May 20, 2019
6 min read

Gauge expected outcomes before conducting penetrating keratoplasty

In some cases, a primary keratoprosthesis may be the best choice.

You've successfully added to your alerts. You will receive an email when new content is published.

Click Here to Manage Email Alerts

We were unable to process your request. Please try again later. If you continue to have this issue please contact

Welcome to another edition of CEDARS/ASPENS Debates. CEDARS/ASPENS is a joint society of cornea, cataract and refractive surgery specialists, here to discuss some of the latest hot topics in ophthalmology.

Managing the cornea and rehabilitating vision in patients with high-risk corneas is always a challenge. These corneas have a high likelihood of graft failure, yet these patients are often desperate to regain lost vision.

This month,Gregg J. Berdy, MD, FACS, and Christopher E. Starr, MD, discuss the use of transplantation for these patients, as well as the use of the keratoprosthesis. We hope you enjoy the discussion.

Kenneth A. Beckman, MD, FACS

Consider PK before keratoprosthesis

Gregg J. Berdy, MD, FACS
Gregg J. Berdy

The Boston keratoprosthesis was developed to be used in situations in which the cornea and ocular surface are compromised, a corneal transplant would fail or the outcomes would lead to unwanted effects, such as a nonhealing abrasion or melt. Typically, a keratoprosthesis is used in late-stage disease or when there is no other alternative. This is only about 2% of what we see, but it is a special group that needs our attention.

A keratoprosthesis should be considered only if a corneal transplant would not result in good vision and good epithelial healing. These situations may include previous multiple corneal transplants, penetrating keratoplasty failures/rejections, vascularization of more than three-fourths of the cornea, Stevens-Johnson syndrome, toxic epidermal necrolysis syndrome, cicatricial pemphigoid, marked scarring, symblepharon and ankyloblepharon. If the patient has the potential to see and a PK will not work, then one could consider utilizing a keratoprosthesis.

Keratoprostheses (KPros) have gotten much better over the past decade, but they can still fail. Infections and leakage can occur, and membranes and fibrosis may develop behind the prosthesis. Additionally, if these complications occur, a repeat KPro surgery may be required.

Another drawback to using a KPro is that it requires protection. We always place a large bandage contact lens over the prosthesis to protect the surface. If a bandage contact lens cannot fit or remain in the eye or if the eyelids will not close, then it would not be a good idea to utilize a keratoprosthesis. Additionally, it is difficult to check IOP with a keratoprosthesis, but with a corneal transplant it is possible to do so. Finally, reimbursement to a surgical center may sometimes be an issue with KPro.

The other alternative for improving vision is a full-thickness PK. If I can improve the existing problems that would lead me to even consider using a keratoprosthesis, then I will perform a keratoplasty because in our hands we are more facile about doing this surgery, the vision is better and we do not need to place a plastic prosthesis in the eye.


In a patient who has end-stage limbal stem cell deficiency, it may not be a good idea to perform a corneal transplant, and a keratoprosthesis could be a viable alternative. But why not consider a keratolimbal allograft (KLAL) to help re-epithelialize the cornea? If the corneal epithelium is restored to a normal functioning surface but the cornea remains cloudy or edematous, then one could perform a PK, which would function better than the “end-game” keratoprosthesis in the original scenario.

In patients who have Stevens-Johnson syndrome, toxic epidermal necrolysis or bullous pemphigoid, one can attempt to rebuild the ocular surface with a mucous membrane graft with either buccal or nasal mucosa; that could work well enough to allow a PK to succeed. Or, one can perform a KLAL and then follow with a PK if necessary. Granted, these options involve multiple surgeries, but re-establishing the ocular surface to close to normal and performing a PK may be preferable to putting a piece of plastic in the eye.

I particularly like corneal transplants if I can rebuild the surface and address the ocular surface disease that usually exists in these cases. It is important to manage the aqueous component, the meibomian gland dysfunction and the mucous goblet cell deficiency. If you can fix the surface, then you can proceed with a more traditional PK surgery.

It is important to note that while we just use steroid eye drops for immune suppression with corneal transplants, a KLAL is treated like a solid tissue transplant. Patients must be on oral immunosuppressants for up to 2 years, white blood cell counts must be followed, and other specialists may need to be involved.

Disclosure: Berdy reports no relevant financial disclosures.

KPro can be first corneal procedure

Christopher E. Starr, MD
Christopher E. Starr

Until 10 to 15 years ago, the indications for Boston keratoprosthesis type 1 were appropriately more selective and restrictive. The patient had to have two or more failed corneal transplants, a poor prognosis for repeat transplant, limbal stem cell failure with resultant corneal opacification, a healthy tear film, normal eyelids without exposure and a visual acuity of 20/200 or worse in both eyes. In addition, keratoprosthesis was strictly performed in only one eye.

Over the years, thanks to the efforts of Claes H. Dohlman, MD, PhD, and Massachusetts Eye and Ear Keratoprosthesis Service, the procedure and outcomes have improved so dramatically that these indications have eased somewhat.


We now have threadless snap-on designs. Thinner titanium back plates are available in addition to PMMA in various sizes; the titanium reduces the risk for formation of retroprosthetic membrane. More holes in the back plate allow for better nutrition, better corneal survival and reduction in corneal melts. A major advance was the addition of vancomycin as a daily topical antibiotic, which has dramatically reduced the rate of gram-positive endophthalmitis.

We have better contact lenses and concomitant retinal procedures. We have much improved glaucoma procedures and better ways of monitoring patients after keratoprosthesis in terms of assessing IOP, optic nerve changes and damage via OCT and other objective measures.

a patient with a high risk for PK failure who underwent KPro
One-day postoperative image from a patient with a high risk for PK failure who underwent KPro, which resulted in 20/20 uncorrected visual acuity.

Source: Christopher E. Starr, MD

The Boston KPro type 1 can be the first corneal procedure in patients who are at high risk for transplant failure. It is costly, risky and not in the patient’s best interest to do multiple high-risk transplants, wait for them to inevitably fail and then do a keratoprosthesis. Multiple prior intraocular surgeries may reduce the success rate and increase complications with KPro surgery. In select patients with incomplete lid closure or exposure, one can combine a type 1 keratoprosthesis with a lid procedure such as a tarsorrhaphy to reduce exposure and the risk for corneal melting.

Another shift in modern indications is that these procedures can be done bilaterally if the first eye is successful, and they can also be done unilaterally in people with reasonable or good vision in the other eye. Both of these scenarios can improve depth perception, improve or restore binocularity and ocular fusion, and greatly increase patients’ quality of life.

Patients who would do well with traditional keratoplasty (eg, penetrating, deep anterior lamellar or endothelial), such as those with Fuchs’ dystrophy, keratoectasia, stromal dystrophies or pseudophakic bullous keratopathy, should still be considered for those procedures first. However, patients with severe corneal neovascularization with deep vessels, multiple prior intraocular procedures, glaucoma or elevated IOP, autoimmune conditions, chemical burns, severe limbal stem cell deficiency, corneal keloid, some pediatric corneal opacities, neurotrophic corneas from herpes infection, previous rejection episodes or severe ocular surface disease are at high risk for PK failure or rejection. For some of these patients, primary KPro may be a more cost-effective, long-term option and may provide faster visual rehabilitation and reduced risks from multiple complex surgeries.


Complications, of course, can occur postoperatively in KPro patients, and close lifetime follow-up is required, typically every 3 months. The most common complication is retroprosthetic membrane, which occurs in about 40% of cases. However, when detected early, this is easily treated in the office with Nd:YAG laser. Endophthalmitis used to be fairly common, but the incidence is down to less than 3% thanks to the use of daily topical anti-infectives and better surveillance. Glaucoma is the most common cause of irreversible vision loss after KPro, and these patients often receive concomitant glaucoma shunts at the initial surgery and are closely monitored postoperatively.

In my opinion, perhaps the most important risk factor for complications after KPro is poor patient compliance (eg, with follow-up visits, daily medication regimen, awareness of early symptoms), and this should always be carefully assessed before committing to KPro surgery. Visual potential should always be assessed preoperatively as well, and if in doubt, a pars plana staging endoscopy can be done to directly visualize and assess the health of the optic nerve, macula and retina.

In the modern era, Boston KPro indications have become less restrictive, and it is no longer considered a “last resort, if all else fails” surgery. I have performed KPros as the first corneal procedures as well as bilateral KPros in carefully selected appropriate patients. These have proven to be among the most rewarding surgeries and most grateful patients in my practice.

Disclosure: Starr reports no relevant financial disclosures.