recurrent erosion of the cornearecurrent erosion ofthe cornea 87 in microformerosion anda larger...

Brit. J. Ophthal. (I976) 6o, 84

Recurrent erosion of the cornea

NICHOLAS BROWN AND ANTHONY BRONFrom the Institute of Ophthalmology, Judd Street, London

In previous studies of superficial comeal disorders(Bron and Brown, 197I) and of superficial lines ofthe cornea (Brown and Bron, I976), we wereimpressed by the frequency of recurrent erosionof the cornea. In this study we have set out toexamine patients presenting with recurrent erosionof the cornea to establish the type and incidenceof the superficial corneal disorders, and to comparethis incidence with that in a control group ofsubjects. Some experience of management ofthese patients is also reported.

Material and methodsMATERIAL

Patients presenting at the Casualty Department atMoorfields Eye Hospital with suspected recurrenterosion of the cornea were referred by the casualtyofficers and examined by one of us. The patient wasadmitted to the study if the diagnostic criteria weresatisfied. These criteria were:

i. A history of initial trauma or abrasion followed bya recurrence of pain with healing, followed by aclinical recurrence at the time of presentation andassociated at some time with the clinical signs ofrecurrent erosion.

2. A typical clinical history of at least two attackssuggestive of recurrent erosion (pain on wakingor being awakened, difficulty in opening the eyelid, watering, light sensitivity), without a historyof trauma but associated with clinical signs ofrecurrent erosion at some time.

For each group, the clinical signs of recurrent erosioncould include not only frank epithelial loss, but alsofocal epithelial oedema, epithelial bulla formation, andepithelial cyst formation in varying combinations. Inthis way, 8o patients were admitted to the study withages ranging from 24 to 73 years.The control group consisted of 2oo patients aged

between 2o and 85 years, who had attended clinics atMoorfields but were not complaining of any symptomsreferrable to the cornea, and had no history of cornealdisease.

METHODS

A proforma was completed for each patient; this con-tained a full clinical history (including details of the

Address for reprints: N. Brown, Institute of Ophthalmology, JuddStreet, London WCiH gQS

menstrual cycle and use of the contraceptive pill) anddetails of the examination of both eyes, including lids,tear films, corneae, and anterior chambers. The comeaewere then re-examined after staining with fluoresceinand with bengal rose. The comeal signs were recordedwith the Zeiss photo slit lamp and with the macrocamera (Brown, I970) at IO X magnification.Each patient was re-examined at least once after

complete healing of the erosion and both corneae wereinspected for superficial dystrophies. The ocular ten-sions were measured at this time.

Despite the traditional name for the disorder, it isrecognized that frank epithelial loss is not required toestablish the diagnosis and that focal epithelial oedemaor bulla formation with or without a breach in theepithelium, and/or the presence of epithelial cysts, arein keeping with the diagnosis when associated with thetypical clinical history. The patients were classed ashaving either macroform or microform recurrenterosion by the criteria of Chandler (I945). Microformrecurrent erosion was recognized by intraepithelialmicrocysts with a minor break in the epithelium (Fig. i).In macroform recurrent erosion, a large area of epithe-lium was separated from the cornea (Fig. 2).

TREATMENT AND FOLLOW-UP

The initial treatment was decided by the CasualtyOfficer and consisted of either chloramphenicol dropsor chloramphenicol ointment with hyoscine 0 5 percent drops, and padding for the eye. Patients with largebags of loose epithelium were treated initially bydebridement, the loose corneal epithelium being care-fully removed at the slit lamp with a Bard Parker no. I5blade, using centripetal strokes to avoid stripping offan unnecessarily large area of epithelium. The eye wasthen treated with drops or ointment and padded asabove.

Eyes which failed to heal on medical treatment andproduced redundant loose epithelium were debridedas above. In those eyes which still failed to heal on thisregime, the debridement was repeated and followed bycarbolization with IOO per cent phenol applied at theslit lamp with a small stick to the dried cornea. Carewas taken to avoid carbolizing the pupillary area of thecornea, but the phenol was otherwise applied to asmuch of the affected area as possible.Once healing was complete the patients were given

sodium chloride 5 per cent ointment to apply eachevening immediately before going to bed. This treat-ment was continued for from 3 to i8 months. Treatmentof longer duration was given in patients who found that

on June 7, 2020 by guest. Protected by copyright.

http://bjo.bmj.com

/B

r J Ophthalm

ol: first published as 10.1136/bjo.60.2.84 on 1 February 1976. D

ownloaded from

http://bjo.bmj.com/

Recurrent erosion of the cornea 85

they suffered relapses if they omitted to apply theirointment.

Treatment was eventually withdrawn from allpatients still attending at i8 months and the relapse ratewas compared with that while on treatment.

Results

AGE AND SEX DISTRIBUTION (Table I)

The age and sex distribution of the patients isshown in Table I, together with the incidence ofmacroform and microform recurrent erosion; it iscommon for both forms to occur in the same indi-vidual at different times (3I per cent of patients).Bilateral recurrent erosion is relatively uncommon(io per cent of patients).

CAUSES OF INITIAL ABRASION (Table II)

The initial abrasion was frequently spontaneous(40 per cent of patients) and the patient was unableto remember any trauma having preceded it. Ofthe identifiable causes (Table II), the most commonwas a scratch by a finger nail, scratches from ababy's finger being particularly prevalent amongwomen. Among men, the causes were also relatedto their occupations, and metallic objects such asscrewdrivers were described as causative. Twocases were seen after the removal of corneal foreignbodies. In the two hereditary cases the patientshad no recollection of trauma.The microform type was the most common in

the spontaneous cases (Table II), whereas macro-form and combined macroforrn and microformerosions were more commonly of traumatic origin.In the eight bilateral cases trauma was reported

Table I Age and sex distribution of 8o patients*

Age group (years) No. of patients

21-30 13

3 I-40 2241-50 22

5I-60 i86I + 5At presentation (mean) 42Percentage of women 48

men 52

Type of recurrent erosion No. Percentage

Macroform only IO 13Microform only 45 56Both types in same individual 25 3'Bilateral cases 8 Io

The youngest was 24, the eldest 73 years

FIG. I Microform recurrent erosion. Macrophotographby retroillumination. Bar gauge 500 [±m

FIG. 2 Macroform recurrent erosion. Stained withfluorescein to show intraepithelial microcysts. Bargauge Ioo .m


http://bjo.bmj.com

/B

r J Ophthalm


ownloaded from

http://bjo.bmj.com/

86 British Yournal of Ophthalmology

Table II Causes of abrasion in 8o patients

Abrasion No. Percentage

Spontaneous 32 40Finger nail I8 22Animal claw 3 4Vegetable matter, for example,

twig I2 15Paper or card 7 9Metallic object 5 6Corneal foreign body 2Contact lens I

Relation of Percentage Percentageerosion of oftype to No. traumatic No. spontaneoustrauma cases cases

Macroformalone 8 10 2 3

Microformalone 20 25 25 31

Combinedform 20 25 5 6

Table III Recurrences in 8o patients

Time of recurrence Percentage

At about the time of waking 76Before waking (awakened by pain) 30After waking 40Difficulty in opening the eye IOAt other times of the day 24

Precipitation

Nil 86Eye rubbing IOBarbiturates 2-5Menstruation I

by only three, and in two of thesetrauma had been to one eye only.

patients, the

FREQUENCY AND PRECIPITATION OF RECURRENCE

(Table III)The interval before the first recurrence after theinitial abrasion varied from 2 days to i6 years,but it may be doubted whether this late recurrencewas related to the initial abrasion after so long atime. The mean interval was i8 months, but thiswas weighted by a few very long durations, and 63per cent occurred within 4 months.

Recurrences, both macro- and microform, oc-curred most commonly at around the time ofwaking (Table III). Some patients were able to

give a clear history of being awakened by the painand, in others, the pain occurred after waking.An interesting symptom was difficulty in openingthe lids; some patients needed to use their fingersto open their eye, and this was commonly followedimmediately by a recurrence.

Precipitating factors were sought in the patient'shistory, but none was elicited in 86 per cent. Eyerubbing was admitted as a cause in io per centand the use of barbiturates was implicated in 3per cent. These patients experienced recurrenceswhenever they resumed taking barbiturates.

Altogether 23 menstruating women were studied,and in only one of these was there a clear historyof recurrences related to the time of onset of eachmenstruation.The general health of the patients appeared to

be unrelated; the only finding which was in any

way consistent was that of a history of atopy in5 per cent. A family history suggesting dominantinheritance of recurrent erosion was obtained intwo patients (3 per cent).The frequency of recurrence was very variable,

ranging from a minor recurrence each morning tomajor recurrences separated by many months. Thefrequency of recurrence varied between patientsand at different periods in the same individual.The recurrences lasted from i to 4 hours in the

microform condition, and from i to 2i days in themacroform condition, which included two cases inwhich the healing was delayed by a secondarybacterial infection.

SYMPTOMS (Table IV)The symptoms of recurrence (Table IV) are

largely as anticipated. Blurring of vision variedfrom absent or minimal in many eyes with micro-form erosion to severe in macroform erosions.

SIGNS

When the eyes were examined at the time of activeerosion, a small epithelial break (Fig. i) was seen

Table IV Symptoms of recurrent erosion in8o patients

PercentageSymptoms of patients

Pain 93Severe pain 10Blurring of vision 6iWatering 72Photophobia 57Redness 10


http://bjo.bmj.com

/B

r J Ophthalm


ownloaded from

http://bjo.bmj.com/


in microform erosion and a larger break, commonlywith loose baggy epithelium around it, in macro-form erosion (Fig. 2). The common site was central,just below the pupil. In both forms microcystswere present in the surrounding epithelium whichusually stained well with fluorescein (Fig. 2). Somedegree of epithelial and stromal oedema was com-mon. The conjunctiva was injected and sometimeschemosed, especially in cases of macroform type.The anterior chamber was normally quiet, butsometimes showed a few cells in macroform cases.

In the healed state the eyelids and conjunctivawere unremarkable and an abnormal intraocularpressure was not found. Tear film thinning,demonstrated by the break-up of the fluoresceinfilm, was found in I4 per cent. This was usuallyrelated to an area where the epithelium was unevenbecause of a superficial corneal dystrophy orepithelial cysts. However, the volume of tears didnot appear to be abnormal and Schirmer's testgave results within the normal range.

SUPERFICIAL CORNEAL DYSTROPHIES(Tables V, VI, and VII)(The term superficial corneal dystrophy will beused for the group of changes occurring in theregion of the corneal epithelium and defined below.Evidence for a genetic aetiology is currently small.See discussion.)

Superficial comeal dystrophies were a frequentfinding (Table V) and commonly bilateral. Theoccurrence of the dystrophies was related to theage of the patient (Table VI). No patient seen inthe present study was below the age of 30 years,but we have occasionally seen affected patientsbelow this age. The frequency increases with age,reaching 78 per cent of patients in the 5I to 6oyear age group.The incidence of superficial corneal dystrophies

in the patients is compared with that in the controlgroup in Table IX. In the unilateral cases, thedystrophy was more frequently found in theunaffected eye than in the eye affected by recurrentcorneal erosion, from which it may be deducedthat the corneal erosion had been responsible forerasing clinical signs of the dystrophy.The types of dystrophies which were found

(Table V), were most commonly the various super-ficial corneal lines, which are listed separately inTable VII. Bleb dystrophy was also common,less commonly found was the more pronouncedlacunar dystrophy of Bietti (I965) which we regardas synonymous with the 'map change' of Guerry(i965).The individual superficial comeal lines have

been described in detail (Brown and Bron, 1976).The most common type is the now familiar 'finger-

Table V Incidence of superficial corneal dystrophies(including all types of lines)

Dystrophy Percentage*

All patients with superficial dystrophyor lines 59Bilateral 66In affected eye only 8In unaffected eye I2

Fine superficial lines 46Bleb dystrophy 28Bietti's lacunar dystrophy (map

dystrophy) 5

*Of 8o patients

Table VI Age distribution of patients with super-ficial corneal dystrophies (excluding demarcation lines)

Age Patientsgroup Percentage(years) No. With dystrophy of age group

21-303 1-404I-50

5I-6+6I +

13 0

22

22

I85

9'5'43

0

4'68786o

Table VII Incidence of fine superficial corneal lines

Corneal lines No. Percentage*

Patients affected 37 46Fingerprint lines 21 26Demarcation lines 17 21Superficial ring lines ISuperficial corneal crazing 2Mare's-tail lines -Tram lines 2Bleb dystrophy and finger-

print lines in association 14 17

*Of 8o patients

print line' (Fig. 3) first described by Guerry(1950). These are fine refractile lines which arebest seen by retroillumination. They are arrangedin a near parallel whorl-like pattern, which givesthem their resemblance to a fingerprint. Tear filmthinning is commonly seen over the line. Twoother fine lines which have the same retroillumina-tion appearance as the fingerprint line, but differin arrangement, are the superficial ring lines andsuperficial comeal crazing. The superficial ringlines form complete rings and the superficial


http://bjo.bmj.com

/B

r J Ophthalm


ownloaded from

http://bjo.bmj.com/

88 British Journal of Ophthalmology

FIG. 5 Demarcation lines marking limit of area ofbleb dystrophy

FIG. 3 Fingerprint lines. Seen by retroillumination offundus. Bar gauge i 000 l±m

FIG. 4 Cogan's epithelial microcysts with adjacentdemarcation line (arrowed). Bar gauge 500 I'm

corneal crazing is found near the limbus anddiffers from the fingerprint line in having a branchedpattern. Neither of these lines was common, butsince each probably represents variants of thefingerprint line, its presence is considered to berelated to recurrent erosion.

'Demarcation lines' (Fig. 4) have a superficialresemblance to fingerprint lines and are oftenfound in the same eye where they separate areasaffected by fingerprint lines, or by bleb dystrophy,from clear areas (Fig. 5). These lines often formcomplete rings of about 5 mm in diameter. Theyoccur in eyes which have had recurrent erosion andare not seen in the fellow eye unless that too hasbeen affected by erosion.

'Mare's-tail lines' (Fig. 6) are recognizablydifferent from fingerprint lines, being seen bestby focal illumination. These lines are bunchedtogether in places and spread out in others; theyshow branching and crossing, which do not occurin fingerprint lines. They appear to be more super-ficial than the other lines and have a significanteffect in thinning the tear film. Although only onepatient is reported with mare's-tail lines in thisseries, we have seen them in other patients withrecurrent erosion and in association with Bietti'slacunar dystrophy which is also related to recurrenterosion.The remaining lines are 'tram lines', described

by Brown and Bron (1976). These rare lines runa long, straight, nearly vertical course, and haverefractile edges with the appearance of paralleltram lines. They were found in two patients inthlis series and in none of the controls. We are


http://bjo.bmj.com

/B

r J Ophthalm


ownloaded from

http://bjo.bmj.com/


FIG. 6 Mare's-tail lines by focal illumination.Bar gauge 500 ,um

unable to say whether their presence is related torecurrent erosion.

Dystrophic changes consisting of fine superficialblebs (Bron and Brown, 1971) were found in28 per cent of patients. The blebs are seen only byretroillumination (Fig. 7) when they have theappearance of pebbled glass. They are subepithelialand do not affect the tear film. The blebs occurredin groups, sometimes covering large areas of thecornea, and were sometimes arranged in a charac-teristic hexagonal net pattern (Bron and Jones,I969). Blebs were often absent from the regionjust below the pupil in eyes which had sufferedrecurrent erosion and in these eyes their limit was

characteristically defined by a demarcation line(Fig. 5).A less common finding (in 5 per cent of the

patients and in none of the controls) was the super-

ficial corneal dystrophy described by Bietti (I965)as lacunar dystrophy of Bowman's membrane(Fig. 8), in which clear lacunae up to 2 mm indiameter are seen on the surface of Bowman'smembrane. The areas between the lacunae are

opalescent and the edges of the lacunae are definedby a denser white line which may have a rolledappearance in focal illumination. The irregularpattern produced by the lacunae gives a map-likeappearance, so that it is probably the same condi-tion as that described by Guerry (I965) andthought by him to be a component of the epithelial

microcystic disorder described by Cogan, Donald-son, Kuwabara, and Marshall (I964). Numerousother workers have since observed these map-likechanges (Wolter and Fralick, I966; Levitt, 197I;Trobe and Laibson, 1972; Luxenberg, Friedland,and Holder, 1975).

MICROCYSTS (Table VIII)Intraepithelial microcysts (Fig. 9) were a commonfinding in healed eyes, being present in 59 per centof the patients (Table VIII), but they were seen inonly 1-5 per cent of the controls (Table X). Theyoccurred in conjunction with other recognizabledystrophies, but were the only recognizable cornealsigns in 21 per cent. This left only i I per cent ofpatients in whom no signs, either cysts or dystro-

Table VIII Microcysts

Microcysts Percentage

Patients affected (total) 59Erosion-affected eye only 49Unaffected eye 24Cogan's microcysts 6Microcysts as only recognizable signs 2aPatients with neither dystrophy nor

microcysts I I


http://bjo.bmj.com

/B

r J Ophthalm


ownloaded from

http://bjo.bmj.com/

go British Journal of Ophthalmology

FIG. 7 Bleb dystrophy by retroillumination offundus. Bar gauge 500 uLm

phies, were visible in the healed state. In mostpatients a retrospective diagnosis of recurrenterosion could be made with confidence whenexamining the patient in the healed state. Clinicalsymptoms, however, were always associated withclinical signs.The microcysts were seen best by retroillumina-

tion, but sometimes contained opaque matter whichmade them visible by focal illumination. (This wasmost marked in those opaque white cysts having theappearance of those described by Cogan and others,1964.) The microcysts, measured from the cornealmacrophotographs, varied from 20 ,um to 74 uim indiameter. The more superficial cysts, which some-times showed a centrally umbilicated appearance,commonly stained with fluorescein, and thosewhich did not stain showed as dark spots in thefluorescein, containing tear film. It was noticedthat the cysts which were first seen after healing ofan erosion showed the same distribution of lightwithin them as that of the background when viewed

by marginal retroillumination, which has beentaken to imply a lower refractive index than thatof the surrounding epithelium (Brown, 197I). Aftera month or more of treatment the cysts tended tobe smaller and more discrete and in many eyeshad disappeared. The remaining cysts sometimeschanged their appearance as seen by retroillumina-tion so that they now reversed the illuminationcompared with that of the background, implyinga refractive index higher than that of the surround-ing epithelium. The appearance was now like thatof small oil droplets.

THE CONTROL GROUP (Table IX)

The findings in the controls are presented inTable IX. A low incidence of superficial dystrophiesof the fingerprint line and bleb type and a fewmicrocysts were found. In none of the controlswas there any history of recurrent erosion, norwere there any symptoms referrable to the cornea.


http://bjo.bmj.com

/B

r J Ophthalm


ownloaded from

http://bjo.bmj.com/

Recurrent erosion of the cornea 9I

Table IX Incidence of superficial cornealdystrophies in 200 control eyes

Superficial corneal dystrophies Percentage

Fingerprint lines I-5Microcysts I5Blebs (including net patterns) 2'0

FOLLOW-UP AND RESPONSE TO TREATMENT (Table X).

Altogether 70 patients attended for follow-up fromI to 48 months (mean i6 months).Healing occurred after the recurrent erosion

with which the patient had presented, in a fewhours in minor microform and up to 3 weeks insevere macroform erosions. In two eyes, taking3 weeks to heal, the healing had been delayed bycorneal infiltration and oedema due to secondarybacterial infection. A healing time of I to 2 dayswas normal for the macroforn erosions. Theinitial treatment given by the Casualty Officer waschloramphenicol drops or ointment and it did notappear to matter which of these was given. De-bridement was performed as an initial treatmentin I2 eyes which had large bags of loose epithelium.In each, this was followed by healing in 2 to 4days (mean 3 i days), and no cases of delayed

Table X Response to prophylactic treatment for3-I8 months (mean io months)

FIG. 8 Bietti's lacunar dystrophy (map dystrophy).Bar gauge 500 ,um

Treatment

NaCl 5 per cent ointmentSymptom-freeImprovedNot improvedInadequate follow-upOintment rejected for stingingRecurrences on treatment:MicroformMacroform

Effect of stopping treatmentfor 3 monthstTreatment stopped in

but 3 persisted against adviceTherefore total in groupRemaining symptom-freeUnchangedDeterioratedRecurrences off treatment:MicroformMacroform

No. ofpatients

6o32I62

55

88

Percentage*

6o5327388

'4I4

32

29

55

'7

I93

FIG. 9 Intraepithelial microcysts. Late stage oftreatment showing reversal of marginal retroillumination.Bar gauge 500 Fm

'7'759

65IO0

*Of 6o patients tOf 29 patients

healing occurred in these patients. Debridementwas also carried out in four eyes showing lack ofspontaneous healing with medical treatment duringwhich a loose bag of epithelium developed. After


http://bjo.bmj.com

/B

r J Ophthalm


ownloaded from

http://bjo.bmj.com/


debridement, the epithelium was able to growsmoothly across the stroma. The incidence ofrecurrence after debridement did not differ fromthat in eyes which had received medical treatmentalone. In one eye a deliberate attempt was made toscarify the surface of Bowman's membrane atdebridement. No recurrence occurred in this eye,but the debridement left a faint superficial nebula.

Carbolization was performed after debridementin seven eyes and was followed by healing in 2 to 4days. There was a single recurrence of a macroformerosion in two eyes and there were several re-currences of microform erosions in one eye whichhad been carbolized. Carbolization appears to bebeneficial in reducing recurrence, but the number ofeyes so treated was too small for the results to besignificant.Once healing was complete, 6o patients were

given sodium chloride 5 per cent ointment toapply each evening immediately before going tobed. The result of this treatment was assessed after3 months (Table X). Most patients were improvedor even symptom free, but in some there wererecurrences. The ointment was well toleratedby most patients, but 8 per cent found that itstung and stopped using it.The effect of deliberately stopping the sodium

chloride ointment was tested in 32 patients (TableX) after treatment for 3 to i8 months. The patients'symptoms tended to deteriorate, particularly withan increased incidence of recurrence of microformero3ions. Thus three patients persisted in usingthe ointment against advice. It appeared thattreatment could be more successfully withdrawn-after a long period (12 months or more), and it wasconsidered proper to withdraw treatment from allpatients once they were able to remain comfortable-without the ointment. Those in whom symptomsrecurred resumed the use of the ointment.

DiscussionCLINICAL FEATURES

Recurrent erosion of the cornea has often beendescribed, and a review of the early literature waswell presented by Chandler (I945). The appearance,of the active erosion has been fully documentedand Chandler was able to divide the erosions, bothsymptomatically and by their appearance, intomacroform and microform. This appears to be auseful classification, but it is now seen that the twodifferent forms can occur in the same corneae atdifferent times, and it is considered that there is nosharp distinction between the two disorders.Chandler found that macroform erosions were more,commonly the result of trauma than microformerosions, which is confirmed by our observations.The existence of bilateral cases was recognized

by Chandler but has been considered rare, so thatFlynn and Esterly (I966) thought it worthwhile toreport a single case. The io per cent incidence ofbilateral cases in our series is rather higher thanthat suggested in the literature.An inherited form of recurrent erosion has been

recognized since Franceschetti (1928) reported afamily with a dominant inheritance. It wouldappear from our experience and that of otherauthors that a family history is uncommon. Twoof our patients had a family history, but in neithercase was there sufficient information to be certainof the pattern of inheritance.The relationship of the time of the recurrence to

the time of waking is interesting. We are in agree-ment with Chandler (I945) that the recurrencemay occur before or after waking, although Szili(I900) maintained that patients always awokefirst. In our patients who awakened first, difficultyin eyelid opening was commonly experienced,which suggests that there was an adhesion betweenthe corneal epithelium and the eyelids. It is obscurewhy this should occur even if the epithelium isfirst lifted from the cornea. Drying of the tear filmwith incomplete lid closure in sleep has beensuggested to account for this (Sturrock, 1975,personal communication). However, most authorstake the opposite view and consider that increasein fluid in the cornea which lifts off the epitheliumcauses the recurrence. Thus Chandler (1945) andTheodore (I964) considered that an endothelialdefect might be important in allowing fluid entryinto the stroma; and Valle (I967) reported a familyshowing dominant inheritance of recurrent erosionin which one member had Fuchs's endothelialdystrophy which Valle considered to be significant.Cogan (I94i) demonstrated the hypotonizity oftears during the night which he considered relevantin allowing fluid entry into the epithelium, andThygeson (I959) considered that persistent epi-thelial oedema was a significant factor.A rational approach to prophylaxis is therefore

to use a lubricating agent at bedtime to preventadhesion between the corneal epithelium and thelid, and an agent to produce corneal desiccation.The use of ointments as lubricants is traditionallyaccepted and methyl cellulose drops have beenrecommended by Guyard and Perdriel (I96I) andby Theodore (I964). Theodore also recommendedlowering intraocular pressure by acetazolamide. Acombined lubricant and desiccating agent isconvenient; Chandler (I945) and Thygeson (I959)used io per cent boric ointment, and Goldman,Dohlman, and Kravitz (I969), and Trobe andLaibson ( 972), like ourselves, used 5 per centsodium chloride ointment.The existence of superficial corneal dystrophies

in relation to recurrent erosion was probably


http://bjo.bmj.com

/B

r J Ophthalm


ownloaded from

http://bjo.bmj.com/


first recognized by Vogt (1930) who observedspots and fine glassy lines which appear in hisillustration to be the same as those now calledfingerprint lines. Bruckner (1950) stated that noslit-lamp abnormality was to be seen in healedrecurrent erosion, but he was able to demonstrateirregularity of the cornea with the Fisher reflecto-graphic technique which is capable of demonstrat-ing such fine abnormalities as the anterior cornealmosaic (Bron, I968), also demonstrable by theproper use of retroillumination.

Several other authors, as reviewed by Chandler(1945), have described white or grey dots in corneaewith healed recurrent erosions and these dots maynow be described as microcysts. Microcysts may berelatively clear, with a low content of cellular debris,and variable in size; they are sometimes fusedtogether and often take up fluorescein stain. Themicrocysts of Cogan's microcystic disorder aremore often densely opaque and discrete; they failto take up stain and may reach a large size (Bronand Tripathi, I973). Although such cysts may occurtogether in the same cornea, their morphogenesismay be different. Like Cogan, Kuwabara, Donald-son, and Collins (i974), we believe that the specialfeatures of the Cogan type of cyst relate to theirsequestration behind anomalous intraepithelialbasement membrane sheets (Bron and Tripathi, inpreparation). Cogan's microcystic dystrophy hasin the past been regarded as benign and responsibleat most for minor irritative symptoms (Cogan andothers, I964; Guerry, I965; Levitt, 197I; Bron andTripathi, I973; Luxenberg and others, I975).Attention has been drawn by Trobe and Laibson(1972) to the association of these cysts with mapdystrophy (which we regard as identical to thelacunar dystrophy of Bietti (I965)) and withfingerprint lines and recurrent corneal erosion.Brodrick, Dark, and Peace (1974) confirmed therelationship between Cogan'¢ microcystic dys-trophy, fingerprint lines, and recurrent cornealerosion. In the present study cysts of the Cogantype and map-like dystrophy are found, but theirincidence is low.

Proper attention has only recently been givento the relationship between the superficial dys-trophies of the fingerprint line, bleb, and net typeand recurrent corneal erosion; in fact Guerry(1950), in his original description of fingerprintlines, considered them to be of no pathologicalsignificance, but a relationship between fingerprintlines and recurrent erosion has been reported byKaufman and Clower (I966), Bron and Brown(97I), Trobe and Laibson (1972), Brodrick andothers (1974), and Brown and Bron (1976), andbetween bleb and net patterns and recurrent erosion(Bron and Brown, 197I).

It appears that the various superficial corneal

disorders which are associated with recurrentcorneal erosion are not discrete entities but repre-sent a spectrum of visible signs with a commonpathological background. Many of our patientsshowed several different forms of disorder in thesame or fellow eye and this relationship Fas beencommented upon previously; Guerry (I965) andLevitt (197I) reported the relationship betweenCogan's microcystic and map-like changes, Bronand Brown (I97I) reported the association offingerprint lines and bleb and net patterns together,and their relationship with recurrent erosion, andTrobe and Laibson (1972) reported the relationshipbetween Cogan's microcystic dystrophy, map-likedystrophy and fingerprint lines, and recurrenterosion.Most corneal abrasions heal satisfactorily and

are not followed by recurrent erosion. Jackson(I960) saw one recurrent erosion in ISo abrasionsand Guyard and Perdriel (I96I) saw no recurrencesin 40 abrasions.

It appears that recurrent erosion follows in acornea which is already abnormal. The high inci-dence of superficial corneal dystrophies in ourpatients (Table V) as compared to controls (TableIX) supports this suggestion. That the superficialdystrophies are the precursor of the erosion andnot the result of it, is demonstrated by their higherincidence and greater extent in the asymptomaticeye (Table V). The recurrent erosion appears toerase the dystrophic changes so that they may beabsent or only minimally present in the healedcornea. The dystrophic change has also beenremoved by deliberate debridement as observedfor fingerprint lines by Guerry (1950), and forCogan's microcystic dystrophy by Wolter andFralick (I966) and Cogan and others (I964, 1974).This would occur if the d&fective basement mem-brane (see below) was also shed with the lost epi-thelium.

In our earlier study (Bron and Brown, I971), 40patients ranging in age from 39 to 4I years werefound to exhibit superficial corneal changes of thefingerprint, bleb, and net pattern and the changeswere bilateral in 73 per cent, suggesting a dystrophicor degenerative aetiology. Symptoms of recurrenterosion were seen in 38 per cent of patients. In thepresent study of 8o patients with recurrent erosionsyndrome, the incidence of superficial cornealchanges (fingerprints, blebs, nets, and maps) was59 per cent compared to 3-5 per cent in the controlgroup (lines and blebs). (Cysts are excluded fromthe comparison because they may at times be anon-specific expression of the presence of epithelialoedema. Data are available in Tables VIII and IX.)This strongly supports the contention that thesuperficial corneal changes under discussion areresponsible for defective epithelial adherence and


http://bjo.bmj.com

/B

r J Ophthalm


ownloaded from

http://bjo.bmj.com/


render patients vulnerable to recurrent erosion.It appears that the patients studied by us in the

past (Bron and Brown, I97i) and in the presentstudy are substantially the same group as thatstudied by Trobe and Laibson (1972). Differencesrelate to the low incidence of map change andCogan's cyst change in our studies and to the ab-sence of net and bleb changes in the series studiedby Trobe and Laibson. These differences may re-late in part to matters of definition, since weattempted to distinguish on biomicroscopic groundscysts of the Cogan type from other microcystsassociated with recurrent erosion (Bron andTripathi, I973), and in part to methods of examina-tion. It is important to emphasize the need forcareful examination of the cornea by retroillumina-tion, using first the iris and then the red reflex ofthe fundus with the pupil dilated, in order not tomiss the finest dystrophic changes.

It seems likely that a common histopathologicalfactor in all these disorders is the anomalous pro-duction of basement membrane material and otherconnective tissue sheets both within and sometimesdeep to the corneal epithelium. This has beenobserved both in the absence of clinical map-likechanges (Cogan and others, I964, 1974), and intheir presence (Guerry, i965; King and Geeraets,1972; Tripathi and Bron, 1973; Rodrigues, Fine,Laibson, and Zimmerman, 1974; Fogle, Kenyon,Stark, and Green, I975; Dark, Brodrick, andPeace, 1973), and Brodrick and others (I974),demonstrated the presence of a fibrillo-granularmaterial with intra- and subepithelial extensionin a patient with fingerprint lines. The histopatho-logy of bleb and net patterns is not yet known.The relevance of basement membrane abnor-

malities to recurrent erosion has been clarified inrecent years. Khoudadoust, Silverstein, Kenyon,and Dowling (I968) experimentally showed theimportance of the basement membrane and of thehemidesmosomes of the basal epithelium to epithe-lial adherence. Poor epithelial adherence is impliedby the erosion itself and can be demonstrated bymanipulating a contact lens on the healed corneaof these patients (Lowe, I964, 1970). A defectivebasement membrane was demonstrated in trau-matic recurrent erosion by Goldman and others(I969), and in spontaneous recurrent erosion byTripathi and Bron (I972), who also noted anabsence of hemidesmosomes in relation to affectedbasal cells. Kenyon (I969) suggested that deficien-cies in the formation of basement membrane,hemidesmosomes, and attachment fibrils accountfor the erosive events encountered in Reis-Bickler'sdystrophy, Cogan's microcystic dystrophy, andnon-traumatic recurrent erosion (Fogle and others,1975). Presumably a similar mechanism is involvedin metaherpetic keratitis (Kaufman, I964).

It would appear that the biomicroscopic appear-ance of fingerprint lines, of map changes, andprobably of the net and bleb patterns relate to thelocation, amounts, and optical properties of ano-malous basement membrane and other connectivetissue material situated in the affected epithelium,and that recurrent erosion episodes are relatedpartly to the presence of subepithelial-connectivetissue sheets and to a disturbance of the epithelialadhesion mechanism.

RESPONSE TO TREATMENT

In the initial treatment it would appear to make nodifference whether the antibiotic is given as dropsor an ointrnent, which accords with the experimentalwork of Hanna, Fraunfelder, Cable, and Hard-berger (I973). The complication of entrapment ofointment beneath the epithelium (Fraunfelder,Hanna, Cable, and Hardberger, 1973) was not seen.However, it was noticed that the refractile appear-ance of microcysts as seen on retroilluminationcould change so that they came to resemble oildroplets.

Simple debridement, without scarification orcarbolization of the surface of Bowman's membrane,appears to be beneficial in assisting healing of amacroform erosion, but not in preventing recur-rence. This is not surprising since the problemappears to be one of epithelial adherence ratherthan of epithelial resurfacing. Some form of phy-sical or chemical injury to the cornea has commonlybeen recommended. Chemical cautery has fre-quently been advocated (Chandler, 1945), and ourown experience with carbolization in a smallnumber of patients would appear to support this.However, carbolization does leave a faint super-ficial nebula, and is best avoided except for highlyintractable recurrent erosions. Debridement withdeliberate scarification of Bowman's membranehas been recommended (Kaufman and Clower,I966; Trobe and Laibson, 1972) and is probablyuseful in preventing recurrence, as in the only caseso treated by us. However, this also leaves a super-ficial scarring.

Successful prophylactic treatment with hyper-tonic agents has been described by Chandler(I945), Thygeson (I959), Goldman and others(I969), and Trobe and Laibson (1972). The hyper-tonic agent which we used, sodium chloride oint-ment 5 per cent, was shown to be effective indehydrating the human cornea by Luxenberg andGreen (1971) who obtained a reduction of 20 percent in the thickness of oedematous corneae. Thesodium chloride ointment was well tolerated bythe patients and appears to have been effective inpreventing recurrence, as witnessed by the in-creased incidence of recurrence once it had been


http://bjo.bmj.com

/B

r J Ophthalm


ownloaded from

http://bjo.bmj.com/


withdrawn (Table IX). The reduction in size andthe increase in refractility of the microcysts withtheir eventual disappearance in some eyes suggestsdesiccation.

Patients could be successfully withdrawn fromtreatment, particularly after long periods, so thatit would appear that the use of the ointment iscurative in the long run, or that recurrent erosionis a self-limiting disorder. The ointment is a logicalchoice (as discussed above), but we cannot becertain at present whether the lubricant or desiccantproperties are the more important. A future trialagainst a simple lubricant ointment should beinformative.

Conclusions

Recurrent erosions occur in corneae with pooradhesion between the epithelium and Bowman'smembrane, and defective basement membranecomplexes are implicated. The defect is commonlypresent in the cornea before the initial trauma aswitnessed by the high incidence of superficialcorneal dystrophies.The examination of the cornea in the healed

state commonly reveals a superficial dystrophy, orthe presence of microcysts. The diagnosis ofrecurrent erosion can thus be made on slit-lampexamination in most healed cases and less relianceneed be placed on the verbal history alone.

Fluid, collecting in or beneath the epitheliumduring the night, with resultant lifting of theepithelium and adherence of the epithelium tothe tarsal conjunctiva are two possible factorsleading to recurrence.

In the initial treatment it does not matter whetherdrops or ointment are used. Debridement of looseepithelium assists healing and resolves pain sooner,but does nothing to prevent recurrences. Recur-rence is best prevented with a lubricant and desic-cant substance (sodium chloride 5 per cent oint-ment) applied at bedtime. This can be discontinuedafter several months once the patient remainssymptom free on withdrawal. No precise periodfor prophylactic treatment can be given and it mustbe adjusted for the individual patient. In eyeswhich continue to be severely affected by recur-rence, in spite of prophylactic treatment, epithelialadhesion can be improved by carbolization ordebridement with scarification, but these proce-dures may cause scarring.

Summary

Altogether, 8o patients aged between 24 and 73years with recurrent erosion of the cornea havebeen studied and compared with a control groupof 200. The patients' erosions were divisible intomacroform and microform types. The macroformoccurred in io per cent, the microform in 56 percent, and both types in the same patients in 31per cent. The macroform was more commonlyrelated to trauma than the microform. However,many (40 per cent) were spontaneous in origin.The most common cause of the initial trauma was afinger nail.The recurrences occurred at around the time of

waking, either just before or just after. Difficultyin opening the eye occurred in io per cent. Therewas little evidence of precipitating factors, but eyerubbing was admitted by io per cent and barbi-turates were implicated in 3 per cent.The corneae were examined in the healed state,

when a high incidence (59 per cent) were found tohave superficial corneal dystrophies of the finger-print lines, bleb, and Bietti's lacunar (map-like)types. These are considered individually, particularattention being paid to the distinctions between thevarious types of line resembling the fingerprintline. Epithelial microcysts were also a commonfinding (59 per cent) and were sometimes of theCogan type. In only i i per cent of patients werethere no corneal signs in the healed state. The needfor careful examination of the cornea by retro-illumination, using both the iris and the fundus, isstressed. The control group, in contrast, showed avery low incidence of dystrophies and cysts.Treatment was given initially with either drops

or ointment and no differences in healing werefound. Debridement was performed in I2 eyesas an initial treatment and also in four eyes whichwere not healing on medical treatment. Debride-ment assisted healing, but did not prevent recur-rence. One eye was treated with debridement andscarification and seven with carbolization. Theseprocedures appeared to reduce the recurrence rate.Sodium chloride ointment 5 per cent was found

useful as a prophylactic taken at bedtime, and therecurrence rate increased when it was withdrawn.

We should like to thank Professor Barrie Jones for hisadvice and for providing us with the facilities to examinehis patients.

ReferencesBIETTI, G. B. (1965) Arch. Ophtal. (Paris), 25, 37BRODRICK, J. D., DARK, A. j., and PEACE, W. (I974) Arch. Ophthal., 92, 483BRON, A. J. (I968) Brit. J. Ophthal., 32, 659

and BROWN, N. A. (I97I) Trans. ophthal. Soc. U.K., 86, I3and JONEs, D. B. (I969) Brit. J. Ophthal., 53, 490


http://bjo.bmj.com

/B

r J Ophthalm


ownloaded from

http://bjo.bmj.com/


and TRIPATHI, R. C. (I973) lbid., 57, 36IBROWN, N. A. (I970) Ibid., 54, 697

-(I1971) Ibid-, 55, 517and BRON, A. J. (1976) Amer. J. Ophthal. (in press)

BRUCKNER, VON A. (1950) Ophthalmologica (Basel), I20, 38CHANDLER, P. A. (I945) Amer. 3t. Ophthal., 28, 355COGAN, D. G. (I94I) Arch. Ophthal., 25, 941

, DONALDSON, D., KUWABARA, T., and MARSHALL, D. (1964) Trans. Amer. ophthal. Soc., 62, 213, KUWABARA, T., DONALDSON, D. D., and COLLINS, E. (1974) Arch. Ophthal., 92, 470

DARK, A., BRODRICK, J. D., and PEACE, G. W. (973) IRCS Med. Sci., I (I) (73-3). The Eye (24); Review No. 24-2-IFLYNN, M. A., and ESTERLY, D. B. (I966) Amer. J. Ophthal., 62, 964FOGLE, J. A., KENYON, K. R., STARK, W. J., and GREEN, W. R. (1975) Amer. J7. Ophthal., 79, 925FRANCESCHETTI, A. (I928) Z. Augenheilk., 66, 309FRAUNFELDER, F. T., HANNA, C., CABLE, M., and HARDBERGER, R. E. (1973) Amer. J. Ophthal., 76, 475GOLDMAN, J., DOHLMAN, C., and KRAVITZ, j. (I969) Trans. Amer. Acad. Ophthal. Otolaryng., 73, 471GUERRY, D. (1950) Amer. 7. Ophthal., 33, 324

(I965) Trans. Amer. ophthal. Soc., 63, 320GUYARD, M. M., and PERDRIEL, G. (I96I) Bull Soc. Ophtal. (Paris), 7-8, 579HANNA, C., FRAUNFELDER, F. T., CABLE, M., and HARDBERGER, R. E. (1973) Amer. Jr. Ophthal., 76, 193JACKSON, H. (I960) Brit. med. J., 2, 713KAUFMAN, H. E. (I964) Amer. Y. Ophthal., 57, 983

and CLOWER, J. W. (I966) Ibid., 6I, 227KENYON, K. R. (I969) Ibid., 65, 339KHOUDADOUST, A. A., SILVERSTEIN, A. M., KENYON, K. R., and DOWLING, J. E. (I968) Ibid., 65, 339

KING, R. G. JR, and GEERAETS, R. (1972) Med. Coll. Va. Qrtly., 8, 241LEVITT, J. M. (I97I) Amer. 7. Ophthal., 72, 38ILOWE, R. F. (I964) Ibid., 57, 397

(I970) Brit. J. Ophthal., 54, 805LUXENBERG, M. N., FRIEDLAND, B. R., and HOLDER, J. M. (1975) Arch. Ophthal., 93, 107

and GREEN, K. (I97i) Amer. J. Ophthal., 71, 847RODRIGUES, M. M., FINE, B. S., LAIBSON, P. R., and ZIMMERMAN, L. E. (1974) Arch. Ophthal., 92, 475SZILI, A. SR (1900) v. Graefes Arch. Ophthal., 51, 486THEODORE, F. M. (I964) Eye, Ear, Nose Thr. Monthly, 43, 62, (2)THYGESON, P. (I959) Amer. 7. Ophthal., 47, 48 (pt. II)TRIPATHI, R. C., and BRON, A. J. (1972) Brit. J7. Ophthal., 56, 73

and (I973) Ibd., 57, 376TROBE, J. D., and LAIBSON, P. R. (1972) Arch. Ophthal., 87, 378VALLE, 0. (I967) Acta ophthal. (Kbh.), 45, 829VOGT, A. (I930) 'Lehrbuch und Atlas der Spaltlampen-mikroscopie', vol. I, p. 244. Springer, BerlinWOLTER, j., and FRALICK, F. (x966) Arch. Ophthal., 75, 380


http://bjo.bmj.com

/B

r J Ophthalm


ownloaded from

http://bjo.bmj.com/

recurrent erosion of the cornearecurrent erosion ofthe cornea 87 in microformerosion anda larger...

Documents