Acute corneal necrosis after excimer laser keratectomy for hyperopia

Acute Corneal Necrosis after Excimer Laser Keratectomy for Hyperopia Holger Mietz, MD,1 Maria Severin, MD,1 Peter Seifert, PhD,2 Peter Esser, MD,1 Gu¨nter K. Krieglstein, MD1 Objective: To describe a new, rare clinical complication after routine excimer laser photorefractive keratectomy to correct hyperopia. Design: Case report with clinicopathologic correlation. Main Outcome Measures: Four weeks after treatment with excimer laser, a perforating keratoplasty was performed for persistent corneal opacities. The corneal button was examined using light and electron microscopy. Special immunohistochemical stains were used to detect apoptosis. Results: The patient developed corneal opacities, endothelial precipitates, and a fibrinous exudate in the anterior chamber after the laser treatment. The changes did not respond to therapy directed against bacteria, fungi, and Acanthamoeba. All examinations and special stains were negative for micro-organisms. By light microscopy, an anterior zone of corneal necrosis was present with a moderate amount of acute inflammatory cells. At the interface between necrotic and viable corneal stroma, keratocytes with typical features of apoptosis were detected by immunohistochemistry and electron microscopy. Conclusion: This is the first full histopathologic report of a case of acute corneal necrosis with signs of apoptosis after excimer laser therapy of the cornea. Surgeons should be aware of this rare but potentially severe complication. Ophthalmology 1999;106:490 – 496 Excimer laser photorefractive keratectomy is a relatively new surgical procedure with a high accuracy that reshapes the anterior corneal surface in a nonthermal fashion to improve or correct the overall refractive system of the eye.1 The procedure is presumed to be safe, with a high success rate and few complications. Although this type of excimer laser surgery, which may be equal or superior to radial keratotomy,2 has been especially popular in several European countries over the past 5 to 10 years, larger clinical trials to obtain the U.S. Food and Drug Administration approval are now under way in the United States. Common complications of excimer laser photorefractive keratectomy include slowly healing corneal erosions, corneal haze, and a regression of the refractive correction.1 Rare complications of this surgical procedure include abnormal epithelial healing, recurrent epithelial breakdown, loss of corneal sensitivity, sterile corneal infiltrates, reactivation of herpes simplex keratitis, and corneal ulceration secondary to infections.1 In this article, we present the first case of acute corneal necrosis manifested by severe corneal stromal opacities, endothelial precipitates, and an anterior chamber reaction with a fibrinous exudate.

Originally received: April 27, 1998. Revision accepted: August 31, 1998. Manuscript no. 98217. 1 Department of Ophthalmology, University of Cologne, Cologne, Germany. 2 Alfried-Krupp-Laboratories, Department of Ophthalmology, University of Bonn, Bonn, Germany. Address correspondence to Holger Mietz, MD, Department of Ophthalmology, University of Cologne, 50924 Koeln, Germany.

490

Case Report A 56-year-old white man presented for excimer laser therapy for hyperopia and astigmatism of the left eye. The patient had no known systemic illnesses or rheumatic conditions. His refraction was ⫹ 2.25 ⫹ 0.25 ⫻ 151 in the right eye and ⫹ 1.75 ⫹ 0.75 ⫻ 51 in the left eye. Best-corrected visual acuity was 20/15 in both eyes. Informed consent for the laser treatment was obtained. The laser used was a VisuMed MEL60 (Meditec, Mu¨nchen, Germany). The outer diameter of the treatment zone was 9.0 mm, and the central zone had a diameter of 6.0 mm. After topical anesthesia with proxymetacaine 5 mg/ml, the epithelium was removed with a single-use sterile blade, and the excimer laser performed 47 scans in removing the superficial corneal stroma as determined by the laser system. With each scan, 1 ␮m of tissue was removed. Afterward, a disposable soft contact lens (AcuVue, Johnson & Johnson, Norderstedt, Germany) was placed on the cornea, and antibiotic eyedrops (kanamycin, 5 mg/ml) were administered every 2 hours. The patient had moderate pain directly after the procedure that worsened overnight to such an extent that he was admitted to the hospital emergency department the next day. On examination, visual acuity of the left eye was reduced to hand motion. There was a marked redness and chemosis of the conjunctiva of the left eye. The soft contact lens was in place. The corneal epithelium was almost completely absent. A large opacity was noted centrally in the anterior stroma, together with a similar, but smaller, localized peripheral opacity superiorly. In addition, multiple endothelial precipitates were present centrally. A moderate amount of cells (⫹⫹) and glare (⫹) were observed in the anterior chamber with a marked fibrinous exudate, and a whitish hypopyon measured approximately 1.5 mm (Fig 1). The iris was barely visible. Ultrasound B-scan of the posterior segment disclosed an attached retina and the vitreous with no opacities. From the clinical impression, an acute infection of the cornea was assumed and the soft contact lens removed. Topically, forti-

Mietz et al 䡠 Corneal Necrosis fied antibiotic eyedrops (azidamfenicol, 10 mg/ml) were given every hour. The patient began receiving intravenous antibiotics (ceftazidime 3 ⫻ 2 g; vancomycin 2 ⫻ 500 mg) and steroids (methylprednisolone 500 mg). The patient continued to suffer from severe pain. On the next day, topical antimycotic therapy was added five times daily (natamycin 10 mg/g). On the third day after laser therapy, a corneal scraping was taken to rule out a mycotic infection. The hypopyon combined with the fibrinous exudate appeared slightly smaller and denser. The intravenous antibiotics (ceftazidime, vancomycin) were given for a total of 5 days followed by oral cefuroxime 2 ⫻ 500 mg. An oral antimycotic therapy with flucanozol 3 ⫻ 100 mg began at day 5 and was given for 2 weeks. Steroids were administered for a total of 20 days in a tapering fashion, starting with 500 mg of methylprednisolone. In addition, starting 14 days after laser treatment, topical therapy with propamidine (Brolene, Rhone-Poulenc, UK) five times per day was added for a total of 2 weeks since Acanthamoeba was also considered as a source of infection. The soft contact lens and the solution that had contained the soft contact lens were examined for possible bacterial or fungal infection. The corneal epithelium healed slowly over time, covering the initial defect, but the corneal opacities did not respond to the initiated therapy, and the hypopyon decreased only moderately in size. Therefore, a penetrating keratoplasty was performed 24 days after the initial laser therapy. The complete opaque central area was removed together with the organized hypopyon and fibrinous exudate of the anterior chamber. The corneal button was dissected into several portions, and native material was directly placed on appropriate plates to rule out a possible infection caused by Acanthamoeba. The fibrinous exudate was submitted for detection of bacteria and fungi. After surgery, the clinical course was uneventful. The graft healed well into place. Seventeen months after surgery, the graft remained clear and the visual acuity was 20/40 with a refraction of ⫹ 0.25 ⫹ 2.25 ⫻ 60 (Fig 2). So far, no recurrent episodes of corneal or intraocular inflammation are noted. Vision was slightly decreased because of cataractous lens changes.

Materials and Methods The half corneal button that was sent for histopathologic examination after penetrating keratoplasty was fixed in 4% paraformaldehyde at 4° C and embedded in paraffin. Four micrometer-thick sections were cut and stained with hematoxylin and eosin (H&E), periodic acid-Schiff (PAS), Gram, GMS (Grocott-Gomori methenamine silver nitrate), von Kossa, and chloracetate esterase. The experiments for the detection of loss of viability and apoptosis were performed as previously outlined.3 DNA breaks were detected on a single-cell level by in situ DNA labeling.4 After deparaffinization and hydration, the sections were digested with proteinase K (20 ␮g/ml) for 15 minutes at room temperature. The specimens were equilibrated with terminal transferase buffer (TT buffer, 30-mmol Tris, pH ⫽ 7.2, 140-mmol sodium cacodylate, 1-mmol cobalt chloride) for 10 minutes and then exposed to 0.25 U/␮l terminal transferase and 50-␮M biotinylated desoxy-uridine triphosphate in TT buffer for 60 minutes at 37° C in a moist chamber. The reaction was stopped by immersion in 300-mmol sodium chloride/30-mmol sodium citrate for 15 minutes. The specimens were washed several times in water and blocked with 2% bovine serum albumine. The slides were incubated with streptavidin–alkaline phosphatase (1:500) for 30 minutes, washed extensively, developed with nitroblue tetrazolium chloride, and counterstained with hematoxylin. Specificity of staining was as-

certained by omitting cobalt chloride, the cofactor for terminal transferase. Human thymus served as positive control. For transmission electron microscopy, the paraffin-embedded specimen was re-embedded. Semithin sections were stained with methylene blue for orientation, and ultrathin sections were contrasted with uranyl acetate and lead citrate and examined in a Zeiss EM109 electron microscope. For control, a similar paraffin-embedded corneal button with bullous keratopathy was re-embedded for electron microscopy and similarly processed and examined. For the detection of herpes simplex virus (HSV) DNA, serial 10-␮m-thick sections were cut from the paraffin block. The sections were extracted twice with acetone before digestion with proteinase K. An aliquot of the extract was analyzed by double polymerase chain reaction (PCR) with nested primers for the detection of HSV1 genomes. A 159-base-pair piece from the region UL-42 was amplified. The PCR products were analyzed by agarose gel electrophoresis. To detect contamination, several negative control specimens containing corneal tissue from eyes with bullous keratopathy were included in each amplification run. Positive control specimens consisted of HSV1 infected cells. Each sample was analyzed in duplicates. Microbiologic examinations of the different specimens were performed as follows. Portions of the storing solution of the contact lens and the organized exudate from the anterior chamber, which was obtained during keratoplasty, were put in blood agar at 37° C for 2 days for detection of bacteria. For the detection of fungi, portions of both specimens were placed in malt extract agar at 30° C for 7 days. The soft contact lens itself was placed in chocolate agar for 14 days at 37° C for detection of bacteria. The corneal scraping taken on day 2 after the laser surgery was sent to a mycologic laboratory (Robert-Koch Institute, Department of Clinical Mycology, Berlin, Germany). Portions of the specimen were examined by Gram stain and cultivated at both 26° C and 37° C with the following media: blood (sheep) agar, Sabouraud’s dextrose agar (supplemented with different antibiotics), nigerseed (Staib) agar, and beerwort agar. After streaking onto the different media, the remaining specimen was given into brain heart infusion medium. Media were examined daily for visible growth for a total of 10 days. Sections from the corneal button, which was excised during keratoplasty, were examined for Acanthamoeba by placing a sample of native tissue on an agar plate coated with a layer of Escherichia coli bacteria for 4 weeks. This was done during surgery in the operating room.

Results Examination of the soft contact lens and the solution that had contained the soft contact lens for storage showed that both were negative for bacteria and fungi. Examination results of the fibrinous exudate from the anterior chamber for bacteria and fungi were also negative, as was the tissue from the corneal button examined for growth of Acanthamoeba. The examinations of the corneal scraping showed no positive results either in the Gram stain or in all cultures. Examination results of sections from the corneal button by light microscopy showed an intact but mostly thinned and irregular epithelium with mild intracellular and intercellular edema and small areas of bullous separation. There was no maturation present from the basal to the superficial cell layers, and there were no more than three to four layers of epithelial cells. Bowman’s layer was completely absent. For approximately two thirds of the length of the button was a zone in the anterior stroma that stained less with H&E and the PAS stains (Figs 3, 4). In that area, there were fewer

491

Ophthalmology Volume 106, Number 3, March 1999

Figure 1. Clinical appearance of the cornea and anterior segment of an eye after excimer laser photorefractive keratectomy 10 days after surgery. The conjunctiva is markedly hyperemic. There are multiple opaque areas of the anterior stroma centrally with an additional small, localized area in the periphery at the 12-o’clock position. An organized fibrinous exudate is present in the anterior chamber inferotemporally. Figure 2. The same eye 17 months after penetrating keratoplasty. The graft is clear, and there is no sign of infection. The single sutures are still in place. Visual acuity is 20/40. There are few posterior synechiae. Figure 3. Low-power view of the cornea. The epithelium is markedly thinned. The anterior stroma stains less in most areas compared to the midstroma and deep stroma (stain, periodic acid-Schiff; original magnification, ⫻10).

492

Mietz et al 䡠 Corneal Necrosis Figure 4. Higher power view of the anterior stroma. There is loss of the surface maturation of the epithelium. Bowman’s layer is absent. The anterior stromal lamellae appear to be less organized and stain pale (stain, periodic acid-Schiff; original magnification, ⫻50). Figure 5. To rule out an infection caused by bacteria, a Gram stain was performed. High-power view. No micro-organisms are present (stain, Gram; original magnification, ⫻50). Figure 6. The special stain for the detection of neutrophils shows a moderate amount of acute inflammatory cells scattered throughout the anterior stroma. The positive cells have a markedly red staining cytoplasm (stain, chloracetate– esterase; original magnification, ⫻100). Figure 7. Immunohistochemical stain for the detection of apoptosis. A positive reaction is present in a zone posterior to the anterior area of corneal necrosis, which stained pale in periodic acid-Schiff (Figs 3, 4) (stain, TUNNEL—reaction; original magnification, ⫻25; inset, ⫻100). Š

keratocytes as compared to deeper stromal layers. The collagen fibers appeared to be arranged irregularly. No bacteria were noted in the Gram stain (Fig 5). There was a mild-to-moderate infiltrate of acute inflammatory cells that did not form a wall or a ringshaped infiltrate surrounding a central zone. The character of these acute inflammatory cells proved to be neutrophils with the chloracetate esterase stain, which is specific for these cells (Fig 6). In the H&E and PAS stains, there were also small round amorphic bodies observed, which stained negative for calcium with the von Kossa stain and resembled cell nuclear fragments. Along the interface between the anterior, less-staining stromal zone and the regularly appearing middle and posterior portion of the corneal stroma, several larger keratocytes were present with enlarged nuclei. These nuclei demonstrated a positive TUNNEL staining, thereby identifying apoptotic cell death in that area (Fig 7). Other keratocytes in the same specimen as well as those in a cornea with bullous keratopathy did not show a positive TUNNEL staining. The human thymus, which served as a positive control specimen, stained for the reaction, while the negative control specimens were negative. The analysis of corneal tissue by PCR to detect a possible infection by the herpes simplex virus (HSV1) was negative, as were the negative control specimens. Examination of the corneal button by electron microscopy also revealed an irregular, thinned corneal epithelium with loss of surface maturation and absence of Bowman’s layer. The lamina basalis of the basal epithelial cells was fragmented. The collagen fibrils appeared to be arranged in an unorganized fashion with several whirl-like areas. These fibrils were smaller in diameter in the anterior stroma compared to those of Bowman’s layer in the control. In some zones, small, round amorphic structures were present. There were also plasma cells, neutrophils, and cells with pyknotic nuclei present (Fig 8). In an area representing the interface between the anterior stromal zone and the deeper layers, cell nuclei undergoing apoptosis were observed (Fig 9).

Discussion Both radial keratotomy and excimer laser photorefractive keratectomy bear the potential risk of postoperative infections.5– 8 These infections may lead to a localized or generalized keratitis, endophthalmitis, or orbital cellulitis.9 Although the reported cases seem to be rare, resulting in a low incidence of this complication, each is a severe disease, and some are even sight threatening. This is especially important for surgical procedures that are performed on eyes that have a best-corrected visual acuity of 20/20 or better. In radial keratotomy, deep corneal incisions are produced. Therefore, not only is the corneal epithelium, Bow-

man’s layer, or the superficial stroma exposed, but some incisions reach deeper to the level of the midstroma or the posterior stroma. In some instances, inadvertent or unnoticed perforations of the entire cornea, best demonstrated by flat anterior chambers, occur and are presumably the reason for endophthalmitis.10 In a few cases, an infectious endophthalmitis developed despite the fact that no perforations were made.9 In addition, reoperations and postoperative contact lens wear have been proposed as risk factors for infection.8 In radial keratotomy, most infections developed within the first 2 weeks after treatment.8 In excimer laser photorefractive keratectomy, the surgical procedure does not usually reach deep into the corneal layers. The exposed surface of the cornea, which is not protected by epithelium in the immediate postoperative period, is much larger compared to incisions performed in radial keratotomy. In addition, in many instances, soft contact lenses are used to cover the corneal defect and to decrease the postoperative pain and discomfort of the patient. The use of the contact lens also bears a risk for infections in itself.11,12 For this reason, some surgeons tend not to use soft contact lenses for that purpose. Infections have been reported to develop as early as 3 days after surgery7 as well as 6 months later.13 The effect of the excimer laser beam on corneal tissue has been the subject of several studies, while yet it remains poorly understood. From a theoretical standpoint, the excimer laser simply removes tissue by vaporization with a minimal effect on the underlying tissue. In histopathologic studies using animal models, it has been shown that the area of surrounding tissue damage is small. This may not necessarily be the case for the use of excimer lasers in treating corneal tissue. However, it could be suspected based on several observations. First, a regression of the refractive effect is noted in many cases, especially when higher refractive errors are treated, which means that more laser energy is used.14 This regression has to be connected with some type of stromal response either mediated by changes of the stromal collagen, the keratocytes, or both. Second, the so-called “corneal haze” is a response of the tissue to the laser beam. The exact nature of the phenomenon has not been elucidated yet, but can be graded and resolves with and without specific treatment in most cases. Topical application of steroids is usually the treatment of choice, which may in itself increase the risk of postoperative corneal infections.

493

Ophthalmology Volume 106, Number 3, March 1999 Moreover, in a rabbit study, the excimer laser caused a higher incidence of apoptosis in keratocytes compared to control eyes.15 This effect can only be interpreted as a lasting effect of the therapy in the tissue. In this regard, it is noteworthy that the CD95 system is expressed in the cornea and has been attributed to modulate keratocyte apoptosis after corneal injury.16 It may also be suggested that the laser beam has an effect on the genome of the keratocytes, since reactivation of herpes simplex virus after excimer laser therapy has been reported several times.1 The exact etiology for this reactivation remains unknown. Laser-induced DNA damage has also been observed.17 Another rarely occurring complication is the sterile corneal infiltrate, which develops in the superficial stroma close to the area of treatment. The infiltrates can develop rapidly after surgery and multiply. An association with the use of therapeutic soft contact lenses has been made.1 Interestingly, the infiltrates resolve without antibiotic therapy within a few weeks. In addition, sterile infiltrates have been reported to occur in association with the use of topical corticosteroids and nonsteroidal anti-inflammatory drugs. In a recent survey,18 the incidence was approximately 1:250, and in no instance were cultures positive for micro-organisms. In most cases, scarring of the corneal stroma persisted and reduced the central visual acuity. To the best of our knowledge, eight cases of corneal infections after photorefractive excimer laser therapy have been reported.6,7,11,13,19 In one case, Staphylococcus epidermidis,19 Pseudomonas aeruginosa,6 and Staphylococcus aureus7 were detected. In four cases,11 a fungal infection was suspected but unproved, and in one case,13 nothing was mentioned regarding examinations for micro-organisms. Therefore, there have not been any proven cases that are associated with fungi or Acanthamoeba. The number of unreported cases is unknown and can only be speculated on. In each documented case, a soft contact lens or a collagen shield19 was used to reduce patient discomfort after laser surgery. Figure 8. Transmission electron micrograph of the anterior cornea of the treated specimen (A, C, E) and a control with bullous keratopathy (B, D, F). A, the corneal epithelium is thinned and irregular with loss of surface maturation (between arrows). Bowman’s layer is absent. The anterior stromal lamellae are arranged irregularly with a decreased amount of keratocytes and an invasion of acute and chronic (mast cell, arrowhead) inflammatory cells. B, control. Basal epithelial cell layer, Bowman’s layer, and regularly arranged stromal lamellae with interspersed keratocytes (arrowheads) (A, B, original magnification, ⫻640). C, transition zone between epithelium and stroma. The epithelial basement membrane (lamina basalis) is largely absent and only present in a few fragmented areas (arrowheads). D, control. Hemidesmosomes among epithelial cells are present. The lamina basalis is intact (arrowheads). The randomly arranged small collagen fibrils (right and lower side of the micrograph) are part of Bowman’s layer (C, D, original magnification, ⫻13,450). E, anterior stroma. The bundles of collagen fibers are arranged in a whirl-like pattern. They have lost most of their regular arrangement. F, control. Regularly appearing layers of collagen fiber bundles of the anterior stroma (E, F, original magnification, ⫻7900).

494

Figure 9. Nuclei of keratocytes in the transition zone between affected and regularly appearing corneal stroma. Upper left corner (arrowheads): normal nucleus. Lower right corner (arrows): largely condensed chromatin in the nucleus of a keratocyte as a morphologic feature compatible with the process of apoptosis (original magnification, ⫻2000).

Mietz et al 䡠 Corneal Necrosis The case presented in this report is distinct from those reported so far. The complication developed almost immediately after excimer laser surgery. It consisted of a dense corneal infiltrate with an endothelial reaction and a marked fibrinous exudate in the anterior chamber associated with severe pain. Clinically, the presentation led to the impression of an infectious keratitis, most presumably caused by fungi, or, because of the pain, by Acanthamoeba. However, all examinations, in an effort to reveal a causative agent, produced negative results. In addition, the corneal button, which became available for histopathologic examination 4 weeks after the laser therapy, was uncharacteristic of an acute infection and showed more of a picture of diffuse necrosis with apoptotic cell nuclei in the periphery of the affected corneal tissue. From the histopathologic observations, it is impossible to rule out the possibility of a healed bacterial ulcer. However, for several reasons, this is unlikely. First, all examination results to reveal any organisms were negative. Second, no micro-organisms were seen on histopathologic examination, including special stains and transmission electron microscopy. Third, from a clinical standpoint, it is unclear why the corneal opacities did not respond at all to the antibacterial treatment. If micro-organisms had responded to therapy, the opacities should have at least decreased in size. In addition, the histopathologic appearance of the corneal button with a moderate amount of acute inflammatory cells, scattered diffusely through the anterior corneal stroma, does not fit to an acute or latent infection with a peripheral wall and a central area of necrosis, which would more likely be present. For these reasons, it seems more likely that the corneal reaction was a sequelae of the laser therapy itself and not related to an infection. The exact classification of the nature of this lesion is difficult, since such clinical findings have to date not been reported. Yashima et al20 described an acoustic effect of an argon–fluoride excimer laser that was used to ablate the stratum corneum of the skin in hairless rats. For ablations through air or water, damage zones in the underlying tissue were seen on transmission electron microscopic examination. In superficial and deeper layers of the dermis, intracellular and intranuclear changes of keratinocytes and fibroblasts compatible with features of apoptosis were seen in both groups, whereas the changes appeared to be more pronounced in those cases in which the ablation was performed through water. In a different investigation, lasers were used to create full-thickness skin wounds. In those lesions performed with a 193-nm excimer laser, a thin but distinct zone of necrosis in the underlying tissue was present immediately after surgery.21 Both studies imply that the excimer laser may be able to have effects on tissue that is not ablated but close to the treatment zone. Holme et al22 performed studies on rabbit corneas using en-face and tangential excimer laser ablation techniques. They described the possibility of an inhomogeneous energy pattern of the laser beam within the treatment area. This means that the ablation surface is not smooth and the tissue effect different within the area of treatment. This effect is larger using the en-face technique, which is used clinically.

The findings were supported by histopathologic examinations. In summary, several findings point into the direction that in this case, the laser energy caused a thin but distinct zone of tissue necrosis. According to the surgeon who performed the laser procedure, the energy settings were not inadvertently high or changed throughout the procedure. It may also be that the cornea itself was more susceptible to damage than most corneas. It may additionally be speculated that this sort of stromal reaction occurs more frequently than thought so far and that the changes are sometimes misinterpreted as infections. We believe that corneal surgeons should be aware of this potential risk of excimer lasers used for photorefractive keratectomy. These lasers may have an effect on deeper corneal layers, which in a few cases may lead to tissue necrosis and nonclearing central opacifications.

References 1. Seiler T, McDonnell PJ. Excimer laser photorefractive keratectomy. Surv Ophthalmol 1995;40:89 –118. 2. Waring GO III, Lynn MJ, McDonnell PJ. Results of the prospective evaluation of radial keratotomy (PERK) study 10 years after surgery. Arch Ophthalmol 1994;112:1298 – 308. 3. Esser P, Heimann K, Bartz–Schmidt KU, et al. Apoptosis in proliferative vitreoretinal disorders: possible involvement of TFG-beta-induced RPE cell apoptosis. Exp Eye Res 1997;65: 365–78. 4. Gavrieli Y, Sherman Y, Ben–Sasson SA. Identification of programmed cell death in situ via specific labeling of nuclear DNA fragmentation. J Cell Biol 1992;119:493–501. 5. Heidemann DG, Dunn SP, Haimann M. Endophthalmitis after radial keratotomy enhancement. J Cataract Refract Surg 1997; 23:951–3. 6. Wee WR, Kim JY, Choi YS, Lee JH. Bacterial keratitis after photoreactive keratectomy in a young, healthy man. J Cataract Refract Surg 1997;23:954 – 6. 7. Fulton JC, Cohen EJ, Rapuano CJ. Bacterial ulcer 3 days after excimer laser phototherapeutic keratectomy. Arch Ophthalmol 1996;114:626 –7. 8. Jain S, Azar DT. Eye infections after refractive keratotomy. J Refract Surg 1996;12:148 –55. 9. McLeod SD, Flowers CW, Lopez PF, et al. Endophthalmitis and orbital cellulitis after radial keratotomy. Ophthalmology 1995;102:1902–7. 10. Gelender H, Flynn HW Jr., Mandelbaum SH. Bacterial endophthalmitis resulting from radial keratotomy. Am J Ophthalmol 1982;93:323– 6. 11. Faschinger C, Faulborn J, Ganser K. Infektio¨se Hornhautgeschwu¨re— einmal mit Endophthalmitis—nach PRK mit Einmalkontaktlinse [Eng. Abstr.]. Klin Monatsbl Augenheilkd 1995;206:96 –102. 12. Alfonso E, Mandelbaum S, Fox MJ, Forster RK. Ulcerative keratitis associated with contact lens wear. Am J Ophthalmol 1986;101:429 –33. 13. Lavery FL. Photorefractive keratectomy in 472 eyes. Refract Corneal Surg 1993;9(2 Suppl):S98 –S100. 14. Sher NA, Barak M, Daya S, et al. Excimer laser photorefractive keratectomy in high myopia. A multicenter study. Arch Ophthalmol 1992;110:935– 43.

495

Ophthalmology Volume 106, Number 3, March 1999 15. Gao J, Gelber–Schwalb TA, Addeo JV, Stern ME. Apoptosis in the rabbit cornea after photorefractive keratectomy. Cornea 1997;16:200 – 8. 16. Wilson SE, Li Q, Weng J, et al. The Fas–Fas ligand system and other modulators of apoptosis in the cornea. Invest Ophthalmol Vis Sci 1996;37:1582–92. 17. Schulte–Frohlinde D, Simic MG, Gorner H. Laser-induced strand break formation in DNA and polynucleotides. Photochem Photobiol 1990;52:1137–51. 18. Sher NA, Krueger RR, Teal P, et al. Role of topical corticosteroids and nonsteroidal antiinflammatory drugs in the etiology of stromal infiltrates after excimer photorefractive keratectomy. J Refract Corneal Surg 1994;10:587– 8.

19. McDonald MB, Frantz JM, Klyce SD, et al. Central photorefractive keratectomy for myopia. The blind eye study. Arch Ophthalmol 1990;108:799 – 808. 20. Yashima Y, McAuliffe DJ, Jacques SL, Flotte TJ. Laserinduced photoacoustic injury of skin: effect of inertial confinement. Lasers Surg Med 1991;11:62– 8. 21. Green HA, Burd EE, Nishioka NS, Compton CC. Skin graft take and healing following 193-nm excimer, continuous-wave carbon dioxide (CO2), pulsed CO2, or pulsed holmium:YAG laser ablation of the graft bed. Arch Dermatol 1993;129:979 – 88. 22. Holme RJ, Fouraker BD, Schanzlin DJ. A comparison of en face and tangential wide-area excimer surface ablation in the rabbit. Arch Ophthalmol 1990;108:876 – 81.

GLAUCOMA RESEARCH FOUNDATION SEEKS INTERNATIONAL OPHTHALMOLOGY FELLOWS The Glaucoma Research Foundation (GRF), a U.S.-based nonprofit organization, seeks applicants for a 1-year program providing ophthalmologists from developing countries with research training opportunities in the United States. GRF places up to three ophthalmologists in U.S. research centers where at least two thirds of their time is spent in clinical or laboratory research. Candidates should be assistant professors with a strong interest in glaucoma research. Candidates must be fluent in written and spoken English. Candidates must return to their home countries to advance glaucoma research and patient treatment at the completion of the fellowship. Candidates must be nominated by department chairs and complete an application form, which includes a curriculum vitae and a letter confirming their academic position. Visit http://www.glaucoma.org/fellowship.html or contact GRF to request an application form. The deadline is July 15, 1999 for the fellowship beginning July 1, 2000. Glaucoma Research Foundation 200 Pine Street, Suite 200 San Francisco, CA 94104 USA Phone (415) 986-3162, FAX (415) 986-3763

496