British Journal of Ophthalmology 1995; 79: 257-260

Comeal epithelium following penetratingkeratoplasty

Kazuo Tsubota, Yukihiko Mashima, Hiroyuki Murata, Masakazu Yamada, Naoki Sato

Department ofOphthalmology, TokyoDental College, Chiba,JapanK Tsubota

Department ofOphthalmology, KeioUniversity School ofMedicine, Tokyo,JapanK TsubotaY MashimaH MurataM YamadaN Sato

Correspondence to:Kazuo Tsubota, MD,Department ofOphthalmology, TokyoDental College, IchikawaGeneral Hospital, 5-11-13Sugano, Ichikawa, Chiba,272, Japan.

Accepted for publication20 October 1994

x

AbstractAims-This study was designed to observeany changes to the corneal epitheliumafter penetrating keratoplasty.Methods-The corneal epithelia of 26patients were observed by specularmicroscopy 1 week, 1 month, 3 months,and 6 months following penetratingkeratoplasty.Results-After re-epithelialisation was

confirmed by biomicroscopy 1 week aftersurgery, specular microscopy revealedmany abnormal cells, including spindleshaped cells, nucleated cells, large cells,as well as irregular cell configurations.Although these abnormal findings tendedto decrease with time, they were stillpresent in some cases as much as 6months postoperatively. Computerisedmorphometric analysis yielded mean cellareas of 1121 (SD 168) pLm2, 1139 (675)pum2, 1712 (496) pim2, and 1400 (377) jim2 at1 week, 1 month, 3 months, and 6 monthsrespectively, all significantly greater thanthat of age matched controls (710 (151)pum2). The shape factor decreased withtime, but was still greater than the controllevel at 6 months.Conclusion-This study demonstratesthat epithelial abnormalities persistlonger than expected after penetratingkeratoplasty, and that these subtlechanges can be detected by specularmicroscopic observation, potentiallyallowing for modification and enhance-ment ofthe wound healing process.

(Br_r Ophthalmol 1995; 79: 257-260)

X2

Idx =x2X2 (y2-yl)dxx

y Flat rate = 'max - 'min'max +lmin

(0=1)

Figure 1 The definition offlat rate.

Refractory superficial punctate keratopathy,irregular astigmatism, persistent healingproblems, and recurrent erosions may beencountered in the corneal epithelium follow-ing penetrating keratoplasty (PK). When theocular surface is compromised, poor epithelialrecovery may be the major cause ofgraft failure,which may then require keratoepithelioplasty.1In many cases, the epithelial rejections andincreased rate of infection are still the majorcomplications of PK.23The epithelium after PK has been reported

to be abnormal in cellular shape,4 oxygenconsumption,5 sensitivity,6 and ultrastruc-ture.7 Only a few studies have examinedepithelial changes following PK.4 8-12 No studyhas ever employed computerised morpho-metric analysis. Although the corneal endothe-lium has been widely investigated using thespecular microscope and morphometric analy-sis, the comeal epithelium following PK hasheretofore not been suitable for such study.The recent development of the specular

microscope contact lens (SM lens) makes itpossible to observe the corneal epithelium atthe cellular level.'3 14 The technique is non-invasive and the corneal epithelium can befollowed prospectively. This study follows theprocess of recovery of the comeal epitheliumafter using specular microscopy and aftermorphometric analysis as first described byMathers et al.15

Materials and methods

PATIENT SELECTIONTwenty patients (10 male, 10 female) at KeioUniversity and six patients (three male, threefemale) at National Tochigi Hospital on whomPKs were performed were enrolled in thisstudy. The average age was 56-3 (19-5) years(range 12 to 84). The preoperative diagnoseswere: herpetic keratitis (n= 7), interstitialkeratitis (n= 5), comeal haziness of unknownaetiology (n=4), keratoconus (n=4), bullouskeratopathy following cataract extraction(n=2), leukoma due to injury (n= 1), gela-tinous drop-like dystrophy (n= 1), band kerato-pathy (n= 1), and macular dystrophy (n= 1).As controls, age and sex matched phakic

myopic patients, five male and five female,average age 53-5 (14-2) years and aphakicpatients, five male and five female, average age59.3 (12-5) years were used. The aphakicpatients had the same level of steroid eyedropsas well as antibiotic eyedrops and the cornealepithelium was observed at 1 month aftercataract surgery with phacoemulsification.

257

on April 28, 2021 by guest. P

rotected by copyright.http://bjo.bm

j.com/

Br J O

phthalmol: first published as 10.1136/bjo.79.3.257 on 1 M

arch 1995. Dow

nloaded from

Tsubota, Mashima, Murata, Yamada, Sato

Fzg 2A

~~~~~~~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ .. :........ET!

............ .*~~~~~~~~~~~~~~~~~~~~~~~:..,1.....A, ~ ~I

Fsig 2B

Fig ZC

rPg Zu

Figure 2 Elongated comeal epithelial cells following penetrating keratoplasty. All barsequal 100 um. (A) 1 Week after surgery (54-year-oldfemale, original disease wasmacular dystrophy). Many small elongated cells were clearly observed. (B) 1 Week aftersurgery (55-year-old male, original disease was leukoma). Elongated cels were observed.(C) 1 Week after surgery (45-year-oldfemale, original disease was herpes simplex). Therewere many elongated cells observed. (D) 1 Month after surgery (same case as Fig 2C).There are still large, elongated cells observed.

SURGICAL METHODSDonor buttons were taken from the epithelialside of preserved eyes using a Draeger electrictrephine. The same or a slightly smallertrephine was used for the recipient cornea. Theaverage preservation periods were 10 5 (5-3)hours, during which the corneal epitheliumremained intact, as determined by bio-microscopy. The donor cornea was tem-porarily sutured with eight 9-0 silk interruptedsutures, and then sutured with 10-0 nylonrunning sutures. During the procedure,hyaluronic acid was used to protect theendothelium. The 9-0 silk sutures wereremoved at the end of the surgery.

During the follow up, topical steroids(0 10% dexamethasone, four times daily) andantibiotics (gentamicin, four times daily) wereused for up to 1 month. If epithelial rejection,chronic keratitis, or any epithelial problemsoccurred, the patients were excluded from thestudy.As controls, 20 age and sex matched normal

individuals (10 male, 10 female, average age54-8 (13-6) years), were recruited for thisstudy.

SPECULAR MICROSCOPYA special contact lens was used to take epithe-lial pictures13 14 at 1 week, 1 month, 3 months,and 6 months after surgery. In each patient,several pictures were taken of the centralcornea, and of the peripheral corneal epithe-lium when possible. It is not certain, but prob-able that determining factors of when this ispossible include the status of the tear, status ofthe mucin, and status of the cornea shape.Kodak Tri-Max ASA 400 film was used andthe pictures were magnified 200 times.

Epithelial pictures were analysed qualita-tively and quantitatively. The presence orabsence of the following abnormal cells weredetermined in each of the subjects.9 14

(1) Spindle-shaped cell: cells in which thelongest axis is more than twice as long as theshortest axis of the cell.

(2) Extra large cell: cells with area morethan 2000 pLm2.

(3) Nucleated cell: cells with a centralround reflex.

(4) Extra bright cell: cells with especiallybright specular reflexes.

(5) Colony of small cells: a group of smallcells each with area less than 500 pLm2.

Quantification of these abnormalities is diffi-cult, so only the presence or absence was noted.For the quantitative analysis, morphometricanalysis of the mean cell area and coefficient ofvariation of the mean cell area (CV) and the flatrates were calculated (Fig 1). In nine out of 26post PK and 10 of each control cases, morpho-metric analysis was performed using a com-puterised system (Sun Contact Lens, Osaka,Japan). For each analysis, at least 50 cells weredigitised for calculation. If the epitheliumpicture was obscure for this analysis, the caseswere excluded from the study. We selected ninecases where good quality pictures were clear atany observation point.

258

...

W,

on April 28, 2021 by guest. P

rotected by copyright.http://bjo.bm

j.com/

Br J O

phthalmol: first published as 10.1136/bjo.79.3.257 on 1 M

arch 1995. Dow

nloaded from

Corneal epithelium following penetrating keratoplasty

307

a)

co0

0

z

20

10

o1 week 1 month 3 months 6 months

Time after surgeryFigure 3 The time course of epithelial abnormalitiesfollowing PK The presence of elongated cells decreased withtime, but did not disappear in 6 months. The large cellspeaked at 1 month, followed by a gradual decrease. Theirregular pattern could be observed throughout theobservation period. 0 = irregular pattern; *= elongated cell;A= nucleated cell; O=large cell.

Fig 4A

Fig 4B

Figure 4 Extraordinarily large corneal epithelial cells following PK All bars equal100 ,um. (A) 1 Month after surgery (56-year-old male, original disease was unknownhaziness). Extra large cells were observed. (B) 3 Months after surgery (44-year-old female,original disease was herpes simplex). The cells were slightly smaller than at 1 month.(C) The epithelium surrounding the suture (same case as Fig 4A). The epithelial cells wereextraordinarily large.

ResultsThe corneal epithelium following PK showed avariety of changes. The most prominentfeature was the presence of spindle-shapedcells, which ranged from moderately to veryelongated and were observed mainly in thecentral cornea. The elongation was greater inthe first few weeks after surgery and thengradually decreased. A typical case is shown inFigure 2A-D. These elongated cells could befurther distinguished on the basis of theirspecular reflexes: bright, medium, and dark.These spindle-shaped cells were observed in22 of 26 cases in the first postoperative weekbut only in 14 cases after 6 months (Fig 3).

Extra large cells, up to 8000 pLm2 near thesuture (Fig 4A-C) were observed mainly in theperiphery of post PK corneas. In contrastwith the spindle-shaped cells, most of these didnot appear until 1 month after the surgery(Table 1).Nucleated cells were observed in all cases

with the nucleus appearing as a bright spot inthe dark cells and as a dark spot in the brightcells. Nuclei were noted in normal shapedcells, as well as spindle-shaped and extra largeones. The time course of these cells are shownin Figure 3. Colonies of small cells were some-times seen in these series.Morphometric analysis of the post PK

epithelia showed considerable changes in thewound healing process. The mean cell areaincreased between the first week and 3 monthsand then began decreasing, yet still not return-ing to normal at 6 months (Table 1).Compared with the control corneal epithelialcell area of 710 (151) pum2 in the phakiccontrol group and 721 (126) jim2 in theaphakic control group, the corneal epitheliumwas still large.The CV of the mean cell area showed no

significant trend.The flat rate showed an increased ratio of

0-86 (0 06) in the first week, and at 6 monthswas still 0-52 (0-09) greater than the 0 30(0.04) and 0-31 (0.04) of the controls.

DiscussionThe most prominent epithelial change was theappearance of spindle-shaped cells, which mayresult from migration from the periphery. Suchelongated cells have been reported in postPK8 9 as well as in various cell cultures.16 17 Insome cases, these elongated cells persisted formore than 6 months. Since the corneal epithe-lium is regenerating every day, with strongmitotic activity, wound healing followingsurgical intervention should not be expected tocontinue for such a long period. Dohlman hasreported that the donor epithelium disappearswithin 6 to 10 days after surgery. 18 If this is thecase, the epithelium should have been entirelyreplaced with recipient cornea within 2 weeks.However, Silverstein et al have demonstratedthe survival of donor epithelium for up to6 months.19 During the dilution of donorepithelium, subclinical rejection or naturalreplacement by the recipient epitheliumoccurs. In this instance, accelerated epithelial

259

0

zz--::4

on April 28, 2021 by guest. P

rotected by copyright.http://bjo.bm

j.com/

Br J O

phthalmol: first published as 10.1136/bjo.79.3.257 on 1 M

arch 1995. Dow

nloaded from

Tsubota, Mashima, Murata, Yamada, Sato

Table 1 Morphometric analysis of corneal epithelium following penetrating keratoplasty

1 Week I Month 3 Months 6 Months Phakic control Aphakic control

Mean cell area (,m2) (SD) 1121 (168)* 1139 (675) 1712 (496) 1400 (377) 710 (151) 721 (126)CV (C%) (SD) 31-5 (9 2) 31-0 (8-1) 32-2 (8 6) 32-1 (7 5) 29-2 (7 9) 32-5 (8-1)Flat rate (SD) 0-86 (0 006)t 0.70 (0-15) 0-52 (0-16) 0-52 (0-09) 0 30 (0-04) 0-31 (0 04)

*p<0.05 significantly different from control; tp<0 01.

wound healing may indeed continue. Lemphas also reported the presence of elongatedcells following PK,10 which was attributed tothe sutures. In this study, all of the corneas hadsutures, and we had similar findings. Thechanges following suture removal should bestudied in the future.

In addition to the presence of spindle-shaped cells, several other epithelial changeswere noted, including extra large cells,nucleated cells, and small cell formations.These changes have also been observed inother conditions, such as keratoconus, epithe-lial defects, diabetic cornea, post epikera-tophakia, and in normal contact lenswearers.20-22 These epithelial changes maythus be non-specific qualitative reactions toany corneal insult, although they may differquantitatively depending on the disease orprocedure.

Corneal hypoaesthesia following PK is wellknown.1 2 Denervation of the cornea decreasesmitotic activity in the epithelium, retardshealing, and impairs cellular adhesion.23-26This may also explain why epithelial cellschange shape.The specular microscope has been widely

used to study the corneal endothelium follow-ing PK.27-29 The numbers of endothelial cellsdecrease while the mean cell area and the asso-ciated coefficient of variation increase over aperiod of 3 years but most prominently withinthe first month. In contrast, epithelial studiesfollowing PK have not been performed exten-sively, primarily because of the difficulty inphotographing the epithelium. The recentdevelopment of the SM lens has greatly facili-tated this process,14 permitting non-invasive,long term monitoring of the epithelium. Thismethod may be used to evaluate subtleepithelial changes at the cellular level enhanc-ing understanding of the healing process andpotentially enabling its modification. Thistechnique is clearly also applicable to the studyof a wide range of corneal diseases, and hasalready been employed in our laboratory toevaluate keratoconus.22This study was supported by a grant from the KowaFoundation. This paper was presented in part at the 13thJapanese Corneal Conference, Hokkaido, Japan, February,1989.

1 Thoft RA. Keratoepithelioplasty. Am Ophthalmol 1984;97: 1-6.

2 Alldredge OC, Krachmer JH. Clinical types of cornealtransplant rejection. Arch Ophthalmol 1981; 99: 599-604.

3 Fong LP, Ormerod LD, Kenyon KR, Foster CS. Microbialkeratitis complicating penetrating keratoplasty.Ophthalmology 1989; 95: 1269-75.

4 Brown AC, Lohman LE, Rao GN, Aquavella JV. Humancorneal epithelial appearance following penetratingkeratoplasty. Invest Ophthalmol Vis Sci 1982; (suppl): 24.

5 Vannas A, Holden BA, Sweeney DF. Epithelial metabolismofthe corneal graft is abnormal. BrJr Ophthalmol 1987; 71:593-97.

6 Rao GN, John T, Ishida N, Aquavella JV. Recovery ofcorneal sensitivity in grafts following penetrating kerato-plasty. Ophthalmology 1985; 92: 1408-11.

7 Kinoshita S, Friend J, Thoft RA. Sex chromatin of donorcorneal epithelium in rabbits. Invest Ophthalmol Vis Sci1981; 21: 434-41.

8 Khodadoust AA, Silverstein AM. The survival and rejectionof epithelium on experimental corneal transplants. InvestOphthalmol Vis Sci 1969; 8: 169-79.

9 Tsubota K, Yamada M, Naoi S. Specular microscopicobservation of corneal epithelial abnormalities.Ophthalmology 1991; 98: 184-91.

10 Lemp MA, Mathers WD. Corneal epithelial cell movementin humans. Eye 1989; 3: 438-45.

11 Lemp MA. The surface of the corneal graft: in vivo colorspecular microscopic study in the human. Trans AmOphthalmol Soc 1990; 87: 619-57.

12 Marechal-Courtois CH, Delcourt JCL. Wide-field specularmicroscopy of corneal epithelium. CLAO J 1986; 12:165-72.

13 Tsubota K. A contact lens for specular microscopicobservation. Am J Ophthalmol 1988; 106: 627-8.

14 Tsubota K, Yamada M, Naoi S. Specular micrscopic obser-vation of normal human corneal epithelium.Ophthalmology 1992; 99: 89-94.

15 Mathers WD, Lemp MA. Vortex keratopathy of the cornealgraft. Cornea 1991; 10: 93-9.

16 Jumblatt MM, Neufeld AH. A tissue culture assay ofcorneal epithelial wound closure. Invest Ophthalmol Vis Sci1986; 27: 8-13.

17 Tsai RJF, Tseng SCG. Substrate modulation of culturedrabbit conjunctival epithelial cell differentiation andmorphology. Invest Ophthalmol Vis Sci 1988; 29: 1565-76.

18 Dohlman CH. On the fate of the corneal graft. ActaOphthalmol 1975; 35: 286-302.

19 Silverstein AM, Rossman AM, Leon AS. Survival of donorepithelium in experimental corneal xenografts. Am JOphthalmol 1970; 69: 448-53.

20 Tsubota K, Chiba K, Shimazaki J. Corneal epithelium indiabetic patients. Cornea 1991; 10: 156-60.

21 Tsubota K. Corneal epithelium following intraepikera-tophalia. J Cataract Refract Surg 1991; 17: 460-5.

22 Tsubota K, Mashima Y, Murata H, Sato N, Ogata T.Corneal epithelium in keratoconus. Cornea 1995; 14:77-83.

23 Mathers WD, Jester JV, Lemp MA. Return of humancorneal sensitivity after penetrating keratoplasty. ArchOphthalmol 1988; 106: 210-1.

24 Mishima S. The effects of the denervation and the stimula-tion of the sympathetic and the trigeminal nerve on themitotic rate of the corneal epithelium in the rabbit. Jpn JOphthalmol 1957; 1: 65-73.

25 Beuerman RW, Schimmelpfennig B. Sensory denervationof the rabbit cornea affects epithelial properties.Exp Neurol 1980; 69: 196-201.

26 Beuerman RW, Schimmelpfennig B. Anatomy of thedenervated corneal epithelium. Invest Ophthalmol Vis Sci1970; (suppl): 126.

27 Matsuda M, Yee RW, Glasser DB, Geroski DH, EdelhauserHF. Specular microscopic evaluation of donor cornealendothelium. Arch Ophthalmol 1986; 104: 259-62.

28 Bourne WM, Doughman DJ, Lindstrom RL, Kolb MJ,Mindrup E, Skelnik D. Increased endothelial cell lossafter transplantation of corneas preserved by modifiedorgan-culture technique. Ophthalmology 1984; 91: 285-9.

29 Matsuda M, Suda T, Manabe R Long-term observations ofthe graft endothelium with different postoperativecourses. Jpn Y Ophthalmol 1983; 27: 556-66.

260

on April 28, 2021 by guest. P

rotected by copyright.http://bjo.bm

j.com/

Br J O

phthalmol: first published as 10.1136/bjo.79.3.257 on 1 M

arch 1995. Dow

nloaded from