1. International Outcomes of the Boston Type I Keratoprosthesis in Stevens-Johnson Syndrome
April 15-17, 2015
World Cornea Congress VII
San Diego, California
Jamie K. Alexander, M.D.
Samar K. Basak, M.D.
Ma. Dominga B. Padilla, M.D.
Fei Yu, Ph.D.
Anthony J. Aldave, M.D.
The authors have no relevant financial interests to disclose.

2. Purpose
Stevens-Johnson syndrome (SJS) can devastate the ocular surface. There is currently no reliable, sustainable method for restoring sight in these patients.
While initial studies of the Boston type I keratoprosthesis (KPro) indicated that its results were poorest in patients with SJS, advances in KPro design and post-operative management have improved outcomes, including those in SJS.
Using international data and the world’s most widely used keratoprosthesis, we present the largest study to date of the KPro in SJS. We compare the device’s outcomes in SJS to those of patients without SJS and seek to determine the factors influencing outcomes of the KPro in SJS.

3. Methods Retrospective, interventional, comparative case series
Study approval: Institutional Review Board at UCLA Ethics Committee at Disha Eye Hospitals Institutional Ethics Review Committee at St. Luke’s Medical Center
Data was collected from 3 international academic medical centers on all patients undergoing KPro surgery, including preoperative characteristics, intraoperative data, and postoperative outcomes:
Statistical Analysis: SAS software version 9.4 (SAS, Inc., Cary, NC) (p=0.05)
Stein Eye Institute Los Angeles, California
May 2004 – Jul 2013
Disha Eye Hospitals
Kolkata, India
Mar 2009 – Jan 2014
St. Luke’s Medical Center Global City
Manila, Philippines
Mar 2009 – Jan 2014

4. Results: Preoperative Characteristics
SJS
Non-SJS
P value
No. of procedures 40
194
No. of eyes 27
182
No. of patients 26
175
Gender (female; male)
14; 12
75; 100
0.3
Mean age (range), yrs
50 (26-72)
60 (3-95)
<0.001 
Follow-up (mos)
Mean (median)
17.6 (12.5)
29.3 (23.5)
 0.001
Standard deviation (range)
16.2 ( )
22.8 ( )
Indication for keratoprosthesis
Failed corneal transplant --
109 (56%)
Stevens-Johnson syndrome
27 (68%)
Limbal stem cell deficiency
24 (12%)
Chemical injury
18 (9%)
Repeat keratoprosthesis
13 (33%)
15 (8%)
<0.001
Corneal vascularization
11 (6%)
Mucous membrane pemphigoid
4 (2%)
Aniridia
Thermal Burn
3 (2%)
AKC
2 (1%)
Other
Glaucoma
7 (26%)
130 (71%)
Previous glaucoma surgery
4 (15%)
45 (25%)
0.334
Previous corneal transplant(s)
18 (67%)
42 (23%)
<0.001* 1
5 (19%)
32 (18%) 2
2 (7%)
54 (30%) 3
33 (18.1%)
> 4
2 (7.4%)
21 (11.5%)
40 (27, 26), 194 (182, 175)
*Comparing <1 to >2 previous corneal transplants

5. Results: CDVA for Most Recent KPro
Baseline
6 Months
1 Year
2 Years
3 Years
4 Years
5 Years
At last follow-up
No. of eyes
SJS 27 23 18 9 3 1 22
Non-SJS
182
169
142 99 68 43 25
162
20/20-20/50
0 (0%)
14 (61%)
7 (39%)
5 (56%)
1 (33%)
1 (100%) --
9 (41%)
1 (1%)
53 (31%)
45 (32%)
31 (31%)
21 (31%)
15 (35%)
9 (36%)
49 (30%)
20/60-20/100
5 (22%)
3 (33%)
2 (67%)
8 (36%)
49 (29%)
34 (24%)
18 (18%)
9 (13%)
5 (12%)
2 (8%)
33 (20%)
20/125-20/200
3 (13%)
4 (22%)
1 (11%)
4 (18%)
9 (5%)
18 (11%)
16 (11%)
13 (19%)
4 (16%)
20 (12%)
20/250-20/400
7 (4%)
6 (4%)
4 (4%)
3 (4%)
4 (9%)
12 (7%) CF
8 (30%)
53 (29%)
27 (16%)
21 (15%)
11 (11%)
6 (9%)
3 (7%)
19 (12%) HM
9 (33%)
1 (5%)
61 (34%)
8 (6%)
12 (12%)
10 (15%)
9 (21%)
1 (4%)
11 (7%) LP
10 (37%)
51 (28%)
4 (2%)
7 (5%)
2 (2%)
1 (2%)
9 (6%)
NLP
2 (1%)
5 (4%)
3 (3%)
>20/200
22 (96%)
18 (100%)
9 (100%)
3 (100%)
21 (96%)
10 (6%)
120 (71%)
95 (66.9%)
67 (68%)
43 (63%)
25 (58%)
15 (60%)
102 (63%)
p-value
0.367
0.01
0.002
0.055
0.547
0.001

6. Results: Complications and Secondary Surgical Procedures
SJS
Non-SJS
p-value
No. of eyes with >1 mo of follow-up 27
181
Retroprosthetic membrane
10 (37%)
89 (49%)
0.30
YAG laser membranotomy
9 (33%)
68 (38%)
0.83
Surgical membranectomy
1 (4%)
13 (7%) 1
Sterile corneal stromal necrosis
16 (59%)
25 (8%)
<0.001
Keratoprosthesis replacement
8 (4%)
Elevated IOP (>25 mmHg)
5 (189%)
32 (18%)
Glaucoma surgery
11 (6%)
Corneal infiltrate
8 (30%)
18 (10%)
0.009
Persistent epithelial defect
44 (24%)
Tarsorrhaphy
14 (52%)
22 (12%)
Infectious endophthalmitis
4 (2%)
Vitreous tap and injections
Retinal detachment
3 (11%)
21 (12%)
Repair of retinal detachment
14 (8%)
0.70
Sterile vitreitis
15 (8%)
0.23
Vitreous tap and injection
9 (5%)
0.61
Cystoid macular edema
2 (7%)
23 (13%)
0.75
Intravitreal injection
17 (9%)
0.48

7. Results: Keratoprosthesis Retention Rate
SJS
Non-SJS
p-value
Mean follow-up of all procedures in mos (range)*
17.6 ( )
29.3 ( )
 0.001  
Retention failure
Keratoprostheses removed (eyes)
14 of 27 (51.9%)
28 of 182 (15.4%)
<0.001
Keratoprostheses removed (procedures)
18 of 40 (45.0%)
32 of 194 (16.5%)
Retention failure rate (procedures)**
18 per eye-years, 0.306/eye-year
32 per eye-years, 0.068/eye-year
<0.001 
Time to failure in mos (mean, range)*
14.1 ( )
19.9 ( )
0.011 
Retention success
Keratoprostheses retained (eyes)
22 of 27 (81.5%)
162 of 182 (89.0%)
0.335
Keratoprostheses retained (procedures)
22 of 40 (55.0%)
162 of 194 (83.5%)
Mean follow-up of retained procedures in mos (range)*
20.5 ( )
31.1 ( )
 0.084
*Kruskall-Wallis test
** Log-rank test

8. Results: Kaplan-Meier Survival of Last KPro

9. Keratoprostheses Retained Keratoprostheses Removed
Results: Characteristics of Patients with SJS Associated with KPro Retention and Failure
Keratoprostheses Retained
Keratoprostheses Removed
p-value
Number (eyes; procedures)
13; 22
14; 18
Pre-operative characteristics (eyes)
Time from SJS to initial procedure (yrs)
13.72
12.02
0.91
Perforation at or before time of KPro
4 (31%)
7 (50%)
0.44
Bulbar conjunctival keratinization
1 (8%)
6 (43%)
0.08
Lid malpositioning*
3 (23%)
4 (29%) 1
Peri-operative characteristics (procedures)
Tarsorrhaphy at time of keratoprosthesis
7 (32%)
5 (28%)
Preoperative immunosuppression
3 (13%)
2 (11%)
Postoperative immunosuppression
13 (59%)
8 (44%)
0.53
Indication for removal (procedure)
Sterile corneal necrosis - 14
Infectious corneal necrosis** 4
*requiring previous, concurrent or subsequent oculoplastics procedures
**organisms including: mixed (1), Candida (2), budding yeast without speciation (1) 

10. Results: Repeat KPro Results in SJS (example)
Slit lamp photomicrograph of a 67 year-old man with SJS. Donor corneal necrosis developed 5 months following KPro surgery, necessitating replacement of the KPro. At the time of the second surgery, the donor cornea was covered with mobilized bulbar conjunctiva and extensive medial and lateral tarsorrhaphies were performed. Two years later, the donor cornea remains covered by conjunctiva and the patient maintains 20/20 CDVA.
Results: SJS Patients in which KPro Not Retained
Pre-op CDVA
# Kpros
Time Kpro retained (mos)
Procedure(s) Performed Following Removal of Last Type I Kpro
Follow-up After Removal of Last Type I KPro (mos)
Current CDVA HM 2
2.2
Boston type II Keratoprosthesis
99.7
20/40 LP 1 6
Penetrating keratoplasty
4.9
5.8
Penetrating keratoplasty (x 2) 42
7.7
(repeat keratoprosthesis planned)
1.1 CF
5.7

11. Conclusions
The Boston type I keratoprosthesis is the most commonly used device of its kind world-wide. This study retrospectively analyzed the international outcomes of 40 KPro procedures performed in 27 eyes of 26 patients with SJS in comparison to those of 194 procedures performed in 182 eyes of 175 patients without SJS.
Demographics
Patients undergoing KPro for SJS are more likely to be younger and less likely to have a preoperative diagnosis of glaucoma or a history of multiple previous corneal transplants compared to patients undergoing KPro for other indications. KPro procedures performed in SJS patients are significantly more likely to be performed as repeat implantations of the device (33% vs. 8%, p<0.001).
Visual Outcomes
Nearly all eyes undergoing KPro with SJS have a final CDVA > 20/200 when the KPro is retained (96%).
- This is significantly higher than that of eyes without SJS (63%, p=0.001)
- This is likely due a significantly lower incidence of preoperative glaucoma.
40 (27, 26), 194 (182, 175)

12. Conclusions - No sight-threatening complications are more common.
Several post-operative complications are more common in SJS patients following KPro: sterile corneal stromal necrosis, persistent epithelial defect, and corneal infiltrate.
- No sight-threatening complications are more common.
- No endophthalmitis was encountered in SJS patients in this study.
Retention
Although retention failure is higher in SJS (relative risk 4.5), eyes with SJS are no less likely to ultimately retain KPro following repeat implantation (82 vs. 89%, p=0.34).
Characteristics Affecting KPro Failure/Success in SJS
Based on available data, there are no identifiable pre- or postoperative characteristics that lead to KPro failure or success in SJS.
- This may be due to the heterogeneous nature of the manifestations of SJS.
The Boston type I keratoprosthesis is an effective means to restore vision to patients blinded by Stevens-Johnson syndrome.