1392– 403. 30. Lass JH, Mack RJ, Imperia PS, et al. An in vitro analysis of aminoglycoside corneal epithelial toxicity. Curr Eye Res 1989; 8:299 –304. 31. Petroutsos G, Guimaraes R, Giraud J, Pouliquen Y. Antibiotics and corneal epithelial wound healing. Arch Ophthalmol 1983;101:1775– 8. 32. Dellaert MM, Casey TA, Wiffen S, et al. Influence of topical human epidermal growth factor on postkeratoplasty re-epithelialisation. Br J Ophthalmol 1997;81:391–5. 33. Mannis MJ, Zadnik K, Miller MR, Marquez M. Preoperative risk factors for surface disease after penetrating keratoplasty. Cornea 1997;16:7–11. 34. Scharf BH. Early postoperative complications. In: Krachmer JH, Mannis MJ, Holland EJ, eds. Cornea. vol 3. Surgery of the Cornea and Conjunctiva. St Louis, MO: Mosby; 1997:1659 –73.
After keratoplasty, the donor corneal epithelium is replaced by the recipient epithelium. This leads to graft reepithelialization, which usually takes no longer than 1 week after transplantation.1 Rapid reepithelialization of the graft is essential for recovery of corneal transparency, for recovery of the epithelial barrier function, and for healing. Conversely, a chronic epithelial defect is a risk factor for inflammation, infection, corneal scarring, corneal thinning, perforation, and graft failure. The objectives of the present study were to analyze graft reepithelialization after penetrating keratoplasty using organ-cultured donor tissue and to evaluate predictive factors for graft reepithelialization time. Originally received: October 27, 2005. Accepted: June 12, 2006. Manuscript no. 2005-1026. From the Department of Ophthalmology, Centre Hospitalier National d’ Ophtalmologie des XV–XX, Paris VI University, Paris, France. Supported by the Paris VI University (Université Pierre et Marie Curie). The authors have no proprietary, commercial, or financial interests in any of the products described in the article. Correspondence and reprint requests to Vincent M. Borderie, MD, PhD, Service V, Centre Hospitalier National d’Ophtalmologie des XV–XX, 28 rue de Charenton, 75012 Paris, France. E-mail: [email protected]. © 2006 by the American Academy of Ophthalmology Published by Elsevier Inc.
Patients and Methods Study Design We studied 1138 consecutive penetrating keratoplasty procedures using organ-cultured donor tissue carried out in 1003 patients between December, 1992, and December, 2004. One hundred thirty-five of 1138 (12%) patients received 2 grafts in the same or contralateral eye during the study period. Because these were not independent observations, we analyzed our data including only the first graft performed during the study period for these 135 patients. Therefore, 1003 penetrating keratoplasty procedures performed in 1003 patients were included in the present study. In accordance with French law, an institutional review board/ethics committee approval was not required for this study because no modifications to French standards of treatment and follow-up were made.
Donor Corneas Donor corneas were stored as previously described.2 They featured an average donor age of 69.7⫾14.0 years (mean⫾standard deviation) and an average death-to-storage time of 29.1⫾13.0 hours. They were stored in organ culture medium for an average time of 19.8⫾4.4 days, followed by incubation in deswelling medium for an average time of 2.0⫾0.6 days. The donor cause of death was cancer in 47.6%, cardiovascular diseases in 28.5%, respiratory diseases in 9.8%, alcoholism in 6.4%, infectious diseases in 5.3%, ISSN 0161-6420/06/$–see front matter doi:10.1016/j.ophtha.2006.06.020
2181
Ophthalmology Volume 113, Number 12, December 2006 Table 1. Patient Characteristics Characteristic Preoperative diagnosis Keratoconus Bullous keratopathy Fuchs’ dystrophy Other corneal dystrophies Corneal scar, corneal ulcer, interstitial keratitis Trauma Regraft Postoperative lens status Phakic Posterior chamber intraocular lens Anterior chamber intraocular lens Aphakic Preoperative intraocular pressure ⱕ20 mmHg and no history of glaucoma ⬎20 mmHg and/or history of glaucoma Recipient rejection status Low risk High risk Previous glaucoma surgery None One or more procedure Use of antiglaucoma drugs None One or more drug
nⴱ
%
214 429 101 27 110 42 80
21.4 42.8 10.0 2.7 11.0 4.2 7.9
329 284 301 89
32.8 28.3 30.0 8.9
783 220
78.1 21.9
770 233
76.8 23.2
All transplants were performed at a single institution by 1 of 11 surgeons. Two surgeons (VMB and LL) carried out 90% of the transplants. The surgical procedure has been reported elsewhere.3 All donor buttons were punched from the posterior corneal surface by using the Hanna device (Moria, Antony, France), and the donor corneal epithelium was not removed. The average recipient trephination size was 7.93⫾0.21 mm. The average donor trephination size was 8.22⫾0.22 mm. For graft suturing, a 16- to 24-bite running suture was used in 19.7% of patients, 24 interrupted sutures were used in 12.2% of patients, and a combination of 8 interrupted sutures and a 16-bite running suture was used in 68.1% of patients.
930 73
92.7 7.3
Medical Treatment
799 204
79.7 20.3
1003 patients (1 graft per patient was included).
and miscellaneous diseases in the remaining 2.4%. Donor corneas were harvested by enucleation between December, 1992, and January, 1997, whereas in situ excision was used between February, 1997, and December, 2004, after French law was modified. They were incubated in 50 ml or 100 ml of organ culture medium at 31° C for 14 to 35 days. This organ culture period was followed by a 1- to 4-day deswelling period. The organ culture medium consisted of Iscove modified Dulbecco medium (Eurobio, Les Ulis, France), fetal calf serum, ␣-thioglycerol, -2-mercaptoethanol, penicillin, streptomycin, and amphotericin. The deswelling medium was made of the same medium supplemented with Dextran T40 (Eurobio). At the end of the organ culture period, the donor corneal endothelium was assessed by means of light microscopy after trypan blue staining, and the donor corneal epithelium was evaluated macroscopically for the absence of ulcer. ABO blood group–matched tissue was used in 68.1% of grafts, Rhesus D blood group–matched tissue was used in 91.7% of grafts, and sex (HY)-matched tissue was used in 72.4% of grafts.
Recipients The mean age of the patients was 59.0⫾21.9 years. Patient characteristics are shown in Table 1. High-risk recipients were defined as having a vascularized cornea (2 or more quadrants of corneal vascularization) or a history of irreversible corneal allograft rejection in the operated eye. Patients with complete limbal stem cell deficiency, severe dry eye (i.e., keratoconjunctivitis sicca with filamentary keratitis, severe punctuate keratitis leading to stromal involvement, or chronic geographic ulcers), severe lid abnormalities (i.e., lid trauma, entropion, or ectropion that cause repercussions on the cornea), or complete corneal anesthesia were not included in this study because they were not considered as candidates for keratoplasty before surgical treatment of the condition (if available). No patients with severe dry eye or complete corneal anesthesia were grafted during the study period. Conversely, 13
2182
patients with limbal stem cell deficiency underwent limbal transplantation (8 autografts and 5 allografts) followed by penetrating keratoplasty and were included in the study. Similarly, 15 patients with severe lid abnormalities underwent lid surgery followed by penetrating keratoplasty and were included in the study.
Surgical Procedures
At the end of surgery, all patients received dexamethasone and neomycin eyedrops. In addition, in 48.2% of eyes, the ocular surface was coated (the ocular surface was covered by sodium hyaluronate before patching the eye) with 1% sodium hyaluronate (Healon; Pharmacia, Saint Quentin en Yvelines, France); in 33.6%, it was coated with a dexamethasone and oxytetracycline ointment (Sterdex; Ciba Vision Ophthalmics, Toulouse, France); in 1.9%, it was coated with both sodium hyaluronate and dexamethasone ointment; and the remaining 16.3% of eyes received nothing other than eyedrops at the end of the surgical procedure. No randomization was carried out for using sodium hyaluronate or dexamethasone ointment. The decision to use sodium hyaluronate or dexamethasone ointment was made with no precise guidelines. Before December, 1992, the routine treatment was dexamethasone ointment coating, which was used in all keratoplasty patients at the end of surgery. The use of sodium hyaluronate was started at the beginning of the study, and it was more frequent at the end of the study than at the beginning. Patients who were considered at risk of postoperative epithelial keratopathy were more likely to receive sodium hyaluronate. Eyes were patched from the end of surgery until the first postoperative day. During hospitalization, 95.1% of patients received dexamethasone (with neomycin) eyedrops and all patients received norfloxacin eyedrops. For statistical analysis, the steroid regimen (Table 2) was considered as a quantitative variable ranging from 0 (i.e., no steroid eyedrops) to 16 (i.e., eyedrops every hour starting at 7 AM and until 10 PM). The regimen was chosen according to recipient rejection risk and preoperative diagnosis. Table 2. Postoperative Dexamethasone Regimen Regimen
n
%
None 3 times daily 4 times daily 5 times daily 6 times daily 8 times daily 10 times daily 16 times daily
51 274 292 87 134 78 11 76
5.1 27.3 29.1 8.7 13.3 7.8 1.1 7.6
1003 patients (1 graft per patient was included).
Borderie et al 䡠 Graft Reepithelialization after Penetrating Keratoplasty Dexamethasone and neomycin eyedrops were maintained for at least 12 months (with progressive decrease in steroid regimen), except in patients with dexamethasone-induced increase in intraocular pressure. In these patients, these eyedrops were replaced by fluorometholone eyedrops.
Table 4. Influence of Qualitative Variables on Graft Reepithelialization Time in Univariate Analysis (Wilcoxon Test and Kruskal-Wallis Analysis of Variance; only significant variables are shown) Variable
Main Outcome Measure Corneal epithelial healing was assessed daily by slit-lamp examination after fluorescein staining. Patients were hospitalized up to the time of graft reepithelialization, which was the main outcome measure. However, patients with postoperative chronic epithelial defects (i.e., persistent epithelial defects after the second postoperative week) were followed up as outpatients. They were treated with various methods (i.e., sodium hyaluronate eyedrops, vitamin A ointment, bandage soft contact lens, lateral tarsorrhaphy, or amniotic membrane transplantation), but eyes were not patched.
Statistical Analysis The observed distribution of graft reepithelialization time was compared with the corresponding theoretical normal distribution by means of the chi-square test. Because significant differences were found between the observed distribution and the theoretical distribution (data not shown), nonparametric tests were used for the subsequent statistical analysis. A univariate analysis was performed first with variables that could influence graft reepithelialization. The influence of quantitative variables on graft reepithelialization time was analyzed with the Spearman regression test. The influence of qualitative variables on graft reepithelialization time was analyzed with the Wilcoxon test and the Kruskal-Wallis analysis of variance. We then performed a multivariate analysis by multiple regression, including variables that were significant at a univariate level (P⬍0.05). We studied both donor variables (i.e., donor age, donor cause of death, death-to-storage time, storage time, deswelling time, ABO compatibility, rhesus D compatibility, HY compatibility), recipient variables (i.e., recipient age, preoperative diagnosis, lens status, recipient rejection status, previous glaucoma, previous glaucoma surgery, previous use of antiglaucoma drugs), surgical variables (i.e., recipient trephination size, suture method, combined procedures), and medical treatments (i.e., use sodium hyaluronate at the end of surgery, use of dexamethasone ointment at the end of surgery, and postoperative steroid regimen). Statistical analysis was carried out using Statistica software version 6.1 (StatSoft France, Maisons-Alfort, France).
Results The average graft reepithelialization time was 4.6⫾13.2 days (range, 1–210; median, 2.5). Complete corneal epithelial healing was obtained in 1 day in 28.5% of patients, in 3 days in 65.8%, in
Table 3. Correlation between Quantitative Variables and Graft Reepithelialization Time in Univariate Analysis (Spearman Regression; only significant variables are shown) Variable
rs
P Value
Death-to-storage time Storage time Deswelling time Recipient trephination size Dexamethasone regimen
0.07 0.11 0.11 0.16 ⫺0.31
⬍0.05 ⬍0.05 ⬍0.05 ⬍0.05 ⬍0.05
Rejection risk Low risk High risk Suture method Running suture Interrupted sutures Mixed suturing Use of dexamethasone ointment Yes No Use of sodium hyaluronate Yes No
Mean Reepithelialization Time (days)
P Value
3.7 7.5
⬍0.05
4.8 6.6 4.2
⬍0.001
5.6 4.0
⬍0.001
3.2 6.0
⬍0.001
7 days in 93.6%, and in 14 days in 97.0%. Postoperative chronic epithelial defects occurred in 3.0% of patients. They healed within 3 to 30 weeks after surgery. In univariate analysis, death-to-storage time, storage time, deswelling time, rejection risk, trephination size, suture method, use of dexamethasone ointment, use of sodium hyaluronate, and steroid regimen all significantly influenced the graft reepithelialization time (P⬍0.05; Tables 3, 4). In multiple regression, only the steroid regimen ( ⫽ 0.09, P ⫽ 0.022; Fig 1), use of sodium hyaluronate ( ⫽ ⫺0.17, P ⫽ 0.001), and rejection risk ( ⫽ 0.07, P ⫽ 0.045) significantly influenced the graft reepithelialization time. Graft epithelial healing time was shorter in patients who received sodium hyaluronate at the end of surgery, and it was longer in high-risk patients and in patients treated with frequent steroid regimen. In the 148 low-risk patients who received sodium hyaluronate at the end of surgery and who were treated after surgery with dexamethasone eyedrops no more than 4 times daily, complete corneal epithelial healing was obtained in 1 day in 58.1%, in 3 days in 88.5%, and in 7 days in 97.3%. Among 30 patients who had chronic epithelial defects, 17 (57%) were high-risk recipients, 10 (33%) received sodium hyaluronate at the end of surgery, and 14 (47%) were treated with dexamethasone eyedrops after surgery more than 4 times daily. Rejection risk and use of sodium hyaluronate seemed to be statistically independent (2 ⫽ 2.7, P ⫽ 0.10), whereas patients who received sodium hyaluronate, as well as high-risk patients, were treated significantly more frequently with a steroid regimen (P⬍0.01, Wilcoxon test).
Discussion In this series of 1003 penetrating keratoplasties using organcultured donor tissue, complete corneal epithelial healing was obtained during the first postoperative week in 93.6% of patients, which is similar to what was observed by Lindstrom et al7 (i.e., intact epithelium or only small defects by 7 days after surgery in 42 of 47 grafts; 90%) and in other studies.4 – 6,8 In a series of 102 grafts, Lass et al4 reported persistent epithelial defects beyond 1 week in 11 eyes (11%), which corresponds to a 89% rate of epithelial healing at 7 days. In another study published by Lass et al,6 this early postoperative complication was observed in 2 of 42
2183
Ophthalmology Volume 113, Number 12, December 2006
Figure 1. Bar graph showing graft reepithelialization time according to the steroid regimen (multiple regression:  ⫽ 0.09, P ⫽ 0.022). SEM ⫽ standard error of the mean.
eyes (5%). Very similar results were published by Lass et al8 in 1 further study including 210 grafts.8 In this latter series, 90.5% of eyes had intact epithelia by 1 week. This was reported in 88% of postkeratoplasty patients in the study by Naor et al.9 In the present study, intact graft epithelium was found in 28.5% of patients on day 1 and in 65.8% on day 3. In a series of 32 grafts, Bourne et al10 reported a mean rate of intact epithelium on day 1 of 64.5%. In the study by Kim et al,11 13 of 40 transplanted corneas (32%) had no epithelial defect on the first postoperative day. More recently, in a series of 91 penetrating keratoplasties, Machado et al12 reported that 65% of the patients had epithelial defects on the first postoperative day. This feature was observed in 26 of 84 patients (69%) in the study published by Chou et al.13 In the present study, the median and mean graft reepithelialization time were 2.5 days and 4.6 days, respectively, which is slightly longer than what was reported by Meyer and Bobb14 (i.e., a 2-day median epithelial healing time; n ⫽ 66) and by Sugar et al15 (i.e., a 2.5-day mean epithelial healing time; n ⫽ 39). In the present study, organ culture was used for donor tissue storage. Organ-cultured corneas feature a decrease in the number of epithelial cell layers (down to 2 cell layers), with no macroscopic epithelial ulcers at the end of storage.16 –18 However, fluorescein staining was not used to assess precisely the donor epithelial status before and after storage in the present study. Organ culture media are devoid of chondroitin sulfate, which is present in many cold storage media. Thus, comparing our results with those reported previously with 4° C-stored corneas is only indirect. Our study includes a large number of patients, allowing us to investigate risk factors in a multivariate analysis. We found 3 significant factors related to graft reepithelialization time that were statistically independent of each other. Furthermore, multiple regression permitted adjustment for the effect of the significant factors. Conversely, most studies in
2184
the literature include a lower number of patients (e.g., no more than 100 patients), which may allow the effect of a particular factor to be demonstrated, but makes adjustment of the significant factor on other factors more difficult to achieve. In univariate analysis, several factors had a significant influence on graft reepithelialization time. Three donor factors (death-to-storage time, storage time, and deswelling time) were significant, indicating presence of a relationship between donor death-to-surgery time and reepithelialization time (the longer the time between donor death and surgery, the longer the graft epithelial healing time). Donor cause of death (which could influence the state of epithelium before storage) had no significant influence on graft reepithelialization. In a series of 84 penetrating keratoplasties, Chou et al13 reported a significant association of longer death-toenucleation times with increased risk of epithelial defects. In a series of 91 penetrating keratoplasties, Machado et al12 reported that patients with macroepithelial defects in the first postoperative month received older donor tissue with longer preservation-to-surgery time. Kim et al11 reported that the degree of epithelial defect on day 1 after penetrating keratoplasty was associated significantly with the interval from storage to surgery. In a series of 66 penetrating keratoplasties, Meyer and Bobb14 demonstrated that postoperative epithelial healing time was prolonged significantly when donor corneal epithelium was missing at keratoplasty. It is well established that corneal tissue storage induces epithelial damage, which increases with storage time.17,19,20 This was confirmed in a large series of 1101 donor corneas (Cornea Donor Study). In this study, fewer epithelial abnormalities were observed as death-to-preservation time decreased.21 This has been demonstrated after hypothermic storage and after organ culture. In corneas stored at 4° C, preservation induces time-dependent epithelial cell apoptosis.22,23 It is likely that a donor corneal transplant with intact epithelium will facilitate postoperative epithelial healing as
Borderie et al 䡠 Graft Reepithelialization after Penetrating Keratoplasty compared with a graft with impaired epithelium.23 However, this effect of donor epithelium condition does not seem to be strong, because donor factors in our study that were significant in univariate analysis were no longer significant in multivariate analysis, showing that the influence of recipient and therapeutic factors was more important. Several medical treatments seemed to have a significant influence on graft reepithelialization in the present study. The use of sodium hyaluronate at the end of surgery to coat the ocular surface seemed to permit significant decrease in graft reepithelialization time both in univariate and multivariate analysis. This factor had the strongest effect in multiple regression ( ⫽ ⫺0.17). Coating the ocular surface with sodium hyaluronate at the end of surgery is considered to be an advisable intraoperative prevention of postoperative epithelial defects.24 However, its clinical effect was not demonstrated. Sodium hyaluronate 0.1% provides better protection of the chick corneal epithelium against dryness than hydroxyethylcellulose 0.1% or phosphate-buffered saline.25 It was demonstrated previously that intraoperative hydration of the ocular surface with sodium hyaluronate increases graft epithelial healing 1 week after surgery.26 However, this way of using sodium hyaluronate during surgery induces difficulties in graft suturing. We used sodium hyaluronate only at the end of surgery, which had no effect on the suturing. The postoperative dexamethasone regimen also had a significant influence on graft reepithelialization in our study, both in univariate and multivariate analysis. The use of dexamethasone ointment at the end of surgery to coat the ocular surface also was associated with a negative effect on graft reepithelialization, but significance was reached only in univariate analysis and not in multivariate analysis. Diclofenac sodium eyedrops, a nonsteroidal antiinflammatory drug, have been shown to be toxic to the corneal epithelium after keratoplasty.27 Sugar et al15 reported a trend, although not significant, toward increase in corneal epithelial healing time in postkeratoplasty patients treated with topical prednisolone acetate as compared with untreated postkeratoplasty patients. In human corneal epithelial cell culture, dexamethasone increases cell proliferation and it induces cell apoptosis. This results in decreased cell proliferation for high concentrations of dexamethasone.28 The negative effect of a frequent use regimen of dexamethasone on graft reepithelialization cannot be considered as an argument not to use corticosteroids after keratoplasty because they may be necessary to prevent rejection. Furthermore, frequent use regimens are useful in patients at high risk of rejection.29 In the present study, 95% of patients received neomycin eyedrops after surgery. Aminoglycosides were demonstrated to be toxic to corneal epithelial cells in vitro. A 30or 60-minute exposure of cultured rabbit corneal epithelial cells to any of 4 aminoglycosides (tobramycin, amikacin, neomycin, and gentamicin) inhibits cell growth in a non– dose-dependent fashion.30 In vivo neomycin was shown to retard normal epithelial healing of deepithelialized corneas.31 It may be possible that neomycin induced a delay in graft reepithelialization in our patients. However, because most patients received neomycin, it is not possible to demonstrate this effect statistically. Among other pharmaceutical agents,
human epidermal growth factor has been evaluated for corneal epithelial healing after keratoplasty. No significant acceleration of corneal reepithelialization could be demonstrated with the use of recombinant human epidermal growth factor.32 Among surgical factors, larger trephination sizes and use of interrupted sutures were associated with longer reepithelialization times. This could be explained by larger area of transplant to be reepithelialized after keratoplasty and, more importantly, surgical damage associated with interrupted sutures. However, these factors no longer were significant in multivariate analysis. One recipient factor (i.e., rejection risk) had a significant influence on graft reepithelialization time both in univariate and multivariate analysis. Grafts of high-risk patients were associated with longer reepithelialization times. These patients featured stromal vascularization or previous graft rejection in the same eye. The latter corneal condition often is associated with corneal vascularization. It seems likely that the donor corneal condition associated with longer graft reepithelialization is the result of impaired limbal stem cell function. However, except for patients with severe limbal deficiency who are not candidates for keratoplasty without associated limbal cell transplantation, it is currently not possible to assess limbal function reliably before keratoplasty. High-risk patients may have a certain level of limbal function impairment, because corneal vascularization is part of the limbal deficiency syndrome. Recipient factors rarely were shown to influence graft reepithelialization. A study by Mannis et al33 including 120 keratoplasties showed that surface keratopathy was associated with older recipients. Abnormalities of the tear film or the lids and corneal hypoesthesia also may influence graft reepithelialization.25 These may be more frequent in high-risk recipients than in low-risk recipients. Patients with complete limbal stem cell deficiency, severe dry eye, severe lid abnormalities, or complete corneal anesthesia were not included in our study. However, tear film break-up time, Schirmer test, corneal sensation, and lid condition were not recorded in all our patients before transplantation. For this reason, these measures could not be included in the statistical analysis. Penetrating keratoplasty performed during an active stage of herpetic disease also is a strong risk factor for postoperative persistent epithelial defect.34 In the present study, all patients with herpetic keratitis were grafted when inflammation was controlled. In conclusion, the use of sodium hyaluronate at the end of surgery to coat the ocular surface may be of value because it decreased graft reepithelialization time after penetrating keratoplasty in the present nonrandomized, retrospective study using organ-cultured tissue. A randomized trial including both organ-cultured and 4° C-stored corneas is needed to draw firm conclusions. Graft reepithelialization time was longer in high-risk recipients and it increased with high-frequency dexamethasone eyedrops regimen. Corticosteroids should be used at the lowest regimen, permitting full control of the host immune response to transplantation.
2185
Ophthalmology Volume 113, Number 12, December 2006
References 1. Hori J, Streilein JW. Dynamics of donor cell persistence and recipient cell replacement in orthotopic corneal allografts in mice. Invest Ophthalmol Vis Sci 2001;42:1820 – 8. 2. Borderie VM, Scheer S, Touzeau O, et al. Donor organ cultured corneal tissue selection before penetrating keratoplasty. Br J Ophthalmol 1998;82:382– 8. 3. Borderie VM, Touzeau O, Allouch C, et al. The results of successful penetrating keratoplasty using donor organcultured corneal tissue. Transplantation 1999;67:1433– 8. 4. Lass JH, Reinhart WJ, Bruner WE, et al. Comparison of corneal storage in K-Sol and chondroitin sulfate corneal storage medium in human corneal transplantation. Ophthalmology 1989;96:688 –97. 5. Lass JH, Reinhart WJ, Skelnik DL, et al. An in vitro and clinical comparison of corneal storage with chondroitin sulfate corneal storage medium with and without dextran. Ophthalmology 1990;97:96 –103. 6. Lass JH, Bourne WM, Musch DC, et al. A randomized, prospective, double-masked clinical trial of Optisol vs DexSol corneal storage media. Arch Ophthalmol 1992;110:1404 – 8. 7. Lindstrom RL, Kaufman HE, Skelnik DL, et al. Optisol corneal storage medium. Am J Ophthalmol 1992;114:345–56. 8. Lass JH, Musch DC, Gordon JF, Laing RA, Corneal Preservation Study Group. Epidermal growth factor and insulin in corneal preservation: results of a multi-center trial. Ophthalmology 1994;101:352–9. 9. Naor J, Slomovic AR, Chipman M, Rootman DS. A randomized, double-masked clinical trial of Optisol-GS vs Chen Medium for human corneal storage. Arch Ophthalmol 2002; 120:1280 –5. 10. Bourne WM, Nelson LR, Maguire LJ, et al. Comparison of Chen Medium and Optisol-GS for human corneal preservation at 4 degrees C: results of transplantation. Cornea 2001;20: 683– 6. 11. Kim T, Palay DA, Lynn M. Donor factors associated with epithelial defects after penetrating keratoplasty. Cornea 1996; 15:451– 6. 12. Machado RA, Mannis MJ, Mandel HA, et al. The relationship between first postoperative day epithelial status and eventual health of the ocular surface in penetrating keratoplasty. Cornea 2002;21:574 –7. 13. Chou L, Cohen EJ, Laibson PR, Rapuano CJ. Factors associated with epithelial defects after penetrating keratoplasty. Ophthalmic Surg 1994;25:700 –3. 14. Meyer RF, Bobb KC. Corneal epithelium in penetrating keratoplasty. Am J Ophthalmol 1980;90:142–7. 15. Sugar A, Bokosky JE, Meyer RF. A randomized trial of topical corticosteroids in epithelial healing after keratoplasty. Cornea 1984;3:268 –71. 16. Borderie VM, Kantelip BM, Delbosc BY, et al. Morphology, histology and ultrastructure of human C31 organ-cultured corneas. Cornea 1995;14:300 –10.
2186
17. Van Horn DL, Doughman DJ, Harris JE, et al. The ultrastructure of human organ-cultured cornea. II. Stroma and epithelium. Arch Ophthalmol 1975;93:275–7. 18. Borderie VM, Baudrimont M, Lopez M, et al. Evaluation of the deswelling period in dextran-containing medium after corneal organ culture. Cornea 1997;16:215–23. 19. Means TL, Geroski DH, L’Hernault N, et al. The corneal epithelium after Optisol-GS storage. Cornea 1996;15:599 – 605. 20. Leibowitz HM, Moore TE. Keratoplasty. In: Leibowitz HM, Waring GO III, eds. Corneal Disorders: Clinical Diagnosis and Management. 2nd ed. Philadelphia, PA: W.B. Saunders; 1998:842– 69. 21. Cornea Donor Study Group. Baseline donor characteristics in the Cornea Donor Study. Cornea 2005;24:389 –96. 22. Komuro A, Hodge DO, Gores GJ, Bourne WM. Cell death during corneal storage at 4 degrees C. Invest Ophthalmol Vis Sci 1999;40:2827–32. 23. Nelson LR, Hodge DO, Bourne WM. In vitro comparison of Chen medium and Optisol-GS medium for human corneal storage. Cornea 2000;19:782–7. 24. Nelson JD. Epithelial problems. In: Brightbill FS, ed. Corneal Surgery: Theory, Technique, and Tissue. 2nd ed. St Louis, MO: Mosby; 1993:235– 46. 25. Wysenbeek YS, Loya N, Ben Sira I, et al. The effect of sodium hyaluronate on the corneal epithelium: an ultrastructural study. Invest Ophthalmol Vis Sci 1988;29:194 –9. 26. Reed DB, Mannis MJ, Hills JF, Johnson CA. Corneal epithelial healing after penetrating keratoplasty using topical Healon versus balanced salt solution. Ophthalmic Surg 1987;18:525– 8. 27. Shimazaki J, Saito H, Yang HY, et al. Persistent epithelial defect following penetrating keratoplasty: an adverse effect of diclofenac eyedrops. Cornea 1995;14:623–7. 28. Bourcier T, Forgez P, Borderie V, et al. Regulation of human corneal epithelial cell proliferation and apoptosis by dexamethasone. Invest Ophthalmol Vis Sci 2000;41:4133– 41. 29. Collaborative Corneal Transplantation Studies Research Group. The Collaborative Corneal Transplantation Studies (CCTS): effectiveness of histocompatibility matching in high-risk corneal transplantation. Arch Ophthalmol 1992;110:1392– 403. 30. Lass JH, Mack RJ, Imperia PS, et al. An in vitro analysis of aminoglycoside corneal epithelial toxicity. Curr Eye Res 1989; 8:299 –304. 31. Petroutsos G, Guimaraes R, Giraud J, Pouliquen Y. Antibiotics and corneal epithelial wound healing. Arch Ophthalmol 1983;101:1775– 8. 32. Dellaert MM, Casey TA, Wiffen S, et al. Influence of topical human epidermal growth factor on postkeratoplasty re-epithelialisation. Br J Ophthalmol 1997;81:391–5. 33. Mannis MJ, Zadnik K, Miller MR, Marquez M. Preoperative risk factors for surface disease after penetrating keratoplasty. Cornea 1997;16:7–11. 34. Scharf BH. Early postoperative complications. In: Krachmer JH, Mannis MJ, Holland EJ, eds. Cornea. vol 3. Surgery of the Cornea and Conjunctiva. St Louis, MO: Mosby; 1997:1659 –73.