Corneal perforation and severe thinning of the stroma (descemetocele), caused by infectious, immune, and traumatic disorders, are ophthalmic emergencies that require prompt intervention to retain ocular integrity and visual function. A delay in treatment may result in endophthalmitis, loss of vision, and even loss of the eye. Current management includes placement of bandage contact lens, application of tissue adhesives, and surgical interventions such as amniotic membrane transplantation, corneal patch graft, or transplantation.1–3
Cyanoacrylate tissue adhesive (CTA) combines cyanoacetate and formaldehyde, and polymerizes upon contact with fluid providing tectonic strength.4 The United States Food and Drug Administration approved CTA for topical skin approximation in 1998.5 Its use in corneal perforation was first reported in 1968, and it is now a first-line, albeit off-label, treatment for corneal perforation, especially when grafting is not feasible or immediately necessary.4,6,7 Previous studies have reported various success rates of CTA in sealing perforation acutely,8–12 but the retention of CTA and maintenance of globe integrity over time has not been thoroughly examined. In addition, factors associated with glue failure have not been systematically analyzed. In this study, we review the use of CTA in 140 eyes with corneal perforation and thinning at the Massachusetts Eye and Ear Infirmary, Boston, between 2001 and 2018 and report the clinical characteristics and outcomes.
We performed a retrospective review of clinical charts of all patients treated with CTA for corneal perforation or thinning at the Massachusetts Eye and Ear Infirmary, Boston, MA, between January 2001 and January 2018. Institutional Review Board/Ethics Committee approval was obtained and the described research adhered to the tenets of the Declaration of Helsinki.
Subjects were identified with Current Procedural Terminology and diagnosis codes and electronic and paper medical records were reviewed. Cases with incomplete data or follow-up shorter than 1 week were excluded. The following data were collected from Research Electronic Data Capture (REDCap, Vanderbilt) by 2 authors (R.B.S and A.Y.): demographic information, medical and surgical histories, detailed ophthalmic history, systemic and ophthalmic medications, ophthalmic comorbidities, best-corrected distance visual acuity (BCVA), intraocular pressure, ocular examination, etiology, location and size of the perforation/thinning, total number of CTA applications, subsequent medical and surgical treatments, and the final outcome. The data were then independently verified by a third author (J.Y.). A combination adhesive of 2-Octyl Cyanoacrylate and n-Butyl Cyanoacrylate (MSI-Epiderm Glue+Flex; Medislav Inc, Canada) was used and it was applied per the treating ophthalmologist's preference. CTA was reapplied when the globe was no longer intact (judged clinically by positive Seidel test, hypotony, and/or shallow anterior chamber). Subsequent follow-up, reapplication of CTA, and medical and surgical interventions were decided by the treating physician(s).
Visual acuity was converted from Snellen chart to logarithm of the minimum angle of resolution as previously described.13 Visual acuities of light perception and no light perception were not converted. Continuous variables were reported as mean (SD) or median (interquartile range), and categorical variables were reported in numbers (percentage).
Survival Analysis of Glue Application Success
GraphPad Prism software version 5.00 for Windows was used for Kaplan–Meier curve survival analysis (GraphPad Software, La Jolla, CA). Both success of a single (initial) CTA application and multiple CTA applications were analyzed. In the case of a single CTA application, success was defined as an intact globe without additional intervention. Each eye was considered 1 sample (N = 140) and subsequent glue application and surgical intervention were counted as “events.” Repeat application was counted as a single event regardless of the times performed. In the case of multiple applications, success of multiple CTA applications was defined as an intact globe without surgical intervention. Each eye was considered 1 sample (N = 140) and subsequent surgical interventions were counted as “events.”
Logistic Regression Analysis
To assess factors correlated with CTA failure, logistic regression was generated using SAS software 9.4 for Windows (SAS Institute Inc, Cary, NC). CTA failure was defined as an open globe in need of surgical intervention within 1 month of the initial CTA applications (regardless of the total number of CTA applications). Univariable analyses were performed using the following variables: age, sex, systemic conditions, systemic autoimmune conditions, use of systemic immunosuppression, significant ocular surface diseases (OSD), use of ophthalmic steroids, location, size, and etiologies of perforation/thinning, and total number of CTA applied. No mathematical transformation of variables was performed. In the final multivariable regression model, perforation (vs. thinning), location, use of topical steroid, sterile melt (vs. other causes), and single CTA application (vs. multiple) were included based on clinical relevance and to prevent overfitting of the model. Two-sided P value <0.05 was considered statistically significant.
A total of 137 patients were included in the study with a median age of 63 years and 69 (50%) were women. Baseline characteristics are shown in Table 1. Systemic conditions were present in 115 patients (84%). Hypertension was the most common systemic condition (54 patients), followed by autoimmune diseases (46 patients) including rheumatoid arthritis (RA), systemic lupus erythematosus, scleroderma, polymyalgia rheumatica, and granulomatosis with polyangiitis. Twenty-eight patients (20%) were on systemic immunosuppression including corticosteroids and/or steroid-sparing agents at the time of presentation. Three patients had bilateral perforation or thinning that required bilateral CTA application: 1 patient had an RA-associated peripheral ulcerative keratitis, 1 patient had severe alkali burn, and 1 patient had biopsy-proven mucous membrane pemphigoid. Baseline characteristics of the 140 glued eyes are shown in Table 1. One hundred three patients (74%) had documented OSD with eyelid disorder (41), neurotrophic keratopathy (41), and dry eye disease (34) being the most common. Most patients (129, 92%) were on topical antimicrobial treatment and 98 patients (70%) were on topical corticosteroid at the time of CTA application. Forty-seven patients (34%) were on glaucoma mediations.
The clinical characteristics of perforation/thinning and CTA application are shown in Table 2. Eighty-nine eyes (64%) presented with perforation (Seidel positivity) and 51 (36%) with severe thinning. The location was central or paracentral in 82 eyes (59%) and peripheral in 57 (41%). Median size of perforation was 3.1 mm2. Causes of perforation or thinning included microbial infection, RA-associated melt, graft-versus-host disease-associated melt, other noninfectious melts, laceration or mechanical trauma, keratoprosthesis melts, and burn; several cases were multifactorial. Sixty-seven of the 77 microbial cases (85%) were confirmed by gram stain, microbial culture, confocal microscopy, or pathology. The median and mean numbers of total CTA applications were 1 and 1.6, with 88 eyes (62.9%) requiring more than 1 application. Median BCVA before and after CTA application was 2.3 and 2.0 in logarithm of the minimum angle of resolution (P = 0.44, Wilcoxon matched pairs test; see Supplemental Figure 1, Supplemental Digital Content 1, http://links.lww.com/ICO/A767).
The Kaplan–Meier curve of CTA success with single and multiple applications is shown in Figure 1. Success was defined as an intact globe without additional intervention (repeat glue application or surgery) in the case of single CTA application and without surgical intervention in the case of multiple CTA applications. The success rate of single and multiple CTA applications was 55% and 72% at 10 days, 39% and 61% at 30 days, and 28% and 46% at 90 days, respectively. The median retention of CTA was 58 days. Logistic regression was performed to identify factors associated with CTA failure (defined as open globe in need of surgical intervention within 1 month of initial CTA application). In univariable analysis (see Supplemental Table 1, Supplemental Digital Content 1, http://links.lww.com/ICO/A767), although age, sex, systemic conditions, OSD, microbial etiology, laceration, keratoprosthesis, and history of corneal grafting were not significantly correlated with CTA failure, size of perforation/thinning (see Supplemental Figure 2, Supplemental Digital Content 1, http://links.lww.com/ICO/A767) and single application were correlated with higher failure rates. The final multivariable model is shown in Table 3. Single application (vs. multiple) was significantly correlated with a higher failure rate (odds ratio 3.09, P = 0.01). Perforation (vs. thinning) was associated with a higher failure rate (odds ratio 2.18), but statistically nonsignificant (P = 0.06). The etiologies and location of perforation/thinning and the use of topical corticosteroid were not significantly associated with glue failure.
Within 1 month of CTA application, complications associated with CTA such as corneal neovascularization and stromal inflammation were present in 17 and 15 eyes, respectively, of 61 eyes with recorded CTA details (Table 4). The most common surgical intervention (Table 4) was penetrating keratoplasty (43 of 52 eyes), followed by keratoprosthesis (5), enucleation/evisceration (3), and amniotic membrane transplant (1). No lamellar keratoplasty was performed in this cohort.
It has been 50 years since Webster et al6 first reported the use of CTA in sealing corneal perforations clinically. Despite CTA's widespread use as a first-line treatment for corneal perforation and several previous reports on its success in sealing the globe acutely,8,10,11 the retention of CTA and maintenance of globe integrity over time, as well as factors associated with glue failure have not been systematically analyzed. The current study of 140 eyes, to our knowledge, is the largest case series reporting the outcomes of CTA application and the first to identify clinical characteristics associated with glue failure. Most patients (84%) presented with systemic conditions; one-third had autoimmune diseases and 20% were on a systemic immunosuppressive regimen. Ophthalmic morbidities were prevalent (74%) and the overwhelming majority were on topical antimicrobial and corticosteroid treatments. A single CTA application was successful in restoring globe integrity in approximately half of the eyes at 10 days after application, but nearly two-thirds of the cases required a secondary procedure such as reapplication of CTA or surgery within 30 days. Multiple CTA applications improved globe stability compared with a single application, but the maintenance of an intact globe declined with time. Perforation (vs. thinning), large size of perforation/thinning, and single application (vs. multiple) were associated with CTA failure; in contrast, sex, age, autoimmune conditions, use of topical steroids, location, and etiologies of perforation/thinning were not significantly correlated with CTA failure.
Corneal melting and subsequent perforation are the end results of many ocular surface disorders, including infectious, immune, and traumatic etiologies.3 Several earlier studies have examined these causes and the use of CTA (see Supplemental Table 2 for a brief summary of previous studies on the use of CTA in corneal thinning and perforation, Supplementary Digital Content 1, http://links.lww.com/ICO/A767).8,11,14 Our collective understanding of these conditions and treatments has improved significantly in the past 30 years, warranting an updated assessment of the nonsurgical management of corneal perforation.15–19 All but 5 cases in our series were between 2008 and 2018, representative of current population and practice. In our series, approximately half of the eyes received CTA application because of infectious keratitis, predominantly bacterial ulcers (49% of 140 eyes). CTA application due to bacterial infection in our series is higher than previously reported (range 20%–29%; see Supplemental Online Data, Supplemental Digital Content 1, http://links.lww.com/ICO/A767).8,10,11,14 On the other hand, we have a lower rate of perforation/thinning associated with RA (10%) compared with previous reports (15%–48%).8,10,14 We speculate that more widespread and earlier use of disease-modifying antirheumatic drugs in rheumatoid diseases may account for the differences. Our series also examined ocular surface comorbidities and found 75% of patients presented with significant OSD, of which lid abnormality and associated exposure, neurotrophic keratopathy, and dry eye disease were most common. This reinforces the critical value of adequately managing these comorbidities to prevent corneal melt and perforation.
The success of CTA application in sealing corneal perforation varied significantly across studies (success rate of 29%–86%), depending on the etiologies of perforation, indications for application, and definition of success.8–12,20 Our series demonstrates that multiple applications of CTA are often required, suggesting that failure to seal perforation after 1 application need not necessarily be interpreted as a requirement to proceed with surgery and that repeat glue attempts may be made. More importantly, our study followed the maintenance of globe integrity after glue application and showed a gradual decline over time. For instance, the success rate decreased from 72% at 10 days, after glue application, to 61% and 46% at 30 and 90 days, respectively. This reinforces the importance of close monitoring and high clinical vigilance after CTA application. Although the “re-glue” rate was higher in our series (63%) compared with previous studies, the retention of CTA was similar, with a median duration of 58 days.8–12,14,20 The longest duration of CTA in our series was 2016 days (about 5.5 years), slightly longer than a previous report.21 One eye with significant lagophthalmos and exposure after severe facial alkali burn underwent a total of 11 CTA applications within 5 months and achieved a final BCVA of 20/25 without surgical intervention.
Complications associated with CTA have been reported including increased ocular surface inflammation, corneal neovascularization, and giant papillary conjunctivitis.20,22,23 Although various forms of cyanoacrylate have been shown to inhibit gram-positive bacterial growth in vitro,24,25 infiltrates could develop underneath the glue. The opacity of CTA makes diagnosing these new or recurrent infections more challenging and requires high clinical suspicion.8,14,26 Our cohort demonstrated cases of persistent neovascularization and stromal inflammation after CTA application; however, it was not clear if the stromal inflammation in these cases was infectious or sterile based on clinical documentation. CTA can be applied directly from an applicator or syringe, or indirectly on a trephined sterile drape.27 When CTA alone is insufficient to seal the wound, combination with suture, tectonic drape patch, paper, or rigid gas permeable contact lens has been reported to secure the globe.28–30 In our recent experience, we have observed additional stability and security by placing a second contact lens, which likely decreases the movement of the underlying contact lens and subjacent CTA with blinking. Because CTA is nonbiodegradable, potentially toxic, and proinflammatory, use of alternative tissue adhesives in sealing corneal perforation and thinning has been explored. In a randomized controlled trial by Sharma et al,20 fibrin glue showed similar efficacy in sealing perforation up to 3 mm in diameter with fewer side effects compared with CTA. Although fibrin glue is biocompatible and widely used in ocular surface surgery, it is designed for adhering soft tissues to one another, and not for filling in stromal defects or for applying to wet surfaces. In addition, fibrin glue takes longer to form plugs, lacks bacteriostatic properties, and has much lower bursting pressure compared with CTA.25 New ways to restore corneal stroma and close corneal wounds include bioengineered collagen-based materials and synthetic polyethylene glycol derivatives.4,7,31 Their clinical use in corneal perforation and thinning is a subject of extensive ongoing investigation.
We performed a regression analysis to identify factors associated with CTA failure. Although large wounds and single (vs. multiple) application are significantly associated with failure in both univariable (see Supplemental Online Data, Supplemental Digital Content 1, http://links.lww.com/ICO/A767) and multivariable analyses, perforation (vs. thinning) is associated with failure but statistically nonsignificant (P = 0.06). Of note, autoimmune conditions, existing OSD, the use of systemic immunosuppression, the use of topical corticosteroids, etiologies, and the location of perforation are not significantly associated with CTA failure. Garg et al12 have reported a high success rate of 76% using CTA in fungal keratitis–related perforation, whereas Moorthy et al noted a much lower success rate of 37% in herpetic keratitis–related perforations.9 Our study has too few cases of fungal and herpetic infections to compare with these series. We examined infectious etiology, trauma, keratoprosthesis, and corneal graft as potential factors in CTA failure and found no correlations in all analyses. In the multivariable analysis, immune-mediated sterile melt had an odds ratio of 1.68 for CTA failure, but this was not statistically significant.
The current study has several limitations. First, the study is retrospective and noncomparative in nature with inherent bias. To our knowledge, the aforementioned study by Garg et al comparing fibrin glue and CTA is the only randomized controlled trial examining CTA.20 Future studies comparing CTA with the new generation of ophthalmic tissue adhesives are warranted. Second, our practice is an academic tertiary eye care referral center and the care may not be truly reflective of that seen in the community. Last, donor tissue and operating room availability is rarely a limiting factor in our practice; hence, the decision to intervene surgically at our center may differ from community practices.
In the aggregate, our data demonstrate that multiple CTA applications are often required to securely seal corneal perforations. Although cyanoacrylate glue is moderately successful in the very short term, the maintenance of globe integrity decreases with time; thus, the need for surgical intervention remains high. Our study suggests that eyes that have had cyanoacrylate glue application should be monitored closely, and repeat glue application needs to be considered before proceeding to surgery.
The authors thank Dr. Ahmad Kheirkhah for his assistance in the study.
1. Portnoy SL, Insler MS, Kaufman HE. Surgical management of corneal ulceration and perforation. Surv Ophthalmol. 1989;34:47–58.
2. Lekskul M, Fracht HU, Cohen EJ, et al. Nontraumatic corneal perforation
. Cornea. 2000;19:313–319.
3. Jhanji V, Young AL, Mehta JS, et al. Management of corneal perforation
. Surv Ophthalmol. 2011;56:522–538.
4. Guhan S, Peng SL, Janbatian H, et al. Surgical adhesives in ophthalmology: history and current trends. Br J Ophthalmol. 2018;102:1328–1335.
5. Mattamal GJ, US FDA perspective on the regulations of medical-grade polymers: cyanoacrylate
polymer medical device tissue adhesives. Expert Rev Med Devices. 2008;5:41–49.
6. Webster RG Jr., Slansky HH, Refojo MF, et al. The use of adhesive for the closure of corneal perforations. Report of two cases. Arch Ophthalmol. 1968;80:705–709.
7. Bhatia SS, Ocular surface sealants and adhesives. Ocul Surf. 2006;4:146–154.
8. Weiss JL, Williams P, Lindstrom RL, et al. The use of tissue adhesive
in corneal perforations. Ophthalmology. 1983;90:610–615.
9. Doughman S, Jhanji V, Constantinou M, et al. Clinical experience with N-butyl cyanoacrylate tissue adhesive
in corneal perforations secondary to herpetic keratitis. Cornea. 2010;29:971–975.
10. Loya-Garcia D, Serna-Ojeda JC, Pedro-Aguilar L, et al. Non-traumatic corneal perforations: aetiology, treatment and outcomes. Br J Ophthalmol. 2017;101:634–639.
11. Leahey AB, Gottsch JD, Stark WJ. Clinical experience with N-butyl cyanoacrylate
(Nexacryl) tissue adhesive
. Ophthalmology. 1993;100:173–180.
12. Garg P, Gopinathan U, Nutheti R, et al. Clinical experience with N-butyl cyanoacrylate tissue adhesive
in fungal keratitis. Cornea. 2003;22:405–408.
13. Holladay JT, Visual acuity measurements. J Cataract Refract Surg. 2004;30:287–290.
14. Moschos M, Droutsas D, Boussalis P, et al. Clinical experience with cyanoacrylate tissue adhesive
. Doc Ophthalmol. 1996;93:237–245.
15. Acyclovir for the prevention of recurrent herpes simplex virus eye disease. Herpetic Eye Disease Study Group. N Engl J Med. 1998;339:300–306.
16. Prajna NV, Krishnan T, Mascarenhas J, et al. The mycotic ulcer treatment trial: a randomized trial comparing natamycin vs voriconazole. JAMA Ophthalmol. 2013;131:422–429.
17. Barron BA, Gee L, Hauck WW, et al. Herpetic Eye Disease Study. A controlled trial of oral acyclovir for herpes simplex stromal keratitis. Ophthalmology. 1994;101:1871–1882.
18. Srinivasan M, Mascarenhas J, Rajaraman R, et al. The steroids for corneal ulcers trial (SCUT): secondary 12-month clinical outcomes of a randomized controlled trial. Am J Ophthalmol. 2014;157:327–333 e3.
19. Singh JA, Saag KG, Bridges SL Jr, et al. 2015 American College of Rheumatology Guideline for the treatment of rheumatoid arthritis. Arthritis Care Res (Hoboken). 2016;68:1–25.
20. Sharma A, Kaur R, Kumar S, et al. Fibrin glue versus N-butyl-2-cyanoacrylate
in corneal perforations. Ophthalmology. 2003;110:291–298.
21. Tan J, Wechsler AW, Watson S. Long-term adhesion of cyanoacrylate
on human cornea. Clin Exp Ophthalmol. 2014;42:791–793.
22. Carlson AN, Wilhelmus KR. Giant papillary conjunctivitis associated with cyanoacrylate
glue. Am J Ophthalmol. 1987;104:437–438.
23. Ferry AP, Barnert AH. Granulomatous keratitis resulting from use of cyanoacrylate
adhesive for closure of perforated corneal ulcer. Am J Ophthalmol. 1971;72:538–541.
24. de Almeida Manzano RP, Naufal SC, Hida RY, et al. Antibacterial analysis in vitro of ethyl-cyanoacrylate
against ocular pathogens. Cornea. 2006;25:350–351.
25. Chen WL, Lin CT, Hsieh CY, et al. Comparison of the bacteriostatic effects, corneal cytotoxicity, and the ability to seal corneal incisions among three different tissue adhesives. Cornea. 2007;26:1228–1234.
26. Cavanaugh TB, Gottsch JD. Infectious keratitis and cyanoacrylate
adhesive. Am J Ophthalmol. 1991;111:466–472.
27. Rana M, Savant V. A brief review of techniques used to seal corneal perforation
using cyanoacrylate tissue adhesive
. Cont Lens Anterior Eye. 2013;36:156–158.
28. Khalifa YM, Bailony MR, Bloomer MM, et al. Management of nontraumatic corneal perforation
with tectonic drape patch and cyanoacrylate
glue. Cornea. 2010;29:1173–1175.
29. Sharma A, Mohan K, Nirankari VS, Management of nontraumatic corneal perforation
with tectonic drape patch and cyanoacrylate
glue. Cornea. 2012;31:465–466.
30. Gandhewar J, Savant V, Prydal J, et al. Double drape tectonic patch with cyanoacrylate
glue in the management of corneal perforation
with iris incarceration. Cornea. 2013;32:e137–e138.
31. Matthyssen S, Van den Bogerd B, Dhubhghaill SN, et al. Corneal regeneration: a review of stromal replacements. Acta Biomater. 2018;69:31–41.
cyanoacrylate; tissue adhesive; corneal glue; corneal thinning; corneal perforation; descemetocele
Supplemental Digital Content
Copyright © 2019 Wolters Kluwer Health, Inc. All rights reserved.