CASE PRESENTATION
A 35-year-old woman presented to our practice expressing interest in refractive surgery due to contact lens intolerance. With a refraction of -4.50 sphere, BCVA in her right eye was 1.5-. Her left eye was also correctable to 1.5- with a refraction of -3.50 -0.25 X 5. Pachymetry in each eye was 520 µm.
Preoperative topography with the Pentacam (Oculus Optikgeräte, Wetzlar, Germany) was within normal ranges in both eyes but indicated the presence of asymmetric astigmatism. Pentacam indices were normal in both eyes. The Topolyzer keratograph (Oculus Optikgeräte) indicated that the left eye was normal, but the right eye had grade 1 keratoconus (Figure 1A and B).
We planned and performed topography-guided PRK in both eyes, combined with ultraviolet (UV) crosslinking in the right eye.
A bandage contact lens was placed in both eyes postoperatively. Postop therapeutic regimen was Predmycin P (neomycin sulfate, polymixin B sulfate, prednisolone acetate, polyvinyl alcohol; Allergan, Inc., Irvine, California) four times daily and Acular (ketorolac tromethamine; Allergan, Inc.) as needed for pain.
Postoperative follow-up in the right eye was uneventful. In the left eye, the patient reported pain grade 3+ on postoperative day 2 and a whitish corneal appearance on day 3, with no discharge and moderate redness.
At the postop visit on day 4, the epithelium in the right eye had healed. In the left eye, there was complete erosion with 3+ stromal infiltration (Figure 2), corneal melting, and loss of BCVA to 0.4. The refraction was -2.00 -1.75 X 90; UCVA was 1.5/60.
At that visit, a subconjunctival injection of steroids was given. The medical regimen was also changed. Predmycin P was replaced with Ciloxan (ciprofloxacin; Alcon Laboratories, Inc., Fort Worth, Texas) every hour. Homatropine twice daily was begun, and the ketorolac was stopped.
Possible diagnosis for this complication was a sterile infiltration due to interaction between the steroid and NSAID in the right eye, while the other eye healed normally due to the antiinflammatory action of the UV crosslinking.
Slow healing of the erosion progressed over weeks with 3+ central haze. At 7 months postop, in the left eye there is a central corneal scar and a thin cornea (thinnest point 296 µm). Loss of visual acuity and residual astigmatism also persist. UCVA is 0.4; BCVA is 1- with a refraction of 0.25 -2.25 X 160 (Figures 3 and 4).
What is the best way to continue the management of this patient?
–Case presented by Jérôme C. Vryghem, MD
MICHAEL ASSOULINE, MD, PHD
Several factors may have contributed to the sterile inflammatory infiltrates associated with a delayed healing response seen in this patient: (1) the epithelial toxicity of combined aminoglycosides (neomycin, polymixin B) plus preservative, along with the relative hypoxia induced by the bandage contact lens, and (2) the use of a topical NSAID of the aryl acetic group under contact-lens–induced corneal hypoxia, which may result in a paradoxical inflammatory exacerbation, sterile infiltrate, epithelial toxicity and delayed healing.1
The subsequent nonspecific delayed healing with stromal melt can be managed regardless of the primary cause until epithelial closure is obtained. Infection from an atypical agent should be ruled out by careful sampling at the corneal surface and all other relevant sites, including eyelid margins, throat, nose, and outer ear, and with the dedicated cooperation of a specialized microbiologist.
Although sterile infiltrates induced by topical NSAIDs of the aryl acetic acid group are usually responsive to topical steroids, this benefit must be balanced against the added toxicity to the nonhealing epithelium. Discontinuation of all topical medication except nonpreserved sodium hyaluronate, 12 times a day, is probably the best option. Daily sub-Tenon's injections of steroid may be helpful if significant inflammation is present in the anterior chamber. A bandage lens, even of the oxygen-permeable silicone-hydrogel type, may be more detrimental than useful to the healing response when polymorphonuclear infiltrates or metalloproteinase-induced stromal melting is observed. A bandage contact lens should not, in any case, be associated with topical NSAIDs, with the possible exception of the soluble form of indole acetic acid derivative as an analgesic when required.
Other more aggressive options may be required if the corneal thinning induced by stromal melt exceeds half of the central corneal thickness (Visante OCT; Carl Zeiss Meditec AG, Jena, Germany), and poses a significant perforation threat. A single- or dual-layered amniotic membrane graft can be sutured over the epithelial defect. For deeper defects, an ultrathin layer of cyanoacrylate adhesive can be applied using a sharpened wooden tip to smear minute (5 µL) drops of glue onto the stromal surface. Either amniotic membrane or adhesive shields usually play a significant role in quelling corneal inflammation by limiting the access of tear film polymorphonuclear cells to the corneal stroma. Both must be protected by a silicone-hydrogel lens of sufficient mechanical strength.
The loss of anterior stromal tissue, observed after surface healing is stabilized, can also be managed without regard to the primary cause of the condition.
Induced irregular astigmatism, the main cause for the limitation of vision, can usually be successfully managed with a reverse-geometry rigid gas permeable (RGP) contact lens adapted to an oblate corneal shape (Plateau; Menicon, Nagoya, Japan) or a large diameter RGP lens with an appropriate eccentricity. This conservative approach will improve vision immediately and buy precious time (a few months to 5 years) to let long-term anterior corneal remodeling smooth out the subepithelial defect and improve the final BCVA.
More aggressive surgical approaches may be indicated if contact lens fitting fails: (1) phototherapeutic keratectomy (PTK) with topography- or wavefront-guided ablation, seldom a satisfactory option for irregularities related to stromal defects, especially in such a thin cornea, (2) lamellar keratectomy using the femtosecond laser and simple epithelial healing, which may expose the patient to the risk of long-term ectasia, or (3) deep anterior lamellar keratoplasty using the big-bubble technique to leave an intact Descemet's membrane and thus minimize the risks associated with penetrating keratoplasty (eg, intraocular infection, primary or secondary wound dehiscence, endothelial rejection, endothelial failure). However, neither anatomical nor functional success can be assured with this method, given the usual rate of micro- and macroperforations and usual refractive outcome with current surgical techniques.
SHERAZ M. DAYA, MD, FACP, FACS, FRCS(ED)
This patient developed an infiltrate associated with pain in one of two eyes that underwent PRK followed by UV crosslinking with riboflavin. This was treated as an infection, and her course was complicated by a persistent epithelial defect. She has now been left with severe scarring and is less than 1 year postoperative.
Pain associated with an infiltrate should be treated as an infection; it is best to perform corneal scrapes, culture, and a gram stain to identify the organism and treat appropriately. The patient was treated with intensive antibiotics but continued to have an infiltrate and associated epithelial defect. Of significance was that her treatment regimen included NSAID eye drops. It is possible that she was infected and improved, but the white cell infiltrate was retained as a result of the migratory influence of NSAIDs on white cells. Of note is the improvement following cessation of NSAID and judicial use of steroid. Unfortunately, this patient now has corneal haze or scarring.
I would like to know the patient's ultrasound corneal thickness. Meanwhile, I would place her on intensive loteprednol (Lotemax; Bausch & Lomb, Rochester, New York) in a last-ditch attempt to reduce the corneal haze, which often occurs as a result of disordered lamellae and fluid from mild chronic inflammation even at this late stage.
Once the patient reaches 1 year postoperative, I would consider performing a PTK to smooth out the surface and eliminate the haze/scar. At the same sitting I would perform a live epikeratophakia or automated lamellar therapeutic keratoplasty, transplanting a thick lenticule of tissue—at least 320 µm, if not more—depending on residual stromal depth. The 8.5-mm lenticule is harvested from a donor eye using a keratome and artificial anterior chamber. The recipient undergoes minimal trephination with an 8.25-mm trephine, and the wound is undermined toward the limbus, creating a pocket for the donor tissue. The donor tissue is then sutured with considerable tension using 16 interrupted sutures, as these are likely to loosen rapidly with resolution of corneal edema and rapid healing. The sutures should be left in for at least 1 year, if possible. Later, depending on refractive stability and the degree of refractive error, LASIK can be performed to correct residual error.
The advantages of this technique, which I have used in similar cases, are that the possibility of recurrent haze is eliminated and the procedure is additive, providing greater biomechanical stability and an option for refractive correction later.
A. JOHN KANELLOPOULOS, MD
Differential diagnosis of this case would include the following: (1) corneal melting as an adverse reaction to the use of the topical NSAID, (2) bacterial infection, or (3) toxic keratopathy associated with PRK.
The thickness of the scar is usually underestimated both by Pentacam and ultrasound pachymetry. I would evaluate the scar thickness at the slit lamp and plan to possibly perform a PTK or collagen augmentation (adding synthetic collagen and performing UV collagen crosslinking) before considering proceeding to a penetrating keratoplasty.
Michael Assouline, MD, PhD, practices at the Clinique de la Vision, Paris. Dr. Assouline is a member of the CRST Europe Editorial Board. He may be reached at e-mail: ma@inclo.com.
Sheraz M. Daya, MD, FACP, FACS, FRCS(Ed), is Director and Consultant, Corneoplastic Unit and Eyebank, at the Centre for Sight, Queen Victoria Hospital, East Grinstead, UK. Dr. Daya is the Co-Chief Medical Editor of CRST Europe. He may be reached at e-mail: sdaya@centreforsight.com.
A. John Kanellopoulos, MD, is a Clinical Associate Professor of Ophthalmology at NYU Medical School and the Director of the Laservision.gr Institute, Athens, Greece. Dr. Kanellopoulos is the Refractive Complications Section Editor. He may be reached at +1 917 770 0586 (US); + 30210 7472777 (Greece); www.brilliantvision.com.
Jérôme C. Vryghem, MD, is from the Brussels Eye Doctors, in Brussels, Belgium. Dr. Vryghem is the Refractive Complications Section Editor. He may be reached at tel: +32 2 741 69 99; e-mail: j.c.vryghem@vryghem.be; info@vryghem.be.
