Download The Boston Keratoprosthesis in Stevens

The Boston Keratoprosthesis in Stevens-Johnson
Syndrome
RONY R. SAYEGH, LEONARD P. K. ANG, C. STEPHEN FOSTER, AND CLAES H. DOHLMAN
● PURPOSE:
To evaluate the use of the Boston keratoprosthesis (KPro) in patients with Stevens-Johnson syndrome
(SJS).
● DESIGN: Retrospective, noncomparative, interventional
case series.
● METHODS: Sixteen eyes of 15 patients with SJS underwent KPro surgery at the Massachusetts Eye and Ear
Infirmary from January 2000 through December 2005.
The preoperative, operative, and postoperative findings
were recorded. All patients underwent either the type I
or type II Boston KPro surgery by one surgeon (C.H.D.).
Retention of the prosthesis, best-corrected visual acuity,
the need for surgical revision, and postoperative complications were recorded. The outcomes were compared
with those of an earlier group of patients from the 1990s.
● RESULTS: The mean age of patients was 50 ⴞ 18 years
(range, 23 to 74 years), and the mean duration of their
disease was 10 ⴞ 6.6 years. The mean follow-up period
was 3.6 ⴞ 1.5 years (range, 10.2 months to 5.6 years).
Ten eyes underwent type II KPro surgery, whereas six
eyes underwent type I KPro surgery. Twelve eyes (75%)
achieved a visual acuity of 20/200 or better after surgery,
with eight eyes (50%) achieving excellent vision of
20/40 or better. Visual acuity was maintained at 20/200
or better over a mean period of 2.5 ⴞ 2.0 years.
Preexisting glaucoma was found to be a significant risk
factor for visual loss. There were no cases of KPro
extrusion or endophthalmitis.
● CONCLUSIONS: KPro in SJS has improved, largely
because of the introduction of vancomycin prophylaxis
and better glaucoma treatment. It seems to be superior to
standard penetrating keratoplasty, with or without allografted stem cell transplantation, as judged from the
literature. However, the outcome of the KPro in SJS is still
substantially less favorable than in nonautoimmune diseases. (Am J Ophthalmol 2008;145:438 – 444. © 2008
by Elsevier Inc. All rights reserved.)
S
TEVENS-JOHNSON SYNDROME (SJS) OFTEN CAUSES
sever ocular surface disease and impairment of vision. In these cases, destruction of the corneal
epithelial stem cells located at the corneal limbus results in
conjunctival invasion, corneal neovascularization, chronic
inflammation, and stromal scarring.1 In end-stage cases,
the ocular surface becomes completely dry and the fornices
become obliterated. Glaucoma is a common association.
Corneal transplantation for severe ocular surface disease
in SJS patients is associated with poor prognosis. Penetrating keratoplasty or lamellar keratoplasty in these eyes often
are complicated by prolonged inflammation, persistent epithelial defects, corneal melting, perforation, and eventual
graft failure. Tugal-Tutkun and associates reported that only
two of seven eyes with SJS maintained a visual acuity of
20/200 or better, three to four months after repeated penetrating keratoplasty.2 Solomon and associates demonstrated
that in patients with SJS who underwent keratolimbal allograft followed by corneal transplantation, all grafts had failed
within 14 months.3
Although keratoprosthesis (KPro) surgery has existed for
more than two centuries, it has been only during the past
two decades that significant improvements in techniques
have led to promising results for the treatment of complicated corneal disorders and repeated failed grafts. However, the treatment of severe ocular surface disease arising
from SJS still remains most challenging. Because of the
devastating nature of this condition, there have been few
reports on the use of KPro surgery for the treatment of
SJS.4 – 6
The Boston KPro makes use of a collar-button design
consisting of a polymethyl methacrylate optic that clamps
within it a donor ring of corneal tissue that is then sutured
in place like a traditional graft.7 Recent modifications to
its design and postoperative management of patients have
led to improved clinical outcomes.8 –10 Herein we report
the use of this KPro for the treatment of SJS.
METHODS
Accepted for publication Nov 5, 2007.
From the Department of Ophthalmology, Harvard Medical School,
Massachusetts Eye and Ear Infirmary, Boston, Massachusetts (R.R.S.,
L.P.K.A., C.S.F., C.H.D.); and the Singapore National Eye Centre,
Singapore, Republic of Singapore (L.P.K.A.).
Inquiries to Claes H. Dohlman, Massachusetts Eye and Ear Infirmary, 243
Charles Street, Boston, MA 02114; e-mail: [email protected].
edu
438
©
2008 BY
WE REVIEWED ALL PATIENTS WHO UNDERWENT BOSTON
KPro implantation from January 2000 through December
2005 at the Massachusetts Eye and Ear Infirmary. A
total of 154 patients had undergone KPro surgery for
complicated corneal disease or repeated failed corneal
grafts. Among these patients, 16 eyes of 15 patients had
ELSEVIER INC. ALL
RIGHTS RESERVED.
0002-9394/08/$34.00
doi:10.1016/j.ajo.2007.11.002
FIGURE 1. Images of eyes with end-stage Stevens-Johnson syndrome (SJS). The preoperative eyes manifest a spectrum of
complications. Shown here are corneal scarring, neovascularization, and symblepharon formation. One eye (Top left) still is slightly
wet, whereas the other (Top right) has an end-stage dry keratinized surface. (Middle) Two types of keratoprostheses (KPro) can
be implanted: (Middle left) the type I KPro is best suited for eyes with some residual tear function; (Middle right) and the type II
counterpart has a protruding cylinder that passes through an opening in the lid and therefore is used in severely dry eyes. The
(Middle right) assembly of the type I KPro is depicted, and examples of the postoperative appearance of (Bottom left) type I and
(Bottom right) type II KPro is shown.
a basic cause of SJS, and the charts of these patients
were reviewed retrospectively. All patients were operated on by the same surgeon (C.H.D.). The symptomatology, best spectacle-corrected visual acuity, physical
VOL. 145, NO. 3
BOSTON KPRO
IN
findings, detailed ophthalmic examination results, surgical details, postoperative examination findings, and
ocular complications were recorded on an itemized data
collection form.
STEVENS-JOHNSON SYNDROME
439
The preoperative evaluation included a detailed history,
an assessment of the visual acuity and visual potential,
tonometry, slit-lamp examination, fluorescein staining,
and anterior segment photography. Light projection and
color discrimination were used to evaluate the visual
potential of the eyes. Particular attention was paid to the
presence of glaucoma and ensuring that its treatment was
optimal before KPro surgery. When the posterior pole
could not be visualized, B-scan ultrasonography was performed to exclude preexisting retinal or optic nerve
abnormalities.
To be considered for this operation, the eye to be
operated on had to have a visual acuity at least light
perception but worse than 20/400, and the opposite eye
also had to have poor visual acuity. Eyes with some residual
tear function underwent type I KPro surgery, whereas eyes
that had clinically significantly reduced tear function or
were completely dry were selected for type II KPro surgery.
Glaucoma shunts, when indicated, were implanted either
as a preliminary procedure, in conjunction with the KPro,
or were implanted as a secondary procedure. The axial
length of the eye was determined by A-scan biometry and
the appropriate aphakic or pseudophakic KPro lens powers
were selected. Donor corneal tissue was obtained from the
Tissue Bank International (Baltimore, Maryland, USA).
acetate (1% suspension) eye drops, all administered four times
daily initially. This dosage was tapered to twice daily over one
month, to be continued as such for life. The postoperative
examination included assessment of the visual acuity, intraocular pressure, optic disk, and fundus. Particular attention was
paid to the intraocular pressure and whether signs of inflammation or infection were present. The soft contact lens was
replaced if lost. The intraocular pressure could be estimated
only by digital palpation of the globe. Other outcome measures that were assessed included retention of the prosthesis,
best-corrected visual acuity over time, the need for surgical
revision, and the presence of complications such as retroprosthesis membrane formation, infection, inflammation, or corneal melting. Cumulative vision-year functional success was
defined as the number of years with visual acuity of 20/200 or
better.
● STATISTICAL ANALYSIS: A Kaplan-Meier survival
analysis was used to evaluate the visual outcome of the
procedure. The influence of duration of disease, number of
previous interventions, type of KPro implanted, age at
surgery, and presence of glaucoma on the final visual
outcome was evaluated using the log-rank test.11 Because
of the small sample size, the P value of the two-sided
log-rank test was obtained based on the permutation
procedure.12 All analyses were conducted using Microsoft
Excel for Windows software version 2003 (Microsoft Corp,
Seattle, Washington, USA), with the level of significance
taken as P ⬍ .05.
● SURGICAL PROCEDURE: The surgery was performed
under general anesthesia, unless contraindicated because
of the medical status of the patient. The host corneal bed
was prepared for trephination, which included releasing
any symblepharon or forniceal shortening that was present.
Depending on the size of the cornea, an 8.0- or 8.5-mm
corneal trephine was used. The KPro was assembled as
previously described.7,10 Thus, briefly, the donor corneal
carrier used was oversized by 0.5-mm compared with the
recipient bed. The optic cylinder of the anterior plate was
inserted through a 3-mm trephined central opening in the
corneal button. The posterior plate of the prosthesis was
screwed into firm apposition with the donor tissue, and,
since 2003, a titanium locking ring has been applied to
secure it in place.
Phakic eyes underwent extracapsular lens extraction and
were left aphakic. Preexisting intraocular lenses were left
in place if they were well positioned and fixated. Peripheral
iridectomy was performed routinely. The donor corneal
tissue with the assembled KPro was sutured in place with
16 10-0 or 9-0 nylon sutures. Finally, a Kontur soft contact
lens (Kontur Contact Lens Co, Hercules, California,
USA) 16.0 mm in diameter and with a 9.8-mm base curve
was inserted over the type I KPro. In the type II counterpart, with the protruding anterior cylinder, an opening in
the lid was created, usually during surgery, and the skin was
tightened around the nub (Figure 1). Antibiotic drops and
ointment were applied.
The postoperative medication consisted of topical vancomycin 14 mg/ml, fluoroquinolone 0.5%, and prednisolone
440
AMERICAN JOURNAL
RESULTS
SIXTEEN EYES OF 15 PATIENTS WITH SJS WHO UNDERWENT
KPro surgery were studied. All but one patient had a severe
form of SJS. One patient had the diagnosis of toxic
epidermal necrolysis. Four patients were male and 11 were
female. The mean age was 50 ⫾ 18 years (range, 23 to 74
years), and the mean duration of their disease was 10 ⫾ 6.6
years (range, two to 20 years). Nine eyes (56%) had
previous failed penetrating keratoplasties, with four having
three or more. Three had undergone limbal cell transplantation and one had undergone mucous membrane grafting.
All eyes had a preoperative visual acuity of counting
fingers at 1 foot or worse in the operated eye as well as in
the fellow eye. The mean follow-up period was 3.6 ⫾ 1.5
years (range, 10.2 months to 5.6 years).
Ten eyes (63%) underwent type II KPro surgery,
whereas six eyes (38%) underwent type I KPro surgery.
Seven eyes (44%) had lens extraction at the time of
surgery and were left aphakic. Six eyes (38%) underwent
anterior vitrectomy to remove vitreous in the anterior
chamber that complicated previous cataract surgery.
Ahmed shunts were placed in all 16 eyes (11 during KPro
surgery, four before surgery, and one after surgery). In two
cases, additional shunts had to be added later.
OF
OPHTHALMOLOGY
MARCH 2008
TABLE. Distribution of Eyes with the Various Visual Acuities
before and after Implantation of a Keratoprosthesis and at
the Last Postoperative Follow-up Visit
No. of Eyes
Visual Acuity
20/15
20/20
20/25
20/30
20/40
20/70
20/80
20/100
20/200
20/400
CF @ 1 foot
CF @ 6 feet
HM
LP
NLP
Preoperative
Best Visual Acuity
Achieved
Visual Acuity at
Last Examination
2
4
1
2
1
1
1
2
2
1
8
5
2
1
1
1
1
1
FIGURE 2. Kaplan-Meier analysis of eyes with SJS implanted
with a KPro. Primary outcomes were loss of 20/200 vision or
better (blue), loss of any vision better than preoperative (red),
and loss of light perception (shaded area). The survival curve
from Yaghouti and associates is reproduced for comparison
(dotted gray line).6 Compared with eyes operated on between
1990 and 1997, a longer conservation of visual acuity was
observed in this study. This can be attributed to advances in the
KPro design, surgical technique, and management of postoperative complications, and particularly the long-term use of
broad-spectrum antibiotics and a large diameter soft contact
lens. NLP ⴝ no light perception.
2
2
1
2
4
CF ⫽ counting fingers; HM ⫽ hand movements; LP ⫽ light
perception; NLP ⫽ no light perception.
● VISUAL OUTCOME:
Twelve eyes (75%) achieved a
visual acuity of 20/200 or better (which was the equivalent
of at least two lines of improvement in visual acuity) after
surgery, with eight eyes (50%) achieving excellent vision
of 20/40 or better (Table). Visual acuity in the 12 eyes was
maintained at 20/200 or better over a mean period of 2.5 ⫾
2.0 years after surgery (range, six months to 5.4 years). Four
eyes did not achieve a visual acuity of 20/200 or better after
KPro surgery, despite an uneventful operative and postoperative course. These patients had preexisting ocular disease that was unrecognized until after the KPro surgery.
The conditions included end-stage glaucoma in two eyes of
the same patient, severe chronic uveitis in one eye, and
severe, scarred age-related macular degeneration in one
eye. The eye with macular degeneration retained good
peripheral vision.
At the last follow-up visit, seven eyes (44%) retained
good to excellent vision of 20/70 or better. Five eyes that
had initial improvement in visual acuity subsequently
deteriorated to less than 20/200 vision because of the
following complications: corneal melting and leak in one
eye, retinal detachment in two eyes, end-stage glaucoma in
one eye, and vitreous hemorrhage in one eye. The latter
four eyes (25%) subsequently lost all perception to light.
In terms of cumulative vision-years, 12 eyes (75%) that
were able to reach a visual acuity of 20/200 or better were
followed up for a total cumulative duration of 45.5 years.
These eyes maintained a visual acuity of 20/200 or better
over a period of 30 years cumulatively.
VOL. 145, NO. 3
BOSTON KPRO
IN
Kaplan-Meier survival analysis was used to evaluate the
visual outcome of the procedure, with the primary end
points being vision of 20/200 or worse, vision worse than
the preoperative visual acuity, or loss of light perception.
Patients who died, were lost to follow-up, or had no visual
loss at the completion of the study period were treated as
censored events (Figure 2).
However, the small sample size may reduce the quality
of the Kaplan-Meier graph. Furthermore, it does not
represent the individual changes in visual acuity until the
full outcome is reached, and no inferences can be made on
the learning curve of the surgeon. Therefore, the individual patient functional outcomes also are illustrated as
color-coded bars, starting with the year surgery was performed (Figure 3).
● COMPLICATIONS:
There were no cases of endophthalmitis, but in one patient, severe sterile inflammation
developed immediately after the operation, causing inoperable retinal detachment. Complications that developed
after KPro implantation included skin retraction, the
formation of a retroprosthesis membrane, and worsening of
glaucoma.
● TISSUE MELT AND LEAK: There were no cases of spontaneous extrusion of the implant. However, in four eyes
(25%), aqueous leakage developed that was managed by
replacement of the KPro in two type II KPro eyes and
STEVENS-JOHNSON SYNDROME
441
FIGURE 3. Graph demonstrating the visual acuity in the different eyes with advanced SJS over time, after implantation of a KPro.
This is a clearer and more comprehensive way of showing visual attrition in patients. Each line represents the individual eyes’ visual
acuity over their follow-up times. The major complications and surgical interventions are noted and a correlation with visual acuity
can be made.
placement of the type II KPro, which prevented further
recurrences.
replacement with a standard corneal graft in two other
eyes. In the latter two cases, one patient had little vision
because of underlying nerve damage secondary to glaucoma as well as chronic uveitis, and penetrating keratoplasty was performed to maintain light perception and for
tectonic reasons. The other patient had severe leak and
fistula formation along with choroidal hemorrhage that led
to light perception visual acuity.
● RETROPROSTHESIS MEMBRANES: Retroprosthesis membranes developed in nine eyes (56%), with subsequent
recurrences in three eyes. The average time for the
formation of the first retroprosthesis membrane was 7.8 ⫾
10.7 months after KPro implantation. Of the 12 retroprosthesis membranes, eight were managed easily by neodymium:
yytrium–aluminum–garnet (Nd:YAG) laser membranotomy
(one session was enough except for one case that needed
repeat YAG), three were removed by pars plana vitrectomy, and a last membrane was left untreated for the eye
had lost vision because of a vascular event.
● SKIN RETRACTION AROUND TYPE II KERATOPROS-
In four of the 10 eyes (40%) that underwent type
II KPro surgery, skin retraction developed around the stem
of the KPro. Skin retraction was first noted an average of
18.6 ⫾ 10.0 months (range, 10 to 33 months) after surgery.
These were managed by skin revision and wound closure at
the site of retraction. Two eyes underwent two skin
revisions and one eye required a third skin revision. Skin
retraction continued to develop in the last eye despite two
skin revisions, and subsequently the eye underwent re-
THESIS:
442
AMERICAN JOURNAL
● GLAUCOMA:
Twelve eyes (75%) had preexisting
glaucoma before KPro surgery, reflecting the severity of
anterior segment damage arising from SJS. Of these, six
(50%) had previous failed corneal transplants, and three
OF
OPHTHALMOLOGY
MARCH 2008
had undergone limbal cell transplantation as well. Four
eyes (25%) had glaucoma shunt procedures performed
before surgery, 11 eyes (69%) had shunt implantation at
the time of KPro surgery, and one eye (6%) had shunt
implantation after surgery. Another two shunts were
added after surgery because of failure of pressure control.
In three eyes (19%), glaucoma developed after the
operation. Of these cases of glaucoma, none were
controlled by medication alone, and shunt surgery was
required for medically uncontrolled glaucoma. Endstage glaucoma was a major cause for visual loss in
operated eyes, resulting in one eye failing to achieve
20/200 vision and another two eyes losing best-corrected visual acuity after surgery.
● PROGNOSTIC FACTORS FOR VISUAL LOSS:
We looked
at various possible risk factors for visual loss. The 12
patients with preoperative glaucoma had an average of 1.3
years of good vision (20/200 vision or better). This is in
comparison with an average of 3.6 years in the four
patients with no preoperative glaucoma. Comparing the
survival curves of patients with and without preoperative
glaucoma using the log-rank test results in a P value of
.039, suggesting that the duration of vision is higher for
patients without preoperative glaucoma.
Having had previous surgical interventions (e.g., penetrating keratoplasty, mucous membrane grafts, or limbal transplantations), implantation of a type I KPro (vs a type II
KPro), long duration of the disease (more than 10 years),
and advanced age of the patient (more than 50 years),
were found not to be significantly correlated with a longer
duration of good vision (P ⬎ .05). Caution should be
exerted in interpreting this data, given the small size of the
patient population.
DISCUSSION
STEVENS-JOHNSON SYNDROME IS A RARE AUTOIMMUNE
disease which may lead to significant corneal changes and
conjunctival scarring. It remains the worst prognostic
category for KPro surgery. The prolonged severe inflammation associated with the disease makes the tissue around
the KPro vulnerable to necrosis and melting, which can
result in leakage, choroidals, and retinal detachment.
Infection, in the form of endophthalmitis, also was a major
concern in previous years.13 Yaghouti and associates previously reported the use of the Boston KPro in a series of
patients that included seven eyes with SJS. None of these
seven eyes retained 20/200 vision five years after
implantation.6
The results of the current study, however, are significantly better than the previous results, as demonstrated in
the Kaplan-Meier survival analysis (Figure 2). Many eyes
maintained an improvement over their preoperative vision
at the last follow-up. This can be attributed largely to
VOL. 145, NO. 3
BOSTON KPRO
IN
improvements to the KPro design, the surgical technique,
and postoperative management, which have helped to
improve significantly the visual outcome and to reduce the
incidence of complications. These advances include a
modification of the design to include perforation holes
in the back plate to facilitate the nutrition and hydration
of the corneal graft, which has reduced the incidence of
corneal melting and aqueous leak.10 The recent addition of
broad-spectrum topical antibiotics and the long-term use
of a large-diameter therapeutic soft contact lens (in type I
keratoprostheses) also have drastically reduced the risk of
corneal melting, extrusion of the implant and endophthalmitis9 (Dohlman CH, et al. IOVS 2003;44:ARVO
E-Abstract 1455).
The virtual elimination of postoperative endophthalmitis
has been particularly important. The long-term prophylactic
use of vancomycin eye drops combined with a third- or
fourth-generation fluoroquinolone has had a dramatic effect. Nouri and associates previously reported that in 12%
of patients undergoing KPro surgery in the 1990s (without
vancomycin), endophthalmitis developed, with SJS patients having the highest incidence.13 In our more recent
study in which all patients received vancomycin drops
prophylactically, usually twice daily, in none of the 16 eyes
did acute bacterial endophthalmitis develop, demonstrating the dramatic effect of this new antimicrobial medication regimen in preventing this devastating complication.
Another important problem related to KPro surgery is
glaucoma.6,14,15 Preexisting glaucoma may worsen after
surgery, especially in eyes that are inflamed, as is the case
for SJS. Eyes that did not have a history of glaucoma also
may develop glaucoma, probably because of the inflammation and progressive angle closure from peripheral anterior
synechiae. In our study, we found that preexisting glaucoma was found to be a significant risk factor for eventual
visual loss, which makes control of this condition particularly critical for good long-term visual outcome. Furthermore, the inability to obtain an accurate intraocular
pressure measurement in the presence of a KPro makes
assessment of the glaucoma treatment particularly difficult.14 Optic disk evaluation and sequential visual field
assessment therefore are important in the assessment of
glaucoma control. Whereas antiglaucoma eye drops are
effective for type I implants, oral medications may be the
only effective treatment for type II. In either case, optimization of the intraocular pressure control often is achieved
with glaucoma shunt surgery, such as implantation of an
Ahmed valve shunt.14 However, these chronically inflamed eyes are at a particularly high risk of failure of any
glaucoma filtration surgery, and intraocular pressure control often requires both surgical and medical measures.
Novel glaucoma shunt procedures to drain aqueous to
extraorbital locations may be useful in the glaucoma
management of these patients (Dohlman CH, Grosskreutz
CL, Dudenhoefer EJ, Nouri M, Rubin PAD, unpublished
data, “Connecting Ahmed valve shunt to the lacrimal sac
STEVENS-JOHNSON SYNDROME
443
or nasal sinuses in severe glaucoma,” presented as a poster
at the American Academy of Ophthalmology meeting,
October 2002).
Retroprosthesis membrane formation is a common complication that is associated with all forms of KPro surgery.10,16 These membranes are postulated to form from
inflammation induced by the release of allogeneic proteins
in the anterior chamber and vary from condensations that
coat the posterior surface of the optic to dense vascularized
sheets of fibrotic tissue. In most of our cases, these were
easily managed with Nd:YAG laser membranotomy. Only
a few recurrent membranes that contained vessels or were
too thick for laser treatment required surgical removal by
pars plana vitrectomy.
In this study, we demonstrated that modifications to
the design as well as the postoperative management of
patients implanted with the Boston KPro have to
significant improvements in the functional and anatomic outcome of this procedure over the situation
existing just a decade ago. Operated eyes demonstrated
a longer conservation of visual acuity compared with
previous reports, and there was a dramatic decrease in
serious sight-threatening complications, such as endophthalmitis. Still results are much inferior to those in
patients with nonautoimmune diseases, and expectations must be adjusted accordingly. Patients with SJS
who elect to have the procedure must be prepared for a
life-long follow-up with an experienced KPro surgeon.
THIS STUDY WAS SUPPORTED BY AN ALCON RESEARCH AWARD, FORT WORTH, TEXAS. THE AUTHORS INDICATE NO
financial conflict of interest. Involved in design and conduct of study (R.R.S., L.P.K.A., C.H.D.); all of the authors were involved in the collection,
management, analysis, and interpretation of the data, and preparation, review, and approval of the manuscript. This retrospective chart review was
approved by the Institutional Review Board of the Massachusetts Eye and Ear Infirmary.
We thank Tianxi Cai, ScD, Harvard School of Public Health, Boston, Massachusetts, for her help with the statistical analysis. Drs Rony R. Sayegh
and Leonard P. K. Ang contributed equally to this article.
9. Dohlman CH, Dudenhoefer EJ, Khan BF, Morneault S.
Protection of the ocular surface after keratoprosthesis
surgery: the role of soft contact lenses. CLAO J 2002;28:
72–74.
10. Dohlman CH, Nouri M. Keratoprosthesis surgery. In: Foster
CS, Azar DT, Dohlman CH, editors. Smolin and Thoft’s the
cornea, 4th ed. Philadelphia, Pennsylvania: Lippincott Williams & Wilkins, 2005:1085–1095.
11. Harrington DP, Fleming TR. A class of rank test procedures
for censored survival data. Biometrika 1982;69:553–566.
12. Efron B, Tibshirani R. An introduction to the bootstrap.
London, England; New York, New York: Chapman & Hall
Ltd, 1993:202–219.
13. Nouri M, Terada H, Alfonso EC, Foster CS, Durand ML,
Dohlman CH. Endophthalmitis after keratoprosthesis: incidence, bacterial causes, and risk factors. Arch Ophthalmol
2001;119:484 – 489.
14. Netland PA, Terada H, Dohlman CH. Glaucoma associated
with keratoprosthesis. Ophthalmology 1998;105:751–757.
15. Aquavella JV, Rao GN, Brown AC, Harris JK. Keratoprosthesis. Results, complications, and management. Ophthalmology 1982;89:655– 660.
16. Hicks CR, Hamilton S. Retroprosthetic membranes in
AlphaCor patients: risk factors and prevention. Cornea
2005;24:692– 698.
REFERENCES
1. Foster CS. Immunologic disorders of the conjunctiva, cornea, and sclera. In: Albert DM, Jakobiec FA, editors.
Principles and practice of ophthalmology, 2nd ed. Philadelphia, Pennsylvania: WB Saunders, 1999:821– 823.
2. Tugal-Tutkun I, Akova YA, Foster CS. Penetrating keratoplasty in cicatrizing conjunctival diseases. Ophthalmology
1995;102:576 –585.
3. Solomon A, Ellies P, Anderson DF, et al. Long-term outcome of keratolimbal allograft with or without penetrating
keratoplasty for total limbal stem cell deficiency. Ophthalmology 2002;109:1159 –1166.
4. Dohlman C, Doane M. Keratoprosthesis in end-stage dry eye.
Adv Exp Med Biol 1994;350:561–564.
5. Dohlman CH, Terada H. Keratoprosthesis in pemphigoid
and Stevens-Johnson syndrome. Adv Exp Med Biol 1998;
438:1021–1025.
6. Yaghouti F, Nouri M, Abad JC, Power WJ, Doane MG,
Dohlman CH. Keratoprosthesis: preoperative prognostic categories. Cornea 2001;20:19 –23.
7. Doane MG, Dohlman CH, Bearse G. Fabrication of a
keratoprosthesis. Cornea 1996;15:179 –184.
8. Dudenhoefer EJ, Nouri M, Gipson IK, et al. Histopathology
of explanted collar button keratoprostheses: a clinicopathologic correlation. Cornea 2003;22:424 – 428.
444
AMERICAN JOURNAL
OF
OPHTHALMOLOGY
MARCH 2008