Results of
Penetrating Keratoplasty for the
Treatment of Corneal Perforations Janis R. Nobe, MD; Bonnie T. Moura, MD;
\s=b\ We retrospectively analyzed 46 consecutive cases of penetrating keratoplasty performed as part of the treatment of corneal perforations; the minimum follow-up time after keratoplasty was 7 months. Predisposing conditions leading to perforation were an infectious keratitis in 26 eyes (57%), trauma in 14 eyes (30%), and corneal melt associated with ocular surface disorder in 6 eyes (13%). The success of penetrating keratoplasty in the treatment of corneal perforation depended on the timing of surgery and the cause of the perforation. If the perforation was traumatic in origin, delaying surgery for at least 3 months significantly improved the chances for graft success. Eighty percent of the penetrating keratoplasties delayed 3 months following primary repair of corneal laceration remained clear, and 50% of these patients had a visual acuity of 20/60 or better. If penetrating keratoplasties were performed for an infectious corneal perforation, grafts had a better chance to remain clear if surgery could be delayed. All grafts performed for corneal perforation associated with melting and ocular surface abnormalities failed.
(Arch Ophthalmol. 1990;108:939-941)
unfortuperforation is Corneal nate sequela of infectious sterile or an
inflammatory keratitis, trauma, and dry-eye conditions, including collagen
vascular disease and cicatricial dis¬ orders.12 Alkali injuries to the eye and exposure keratitis can also result in corneal ulcération and perforation.3"5 Since corneal perforation has a high ocular morbidity, prompt diagnosis is crucial. Treatments available for the
Accepted for publication April 3,1990. From the Doheny Eye Institute and the Department of Ophthalmology, the University of Southern California School of Medicine, Los Angeles. Dr Robin is now with the University of Illinois at Chicago Eye Center. Reprint requests to Doheny Eye Institute, 1355 San Pablo St, Los Angeles, CA 90033 (Dr Smith).
Jeffrey
B.
Robin, MD; Ronald E. Smith,
MD
management of corneal perforation include tissue
adhesives,5"7 lamellar keratoplasty,1-8 therapeutic soft con¬ tact lenses,9 conjunctival flaps,10 and penetrating keratoplasty (PK).2·11"13 Penetrating keratoplasty may be in¬ dicated immediately after corneal per¬ foration to restore anatomic integ¬ rity of the anterior segment,21113 al¬ though such procedures have widely varying success rates.2·811 Dohlman et al8 considered the treatment of corneal
perforations by PK to be dangerous due to the high risk of complicating synechiae, glaucoma, uveitis, and eventual graft failure in these in¬ flamed eyes. Similarly, Foster and Duncan,14 in a series of PKs for herpes simplex keratitis, reported no clear grafts after 16 emergency PKs for perforation. Tragakis et al," however, described 10 cases of corneal perfora¬ tion treated by PK as a primary proce¬ dure. Seven grafts (70%) remained clear and eyes achieved a visual acuity
of 20/60 or better after at least 7 months of follow-up. More recently, Hill15 described clarity in 70% of grafts following PK performed in the acute phase of deep, indolent, infectious cor¬ neal ulcers or descemetoceles. To better understand the relation¬ ship between the timing of surgery and graft success, we report herein a ret¬ rospective analysis of 46 consecutive cases of PK performed as part of the treatment of corneal perforation. SUBJECTS AND METHODS
The records of all patients with corneal perforation who underwent PKs at the Los Angeles County/University of Southern California Medical Center or the Doheny Eye Institute, Los Angeles, between July
1979 and December 1988 were reviewed. If PK had been performed more than once in an eye, the results of only the first PK were
analyzed.
Factors analyzed included patient age and sex, predisposing conditions leading to corneal perforation, the timing of PK fol¬
lowing perforation, previous therapy (in
Downloaded From: http://archopht.jamanetwork.com/ by a Oakland University User on 05/31/2015
cases in which PK was delayed), surgical procedure, and graft clarity. All cases of infectious keratitis were treated medically with appropriate antibacterial or antiviral agents. Although several corneal surgeons performed these operations, standard PK techniques were routinely used. Oversized donor buttons were used in all cases (0.25 mm for phakic eyes and 0.5 mm for aphakic and pseudophakie eyes). Graft size was de¬ pendent on the area of original corneal dis¬ ease, and donor button diameters ranged from 7.5 to 10 mm. Interrupted 10-0 nylon
sutures
were
used in all
cases
and knots
routinely buried in host corneal tissue. The minimum follow-up time for all pa¬ tients was 7 months. Statistical signifi¬ cance was determined by the Mantel-Haenswere
zel 2 test and Fisher's Exact Test.
Penetrating keratoplasties following cor¬
neal perforation were classified into one of four categories: group 1, emergent PK, per¬ formed within 24 hours of corneal perfora¬ tion; group 2, urgent PK, performed within 2 to 6 days after perforation; group 3, PK performed at an intermediate stage after perforation (1 week to 2 months), after suc¬ cessful use of tissue adhesive or lamellar keratoplasty as the initial treatment for corneal perforation; and group 4, PK de¬ layed at least 3 months (Table 1). RESULTS
Forty-six eyes of 46 patients were treated with PK following corneal per¬ foration. Twenty-six (57% ) of the pa¬ tients were male and 20 (43%) were female; the mean age was 46 years, with a range of 7 months to 87 years. The predisposing condition leading to perforation was an infectious keratitis in 26 eyes (57%), trauma in 14 eyes (30%), and corneal melt associated with ocular surface disorder in 6 eyes (13%).
Corneal transplantations per¬ formed emergently (group 1) retained clarity in only 1 (9% ) of 11 eyes. If PK was performed for an infectious cor¬ neal perforation, clear grafts were seen more often when the timing of surgery was urgent or intermediate (57% and 31%, respectively) than when it was performed emergently (17%, Table 2). If PK was performed
Table 1.—Cases of Corneal Perforation Treated Case No.
Cause of Perforation
Age, y/ Sex
Group 0.6/Female 87/Female 5/Male
by Penetrating Keratoplasty*
Graft Status 1
Initial
Therapy
VA
(Emergent PK)
Pseudomonas keratitis Culture-negative keratitis
PK
remained clear. The minimum followpatients was 7 months. All PKs performed following cor¬ neal perforation associated with cor¬ neal melting and ocular surface abnor¬ malities failed. Of the six patients, four had rheumatoid arthritis, one had ocular cicatricial pemphigoid, and one had ectodermal dysplasia. In all six cases, further corneal melting oc¬ curred after PK was performed and resulted in a second descemetocele or perforation. In five of the six cases, a repeated PK was performed (one pa¬ tient refused additional surgery), and was followed by continued corneal
LP
Trauma
PK PK
65/Female
Pseudomonas keratitis Corneal melt (pemphigoid)
21/Female
Herpetic
33 /Female
Corneal melt
67/Male
up time for all
PK
keratitis
PK
(ectodermal dysplasia) 80/Female
Trauma
48/Male
Culture-negative keratitis Culture-negative keratitis
PK
48/Female
Trauma
PK
Group 76/Male
2
(Urgent PK)
Corneal melt
G, CL
(rheumatoid arthritis) 13
15
47/Male
Moraxella keratitis
65/Female
Corneal melt
72/Male 5/Female
18
55/Female 65/Male 29/Male 76/Male
Staphylococcal keratitis Herpetic keratitis Streptococcal keratitis
Group
23
20/Female 80/Male
Herpetic Staphylococcal
86/Female
Corneal melt
67/Male 25
55/Female
26
64/Male 62 /Male
28
62/Female
29
33/Male
30
53/Male 53/Male
32
3/Male
33
71/Male
35
30/Male 30/Female 85/Female
melting. Ultimately, two patients de¬ veloped phthisical eyes, two eyes un¬ derwent conjunctival flap, and two eyes were intact with a cloudy graft. Overall, PK was performed alone in
CL
(rheumatoid arthritis) Streptococcal keratitis Herpetic keratitis Streptococcal keratitis
3
20/25
20/200
G, CL G, CL G, CL
28 cases, was combined with cataract extraction in 7 cases, and was per¬ formed with vitrectomy in 8 cases. In 2 other cases, intraocular lenses had to be removed; in 1 emergent case, chor¬ oidal effusions were drained.
20/30
G, CL 20/100
CL
20/40
G, CL G, CL G, CL
(Intermediate PK)
keratitis
COMMENT
keratitis
(rheumatoid arthritis) Herpetic keratitis Culture-negative keratitis Herpetic keratitis Staphylococcal keratitis Streptococcal keratitis
HM
Gonococcal keratitis Mixed-bacterial keratitis Mixed-bacterial keratitis
G, G, G, G, G,
CL CL CL CL CL
Lamellar G, CL
keratoplasty
G, CL G, CL G, CL
Trauma Corneal melt
(rheumatoid arthritis) Pseudomonas keratitis Mixed-bacterial keratitis
Culture-negative
G, CL G, CL G, CL
keratitis
Group Trauma Trauma
4
(Delayed PK)
Primary repair Primary repair 35/MaL. Trauma 39 20/25 Primary repair 40 62/Male Trauma 20/100 Primary repair 20/300 Primary repair 22/Male Trauma Primary repair Trauma 43 29/Male Primary repair 24/Female Trauma 20/100 Primary repair 6/Female Trauma 45 20/400 Primary repair 29/Female Trauma 46 20/50 Primary repair * VA indicates postoperative visual acuity in patients with clear grafts; PK, penetrating keratoplasty; F, failed; C, clear; LP, light perception; G, cyanoacrylate tissue adhesive; CL, bandage soft contact lens; and HM, hand 46/Male
following a traumatic corneal perfora¬ tion, a graft clarity rate of 80% (8/10) was seen if surgery could be delayed 3 months (group 4). None of the three corneal grafts done emergently (group 1) for traumatic corneal perforations
20/40
motions.
Downloaded From: http://archopht.jamanetwork.com/ by a Oakland University User on 05/31/2015
The results of this study confirm that the success of PK in the treatment of corneal perforations depends on the timing of surgery and the cause of the perforation.2516 In our cases, delaying PK for at least 3 months after trau¬ matic corneal perforation signifi¬ cantly improved the chances for graft clarity (P < .05). Eight of the 10 grafts delayed 3 months following primary repair of corneal laceration have re¬ mained clear; 50% of these eyes have a visual acuity of 20/60 or better, and the others have a visual acuity between 20/100 and 20/400, with the reduction due to amblyopia, macular edema, or macular degeneration. Several factors may account for the observed differ¬ ence in graft success for traumatic perforations in group 1 (emergent) vs group 4 (delayed 3 months). Severe trauma is generally associated with a marked degree of inflammation ini¬ tially. This associated inflammation may have adversely affected graft out¬ come. The traumatic corneal perfora¬ tions in eyes in group 4 that underwent primary repair were treated with top¬ ical steroids to decrease inflammation for at least 3 months prior to PK. An¬ other factor relates to the size and ex¬ tent of the corneal lacerations. It can be expected that larger lacerations
Table 2.—Graft
Clarity Rates, Cause
of
Perforation, and Timing of Penetrating Keratoplasty (PK)* No. (%) of Clear Grafts
Cause of Corneal Perforation Infectious keratitis Melt
.-
Group 1, Emergent 1/6(17) 0/3
All
1/11(9)
cases
Group 2, Urgent 4 / 7 (57)
0/2
Trauma
Following
PK
-
Group 3, Intermediate 4 /13 (31 )
0/2
Group 4, Delayed
0/2 0/1
4/9(44)
8/10(80) 8/10(80)
4/16(25)
...
*Eyes in group 1 underwent PK less than 24 hours after corneal perforation; eyes in group 2 underwent PK within 2 to 6 days after corneal perforation; eyes in group 3 underwent PK within 1 week to 2 months after corneal perforation; and eyes in group 4 underwent PK 3 months or more after corneal perforation.
(with their associated intraocular complications) would more frequently require emergent PK. Because of the retrospective nature of this study,
however, this variable could not be properly evaluated and analyzed. None of the eyes with traumatic perfora¬ tions had underlying infections or im¬ munologie disorders that could have
affected graft outcome. Our success rates observed in de¬ layed PK for traumatic corneal perfo¬ ration compare favorably with those of Sharkey and Brown,17 who described eight eyes with opaque corneas from healed corneal lacerations that under¬ went PK. The corneal grafts remained clear in all eyes and visual acuity im¬ proved to 20/50 or better in two of the eyes.
study, all
and perforations.2·5 Fifty-six percent of the corneal perfo¬ rations in our study were associated
descemetoceles with
an
infectious keratitis. While
grafts in this category tended to re¬ main clear more frequently if surgery could be delayed (57% and 31% in groups 2 and 3, respectively) than if surgery were emergent (17%, group 1), the difference was not statistically significant. Additionally, it might be expected that the more severe and ex¬ tensive infections with larger perfora¬ tions would frequently require emer¬ gent PK. Due to the retrospective na¬ ture of the study, however, this variable could not be properly evalu¬ ated and analyzed. Others2·14 have observed that
emer¬
gent PKs in perforated eyes have
higher
rates of
PKs performed on patients with corneal melting and per¬ foration due to immunologie disorders and associated ocular surface abnor¬ malities failed. Since further corneal melting and perforation occurred after the initial PK (independent of the timing of PK after perforation) and resulted in a repeated surgery and failure, it seems likely that the pri¬ mary immunologie processes contin¬ ued after PK. Keratoconjunctivitis sicca and other ocular surface prob¬ lems associated with these immuno¬ logie disorders may also have ad¬ versely affected graft outcome. Kenyon18 emphasized the early ap¬ plication of tissue adhesive in any
graft failure than do those performed several weeks to months after perforation occurs. Fos¬ ter and Duncan,14 in their study of PKs performed for herpetic keratitis, noted that all 16 grafts performed acutely for perforated corneal ulcers failed. Eighty-five percent of the eyes that had PKs performed secondarily, 6 to 24 months after perforation, had clear grafts. These authors emphasized con¬ servative treatment of herpetic perfo¬ rations (tissue adhesive or patch graft), followed by intensive anti-in¬ flammatory therapy to achieve a quiet eye, maintenance of this quiescent
cornea to like a rational
herpetic corneal per¬ forations, and suggested intensive top¬ ical steroid treatment postoperatively to prevent synechiae and graft rejec¬
In
our
noninfected, progressively thinning prior perforation. This seems approach, especially in view of the poor results of PK. Many of our patients had tissue adhesive placed prior to surgery, but only after perfo¬ ration had occurred. Perhaps earlier application of tissue adhesive could have averted the need for PK. How¬ our results suggest that if PK is performed for corneal perforation as¬ sociated with an underlying immuno¬ logie or ocular surface disorder, then graft outcome may be particularly ever,
poor.
Infections have frequently been cited to be the most important under¬ lying factor associated with corneal
state, and then PK to restore vision. In contrast, Patten et al12 reported a 60% success rate for cases of acute
immediate PKs in
tion.
Emergency PK may be necessary to integrity in large corneal perforations. However, in those cases in which temporizing mea¬ sures can be effectively used, our re¬ restore anatomic
sults indicate that a reasonable rate of
graft clarity can be achieved if PK can be delayed. Presently available alter¬ natives to acute PK include therapeu¬ tic contact lenses, lamellar keratoplas¬ ty, tissue adhesives, and primary cor¬ neal repair of traumatic perforations.
Downloaded From: http://archopht.jamanetwork.com/ by a Oakland University User on 05/31/2015
perforation is traumatic in na¬ recommend primary repair and delay of PK for at least 3 months. If the
ture,
we
This study was supported in part by Public Health Service Core Center grant EY03040 from the National Eye Institute, National Institutes of Health, Bethesda, Md, and by a grant from Research to Prevent Blindness Ine, New York, NY.
References 1. Portnoy SL, Insler MS, Kaufman HE. Surgical management of corneal ulceration and perforation. Surv Ophthalmol. 1989;34:47-58. 2. Arentsen JJ, Laibson PR, Cohen EJ. Management of corneal descemetoceles and perforations. Ophthalmic Surg. 1985;16:29-33. 3. Brown SI, Tragakis MP, Pearce DB. Treatment of the alkali-burned cornea. Am J Ophthalmol. 1972;74:316-320. 4. Pfister RR. Chemical injuries of the eye.
Ophthalmology. 1983;90:1246-1253. 5. Hirst LW, Smiddy WE, Stark WJ. Corneal perforations: changing methods of treatment, 1960-1980. Ophthalmology. 1982;89:630-635. 6. Webster RG Jr, Slansky HH, Refojo MF, Boruchoff SA, Dohlman CH. The use of adhesive for the closure of corneal perforations: report of two cases. Arch Ophthalmol. 1968;80:705-709. 7. Boruchoff SA, Refojo M, Slansky HH, WebJr, Freeman MI, Dohlman CH. Clinical applications of adhesives in corneal surgery. Trans Am Acad Ophthalmol Otolaryngol. 1969; ster RG
73:499-505. 8. Dohlman CH, Boruchoff SA, Sullivan GL. A technique for the repair of perforated corneal ulcers. Arch Ophthalmol. 1967;77:519-525. 9. Leibowitz HM, Berrospi AR. Initial treatment of descemetocele with hydrophilic contact lenses. Ann Ophthalmol. 1975;7:1161-1166. 10. Paton D, Milauskas AT. Indications, surgical technique, and results of thin conjunctival flaps on the cornea: a review of 122 consecutive cases. Int Ophthalmol Clin. 1970;10:329-345. 11. Tragakis MP, Rosen J, Brown SI. Transplantation of the perforated cornea. Am J Ophthalmol. 1974;78:518-522. 12. Patten JT, Cavanagh HD, Pavan-Langston D. Penetrating keratoplasty in acute herpetic corneal perforations. Ann Ophthalmol. 1976; 8:287-294. 13. Girard LJ. Emergency keratoplasty in corneal lacerations. Cornea. 1983;2:247-249. 14. Foster CS, Duncan J. Penetrating keratoplasty for herpes simplex keratitis. Am J Ophthalmol. 1981;92:336-343. 15. Hill JC. Use of penetrating keratoplasty in acute bacterial keratitis. Br J Ophthalmol. 1986; 70:502-506. 16. Arentsen JJ, Laibson PR, Cohen EJ. Management of corneal descemetoceles and perforations. Trans Am Ophthalmol Soc. 1984;82:92-105. 17. Sharkey TG, Brown SI. Transplantation of lacerated corneas. Am J Ophthalmol. 1981;91:721\x=req-\ 725. 18. Kenyon KR. Corneal perforations. Oph-
thalmology. 1982;89:634-635.