The Failure Risk of Surgical Therapy for Corneal Perforations Using Cryopreserved Donor Corneas

Background To evaluate the long-term results of therapeutic keratoplasty for a consecutive case series of corneal perforation. Methods The cases comprised 41 eyes (41 patients) that underwent therapeutic keratoplasty using cryopreserved donor grafts at the Keio University Hospital between January 2012 and December 2016. The eyes were evaluated regarding the cause of corneal perforation, surgical procedure, size of the perforation, presence of anterior chamber collapse, visual prognosis, and complications. Results The major causative diseases included herpetic keratitis (n = 11), bacterial and fungal corneal ulcer (n = 4), Mooren’s ulcer (n = 5), severe dry eye (n = 4), and neuroparalytic keratitis (n = 4). Penetrating keratoplasty was performed in 28 eyes (68.2%), lamellar keratoplasty in 8 eyes (19.5%), and keratoepithelioplasty in 5 eyes (12.2%).


Background
Corneal perforation is caused by various conditions such as trauma, infections, in ammatory corneal ulcer, neurotrophic disorders, and lagophthalmos. Corneal perforation is a vision-threatening condition that usually requires emergency surgical intervention to recover the anatomical structure of the eye. 1 Severe sequelae such as angle-closure glaucoma, choroidal detachment, or endophthalmitis may occur, which may lead to permanent blindness. There have been several reports on treatment techniques for corneal perforation, including bandage soft contact lenses, 2-4 cyanoacrylate glue, 5 suture of the perforated area, 4 conjunctival aps, 2 amniotic membrane transplantation (AMT), 2,6−8 tarsorrhaphy, 9 and penetrating or lamellar keratoplasty in cases of large corneal perforations. 2,10−15 Therapeutic keratoplasty is a better surgical technique for closing large corneal perforations. 6,10,16 However, surgical techniques and postoperative management may be di cult. [17][18][19][20] In Japan, there is a shortage of donor corneas available for all corneal diseases including corneal perforation. Therefore, fresh donors are often not available in cases of an emergency, and corneal perforation patients rarely undergo surgery using a fresh donor graft. In this study, we report the clinical outcomes of the therapeutic corneal transplantation with cryopreserved grafts performed in consecutive cases over ve years in the Keio University Hospital, Tokyo, Japan.

Methods
This study included 41 eyes (41 patients, 26 males and 15 females) that underwent therapeutic keratoplasty between January 2012 and December 2016 in the Keio University Hospital. All surgical procedures were performed by corneal specialists (S.S, Y.M, T.K, Y.I, M.F). In this study, clinical data including the cause of the corneal perforation, surgical procedure, anatomical recovery, size of the corneal perforation, size of the donor graft, pre-surgical condition of the iris, depth of the anterior chamber, visual prognosis, and complications were analyzed in a retrospective manner. We divided cases into two groups according to the corneal perforation size (small, < 4 mm vs large, ≥ 4 mm), as well as into two groups according to the donor graft size in cases of penetrating keratoplasty (PKP; small, < 8.5 mm vs large, ≥ 8.5 mm). Tectonic success was de ned as the successful reconstruction of the anatomical integrity of the eye, with no leakage of aqueous humor and no further surgical intervention required for the original corneal perforation. Postoperative results ending in phthisis and enucleation were excluded from tectonic success cases.

Surgical Procedures
Only one case of keratoplasty was performed under general anesthesia, whereas all other cases were done under local anesthesia including retrobulbar and sub-tenon anesthesia. All 41 keratoplasty cases used cryopreserved donor tissue stored in storage medium (Optisol) at -80 °C.
The surgical procedure of PKP was started by marking the center of the recipient cornea, and then a trephination was performed using the Hessburg-Barron vacuum trephine (Katena Products, Inc., Denville, NJ). A 0.25-0.50 mm oversized donor graft was prepared using a donor punch (Katena Products, Inc.).
The recipient's corneal dissection was completed using corneal scissors. The donor tissue was placed on the host bed by an interrupted suture with 10 − 0 mono lament nylon, and all knots were buried.
For the lamellar keratoplasty (LKP), the Hessburg-Barron vacuum trephine was used to partially cut both the recipient and donor corneal stromal tissue. Lamellar dissection of the stroma was done using a disposable lamellar knife (FEATHER Safety Razor Co., Ltd., Osaka, Japan). The donor cornea was sutured with 10 − 0 mono lament nylon sutures to the recipient bed using the same suturing techniques as for the PKP.
Keratoepithelioplasty (KEP) is a surgical procedure used to treat a perforation of peripheral corneal ulcers such as in Mooren's ulcer or rheumatoid arthritis. 21 First, a conjunctival peritomy was performed along the corneal limbus. Hemorrhages were cauterized with bipolar cautery forceps. A corneal incision was made to almost 80% depth of the cornea using a knife to prepare the recipient bed. A trephine 0.25 mm larger than the ulcer was used to make a donor graft in the shape of a fan. An ophthalmic visco-surgical device was used to return the iris to the correct position in the anterior chamber. The graft was xed to the recipient bed using 10 − 0 mono lament nylon sutures.
Subconjunctival injection of dexamethasone and gentamicin was administered after transplantation in all cases except in active bacterial and fungal cases, where only the antibiotic was injected. The postoperative treatment included topical antibacterial drops for infection control. Topical steroid therapy was also maintained except in cases of fungal infections, where the topical steroid was not used immediately after surgery. Topical steroid administration was started from twice a day after corneal epithelialization was achieved without any signs of fungal recurrence. The frequency of topical steroid application was gradually increased to four times a day and maintained for at least 1 year. Oral valacyclovir 1000 mg/day was used for herpetic keratitis for 2-3 weeks postoperatively. Optical PKP was performed if the in ammation was controlled and no signs of recurrence were observed for at least 6 months.

Statistical Analysis
All data are reported as mean ± standard deviation. The Fisher's exact probability test and Chi-squared test were used for group comparisons. All statistical analyses were performed using Microsoft O ce Excel 2016 software, version 1711 (TTT, North Carolina, USA). Categorical variables were presented as frequency (%) and compared between groups using the Chi-square test and Fisher's probability exact test.
Signi cant differences were considered when the P-value was < 0.05.
In PKP cases (28 eyes), the patients were divided into two categories according to the corneal graft size. The donor graft size was 8.5 mm or larger in 10 eyes, and 18 eyes received a donor graft with a size smaller than 8.5 mm. Donor grafts smaller than 8.5 mm achieved a signi cantly higher successful anatomical recovery compared to grafts larger than 8.5 mm (P < 0.01), especially in the infection group (P < 0.01; Table 2). A total of 14 eyes required additional surgery (7 eyes in the infectious group and 7 eyes in the noninfectious group). Among participants of the infectious group, 4 eyes underwent optical PKP with fresh donor grafts, 1 eye underwent a double procedure (PKP with fresh donor graft and unplanned intracapsular cataract extraction), and optical cataract removal and intraocular lens implantation were performed in another eye. An emergency therapeutic PKP with a frozen donor graft for graft destruction due to herpetic ulcer and melting was performed in only 1 eye. Anatomical recovery among cases with required additional surgeries was achieved in 7 eyes (100%) in the infection group, and in the noninfection group, it was achieved in 5 eyes (71.4%) after reoperation (Table 3).  We also evaluated the depth of the anterior chamber, iris incarceration, size of the perforation, and donor size at the time of the perforation and analyzed the correlation of these parameters with the nal VA improvement. However, no statistically signi cant relationship was found between these parameters and the nal VA values (Table 4).  Table 5). Secondary glaucoma was the main complication in previous reports, 9,22 but we experienced in our study population only 1 eye (1.8%) with ocular hypertension that did not respond to glaucoma eye drops and eventually required tube shunt surgery.

Discussion
In this study, the primary goal of sealing the corneal perforation was achieved in 28 eyes (68.3%) of the 41 eyes after the rst operation, and in 34 eyes (82.9%) of 41 eyes on nal observation. We used cryopreserved donor corneas for therapeutic keratoplasty (41 eyes, 100%). In Japan, donor organs including corneas are in short supply. Thus, ophthalmologists do not always have access to fresh donor organs when needed on an emergency basis. A previous report suggested that cryopreserved donor corneas have a high risk for recurrence of reperforation and infection. 4 According to their study, sometimes a second keratoplasty using fresh donor grafts is needed to recover corneal transparency. 4 However, in the report by Hanada et al. with 67% fresh donor corneas used in therapeutic corneal transplantation, the cure rate was 80%, which is not so different from that of our study, and there was also no difference in the rate of complications. 13 If in bacterial or fungal keratitis the infected lesion was incompletely removed, reoccurrence of the infectious keratitis was sometimes observed. A previous report showed that it is necessary to remove the focus of infection by oversizing 0.5 mm to 1.0 mm. 23 Yokogawa et al. 23 reported corneal transplantation for corneal perforation by bacterial and fungal keratitis performed with small grafts, however, a second large graft was eventually required for all cases. Li et al. reported therapeutic corneal transplantation in 116 cases of fungal keratitis and found that graft rejection and secondary glaucoma were more likely to occur in large grafts. 11 Other reports described that large grafts are associated with poor postoperative prognosis and complications. [24][25][26] On the other hand, there are reports that a large graft size does not in uence the postoperative prognosis. 10,27 Considering these study ndings, publications regarding the effects of the graft size on anatomical recovery rate and complications were gathered only for cases of infectious keratitis such as bacterial, fungal, and acanthamoeba keratitis, and they were similar to the results in the present study. The group with grafts 8.5 mm or larger was less controlled with appropriate treatment than that with grafts smaller than 8.5 mm. Among the cases using 8.5 mm or larger grafts in the infection group, all 5 eyes with bacterial or fungal keratitis did not achieve the sealing of the corneal perforation. Koçluk et al. reported 25 cases of bacterial and fungal keratitis divided into two groups according to the time until the surgical intervention (> 15 days vs ≤ 15 days). As a result, there were more large grafts (diameter ≥ 9.5 mm) in the group with the longer time until intervention. 24 Our ndings may demonstrate some important points when sealing corneal perforations in cases of bacterial and fungal keratitis; primarily that infected lesion should be completely removed and secondly that therapeutic keratoplasty should be performed before the focus of infection reaches the vicinity of the limbus.
Among the cases with secondary surgery for infections, both the proportion of optical indications and the cure rate were high. In non-infectious cases, there was a higher incidence of tectonic surgery. After the second transplantation, the cure rate was 100% in infectious cases, of which the visual acuity improved in 71.4%, and there were no cases with decreased visual acuity. This shows that a second optical operation is good option for visual acuity improvement after the infection is under control.
Penetrating keratoplasty is the most common surgical procedure in both infectious and non-infectious diseases with corneal perforations. Recently, some authors reported that LKP and deep anterior lamellar keratoplasty (DALK) for corneal perforation is safer in terms of endothelial rejection. 10,19,21,27−33 Other studies have noted that Descemet stripping automated endothelial keratoplasty (DSAEK) is effective for perforation cases. 33 Nevertheless, a disadvantage of LKP is the possibility that the infected lesion may not be completely removed and interface opacity may cause blurred vision. 21 Shimmura et al. proposed that good visual acuity was achieved by DALK performed on therapeutic transplantation and that it is necessary to select a surgical method with the depth of the infected lesion in consideration. 12 In the current study, LKP was chosen for 2 cases of herpetic and 1 case of bacterial keratitis. Although one herpetic case required optical transplantation, all cases were cured.
In our study, herpetic keratitis was the most common disease in the infection group, whereas Mooren's ulcer was most frequent in the non-infection group. The original diseases recurred in 4 of 11 eyes in herpetic keratitis (36.7%) and 2 of 5 eyes in Mooren's ulcer (40%) after the rst therapeutic keratoplasty. This is a similar recurrence rate compared to previous studies reporting the recurrence in herpetic keratitis and Mooren's ulcer. 23,34,35 Cyclosporine, 36  Mooren's ulcers. 24 We additionally excised the conjunctiva with a depth of 2-3 mm along the corneal limbs and the corneal tissue around the ulcer. We also removed the immune cells at the bottom of the ulcer (Brown). 39 KEP with conjunctival resection was performed in 3 of 5 eyes with Mooren's ulcers and showed no recurrence. However, the other 2 eyes, in which conjunctival resection was not preformed, resulted in recurrence after keratoplasty. We performed conjunctival resection in both cases and consequently controlled the in ammation with surgical success. KEP in addition to conjunctival resection may suppress in ammation by maintaining the space between the cornea and immune cells of the conjunctiva.
Ti et al. reported that 11 eyes (11.9%) in 92 fungal keratitis cases perforated. Perforation risk factors of fungal keratitis included misdiagnosis, extensive infectious area, 11 usage of topical steroids, 10 and delayed treatment. [40][41][42] In our study, when the cornea perforated in cases of bacterial and fungal keratitis, we washed out the anterior chamber during the operation. 22,43 Furthermore, topical steroids were not used for fungal keratitis for at least 2 weeks postoperatively and only if fungal in ltration was not observed. 38,44 Thus, two cases of fungal keratitis achieved anatomical recovery with frozen donor tissue.
In one case, the fungal keratitis recurred after therapeutic penetrating surgery but was nally controlled with conservative eye drop treatment. However, one case of fungal keratitis used topical steroids after misdiagnosis as herpes keratitis and had to undergo therapeutic transplantation due to perforation. This case eventually ended up with phthisis bulbus. In the treatment of fungal keratitis, misdiagnosis and early use of steroids may result in detrimental consequences.
The main complications after therapeutic keratoplasty were recurrence of the original disease (7 eyes; 17.1%) and phthisis bulbi (7 eyes; 17.1%). In a previous report, secondary glaucoma and corneal epithelial defect were the main complications. 9,27 Iris anterior synechiae preoperatively may induce secondary glaucomas. 45 If conservative treatment is not effective and the anterior chamber is not reformed, irreversible iris anterior synechiae may occur. Therefore, if iris anterior synechiae did not return to the original position for more than 2 weeks, we intervened surgically to separate the synechiae to reconstruct the anterior chamber. When in bacterial and fungal keratitis the graft margin was close to the limbus, postoperative in ammation causes trabeculitis and glaucoma. 23 In our study, all cases of large grafts resulted in phthisis bulbi, perhaps because the ciliary body function was weakened.
A limitation of the current study is the low number of cases. However, our results show that it is necessary to correctly diagnose the original cause of the disease resulting in corneal perforation and to carefully determine the time of surgical treatment.

Conclusions
Therapeutic keratoplasty using cryopreserved corneas for corneal perforation is good option for achieving an anatomical cure. As the graft size affects the prognosis, it is important to perform surgical treatment before the infected lesion reaches the limbus, especially in infectious keratitis. We found that a good cure rate was achieved regardless of using cryopreserved donor corneas in all cases. It is necessary to diagnose the cause of the disease at an early stage and provide appropriate conventional and surgical treatments.

List Of Abbreviations
AMT; amniotic membrane transplantation, PKP; penetrating keratoplasty, LKP; lamellar keratoplasty, KEP; keratoepithelioplasty, VA; visual acuity, DALK; deep anterior lamellar keratoplasty, DSAEK; Descemet stripping automated endothelial keratoplasty Declarations Ethics approval and consent to participate The Institutional Review Board of Keio University, Tokyo, Japan approved the protocol (#2017-0219), and all participants undergoing surgery provided written informed consent. The research methods adhered to the tenets of the Declaration of Helsinki.

Consent for publication
Parental consent for publication was obtained for this report.

Availability of data and materials
The datasets used and analyzed during the current study available from the corresponding author on reasonable request.

Competing Interests
Yuichi Uchino and Miki Mizuno declare that they have no competing interests.
Shigeto Shimmura Research funding received from Santen Pharmaceutical Co., Rhoto Pharmaceuticals CO, and Sumitomo Dainippon Pharma Co.
Kazuo Tsubota: Consultant and research funding received from Santen Pharmaceutical Co., Ltd., research funding received from Kowa Company, and holds the patent rights for the method and the apparatus used for the measurement of the functional visual acuity (US patent no: 7470026). Founder of Tsubo Labo, Tokyo, Japan.

Funding
There is no sponsorship of funding arrangements relating to our research.

Authors' contributions
MM had contributed to the conception and design of the study, analysis and interpretation of data, and drafted the article. YU and SS had made contributions to the conception of the study and critical revision of the article for important intellectual content. KT has made contributions to the conception of the work. Figure 1