●

Aif●

●

wcyplfiisiHwM●

cpLtfisbt●

Ltpgo

A

MaeA((

UHe

6

Avellino Corneal Dystrophy Worsening after Laser InSitu Keratomileusis: Further Clinicopathologic

Observations and Proposed Pathogenesis

SHADY T. AWWAD, MARIO A. DI PASCUALE, ROBERT N. HOGAN, STEPHAN L. FORSTOT,

JAMES P. McCULLEY, AND HARRISON D. CAVANAGH

(E

GdsoOtAw

T

cBB((ewc1r

icpksmwibiaaLp

PURPOSE: To study the nature of the deposits invellino corneal dystrophy (ACD) worsening after laser

n situ keratomileusis (LASIK), and suggest a mechanismor histopathogenesis. DESIGN: Interventional case report. METHODS: A 28-year-old woman previously diagnosedith bilateral ACD underwent bilateral LASIK. The

orneal dystrophy progressively worsened bilaterally, oneear later. A penetrating keratoplasty was subsequentlyerformed on the right eye at 31 years of age, and in theeft eye a year later. The clinical and histopathologicndings of the corneal graft of the right eye were reported

n the literature, with positivity to the Masson trichrometain, negative staining with Congo red, and heterozygos-ty for the Arg124His mutation by serum DNA studies.istopathologic studies of the corneal graft of the left eyeere conducted at the University of Texas Southwesternedical Center.

RESULTS: Histopathologic examination of the excisedornea showed the Masson trichrome positive depositsresent from underneath the Bowman layer to theASIK interface, with absence of deposits posterior tohe latter. In contrast to the prior report describingndings in the corneal graft of the left eye, the depositstained lightly with Congo red, but failed to showirefringence under polarized light, or fluorescence withhioflavin T. CONCLUSION: Accelerated deposits developing afterASIK in ACD eyes seem to harbor pre-amyloid fea-ures. The epithelium is likely to be the culprit, in aathway independent of with human transformingrowth hormone beta (TGF-beta), with deposits devel-ping in the anterior stroma and the stromal interface.

ccepted for publication Dec 5, 2007.From the Cornea, External Diseases, and Refractive Surgery (S.T.A.,.A.D., J.P.M., H.D.C.), and Ophthalmic Pathology (S.T.A., R.N.H.);

nd the Department of Ophthalmology, University of Texas Southwest-rn Medical Center, Dallas, Texas; Department of Ophthalmology,merican University of Beirut Medical Center, Beirut, Lebanon

S.T.A.); and Corneal Consulting of Colorado, Littleton, ColoradoS.L.F.).

Inquiries to Harrison D. Cavanagh, Department of Ophthalmology,niversity of Texas Southwestern Medical Center at Dallas, 5323 Harry

cines Blvd, Dallas, TX 75390; e-mail: dwight.cavanagh@utsouthwestern.

du

© 2008 BY ELSEVIER INC. A56

Am J Ophthalmol 2008;145:656 – 661. © 2008 bylsevier Inc. All rights reserved.)

RANULAR CORNEAL DYSTROPHY TYPE II (AVEL-

lino corneal dystrophy, combined granular cor-neal dystrophy with amyloid) is an autosomal

ominant disorder which usually appears during the first orecond decade of life, and is characterized by cornealpacities that resemble rings, disks, stars, or snowflakes.1–5

riginally described in patients and families with ancestryraced to Northern Italy, specifically around the region ofvellino, it has since been reported all around theorld.6–8

METHODS

HE CONDITION, ALONG WITH OTHER TYPES OF GRANULAR

orneal dystrophies, lattice corneal dystrophies, and Reis-uckler dystrophy, results from a specific mutation in theIGH3 gene on chromosome 5q31, dictating a TGF-betaTGFB)-induced cell adhesion protein, keratoepithelin68 Kda), which is mainly expressed by the cornealpithelium and normally present in the corneal stroma,ith highest density in the Bowman layer.9,10,18 –20 In thease of Avellino corneal dystrophy, a mutation in codon24 of BIGH3 gene (histidine replacing arginine) isesponsible for all studied cases.9,10

Phototherapeutic keratectomy has been successful inmproving the visual acuity of visually significant Avellinoorneal dystrophies (ACD). However, recurrences afterhototherapeutic keratectomy (PTK) and photorefractiveeratectomy (PRK) have been documented.11,12 Laser initu keratomileusis (LASIK) in patients with mild tooderate dystrophy has been reported to systematicallyorsen the disease condition, with histopathologic find-

ngs recently being described, but the mechanism has noteen formulated to date.13–19 A previous article publishedn the American Journal of Ophthalmology by Aldave andssociates reported on a 28-year-old woman with anccelerated granular-lattice corneal dystrophy course afterASIK, with a bilateral penetrating keratoplasty (PK)erformed thereafter.16 Histopathologic findings of the

orneal graft of the right eye were presented, which

LL RIGHTS RESERVED. 0002-9394/08/$34.00doi:10.1016/j.ajo.2007.12.008

itsWtTlmpeoaidpL

●

twSvmwbnratimd

feswmc

UDasieaattasibMHStsBiTas

Fcc

V

ncluded negative Congo red staining, with the conclusionhat there is an accelerated TGFB-induced protein re-ponse in patients with ACD after excimer laser ablation.

e had the chance to examine the patient subsequently athe University of Texas Southwestern Medical Center.he second corneal graft, performed in the left eye, was

ater sent to us for histopathological evaluation. Althoughany findings were similar to the ones described by the

revious report, based on the first corneal graft of the rightye, there were important additional findings which rec-ncile the histologic picture with previous findings by Leend associates, and we believe are vital to report tomprove our understanding of the pathophysiology of theisease discussed, the etiology of the deposits seen, and theotential mechanisms of their accelerated formation afterASIK.14

CASE: A 28-year-old woman had been diagnosed sincehe age of 23 as having bilateral granular corneal dystrophyith bilateral central anterior stromal “fleck-like” crystals.he underwent bilateral LASIK, and in the early follow-upisits, few deposits were noted bilaterally anterior to andainly in the interface. A year later, an increase in finehite granular deposits were noted in the flap interface inoth eyes. Over two years, the interface deposits wereoted to be denser in both eyes with decrease in uncor-ected visual acuity (UCVA) and best-corrected visualcuity (BCVA) in both eyes and poor brightness acuityest (BAT) results. Subsequently, the LASIK flap was liftedn the right eye and the interface was scraped and 0.02%itomycin C (MMC) was applied for two minutes, but the

IGURE 1. Avellino corneal dystrophy worsening after laser inornea of the left eye showing extensive granular deposits. (Righoncentrated in the LASIK flap interface.

eposits started to increase again along the interface only T

AVELLINO CORNEAL DYSTROPHYOL. 145, NO. 4

ew months later, and a PK was undertaken in the rightye. The clinical and histopathologic findings were de-cribed in details in the report of Aldave and associates,ho later performed a BIGH3 gene screening using poly-erase chain reaction on a blood sample from the patient

onfirming heterozygousity for the Arg124His mutation.16

At one year post-PK, the patient was referred to theniversity of Texas Southwestern Medical Center atallas for evaluation. BCVA was 20/20 in the right eye,

nd 20/30 in the left eye with the patient complaining ofignificant glare. BAT performed in both eyes at highllumination levels yielded a BCVA of 20/20 in the rightye and 20/400 in the left. Slit-lamp examination revealedclear corneal graft in the right eye, and anterior stromal

nd interface granular deposits involving the visual axis inhe left eye (Figure 1). Confocal microscopy performed onhe left eye showed the deposits to be solely localizednterior to and along the flap interface. The patientubsequently had PK performed in the left eye by hernitial surgeon one month later. The recipient cornealutton was sent to the University of Texas Southwesternedical Center at Dallas for histopathological evaluation.ematoxilin and eosin (H&E) staining, Periodic Acid-chiff (PAS), Congo red, thioflavin T, and the Massonrichrome staining were performed. On H&E and PAStaining, the corneal deposits were found just beneath theowman layer and extended down to and along the flap

nterface, where they were mainly concentrated (Figure 2,op). No deposits were seen posterior to the flap interface,nd none took up PAS stain (Figure 2, Bottom). But theytained deeply red with the Masson trichrome (Figure 3,

keratomileusis (LASIK). (Left) Slit-lamp photography of thehe deposits are scattered in the anterior stroma, but are mainly

situt) T

op), and weakly, but positively with Congo red, but failed

WORSENING AFTER LASIK 657

tacB

R

dPLhtdL

ao

FLtw�(vm

FLtosTid

6

o show apple-green birefringence on cross-polarized light,fter Congo red staining (Figure 3, Middle). No fluores-ence was exhibited on thioflavin T staining (Figure 3,ottom).

DISCUSSION

ECURRENCE OF ACD AS WELL AS GRANULAR AND LATTICE

ystrophies has been documented after treatment by PTK andRK.11,12 Worsening ACD has also been reported afterASIK.13–19 The deposits were described clinically andistologically to be located in the LASIK flap interface andhe stroma anterior to it. Three reports have so farescribed histopathologic findings in eyes with ACD after

IGURE 2. Avellino corneal dystrophy worsening afterASIK: histopathologic findings. (Top) Histologic section ofhe corneal recipient tissue (stain, hematoxylin and eosin;riginal magnification �200). The deposits are seen in theub-Bowman layer, anterior stroma, and the LASIK interface.he size of the deposits tends to increase as they approach the

nterface. (Bottom) Periodic Acid-Schiff section, with theeposits failing to stain (original magnification �200).

ASIK. One of those reports, published by Aldave and m

AMERICAN JOURNAL OF58

ssociates, describes the findings of the right eye corneal graftf our patient.16 The findings consisted of eosinophilic

IGURE 3. Avellino corneal dystrophy worsening afterASIK: special histopathologic staining. (Top) Histologic sec-ion of the central corneal recipient tissue, staining deep redith the Masson trichrome stain (original magnification200). (Middle) The section stains weakly with Congo red

original magnification �200). (Bottom) Staining with thiofla-in T. No fluorescence is exhibited under the fluorescenticroscope (original magnification �200).

aterials staining positive with the Masson trichrome

OPHTHALMOLOGY APRIL 2008

sdioaawbesrpat

wowjudcdetaccysfiarotpa

rLmiofibwklitkebcce

itwetAfilcptdtaadtdbttw

NdbcsttaoftamageIdrldioaLisTai

A

V

tain with no uptake with Congo red stain. Most of theeposits were described to be located in the LASIKnterface, the stroma of the flap, and the anterior portionf the stromal bed. In the second report, by Lee andssociates, the deposits were mainly seen in the interface,nd they stained heavily with the Masson trichrome andith Congo red, interpreted by the authors as showingoth granular and amyloid nature, respectively. However,lectron microscopy performed by the same authorshowed only granular features, not amyloid.14 The thirdeport, by Chiu and associates, describes a similar clinicalresentation with confirmatory deoxyribonucleic acids (DNA)nalysis and histopathology showing positive staining withhe Masson trichrome and negative Congo red staining.18

The histopathologic findings of our patient confirm asell as reconcile the histopathologic features of the previ-us reports and extend their findings farther. The depositsere not solely confined to the interface, but were present

ust underneath the Bowman layer and all along thenderlying stroma anterior to the interface. However,eposits were heavily scattered throughout the interface,entrally, with no involvement of the stromal bed. Theeposits stained weakly with Congo red, but did notxhibit birefringence with cross-polarized light, nor didhey reveal fluorescence after thioflavin T staining. Thebsence of apple-green birefringence and lack of fluores-ence on thioflavin T indicate that the beta-pleated sheetonfiguration, responsible for the birefringence, had notet formed, and that the myriad of individual filaments aretill in disarray, and therefore nonbirefringent.21–23 Thesendings are consistent with the report of Lee and associ-tes, in which the deposits stained strongly with Congoed, but did not show amyloid features of parallel packingf fibrils on electron microscopy.14 Hence, collectively,hese findings indicate that the deposits laid could harborre-amyloid features and could ultimately develop a similarmyloid picture as late stage ACD.

The mechanism for the worsening of ACD after LASIKemains elusive, much as the corneal dystrophies. AfterASIK, the epithelial basement membrane and the Bow-an layer remain intact, and hence there is minimal

ncrease in TGF-beta and other inflammatory products, aspposed to PRK or PTK, where TGF-beta increases in therst few months postoperatively in the ablated area, andecomes undetectable only after six months.24–26 Hence,hether the keratoepithelin source is the epithelium or theeratocytes, the stimulation of the mutated keratoepithe-in protein in ACD corneas after LASIK seems to bendependent of TGF-beta. One prevailing theory for kera-oepithelin-derived corneal dystrophy is that the mutatederatoepithelin protein, secreted predominantly by thepithelium, diffuses posteriorly through the Bowman mem-rane and aggregates with itself and possibly other mole-ules to form the final deposits characteristic of theorresponding disorder.2,9,10,27 An indirect evidence of the

pithelial nature of keratoepithelin and its diffusion ability t

AVELLINO CORNEAL DYSTROPHYOL. 145, NO. 4

s the slow progression from anterior to posterior stroma ofhe lattice, granular, and Avellino corneal dystrophies, asell as the initial confinement of the recurrence to thepithelium after keratoplasty or PTK.28 Additional creditso this theory are provided by ultra-structural studies.27 InCD, the granular lesions appear earliest and more super-

cially, whereas the lattice lesions usually develop at aater stage and are typically found more posteriorly in theornea. Interestingly, in the case of lattice corneal dystro-hy, keratoepithelin forms amyloid early in the course ofhe disease and more anteriorly in the stroma. The intro-uction of antibodies specific for the N terminal and Cerminal portions of the keratoepithelin protein (KE-15nd KE-2, respectively) has elucidated this phenomenonnd contributed significantly to our understanding of theifferent keratoepithelin-derived corneal dystrophies andhe various clinical and histological situations engen-ered.2,10,29 It has been shown that keratoepithelins thatind both KE-15 and KE-2 tend to stain with the Massonrichrome and develop into granular deposits, while kera-oepithelins exclusively binding to KE-2 are the oneshich develop into amyloid deposits.9

Proteolysis in the epithelial area is believed to act on the-terminal of keratoepithelins, leading to configurations that

on’t bind to KE-15 and hence are prone to aggregate to formeta-pleated sheets.2,10,22,30 In a patient with classic latticeorneal dystrophy, frequent epithelial erosions occur withubsequent proteolysis, promoting the synthesis of both mu-ant and wild type keratoepithelin, leading to keratoepithelinhat binds only to KE-2, and hence is receptive to aggregatend form amyloid.2 The quick surge in synthesis and prote-lysis of keratoepithelin favors the molecules, which are nowound in high concentration next to the epithelial entourage,o aggregate and form amyloid products early in life and in thenterior stroma. In ACD, the epithelial proteolysis is slow andainly dependent on aging, as the erosions are infrequent

nd hence the amyloid deposits appear later in life than theranular ones, and are more posterior.2,9 After LASIK, how-ver, the epithelium reactively secretes products like collagenII and undergoes a hyperplastic and hypertrophic response asocumented by confocal microscopy.31–33 This epithelialeactivity could possibly result in an increase in keratoepithe-in synthesis and proteolysis, leading to a relatively rapidevelopment of granular and potentially pre-amyloid depos-ts. The clinical and histological findings in our case and thenes reported so far lend some credit to this theory. Inddition, the total absence of the deposits posterior to theASIK interface and the presence of the deposits in increas-

ng number and size from the Bowman layer to the interfaceuggest a possible epithelial origin for the keratoepithelin.he LASIK interface being a large loose potential space, theggregated keratoepithelin would tend to diffuse alongside oft, polymerizing even further to produce large deposits.

In summary, the pathophysiology of the worsening ofvellino corneal dystrophy after LASIK is likely to be similar

o the original disease, but with an enhanced pace, possibly

WORSENING AFTER LASIK 659

aiooi

kwic

TBYMJ

1

1

1

1

6

ttributable to epithelial reactivity to keratocyte injury, andndependent of TGF-beta. The staining with Congo red inur report and in a previously published one, in the absencef birefringence, or ultra-structural evidence of parallel pack-

ng of fibrils, suggests the production of amyloid-forming l

dystrophy after LASIK. Ophthalmology 2004;111:463–468.

1

1

1

1

1

1

2

2

22

2

2

2

2

2

AMERICAN JOURNAL OF60

eratoepithelin configurations. The latter should be bindingith KE-2, but not with KE-15. Immunohistochemical stud-

es evaluating the binding of those two antibodies to theorneal deposits in ACD eyes after LASIK would shed more

ight on this intriguing phenomenon.

HIS STUDY WAS SUPPORTED IN PART BY NATIONAL EYE INSTITUTE (NEI) GRANTS EY10738 (DR CAVANAGH) AND EY016664,ethesda, Maryland, The Pearle Vision Foundation, Dallas, Texas, and an unrestricted grant from Research to Prevent Blindness, Inc, New York, Nework. The authors indicate no financial conflict of interest. Involved in study design (S.T.A., M.D.P., R.N.H., H.D.C., S.L.F.); material source (S.T.A.,.D.P., R.N.H., H.D.C., S.L.F.); data interpretation (S.T.A., R.N.H., J.P.M., H.D.C.); and preparation and review of the manuscript (S.T.A., H.D.C.,

.P.M., H.D.C.). The study had Institutional Review Board approval and conformed to the Declaration of Helsinki and HIPAA regulation.

REFERENCES

1. Folberg R, Alfonso E, Croxatto JO, et al. Clinically atypicalgranular corneal dystrophy with pathologic features of lat-tice-like amyloid deposits. A study of these families. Oph-thalmology 1988;95:46–51.

2. Konishi M, Yamada M, Nakamura Y, Mashima Y. Immuno-histology of kerato-epithelin in corneal stromal dystrophiesassociated with R124 mutations of the BIGH3 gene. CurrEye Res 2000;21:891–896.

3. Siddiqui N, Afshari NA. The changing face of the genetics ofcorneal dystrophies. Curr Opin Ophthalmol 2002;13:199–203.

4. Holland EJ, Daya SM, Stone EM, et al. Avellino cornealdystrophy. Clinical manifestations and natural history. Oph-thalmology 1992;99:1564–1568.

5. Afshari NA, Mullally JE, Afshari MA, et al. Survey ofpatients with granular, lattice, avellino, and Reis-Bucklerscorneal dystrophies for mutations in the BIGH3 and Gelsolingenes. Arch Ophthalmol 2001;119:16–22.

6. Dolmetsch AM, Stockl FA, Folberg R, et al. Combinedgranular-lattice corneal dystrophy (Avellino) in a patientwith no known Italian ancestry. Can J Ophthalmol 1996;31:29–31.

7. El-Ashry MF, Abd El-Aziz MM, Larkin DF, et al. A clinical,histopathological, and genetic study of Avellino cornealdystrophy in British families. Br J Ophthalmol 2003;87:839–842.

8. Yamamoto S, Okada M, Tsujikawa M, et al. The spectrum ofbeta ig-h3 gene mutations in Japanese patients with cornealdystrophy. Cornea 2000;19:S21–S23.

9. Bron AJ. Genetics of the corneal dystrophies: what we havelearned in the past twenty-five years. Cornea 2000;19:699–711.

0. Korvatska E, Munier FL, Chaubert P, et al. On the role ofkerato-epithelin in the pathogenesis of 5q31-linked cor-neal dystrophies. Invest Ophthalmol Vis Sci 1999;40:2213–2219.

1. Dogru M, Katakami C, Nishida T, Yamanaka A. Alterationof the ocular surface with recurrence of granular/avellinocorneal dystrophy after phototherapeutic keratectomy: reportof five cases and literature review. Ophthalmology 2001;108:810–817.

2. Dinh R, Rapuano CJ, Cohen EJ, Laibson PR. Recurrence ofcorneal dystrophy after excimer laser phototherapeutic ker-atectomy. Ophthalmology 1999;106:1490–1497.

3. Jun RM, Tchah H, Kim TI, Stulting RD, et al. Avellino corneal

4. Lee WB, Himmel KS, Hamilton SM, et al. Excimer laserexacerbation of Avellino corneal dystrophy. J Cataract Re-fract Surg 2007;33:133–138.

5. Roh MI, Grossniklaus HE, Chung SH, et al. Avellinocorneal dystrophy exacerbated after LASIK: scanning elec-tron microscopic findings. Cornea 2006;25:306–311.

6. Aldave AJ, Sonmez B, Forstot SL, et al. A clinical andhistopathologic examination of accelerated TGFBIp deposi-tion after LASIK in combined granular-lattice corneal dys-trophy. Am J Ophthalmol 2007;143:416–419.

7. Banning CS, Kim WC, Randleman JB, et al. Exacerbation ofAvellino corneal dystrophy after LASIK in North America.Cornea 2006;25:482–484.

8. Chiu EK, Lin AY, Folberg R, Saidel M. Avellino dystrophyin a patient after laser-assisted in situ keratomileusis surgerymanifesting as granular dystrophy. Arch Ophthalmol 2007;125:703–705.

9. Wan XH, Lee HC, Stulting RD, Kim T, et al. Exacerbationof Avellino corneal dystrophy after laser in situ keratomileu-sis. Cornea 2002;21:223–226.

0. Streeten BW, Qi Y, Klintworth GK, et al. Immunolocaliza-tion of beta ig-h3 protein in 5q31-linked corneal dystrophiesand normal corneas. Arch Ophthalmol 1999;117:67–75.

1. Yanoff M, Fine BS. Ocular Pathology. St Louis, Missouri:Mosby, 2002:279.

2. Sipe JD. Amyloidosis. Annu Rev Biochem 1992;61:947–975.3. Mahalingam M, Palko M, Steinberg-Benjes L, Goldberg LJ.

Amyloidosis of the auricular concha: an uncommon variantof localized cutaneous amyloidosis. Am J Dermatopathol2002;24:447–448.

4. Lee JB, Choe CM, Kim HS, et al. Comparison of TGF-beta1 in tears following laser subepithelial keratomileusisand photorefractive keratectomy. J Refract Surg 2002;18:130 –134.

5. Kaji Y, Soya K, Amano S, et al. Relation between cornealhaze and transforming growth factor-beta1 after photorefrac-tive keratectomy and laser in situ keratomileusis. J CataractRefract Surg 2001;27:1840–1846.

6. Long Q, Chu R, Zhou X, et al. Correlation betweenTGF-beta1 in tears and corneal haze following LASEK andepi-LASIK. J Refract Surg 2006;22:708–712.

7. Akhtar S, Meek KM, Ridgway AE, et al. Deposits and proteo-glycan changes in primary and recurrent granular dystrophy ofthe cornea. Arch Ophthalmol 1999;117:310–321.

8. Lyons CJ, McCartney AC, Kirkness CM, et al. Granular

corneal dystrophy. Visual results and pattern recurrence after

OPHTHALMOLOGY APRIL 2008

2

3

3

3

3

Ttftfiwp

V

lamellar or penetrating keratoplasty. Ophthalmology 1994;101:1812–1817.

9. Stix B, Leber M, Bingemer P, et al. Hereditary lattice cor-neal dystrophy is associated with corneal amyloid depositsenclosing C-terminal fragments of keratoepithelin. InvestOphthalmol Vis Sci 2005;46:1133–1139.

0. Takacs L, Boross P, Toszser J, et al. Transforming growthfactor-beta induced protein, BetaIG-H3, is present in de-graded form and altered localization in lattice corneal dys-trophy type I. Exp Eye Res 1998;66:739–745.

AJO History of O

Early To

he first real attempts at quantitating intraocular pres-sure were made in 1862 by Albrecht von Graefefollowed by Donders, Snellen and several other oph-

halmologists. These indentation tonometers were plagued byriction and difficulty in repeatable measurements. A manruly ahead of his time, Alfred Weber presented in 1867 therst of the applanation tonometers and explained why itould permit a more accurate means of assessing intraocular

ressure. Although he actually designed an instrument, it was O

AVELLINO CORNEAL DYSTROPHYOL. 145, NO. 4

1. Nakamura K. Interaction between injured corneal epithelialcells and stromal cells. Cornea 2003;22:S35–S47.

2. Patel SV, Erie JC, McLaren JW, Bourne WM. Confocalmicroscopy changes in epithelial and stromal thickness up to7 years after LASIK and photorefractive keratectomy formyopia. J Refract Surg 2007;23:385–392.

3. Dawson DG, Edelhauser HF, Grossniklaus HE. Long-termhistopathologic findings in human corneal wounds afterrefractive surgical procedures. Am J Ophthalmol2005;139:168 –178.

halmology Series

meters

ever made commercially available. It was Maklakov, workingithout knowledge of Weber, who introduced the first prac-cal applanation tonometer in 1885. This is still in use in partsf the old Soviet empire today. The Academy of Ophthal-ology has three Maklakov tonometers manufactured in the

960s in its Museum collection.

Provided by Steven A. Newman, MD, of the Cogan

pht

no

,nwtiom1

phthalmic History Society.

WORSENING AFTER LASIK 661