Br J Ophthalmol 88 PostScript · 2004-03-10 · the Agence Franc¸aise de Se´curite´ Sanitaire des Produits de Sante´. They are very similar in composition, both being based on

Mixed infection (Pseudomonasand coagulase negativestaphylococci) microbial keratitisassociated with extended wearsilicone hydrogel contact lensContact lens induced ulcerative keratitis is aserious complication which can be devastat-ing for the patient if treatment is delayed.Extended wear is the commonest cause ofmicrobial keratitis in contact lens wear.1 Newextended wear silicone hydrogel contactlenses have higher oxygen transmissibilityso that they can be worn continuously for30 days. They can also be used as bandagecontact lenses.

The risk of Pseudomonas microbial keratitiswith overnight wear is significantly increasedby contact lenses with low oxygen transmis-sibility.2 By virtue of high oxygen transmis-sibility, the silicone hydrogel contact lensesare thought to be associated with low risk ofinfectious keratitis.3 4 So far only four cases ofmicrobial keratitis have been reported withtheir use.5 In spite of various claims ofprotection against serious microbial keratitiswith pathogens such as P aeruginosa, we haverecently come across the first case ofPseudomonas keratitis in a patient wearingsilicone hydrogel contact lenses.

Case reportA 23 year old male patient presented with1 day history of severe pain, ocular injection,photophobia, and reduced vision of right eye.He was wearing the day and night siliconehydrogel contact lenses, which was replaced

once every 30 days (Ciba vision Focus day andnight). He has been wearing these contactlenses for 7 months before the presentation.

Examination revealed a visual acuity ofhand movement for the right eye and 6/5 forthe left eye. The right eye had a centralcorneal ulceration of 3 mm in diametersurrounded by severe oedema and a 1 mmhypopyon. Cultures grew P aeruginosa andcoagulase negative staphylococci both sensi-tive to ciprofloxacin and gentamicin. Topicalofloxacin and gentamicin were commencedwith cyclopentolate. Unpreserved predniso-lone eye drops (0.5%) were added after1 week. Two weeks later, the epithelial defecthad resolved leaving behind a central sub-epithelial corneal scar (fig 1). His visionimproved to 6/18 unaided, 6/9 through thepinhole, 1 month after the admission.

CommentThe major barrier to prescribing a continuouswear contact lens is a perceived danger ofmicrobial keratitis. Many factors are involvedin the development of microbial keratitis andthese include bacterial adherence to the lenssurface, formation of bacterial glycocalyx onthe lens, corneal hypoxia, deposits on the lenssurface, and the effect of contact lens onclosed eye environment.6 Silicone hydrogelcontact lenses have high oxygen transmissi-bility and these lenses are colonised bysimilar numbers and type of micro-organismscompared with HEMA based materials.7 Anumber of studies have shown lower risk ofinfectious keratitis with new silicone hydro-gel contact lenses.3 4

However, the use of silicone hydrogelcontact lenses was associated with slightlyhigher levels of visible deposits,3 which mayact as a risk factor for bacterial keratitis. As inour case young male patients were alsoconsidered a risk factor for contact lensinduced microbial keratitis. Our experiencesuggests that extended wear silicone hydro-gel contact lenses are not free of the risk ofmore serious microbial keratitis caused by Paeruginosa and coagulase negative staphylo-cocci. With increasing popularity amongoptometrists and the use of silicone hydrogelcontact lens as a bandage contact lens, such aserious complication cannot be ignored.

As suggested by other authors,6 our experi-ence points towards a multifactorial aetiologyfor microbial keratitis, rather than just oxy-gen transmissibility. Further studies arerequired to find out the safety of the siliconehydrogel contact lenses with regard to devel-opment of microbial keratitis.

P Syam, B Hussain, C HutchinsonDepartment of Ophthalmology, Calderdale Royal

Hospital, Salterhebble, Halifax HX3 0PW, UK

Correspondence to: Mr P P Syam, Department ofOphthalmology, Calderdale Royal Hospital,

Salterhebble, Halifax HX3 0PW, UK;[email protected]

Accepted for publication 21 January 2003

References

1 Dart JKG, Stapleton F, Minassian D. Contactlenses and other risk factors in microbial keratitis.Lancet 1991;338:650–3.

2 Imayasu M, Petroll M, Jester J, et al. The relationbetween contact lens oxygen transmissibility andbinding of pseudomonas aeruginosa to thecornea after overnight wear. Ophthalmology1994;101:371–88.

3 Brennan NA, Chantal Coles ML, Comstock TL,et al. A 1-year prospective clinical trial ofBalafilcon A (Pure Vision) silicone—hydrogelcontact lenses used on a 30-day continuous wearschedule. Ophthalmology 2002;109:1172–7.

4 Nilsson SEG. 7-Day extended wear and 30-daycontinuous wear of high oxygen transmissibilitysoft silicone hydrogel contact lenses. Arandomised one-year study of 504 patients.CLAO J 2001;27:125–36.

5 Lim L, Loughnan MS, Sullivan LJ. Microbialkeratitis associated with extended wear of siliconehydrogel contact lenses. Br J Ophthalmol2002;86:355–7.

6 Dart JKG. Contact lens and prosthesis infections.In: Jaeger E, Tasman W, eds. Duane’sfoundations of clinical ophthalmology, Vol 2.Philadelphia: Lippincott-Raven, 1996:1–30.

7 Keay L, Willcox MDP, Sweeney DF, et al.Bacterial populations on 30-night extended wearsilicone hydrogel lenses. CLAO J 2001;27:30–4.

Controlled study of the influenceof storage medium type onendothelial assessment duringcorneal organ cultureSelection of corneal grafts in eye banks ismainly based on end-of-storage endothelialassessment, which consists of endothelial celldensity (ECD) measurement and, to someextent, cell morphometry. Below a certainECD threshold, generally 2000 cells/mm2, thecornea is deemed unfit for penetratingkeratoplasty. Precise ECD measurement atthe end of storage is thus a key issue for eyebanks, and also for patients, because itinfluences the long term survival of thegraft.1–3

For long term storage in organ culture, themost common method in Europe,4 endothe-lial observation is possible only by trans-mitted light microscopy. The endothelial cellsare exposed to 0.9% sodium chloride orsometimes to 1.8% sucrose, which induce asmall degree of osmotic cell shrinkage anddilatation of the intercellular spaces thusmaking individual cells visible. The cells canthen be counted manually, through a cali-brated reticule or from a photograph; or usingan advanced image analysis system.5–7

Whichever method of count is used, precisiondepends primarily on good visualisation ofthe cell borders. It has long since been shownthat, even under experimental conditions ofperfect cell membrane visualisation usingalizarin red, maximum precision ranges from+5% to 25%.8

Two commercial media are authorised bythe Agence Francaise de Securite Sanitairedes Produits de Sante. They are very similarin composition, both being based on HEPES-buffered Iscove’s Modified Dulbecco’s med-ium containing sodium bicarbonate and 2%fetal bovine serum, with the same pH of 7.25but the osmolality of Inosol (Bausch & Lomb,Chauvin-Opsia, Labege, France) is only305 mosmol/kg (range 295–315) comparedwith 320 mosmol/kg (range 300–340) forCorneaPrep/Max (Eurobio, Les Ulis, France).One has nevertheless acquired the reputationof allowing better visualisation of endothelial

Figure 1 Photograph taken 3 weeks afteradmission showing central corneal scar.

PostScript . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

LETTERS

If you have a burning desire to respond to apaper published in BJO, why not make useof our ‘‘rapid response’’ option?

Log on to our web site (www.bjophthalmol.com), find the paper that interests you, andsend your response via email by clicking onthe ‘‘eLetters’’ option in the box at the top righthand corner.

Providing it isn’t libellous or obscene, itwill be posted within seven days. You canretrieve it by clicking on ‘‘read eLetters’’ onour homepage.

The editors will decide as before whether toalso publish it in a future paper issue.

Br J Ophthalmol 2004;88:579–599 579

www.bjophthalmol.com

on Novem

ber 13, 2020 by guest. Protected by copyright.

http://bjo.bmj.com

/B

r J Ophthalm

ol: first published as 10.1136/bjo.2003.029371 on 18 March 2004. D

ownloaded from

http://bjo.bmj.com/

cells and thus facilitating ECD measurement.We therefore compared the quality ofendothelial cells visualisation in these twocommercial media, using an original imageanalyser specially designed for the assess-ment of corneal endothelium by light micro-scopy.

MethodsWe conducted a randomised prospectivestudy with masked analysis of the results.Donors with history of anterior segmentsurgery were excluded. After procurement ofa pair of corneoscleral discs, one of thecorneas (group A) was immersed in Inosoland the other (group B) in Corneaprep/Maxfor organ culture at +31 C. The media wereallocated to the right or left cornea accordingto a randomisation list. Two consecutiveendothelial examinations were performed ina similar fashion. The initial examination wasdone between the first and fifth days afterprocurement, and the final one two daysbefore cornea delivery.

After the endothelial side was incubatedfor four minutes in 0.9% sodium chloride(Aguettant, Lyon, France), it was observedunder a direct light microscope (Laborlux,Leica, Wetzlard, Germany) with 610original magnification. Ten wide-field(12506950 mm), non-overlapping images ofthe mosaic, contained within the central8 mm, were captured by a black and whitemono CCD camera. The evaluation wasperformed by an experienced technician

masked to storage medium, using aSambacornee analyser (Samba Technologies,Meylan, France), the commercial version ofthe ‘‘tri-image’’ analyser prototype developedby our team and described previously.9 10 Afully automatic and a touch up analysis (withuser intervention to identify cell boundariesmissed or delineated incorrectly in automaticmode) of exactly the same zones of the samethree images, were performed on receipt(initial examination) and on delivery (finalexamination) (Fig 1).

The three images selected by the analyserwere qualitatively assessed on a scale of three(Table 1) by three independent observersmasked to storage medium. Discordant caseswere reviewed.

The normality of the data distribution wastested using both the Lilliefors variant of theKolmogorov-Smirnov test and Shapiro-Wilknormality test, with the cut off for non-normality set at p,0.05. The quantitativevariables (number of cells ‘‘reliably recog-nised’’, ECD, touch up duration) were com-pared using a paired t test in the case ofnormal distribution, and otherwise by a non-parametric test (Wilcoxon). The image qual-ity scores were compared by the x2 test in a362 grid. p,0.01 was deemed significant.

ResultsAs the study design required inclusion ofpaired corneas having had two successiveanalyses, of a series of 77 pairs of corneasprocured consecutively, 30 pairs of 13 women(43%) and 17 men (57%) with a mean age of69 years (range 30–92) were included in thisstudy. Mean time between death and cornea

procurement was 10 hours (range 0, for heartbeating donors-24).

At the initial examination (Table 2), per-formed on an average three days (range 1–5)after procurement, image quality was com-parable between the two groups. Whicheveranalysis mode (automatic or touch up) wasused, all parameters were comparablebetween the two media except for the touchup duration which was slightly shorter, on anaverage by 1 minute (59 seconds, 95% con-fidence interval, CI (19–102)), for the corneasin group A. Compared to the automaticanalysis, the touch up analysis only slightlydecreased the initial ECD value in group A, by154 cells/mm2 (95% CI 36 to 79), or 24.7%(p,0.001) and insignificantly in group B, by101 cells/mm2 (95% CI 239 to 240), or 23.1%(p = 0.150).

Between the initial and final examination(Table 3), performed on an average 13 days(range 8–22) after procurement, image qual-ity deteriorated markedly in group A(p,0.001) but remained stable in group B(p = 0.357). At the final examination, groupA displayed no ‘‘good’’ images against nearlyone in two for group B. Automatic recogni-tion of the cells was thus made much easierin group B, with on an average 238 (46)against only 159 (47) cells in group A(p,0.001). For group B, the need for touchup was reduced, with a mean time gain ofabout 3 minutes (163 seconds, 95% CI 116 to211), p,0.001) and allowing a higher num-ber of cells to be taken into account (456(82%) against 357 (72%) for group A,p,0.001). In both groups, the touch upanalysis reduced considerably the final ECDvalue compared to the automatic analysis, by435 cells/mm2 (95% CI 317 to 552), or213.8% (p,0.001) for group A and by 313cells/mm2 (95% CI 239 to 386), or 210.3%,(p,0.001) for group B.

The two media did not differ in terms ofpreserving endothelial viability: ECD, percen-tage cell loss, and morphometry (all deter-mined in touch up mode) were comparablebetween the two media.

CommentOur randomised, prospective parallel groupstudy, performed masked with an objectiveimage analysis tool, demonstrates that visua-lisation of the endothelial mosaic is betterafter organ culture in CorneaPrep/Max

Table 1 Qualitative grading of the endothelial images

Image quality Score Criteria

Good 2 Excellent view of cell borders; low background noise; cells visible on over2/3 of image area

Average 1 Good view of cell borders; moderate background noise; cells visible on 1/3to 2/3 of image area

Poor 0 Poor view of cell borders; high background noise; cells visible on less than1/3 of image area

For the analyses in tri-image mode, overall quality was graded ‘‘good’’ if the three images obtainedscores of 2/2/2 or 2/2/1, ‘‘average’’ if the scores were 2/1/1, 2/1/0, or 1/1/1, and ‘‘poor’’ if thescores were 1/1/0, 1/0/0, or 0/0/0.

Figure 1 Illustration of the image analysistools specifically developed for easier touchingup of cell borders. Areas outlined in red wereeither ,150 mm2 (black arrow), often artefactslike nucleus, debris, or .1500 mm2

(arrowhead) or with width greater than twicethe length (asterisk), the later two indicatingpoorly separated cells. These cells, assumed tobe poorly recognised, would suggest to thetechnician the need for contour correction or,where appropriate, validation. To calculate theendothelial cell density in automatic mode, onlythe green reliably recognised cells were takeninto account.

Table 2 Initial examination of paired corneas stored in medium A and B

Group A (n = 30) Group B (n = 30) p Value

Image quality (good/average/poor,(%))

56/27/17 43/30/27 0.526

Cells well recognised per se,‘‘automatic’’ mode (n)

298 (68) 147–417/304

275 (89) 113–417/284

0.058

ECD, ‘‘automatic’’ mode(cells/mm2)

3267 (197) 2779–3672/3250

3248 (305) 2445–3753/3249

0.770

Cells counted, ‘‘touch up’’ mode (n) 515 (126) 278–973/521

540 (259) 314–1777/519

0.579

ECD, ‘‘touch up’’ mode (cells/mm2) 3113 (341) 2343–4008/3107

3147 (391) 2461–3893/3190

0.533

Coefficient of variation of cell area,‘‘touch up’’ mode (%)

29.1 (6.1) 21.2–47.2/28.1

29.1 (4.3) 21.7–39.6/29.1

0.986

Hexagonality, ‘‘touch up’’ mode (%) 53.8 (9.6) 32.0–75.0/53.2

51.6 (9.0) 34.8–69.6/51.9

0.212

Touch up duration (seconds) 345 (131) 129–560/334

406 (166) 171–721/382

0.006

Results were expressed as mean (standard deviation), minimum–maximum/median. The automaticanalysis mode provides reliable results compared to the more time consuming touch up one. Thisallowed, in our eye bank, to decide in a few moments whether to continue with organ culture.

580 PostScript


on Novem


http://bjo.bmj.com

/B

r J Ophthalm


ownloaded from

http://bjo.bmj.com/

medium than in Inosol. The former facilitatesECD measurement at delivery, the mainparameter of corneal quality control. Ourstudy shows that: (1) prolonged storage inCorneaPrep/Max caused no deterioration inimage quality, unlike that with Inosol. Bettervisualisation of cell borders at deliveryshortened touch up durations considerably,on average by three minutes per compu-terised analysis. This point should be parti-cularly relevant for the eye banks whichperform endothelial assessment only once,generally after 10–15 days of organ cul-ture,4 11 (2) At the end of organ culture, thecorneas stored in CorneaPrep/Max respondedbetter to osmotic dilation of the intracellularspaces than those stored in Inosol. There wasno such discrepancy at the start of storage.We may assume that this difference inbehaviour is due to the increased osmolalityof CorneaPrep/Max, which is 15 mOsm/lgreater than that of Inosol. During organculture, the very gradual change of ioniccontent between the cells and/or of theintercellular junctions may increase wateregress from the cells and thus promotedilation of the intercellular spaces in thepresence of 0.9% sodium chloride. However,it is likely that over three days these changesdid not have time to occur, which wouldexplain the lack of initial differences in imagequality. Further histological study couldconfirm the nature of these changes in thecells and/or their junctions but whatever theirnature, they do not affect viability.

In 2001, the twenty one French eye banksstored on average 292 corneas (range 32–846).11 Because of their small size, the eyebanks naturally prefer commercial organculture media to their own preparations.The two media authorised for use in Franceare very similar in composition; neither issuperior in terms of preserving endothelialviability, as our study confirms. Until now,medium selection was dictated essentially byeconomic arguments. Our study provides anadditional criterion, namely higher quality ofmosaic visualisation to justify the choice ofmedium. It should be possible to extrapolatethe beneficial impact of good mosaic visua-lisation to manual counting, the method stillused by most French and European eyebanks.

Improving endothelial quality control hasbecome a priority in our eye bank and

research laboratory.7 10 Work is under wayto determine the ideal composition of organculture medium that allows the best endothe-lial visualisation without inducing additionalcell loss.

G Thuret, C Manissolle, S Herrag, N Deb,L Campos-Guyotat, P Gain

’Cell survival and adhesion in cancers and grafts,’ EA3063, Faculty of Medicine, University Hospital of

Saint-Etienne, France

G Thuret, N Deb, P GainOphthalmology Department, Bellevue Hospital,

University Hospital, France

S AcquartCornea Bank/French Blood Centre, Saint-Etienne,

France

Correspondence to: Gilles Thuret, Serviced’Ophtalmologie, Pavillon 50A, Hopital Bellevue,

CHU Saint-Etienne, 42055 Saint-Etienne Cedex 2,France; [email protected]

Accepted for publication 26 May 2003

References

1 Nishimura JK, Hodge DO, Bourne WM. Initialendothelial cell density and chronic endothelialcell loss rate in corneal transplants with lateendothelial failure. Ophthalmology1999;106:1962–5.

2 Bourne WM. Cellular changes in transplantedhuman corneas. Cornea 2001;20:560–9.

3 Thuret G, Chiquet C, Bernal F, et al. Prospectiverandomized clinical and endothelial evaluation oftwo storage times for corneal donor tissue inorgan culture at 31 degrees centigrade. ArchOphthalmol 2003;121:442–50.

4 EEBA. European Eye Bank Association Directory.Eleventh ed., Amsterdam, 2003.

5 Barisani-Asenbauer T, Baumgartner I,Grabner G, et al. Automated digital imageanalysis of organ culture preserved donorcorneas. Ophthalmic Res 1993;25:94–9.

6 Reinhard T, Spelsberg H, Holzwarth D, et al.Wissensbasierte Bildanalyse des Endothels vonHornhauttransplantaten. Klin MonatsblAugenheilkd 1999;214:407–11.

7 Thuret G, Manissolle C, Acquart S, et al. Ismanual counting of corneal endothelial cell

density in eyebanks still acceptable? The Frenchexperience. Br J Opthalmol 2004 (in press).

8 Sperling S, Gundersen HJ. The precision ofunbiased estimates of numerical density ofendothelial cells in donor cornea. ActaOphthalmol (Copenh) 1978;56:793–802.

9 Gain P, Thuret G, Chiquet C, et al. Automatedanalyser of organ cultured corneal endothelialmosaic. J Fr Ophtalmol 2002;25:462–72.

10 Gain P, Thuret G, Kodjikian L, et al. Automatedtri-image analysis of stored corneal endothelium.Br J Ophthalmol 2002;86:801–8.

11 Delbosc B. French directory of corneal storagecentres. Ninth ed. Besancon, 2001.

Prospective case control study ongenetic assocation ofapolipoprotein e2 withintraocular pressureGlaucomas are a leading cause of blindnessthroughout the world. This group of diseaseshas a common characteristic: degeneration ofthe optic nerve that is usually associated withincreased intraocular pressure (IOP).Increased IOP is one of the major risk factorsfor developing glaucomatous damage,whereby the loss of retinal ganglion cells isthe typical pathological finding. However, thepathophysiology of pressure induced glauco-matous optic neuropathy remains unclearand is still a matter for debate. Genomescans have been performed to identify thegenomic locations of glaucoma susceptibilitygenes.1

Apolipoprotein E (APOE), a lipid transport-ing protein produced in the liver and brain, isunique among apolipoproteins in that it hasparticular relevance to nervous tissue. It isinvolved in the mobilisation and redistribu-tion of cholesterol in repair, growth, andmaintenance of myelin and neuronal mem-branes during development or after injury.Recently it has been shown that the APOE e4allele is associated with elevated risk ofnormal tension glaucoma.2 The APOE e2allele was shown to be significantly asso-ciated with an elevated risk of age relatedmacular degeneration (AMD).3

Material and methodsThis prospective case control study included32 controls (IOP ,22 mm Hg, normal opticdisc, normal visual field), 54 patients withocular hypertension (OHT, IOP .21 mm Hg,normal optic disc, normal visual field), 96patients with primary open angle glaucoma(POAG, 55 patients with preperimetric openangle glaucoma (pre-OAG, IOP .21 mm Hg,glaucomatous optic disc, normal visual field),and 41 patients with perimetric open angleglaucoma (OAG, IOP .21 mm Hg, glauco-matous optic disc, visual field defects). Allindividuals included in the study wereunrelated, white, and had open anteriorchamber angles, clear optic media, and avisual acuity of 20/25 or better. Exclusioncriteria were all eye diseases other thanglaucoma, diabetes mellitus, myopic refrac-tive error exceeding 28 diopters, and visualacuity less than 0.7.

The study followed the tenets of thedeclaration of Helsinki for research involvinghuman subjects and informed consent wasobtained from all participants.

All control subjects and patients werethoroughly examined by clinical biomicro-scopy including slit lamp inspection, gonio-scopy and ophthalmoscopy, applanation

Table 3 Final examination of paired corneas stored in medium A and B

Group A (n = 30) Group B (n = 30) p Value

Image quality (good/average/poor, (%)) 0/23/77 43/43/14 ,0.001Cells well recognised per se, ‘‘automatic’’mode (n)

159 (47) 75–248/159

238 (46) 151–324/238

,0.001

ECD, ‘‘automatic’’ mode (cells/mm2) 3143 (159) 2829–3543/3105

3029 (188) 2650–3451/3039

0.011

Cells counted, ‘‘touch up’’ mode (n) 357 (72) 209–559/339

456 (82) 292–608/446

,0.001

ECD, ‘‘touch up’’ mode (cells/mm2) 2708 (320) 2277–3430/2704

2716 (284) 2259–3579/2686

0.934

Overall cell loss, ‘‘touch up’’ mode (%) 12.6 (9.3) 7.2–34.1/12.2

13.1 (8.0) 1.3–27.8/15.6

0.79

Coefficient of variation of cell area, ‘‘touchup’’ mode (%)

28.4 (3.5) 23.7–40.4/27.8

27.2 (3.4) 21.7–34.8/27.3

0.098

Hexagonality, ‘‘touch up’’ mode (%) 50.7 (9.2) 32.5–72.8/49.7

53.0 (6.6) 36.2–66.6/52.6

0.201

Touch up duration (seconds) 551 (134) 361–853/521

388 (120) 226–653/362

,0.001

Results were expressed as mean (standard deviation), minimum–maximum/median. The automaticanalysis was less relevant at delivery than at receipt (see Table 1), consequently, a touch up analysisshould be recommended.

Presented in part at the 15th annual meeting of theEuropean Eye Bank Association held in Brussels on

17–19 January 2003.

PostScript 581


on Novem


http://bjo.bmj.com

/B

r J Ophthalm


ownloaded from

http://bjo.bmj.com/

tonometry, perimetry (Octopus G1 program,3 phases), and pachymetry (Tomey AL-1100).In addition, a 24 hours IOP curve withmeasurements at 7 am, 12 am, 5 pm, 9 pm,12 pm, 7 am was measured in all patients.

For a classification of study groups the 15˚colour stereo photographs were evaluatedqualitatively by two observers. Criteria forthe diagnosis in all glaucomas were increasedIOP and glaucomatous changes of the opticnerve head, including abnormally smallneuroretinal rim area in relation to the opticdisc size, abnormal neuroretinal rim shape,cup:disc ratios being higher vertically com-pared with horizontally, and localised ordiffuse loss of retinal nerve fibre layer.

All subjects were familiar with visual fieldtesting. Subjects with a higher than 12% rateof false-positive or false-negative responseswere excluded. A ‘‘perimetric’’ glaucomatousvisual field was defined as an Octopus G1field with (a) at least three adjacent testpoints having a deviation of equal to orgreater than 5 dB and with one test pointwith a deviation greater than 10 dB lower

than normal; (b) at least two adjacent testpoints with a deviation equal to or greaterthan 10 dB; (c) at least three adjacent testpoints with a deviation equal to or greaterthan 5 dB abutting the nasal horizontalmeridian, or (d) a mean visual field defectof more than 2.6 dB.

For APOE genotyping, genomic DNA wasextracted from anticoagulated blood afterisolation of peripheral lymphocytes followingthe ‘‘salting out’’ method. The APOE geneshows a polymorphism with three alleles (e2,e3, e4) (table 1). As allele e3 is considered tobe the ancestral allele, and e2 and e4 areconsidered as variants on the basis of singlepoint mutations, the e3e3 genotype was usedas reference.

ResultsThe mean (SD) IOP of the 24 h diurnal curvewas significantly higher in subjects with thee2 allele (independent samples t test,t = 22.57, p = 0.011, 17.7 (2.7) v 16.4(2.4) mm Hg) (fig 1). The maximum andminimum IOP were also increased ordecreased, but not significantly (Mann-Whitney U, p = 0.052, 20.5 (3.2) v 19.1(3.0) mm Hg, respectively. p = 0.178, 14.8(2.7) v 14.0 (2.5) mm Hg). This was approxi-mately continuous in the different studygroups (normals, OHT, pre-OAG, and OAG);however it was only significant in the normalsubjects for mean (t = 22.183, p = 0.043, 18.0(4.9) v 14.3 (2.5) mm Hg) and minimum IOP(U = 15, p = 0.031, 15.0 (3.9) v 11.7(2.5) mm Hg). The central corneal thicknesswas not different between the subjects withthe e2 allele and the reference group withe3e3 genotype (t = 20.035, p = 0.973, 587(33) v 586 (51) mm).

DiscussionThe results of this study show a significantassociation between the level of IOP and theAPO e2 allele. This may be supported by therecent findings that the APO e4 allele isassociated with higher risk for glaucomatouschanges that are not related to increasedIOP.2

It is not yet obvious how the APOE allelesmay be a source of genetic risk for glaucomaand increased IOP. It will be intriguing toinvestigate whether there is increased expres-sion of APOE in trabecular meshwork inglaucoma or an isoform dependent expres-sion in different types of glaucoma. Apossible role for ApoE promoter singlenucleotide polymorphisms as modifiers ofthe POAG phenotype has been hypothesised.4

To conclude, we have shown a significantassociation between APOE and glaucoma andIOP. Quite recently it was argued that an IOPreduction of 1 mm Hg from baseline willdecrease the risk of progression by about10%.5 Although in need of confirmation, ourdata emphasise the role of APOE in regula-tion of IOP and may indicate that we haveidentified a susceptibility gene for glaucoma.

As future perspective for the APOE alleles,analysis of a larger number of glaucomapatients—taking into account family history,age, and sex—will provide more detailedinsight.

A Junemann, N Wakili, C Mardin,G O H Naumann

Department of Ophthalmology, Friedrich-Alexander-University, Erlangen-Nuremberg, Germany

S Bleich, K Henkel, G Beck, J KornhuberDepartment of Psychiatry and Psychotherapy,

Friedrich-Alexander-University, Erlangen-Nuremberg,Germany

U ReulbachDepartment of Medical Informatics, Biometry and

Epidemiology, Friedrich-Alexander-University,Erlangen-Nuremberg, Germany

B Rautenstrauss, A ReisDepartment of Human Genetics, Friedrich-Alexander-

University, Erlangen-Nuremberg, Germany

Correspondence to: Dr A Juneman, Department ofOphthalmology, Friedrich-Alexander-University,

Erlangen-Nuremberg, Schwabachanlage 6 Erlangen,Germany; anselm.juenemann@

augen.imed.uni-erlangen

Accepted for publication 7 June 2003

References

1 Wiggs JL, Allingham RR, Hossain A, et al.Glaucoma-wide scan for adult onset primaryopen angle glaucoma. Hum Mol Genet2000;9:1109–17.

2 Vickers JC, Craig JE, Stankovich J, et al. Theapolipoprotein 4 gene is associated with elevatedrisk of normal tension glaucoma. Mol Vis2002;8:389–93.

3 Klaver CCW, Kliffen M, Duijn CM, et al. Geneticassociation of apolipoprotein E with age-relatedmacular degeneration. Am J Human Genet1998;63:200–6.

4 Copin B, Brezin AP, Valtot F, et al. ApolipoproteinE-promoter single-nucleotide polymorphismsaffect the phenotype of primary open-angleglaucoma and demonstrate interaction with themyocilin gene. Am J Hum Genet2002;70:1575–81.

5 Leske MC, Heijl A, Hussein M, et al. Factors forglaucoma progression and the effect of treatment:the early manifest glaucoma trial. ArchOphthalmol 2003;121:48–56.

The safety of anterior chamberparacentesis in patients withuveitisAnterior chamber (AC) paracentesis is avaluable procedure in the management ofuveitis, particularly in diagnosing infectivecauses.1 2 It may also be used therapeuticallyto lower intraocular pressure,3 and it providessamples for clinical research. Nevertheless,there have been isolated reports of ACparacentesis related serious complications,including endophthalmitis and cornealabscess.4 5 As the risk of trauma to the irisand lens are also major concerns, AC para-centesis is often used with reluctance.

Figure 1 Maximal and mean IOP of the24 hour diurnal curve of 130 individuals(normal controls, OHT, pre-OAG, and OAG;+ cases with more than 1.5 box lengths fromthe upper or lower edge of the box. The boxlength is the interquartile range.

Table 1 Distribution of APOE genotypes and allele frequency

Frequency in APOE characteristic (n (female) mean age in years)

Normals 32(17) 54.8

OHT 54 (27)53.3

Pre-OWG 55(25) 54.4

OWG 41(20) 56.3 S

Genotypee2e2 0 1 (0) 0 0 1 (0) 63.0e2e3 3 (2) 11 (5) 12 (7) 6 (2) 32 (16) 56.0e2e4 6 (5) 0 1 (0) 0 7 (5) 53.4e3e3 14 (6) 32 (17) 24 (12) 27 (15) 97 (50) 54.5e3e4 9 (6) 9 (5) 18 (6) 8 (3) 44 (18) 53.8e4e4 0 1 (0) 0 0 1 (0) 43.0

Allele frequencye2 9 13 13 6 41e3 40 84 78 68 270e4 15 11 19 8 53

582 PostScript


on Novem


http://bjo.bmj.com

/B

r J Ophthalm


ownloaded from

http://bjo.bmj.com/

Although there are many studies on theanalysis of aqueous humour obtained fromAC paracentesis, our literature search showedonly one publication on the safety of ACparacentesis.6

The purpose of this study was to describe amethod of AC paracentesis that can be easilyperformed as an outpatient procedure withthe patient sitting at the slit lamp.

Methods and resultsA total of 70 patients (41 male, 29 female)aged 18–83 years (median 39 years) withvarious types of active uveitis attending theBirmingham and Midland Eye Centre under-went AC paracentesis. Fourteen paracenteseswere performed for diagnostic purposes whilethe remainder for experimental analysis aspart of another study. Patients with dilatedand undilated pupils were included. Localresearch ethics committee approval andinformed consent was obtained.

Benoxinate 0.4% eye drops (minims) wereinstilled three times over a 3 minute period,followed by instillation of betadine 5% anti-septic solution that had been drawn up intothe empty benoxinate minim container. Thepatient was positioned at the slit lamp, theupper lid and eyelashes held out of the wayby an assistant. No lid speculum wasrequired.

Of the 70 paracenteses, 48 were performedusing a 27 gauge needle attached to aninsulin syringe, while the remaining 22 wereperformed using an aqueous pipette. Where a27 gauge needle was used, this was insertedat the paralimbal clear cornea in a planeabove and parallel to the iris with the bevel ofthe needle facing forward until the wholebevel penetrated the cornea (fig 1). Underdirect vision, the sampler pulled the plungerof the syringe to aspirate the aqueous. Theaqueous pipette (Visitec, Sarasota, FL, USAdesigned by O’Rourke) consists of a short 30gauge needle mounted inside plastic tubing,which in turn is connected to a soft poly-ethylene suction-infusion bulb. The bulb wassqueezed to create a vacuum and the needleinserted at the limbus as described above(fig 2). When pressure on the bulb wasreleased, aqueous spontaneously filled thepipette. Using either method, the eye can befixed with a pair of forceps at the oppositelimbus, if necessary. After sampling anantibiotic drop was prescribed for three days.The whole procedure takes less than fiveminutes. All patients were re-examined 20minutes after the procedure and 1–2 weekslater.

Two patients had an air bubble inadver-tently injected into the AC. The visual acuity

returned to normal at review in both cases.One patient developed an acute allergicconjunctivitis to betadine, which settled aftertreatment with prednisolone 0.5% eye drops.None of the 70 patients developed detectablecorneal damage, lens changes, orendophthalmitis.

CommentVarious methods for performing AC para-centesis have been described.6–9 However, ourliterature search only identified one systema-tic report investigating the safety of ACparacentesis.6 This technique required thepatient to lie supine under the microscope,needed insertion of a lid speculum andpreincision of the cornea with a 15˚ microsharp blade, and the aqueous was aspiratedusing a 27 gauge needle on a tuberculinsyringe. No serious complications werereported in 361 uveitis patients. A smallhyphaema occurred in 5/72 (6.9%) patientsexamined 30 minutes after the paracentesis.The method described by O’Rouke using theaqueous pipette7 is relatively new and nosystematic analysis of its safety profile hasbeen published. Other methods for AC para-centesis include not using a syringe8 or asyringe with the plunger removed,9 thusavoiding any potential complications asso-ciated with aspirating with a plunger, butcollecting the aqueous specimen may betechnically difficult.

The cases of inadvertent injection of an airbubble into the AC both occurred using theaqueous pipette and was most likely causedby air trapped inside the bulb prior toinserting the pipette into the AC. We recom-mend ensuring the bulb is thoroughlydepressed to evacuate all air before insertingthe needle into the eye. Pressure on the bulbmust also be maintained while the needle isbeing inserted, to avoid air entry.

Our study showed that performing ACparacentesis with the patient sitting at theslit lamp is safe using either the 27 gaugeneedle or the aqueous pipette. Preincisionwith a sharp blade and the use of a lidspeculum is unnecessary.

AcknowledgementsWe are grateful to Robert Harvey and Salman Mirzafor their help in this study.

C M G Cheung, O M Durrani, P I MurrayBirmingham and Midland Eye Centre, Sandwell and

West Birmingham Hospitals NHS Trust, City Hospital,Birmingham, UK

Correspondence to: Professor P I Murray, AcademicUnit of Ophthalmology, Division of Immunity and

Infection, The University of Birmingham, Birmingham

and Midland Eye Centre, Sandwell and WestBirmingham Hospitals NHS Trust, City Hospital,

Dudley Road, Birmingham B18 7QU;[email protected]


References

1 Boer de JH, Luyendijk L, Rothova A, et al. Analysisof ocular fluids for local antibody production inuveitis. Br J Ophthalmol 1995;79:610–6.

2 Boer de JH, Verhagen C, Bruinenberg M, et al.Serological and polymerase chain reactionanalysis of intraocular fluids in the diagnosis ofinfectious uveitis. Am J Ophthalmol1996;121:650–8.

3 Carnahan MC, Platt LW. Serial paracenteses inthe management of acute elevations of intraocularpressure. Ophthalmology 2002;109:1604–6.

4 Helbig H, Noske W, Kleineidam M, et al.Bacterial endophthalmitis after anteriorchamber paracentesis. Br J Ophthalmol1995;79:866.

5 Azuara-Blanco A, Katz LJ. Infectious keratitis in aparacentesis tract. Ophthalmic Surg Lasers1997;28:332–3.

6 Van der Lelij A, Rothova A. Diagnosticanterior chamber paracentesis in uveitis:a safe procedure? Br J Ophthalmol1997;81:976–9.

7 O’Rourke J, Taylor DM, McDonald P, et al. Anaqueous paracentesis pipet. Ophthalmic Surg1991;22:166–7.

8 May DR, Noll FG. An improved approach toaqueous paracentesis. Ophthalmic Surg1988;19:821–2.

9 Grewal SPS. A technique for paracentesis.Ophthalmic Surg 1989;20:525.

Rapid recovery of night blindnessdue to obesity surgery aftervitamin A repletion therapyNight blindness is the most common andearliest symptom of vitamin A deficiency.1

The latter can be caused by general malnutri-tion, malabsorption of vitamin A, or impairedvitamin A metabolism due to liver disease.2

Several surgical methods are currently usedfor the treatment of obesity. In the Scopinaroprocedure, a biliopancreatic bypass is com-bined with a bypass of part of the smallbowel, thus promoting intestinal malabsorp-tion.3

The fat soluble vitamin A can exist asretinol, its ester, and retinoic acid. It hasseveral roles in ocular metabolism: it isessential for corneal and conjunctival epithe-lial cell RNA and glycoprotein synthesis,4

while retinal is the crucial chromophorewhich combines with both rod and coneopsins to form rhodopsin and activated coneopsins, which are essential for phototrans-duction.

Case reportA 39 year old man presented with a 6 monthhistory of night blindness, progressing morerapidly in the past 2 weeks. Three yearsbefore he had undergone a partial gastrect-omy and biliopancreatic derivation for mor-bid obesity (Scopinaro procedure). His meanbody mass index (BMI) decreased from50 kg/m2 to 31 kg/m2 3 years later.

At presentation, visual acuity was 6/5 inboth eyes with a spherical correction of +0.75dioptres. Slit lamp examination and fundu-scopy were unremarkable in both eyes. Con-centric narrowing in both eyes could be seenon Goldmann visual field (VF) analysis (fig 1A).Goldmann-Weekers dark adaptometry (DA)

Figure 1 Anterior chamber paracentesis with27 gauge needle (pig’s eye used fordemonstration).

Figure 2 Anterior chamber paracentesis withaqueous pipette (pig’s eye used fordemonstration).

PostScript 583


on Novem


http://bjo.bmj.com

/B

r J Ophthalm


ownloaded from

http://bjo.bmj.com/

showed a considerable decrease in sensitivity(fig 1B). Electro-oculography was subnormalbefore therapy with a lightPeak/darkTrough

ratio of 166% for the right eye and 146%for the left eye (normal ratio .180%).ISCEV standard electroretinography5

showed only minimal residual scotopicresponses in both eyes. The SRC of vitaminA before therapy was 14 mg/dl (normal range30–80). Vitamin E levels (0.49 mg/dl; normal0.5–1.8) and total protein levels were slightlysubnormal. Vitamin B and vitamin D levelswere normal.

Our patient was given 60 000 IU retinol/day and vitamin E 140 mg/day (Rovigon,Roche).

After only 3 days, partial normalisation ofGoldmann VFs occurred. After 3 days ofvitamin A supplementation, scotopic ERGresponses had already improved to one thirdof normal (fig 2). Subjectively, the patientreported a ‘‘sudden visual recovery’’ 3 daysafter initiation of therapy.

After 5 days of therapy the EOG Lp/Dt ratioreturned to near normal.

Ten days after initiation of treatment, allERG parameters returned to normal (fig 2).Complete normalisation of DA was also seen(fig 1B).

From day 36 Goldmann visual fields wereconsidered to be normal. The ERG (fig 2) andEOG had completely normalised by then.

After 135 days of repletion therapy SRC ofvitamin A was still only 26 mg/dl, whilevitamin E levels returned to normal(0.6 mg/dl). Treatment was maintained.

CommentNormal biliary secretion, fat absorption,dietary protein intake, and the presence ofzinc are necessary for fat soluble vitaminabsorption.6

Vitamin A has a major role in photorecep-tor function because it combines, in theform of its 11-cis isomer, with photoreceptoropsins to form rhodopsin and activated coneopsins.

At presentation, the ERG in our patientshowed a considerable decrease in rod and, toa lesser extent, in cone function.

After 3 days of vitamin A repletion asignificant improvement in the scotopicresponses was noted. All ERG responsesnormalised completely after only 10 days oftherapy. This rapid recovery of all electro-physiological and clinical parameters indi-cates that vitamin A deficiency was still inthe earlier stages. The lag between obesitysurgery and symptoms can be attributed tothe presence of considerable liver stores ofvitamin A when surgery was performed.

Our patient was repleted with 15 timesthe recommended daily allowance (RDA)of retinol (RDA of retinol 4000 IU/day)and 12 times that of vitamin E (RDA ofvitamin E 12 IU/day). Interestingly,vitamin E deficiency seems to decreasethe amount of vitamin A which can bestored in the retina.7 Long term vitaminreplacement therapy is essential after bilio-pancreatic derivation surgery of theScopinaro type.

Only a limited number of reports havedescribed cases of vitamin A deficiencyfollowing bowel surgery for obesity.

In 1999 Smets et al described a caseof night blindness and optic neuropathyafter biliopancreatic bypass with nor-malisation of all electrophysiologicalparameters when retested after 10 months.

No cone dysfunction was reported,despite SRC well below those in ourpatient.8

In all reports of vitamin A deficiencyto date, the rod involvement is seenearlier and is more extensive than cone

Figure 1 (A) Goldmann visual field analysis; although peripheral limits as tested with object V4 ofGoldmann remained normal, only test object I4 was perceived more centrally, indicating loss ofretinal sensitivity. On day 3 of therapy, the patient could already perceive test object I2 and I3illustrating partial normalisation, while, unexpectedly, peripheral limits were more constricted.From day 36 Goldmann visual fields were considered to be normal. (B) Goldmann-Weekers darkadaptometry (DA) showed considerable decrease in sensitivity before repletion therapy; day 1 is atpresentation, before treatment; considerable improvement seen on day 3, with completenormalisation on day 10.

584 PostScript


on Novem


http://bjo.bmj.com

/B

r J Ophthalm


ownloaded from

http://bjo.bmj.com/

involvement.9 The reasons for thisremain obscure, although rods are knownto be more dependent on availabilityof vitamin A from the retinal pigmentepithelium.10

In conclusion, our case proves thatmalabsorption caused by biliopancreaticderivation surgery of the Scopinaro typecan induce vitamin A deficiency withprogressive rod-cone dysfunction, as well

as deficiencies of all fat soluble vitaminsand low plasma proteins.

In the early stages of vitamin A deficiency,recovery of visual function rapidly follows afteroral repletion therapy, and can be nearly com-plete 1 week after initiation of such therapy.

Y Spits, J-J De Laey, B P LeroyDepartment of Ophthalmology and Centre for MedicalGenetics, Ghent University Hospital, Ghent, Belgium

Correspondence to: B P Leroy, Department ofOphthalmology and Centre for Medical Genetics,

Ghent University Hospital, De Pintelaan 185, Ghent,Belgium; [email protected]


References

1 Harris EW, Loewenstein JI, Azar D. Vitamin Adeficiency and its effects on the eye. IntOphthalmol Clin 1998;38:155–61.

2 Groff JL, Gropper SS, Hunt S. Advanced nutritionand human metabolism. St Paul, MN: WestPublishing Company, 1995:295–7.

3 Scopinaro N, Gianetta E, Civalleri D. Two yearsof clinical experience with biliopancreatic bypassfor obesity. Am J Clin Nutr 1980;33:506.

4 Smith J, Steinemann TL. Vitamin A deficiency andthe eye. Int Ophthalmol Clin 2000;40:83–91.

5 Marmor MF, Zrenner E. Standard for clinicalelectroretinography (1999 update). DocOphthalmol 1998;97:143–56.

6 Tsinopoulos I, Nousia-Arvanitakis S, Galli-Tsinopoulou A, et al. Role of electroretinographyin the assessment of retinal function as anindicator of vitamin A status. Doc Ophthalmol2000;101:211–21.

7 Robison WG, Kuwabara T, Bieri JG. Vitamin Edeficiency and the retina: photoreceptor andpigment epithelial changes. Invest Ophthalmol VisSci 1979;18:683–90.

8 Smets RM, Waeben M. Unusual combination ofnight blindness and optic neuropathy afterbiliopancreatic bypass. Bull Soc Belge Ophtalmol1999;271:93–6.

9 Scholl HP, Zrenner E. Electrophysiology in theinvestigation of acquired retinal disorders. SurvOphthalmol 2000;45:29–47.

10 Bridges CDB. Retinoids in photosensitive systems.In: Sporn MB, Roberts AB, Goodman DS, eds. Theretinoids. New York: Academic Press,1984:125–76.

Incision-less frontalis suspensionFrontalis suspensions with alloplastic slingsare well established.1 2 The thick eyebrowskin of infants is prone to scar formation.Forehead scars caused by frontalis suspen-sion procedures can be problematic. Wedescribe a technique of congenital ptosissurgery that avoids eyebrow incisions.

Surgical techniqueThis new procedure utilises a Nylon mono-filament suture for frontalis suspension. TheNylon suture is passed in a circlage fashionvia puncture wounds without making eye-brow incisions. Two puncture sites, approxi-mately 10 mm apart, are marked 3 mmabove the lash line centred over the area ofdesired maximal eyelid elevation. Anothertwo puncture sites are marked above theeyebrow approximately in line with thelateral and medial canthi. The path of thecirclage is marked out by joining the markedpuncture sites. The eyelid and eyebrow areinfiltrated with local anaesthetic with adre-naline (epinephrine).

A Keith needle is dual threaded with a 4/0Nylon and a 4/0 Vicryl suture. It is thenpassed from one eyelid puncture site towardsthe corresponding eyebrow exit site in a sub-orbicularis plane (fig 1, top left) with theglobe protected by a lid guide. From this site,the needle is passed through the needle trackto the adjacent eyebrow puncture site (fig 1,top right) and then down towards theremaining eyelid puncture site. At this pointin the procedure, the ends of the Nylon andVicryl sutures emerge through the two eyelidpuncture sites. The two ends of the Vicryl

Figure 2 ERGs on day 1 (before treatment) and subsequent days as indicated; normal control forcomparison at bottom. At presentation, only minimal residual scotopic responses were seen in botheyes. Amplitudes of oscillary potentials were only residual. Responses to a single bright white flashin the dark adapted eye were minimal, with amplitudes of approximately one quarter of normal fora-wave and 1/13 for b-wave in both eyes. Single flash cone responses still had normal a-waves,while b-waves were two thirds of normal in both eyes. The 30 Hz flicker responses were four fifthsof normal in right eye and normal in left eye. After 3 days of vitamin A supplementation, scotopicresponses improved to one third of normal in both eyes. Ten days after initiation of treatment, allERG parameters were within normal limits, although amplitudes still increased up to day 22.

PostScript 585


on Novem


http://bjo.bmj.com

/B

r J Ophthalm


ownloaded from

http://bjo.bmj.com/

suture are then manoeuvred in a sawingfashion to create friction to release skindimpling at the eyebrow exit sites (fig 1,bottom left). The Vicryl suture is thenremoved and the Nylon suture needle (SHneedle) is passed from one eyelid puncturesite to another via a deep, partial thicknesstarsal passage with the eyelid everted toensure no full thickness penetration (fig 1,bottom right). The two ends of the Nylonsuture, exiting at one eyelid puncture site, aretied and the tension adjusted to achieve thedesired lid elevation and contour.Occasionally, peaking of the eyelid occursand can be managed by slightly enlarging thepuncture site at the tight suture end with aWestcott scissors and gentle spreading toundermine the soft tissues around the suture.This undermining action helps to release thesuture tension on the puncture site tosmoothen out the lid contour but should bedone carefully to avoid cutting the suture.The puncture sites usually do not requireclosure.

CommentWe performed this surgery on three infantswith visually significant congenital ptosis.The mean age and follow up period of theinfants were 5.6 months and 6.9 monthsrespectively. The visual axis was cleared inall patients as measured by an improvementof their margin reflex distance one (MRD1).The lid contour was good in all patients. Anexample is illustrated in figure 2. There wereno intraoperative or postoperative complica-tions. The eyelid puncture sites healed with-out visible scar.

This minimally invasive surgery is scarlessand can be performed with little trauma tothe orbicularis oculi muscle. We realise thatthe results of frontalis suspension usingallogenic material are not permanent1 andmay be associated with late failures.However, this is a simple, safe, temporarymeasure that elevates the eyelid for visualdevelopment until the child is old enough fordefinitive surgery using autologous3 orbanked tissues.4

C-C Yip, R A Goldberg, T L Cook, J D McCannOrbital and Ophthalmic Plastic Surgery Division, Jules

Stein Eye Institute, UCLA School of Medicine, LosAngeles, CA, USA

C-C YipThe Eye Institute, National Health Care Group,

Singapore. Tan Tock Seng Hospital

Correspondence to: Robert A Goldberg, MD, FACS,Jules Stein Eye Institute, Orbital and Ophthalmic

Plastic Surgery Division, 100 Stein Plaza, PO Box957006, Los Angeles, CA 90095-7006, USA;

[email protected]


References

1 Wagner RS, Mauriello JA Jr, Nelson LB, et al.Treatment of congenital ptosis with frontalissuspension: a comparison of suspensorymaterials. Ophthalmology 1984;91:245–8.

2 Steinkogler FJ, Kuchar A, Huber E, et al. Gore-Tex soft-tissue patch frontalis suspensiontechnique in congenital ptosis and inblepharophimosis-ptosis syndrome. Plast ReconstrSurg 1993;92:1057–60.

3 Crawford JS. Repair of ptosis using frontalismuscle and fascia lata: a 20-year review.Ophthalmic Surg 1977;8:31–40.

4 Wilson ME, Johnson RW. Congenital ptosis.Long-term results of treatment using lyophilizedfascia lata for frontalis suspensions.Ophthalmology 1991;98:1234–7.

Spontaneous resolution of sixthnerve palsy with ipsilateralcavernous carotid dolichoectasiaA 73 year old man was evaluated for thesudden onset of binocular horizontal diplopiawhich was worse in left gaze and whichbegan 1 day before initial examination. Healso complained of a dull headache over hisleft brow. He had a medical history of hip andknee surgery and was taking no medications.He was a 50 pack a year smoker but had noother history of vascular disease, includinghypertension and diabetes mellitus. He hadno previous history of strabismus or eyemuscle surgery. His referring ophthalmolo-gist was concerned about giant cell arteritis(GCA) and ordered a Westergren erythrocytesedimentation rate test, which was 15 mm inthe first hour.

Additional history revealed that he had nojaw claudication, scalp tenderness, or othersymptoms of GCA. Visual acuity was 20/25 inboth eyes and his colour vision and con-frontation visual fields were normal. Hispupils were equal in size and briskly reactivewithout a relative afferent pupillary defect. Aleft abduction deficit was noted (fig 1) and,with alternate cover testing, there was a10 prism dioptre esotropia in primary posi-tion and at near, which increased to 20 prismdioptres on left gaze and decreased to 2 prismdioptres in right gaze. He had slowedsaccades of the left lateral rectus muscle.There was no evidence of ptosis or ocularmotor synkinesis. The remainder of hiscranial nerve and dilated fundus examinationwere normal. Magnetic resonance imaging(MRI) (fig 2) and magnetic resonanceangiography (MRA) (fig 3) of the brainrevealed a lateral course of the left cavernouscarotid artery consistent with dolichoectasia.

Figure 1 (Top left) A Keith needle, threaded with a Nylon and a Vicryl suture, is passed from oneeyelid puncture site to the corresponding eyebrow puncture site in a sub-orbicularis plane. (Topright) The Keith needle, loaded with the sutures, is passed from one eyebrow puncture site toanother. (Bottom left) The 4/0 Vicryl suture is manoeuvred in a ‘‘sawing’’ manner with both handsto release the soft tissues at the eyebrow puncture sites to avoid skin dimpling. (Bottom right) The4/0 Nylon suture is passed from one eyelid puncture site to another taking a partial thickness bite.The eyelid is everted during this tarsal passage to ensure no full thickness penetration.

Figure 2 (Left) Preoperative picture of a 1 year old girl with bilateral congenital ptosis and a chin-up position. The child has bilateral poor levator function. (Right). Postoperative picture of the patientafter bilateral frontalis suspension using the described technique. Both the eyelids are adequatelyelevated with a satisfactory contour although the chin-up position is not totally ameliorated.

The authors have no financial interest in this paper.

586 PostScript


on Novem


http://bjo.bmj.com

/B

r J Ophthalm


ownloaded from

http://bjo.bmj.com/

Follow up examination 1 month laterrevealed no history of variability of thediplopia and no change in the ocular mis-alignment; however, over the next 2 monthsthe patient reported a gradual improvementin symptoms. He returned 3 months after theinitial onset of symptoms and his abductiondeficit had resolved. There was no evidence ofan ocular misalignment with alternate covertesting. Repeat MRI/MRA showed no changein the calibre or position of the left cavernouscarotid artery. He has reported no newsymptoms with 1 month of additional followup.

CommentDolichoectasia, or pathological enlargement,of the intracranial arteries is a finding rarelyseen with neuroimaging or arteriography.Arteriosclerosis, with thinning of the mediaand defects in the internal elastic laminae ofthe vessel walls, is thought to predispose toprogressive enlargement of the vessel lumen.1

Ectasia of the intracranial arteries is believedto cause symptoms because of compression ofadjacent structures and/or ischaemia second-ary to intraluminal thrombus formation andblockade of perforating vessels along thelength of the dolichoectatic vessel.1

Dolichoectasia of the cavernous carotidartery has been suggested as an infrequentcause of sixth nerve paresis. One in 23patients with carotid ectasia (in a series ofapproximately 40 000 patients undergoingcarotid arteriography) was found to have anacute sixth nerve palsy with ‘‘good recovery,’’although the clinical course was not specified.2

Ipsilateral dolichoectasia was noted in a59 year old man with seven episodes of sixthnerve paresis, each lasting between 2–5 weeks.3 The authors did not provide anexplanation for the mechanism of recurrence.A single patient with bilateral sixth nerve

paresis was reported to have bilateral carotiddolichoectasia as the underlying cause.4

However, in the discussion the causal relationof the dolichoectasia, presumably from com-pression of the carotid artery, was called intoquestion. In addition, dolichoectasia of thecavernous carotid artery has been noted inpatients without ocular motor deficits.1

This patient’s left sixth nerve paresisspontaneously resolved 3 months after theinitial onset of symptoms. Despite the pre-sence of ipsilateral cavernous carotid doli-choectasia, his clinical course is mostconsistent with that of a vasculopathic sixthnerve paresis. Whether the dolichoectasiawas causative or an incidental finding is notclear in this patient. Arterial dissection in apreviously ectatic vessel has been suggestedas an explanation for the acute onset ofsymptoms in patients with dolichoectasia5;however, no evidence of arterial dissectionwas seen in this patient’s MRI/MRA.Ischaemia of the vaso vasorum of the sixthnerve, perhaps because of intraluminalthrombus formation, may have resulted in avasculopathic sixth nerve palsy, but there wasno evidence of thrombus formation on theMRI/MRA.

Because the causative mechanism inpatients with persistent sixth nerve paresisfrom presumed dolichoectasia is not certaintreatment guidelines are not clear. Monocularocclusion and prism therapy may providetemporary or long lasting relief of diplopia.Neurosugrical intervention to relievemechanical compression between the caver-nous carotid artery is a difficult, potentiallylife threatening, procedure. Extraocular mus-cle surgery may correct the ocular misalign-ment, without treating the underlyingmechanical compression, with uncertain longterm benefit. Spontaneous resolution of theleft sixth nerve palsy in this patient withipsilateral carotid dolichoectasia suggeststhat a period of careful observation should

precede plans for surgical correction of theocular misalignment.

R ForoozanNeuro-Ophthalmology Service, Cullen Eye Institute,Baylor College of Medicine, 6565 Fannin NC-205,Houston, TX 77030, USA; [email protected]


References

1 Anson JA, Lawton MT, Spetzler RF.Characteristics and surgical treatment ofdolichoectatic and fusiform aneurysms.J Neurosurg 1996;84:185–93.

2 Yu YL, Moseley IF, Pullicino P, et al. The clinicalpicture of ectasia of the intracerebral arteries.J Neurol Neurosurg Psychiatry 1982;45:29–36.

3 Blumenthal EZ, Gomori JM, Dotan S. Recurrentabducens nerve palsy caused by dolichoectasia ofthe cavernous internal carotid artery.Am J Ophthalmol 1997;124:255–7.

4 Neugebauer A, Kirsch A, Fricke J, et al. Newonset of crossed eyes in an adult. SurvOphthalmol 2001;45:335–44.

5 Mizutani T, Aruga T. ‘‘Dolichoectatic’’intracranial vertebrobasilar dissecting aneurysm.Neurosurgery 1992;31:765–73; discussion 773.

Intravitreal triamcinoloneacetonide as treatment forextensive exudative retinaldetachmentCoats’ disease or entities like Coats’ diseaseare characterised by a marked exudativeretinal detachment with leakage of peripheralretinal vessels, pronounced subretinal deposi-tion of lipids, and eventual progression tototal retinal detachment. In some situations,iris neovascularisation can occur, suggestingan angiogenetic component in the course ofthe disease. In view of the subretinal exuda-tion from the leaking retinal vessels and thepossibly neovascular aspect in the diseaseprocess, the purpose of this study was toevaluate whether intravitreal triamcinoloneacetonide may be helpful in the treatment ofCoats’ like diseases. Intravitreal triamcino-lone acetonide has recently been shown tohave a pronounced anti-oedematous andpossibly anti-angiogenic effect in diseasessuch as diffuse diabetic macular oedema,proliferative diabetic retinopathy, chronicpre-phthisical ocular hypotony, chronic uvei-tis, and persistent pseudophakic cystoidmacular oedema.1 2

Case reportThe prospective clinical interventional casereport included two patients who presentedwith subtotal exudative retinal detachment.A 39 year old female patient showed anextensive exudative retinal detachmentextending from the temporal periphery ofthe fundus to the macular region. Diagnosedwith Coats’ disease in her early teens, she hadreceived multiple xenon arc coagulations aswell as argon laser coagulations. Her visualacuity was 0.02. Intraocular pressure mea-sured 13 mm Hg. The second patient was a75 year old woman presenting with almosttotal exudative retinal detachment withmarked subretinal deposition of lipids.Visual acuity was 0.05. Intraocular pressuremeasured 21 mm Hg.

Figure 1 Ocular motility testing reveals a left abduction deficit.

Figure 2 T2 weighted MRI reveals a lateralcourse of the left cavernous carotid artery(arrow).

Figure 3 MRA of the circle of Willis shows thelateral course of the left cavernous carotidartery (arrow).

The author has no proprietary interest in any contentswithin this manuscript

PostScript 587


on Novem


http://bjo.bmj.com

/B

r J Ophthalm


ownloaded from

http://bjo.bmj.com/

Under topical anaesthesia, both patientsreceived an intravitreal injection of 25 mgtriamcinolone acetonide, which was trans-conjunctivally applied through the parsplana. Both patients were fully informedabout the experimental character of thetreatment and had given informed consent.The technique has already been described indetail.2 Follow up after the injections were2 years and 10 months, respectively.

After the injection, visual acuity remainedunchanged, and intraocular pressure rangedbetween 10 and 15 mm Hg in the firstpatient. In the second patient, visual acuityeventually decreased to light perception afterthe injection. Intraocular pressure rangedbetween 19 and 25 mm Hg. In both patients,flare in the anterior chamber and in thevitreous cavity, as assessed by slit lampbiomicroscopy, decreased markedly. Uponophthalmoscopy, the extent of exudativeretinal detachment increased slightly, withsubretinal strands being stronger and morevisible.

CommentAlthough intravitreal triamcinolone aceto-nide can markedly reduce retinal oedema ineyes with diffuse diabetic macular oedemaand pseudophakic cystoid macular oedema,intravitreal triamcinolone acetonide was notpronouncedly helpful in reducing subretinaloedema and re-attaching the retina in thetwo patients presented in this study. Thisresult was unexpected in view of the pre-sumed anti-phlogistic and anti-proliferativeeffect of steroids such as triamcinoloneacetonide.1 2 It may be explained by aprevious experimental study in which triam-cinolone acetonide inhibited the proliferationof rabbit dermal and conjunctival fibroblastsin cell culture at 150 mg/l, but paradoxicallyincreased proliferation almost twofold atconcentrations ranging from 1–30 mg/l underidentical culture conditions.3 As long as theinfluence of steroids on the proliferation ofretinal pigment epithelium cells is unclear,intravitreal triamcinolone acetonide maythus cautiously be taken as adjunct treatmentof marked exudative retinal detachment ineyes with a Coats’ like disease. A similarconclusion was drawn in a recent study oneyes with proliferative vitreoretinopathy, inwhich pars plana vitrectomy was combinedwith an intravitreal injection of 25 mg oftriamcinolone acetonide, and in which unex-pectedly, the recurrence rate of proliferativevitreoretinopathy was not markedly dimin-ished.4 Future randomised studies as well asinvestigations evaluating the effect of intra-vitreal steroids combined with other drugssuch as 5-fluorouracil on the proliferation ofretinal pigment epithelium cells and retinaldetachment rate5 may be warranted.

J B JonasDepartment of Ophthalmology, Faculty of ClinicalMedicine Mannheim, Ruprecht-Karls-University of

Heidelberg, Germany

Correspondence to: Dr J Jonas, Universitats-Augenklinik, Theodor-Kutzer-Ufer 1-3, 68167Mannheim, Germany: [email protected].

uni-heidelberg.de

Accepted for publication 1 July 2003

References

1 Machemer R, Sugita G, Tano Y. Treatment ofintraocular proliferations with intravitreal steroids.Trans Am Ophthalmol Soc 1979;77:171–80.

2 Jonas JB, Sofker A. Intraocular injection ofcrystalline cortisone as adjunctive treatment ofdiabetic macular edema. Am J Ophthalmol2001;132:425–7.

3 Blumenkranz MS, Claflin A, Hajek AS. Selectionof therapeutic agents for intraocular proliferativedisease. Cell culture evaluation. Arch Ophthalmol1984;102:598–604.

4 Jonas JB, Sofker A, Hayler J, et al. Intravitrealcrystalline triamcinolone acetonide as additionaltool in pars plana vitrectomy for complicatedproliferative vitreoretinopathy? Acta Ophthalmol2003;(in press).

5 Berger AS, Cheng CK, Pearson PA, et al.Intravitreal sustained release corticosteroid-5-fluoruracil conjugate in the treatment ofexperimental proliferate vitreoretinopathy. InvestOphthalmol Vis Sci 1996;37:2318–25.

Long term efficacy and safety ofbotulinum toxin A injection forcrocodile tears syndromeGustatory lacrimation, also called crocodiletears syndrome (CTS), is an autonomicsynkinesia in which patients tear excessivelyin response to salivary stimuli. It occurs mostcommonly in the setting of idiopathic ortraumatic facial palsy and is thought to resultfrom aberrant reinnervation of the lacrimalgland by salivary efferent fibres from eitherthe seventh or ninth cranial nerve. Manypatients tolerate CTS and require no inter-vention. For patients who cannot tolerateCTS, past treatments have included anti-cholinergic drugs, subtotal resection of thepalpebral lobe of the lacrimal gland, andresection of the tympanic nerve proximal tothe lesser superficial petrosal nerve. None ofthese approaches is optimal because oflimited efficacy, morbidity, or both.

Injection of botulinum toxin A has beenshown to be effective for a host of disorderscharacterised by involuntary muscle spasms,including blepharospasm, hemifacial spasm,and torticollis. Botulinum toxin A also hasbeen used to treat a number of localisedautonomic disorders, including axillaryhyperhidrosis, palmar hyperhidrosis, andFrey syndrome.1 In 1998, Boroojerdi et alreported the successful treatment of CTS byinjection of botulinum toxin A directly intothe lacrimal gland.2 Since then, there havebeen five reports of similar treatments, all ofwhich were successful.3–7 All of these studiesreport complete or near complete resolutionof the syndrome within a week with onlyinfrequent, minor, and reversible complica-tions. We now report a patient with CTS whohas been successfully managed for 3 yearswith injections of botulinum toxin A.

Case reportA 38 year old man presented in July of 2000with a 6 month history of right sided tearingand hyperhidrosis of the auriculotemporalregion when eating or when hungry.Fourteen months previously, he had under-gone a total right parotidectomy for a mixedtumour of the parotid gland. Immediatelyafter surgery, he had a complete right sidedfacial palsy and numbness of the right lowerface. The facial palsy resolved completely amonth later, but the facial numbness per-sisted. Eight months later, the patient beganto experience increased tearing on the rightafter eating, most notable after eating mints.On examination, the patient had normalfacial movement but decreased sensation tolight touch in the region of the second

division of the trigeminal nerve and spasmof the right lower lid on palpation. Inaddition, he perspired from the right side ofthe face and had tearing from the right eyeupon eating. His ocular and neurologicexaminations were otherwise unremarkable.

In light of the bothersome nature of CTS tothis patient, we felt a trial of botulinum toxinA was warranted. Accordingly, after obtain-ing consent, we injected botulinum toxin A(Botox 2.5 U) transconjunctivally into thepalpebral lobe of the right lacrimal glandunder direct visualisation at the slit lampbiomicroscope without complication. Thepatient’s excess epiphora completely resolvedwithin 5 days, and he remained asympto-matic for 11 months. He has subsequentlyrequired injections of botulinum toxin Aevery 7–11 months, experiencing relief ofsymptoms with each injection. There havebeen no complications from any of theinjections.

CommentSeveral different groups have now reported atotal of 12 cases of CTS treated withbotulinum toxin A.2–7 All of the patientsreported have had complete or near completeshort term resolution of symptoms withdoses of botulinum toxin A (Botox) rangingfrom 2.5–60 U. The higher doses seem tohave no additional benefit in terms of efficacyor duration.

Injection of botulinum toxin A for CTSappears to be safe, although minor complica-tions occasionally occur. Two of the patientsinjected transcutaneously developed ptosis,one accompanied by a superior rectus palsy,5 7

whereas two others developed dryness of theinjected eye.7 These complications resolvedover several months. No cases of ptosis orextraocular muscle weakness have beenreported after transconjunctival injection,and our patient has had no complicationswith any of his injections over the last 3years.

As with all new treatments, there areconcerns about long term efficacy and safety.Botox has been found both safe and effectiveat the neuromuscular junction, but its longterm effects on the peripheral autonomicsystem are unknown and one might postu-late that the repeated minor trauma of theinjection could eventually impair lacrimalgland function. In light of these concerns, itis encouraging to be able to report thatrepeated injections of botulinum toxin Acontinue to be effective in controlling thispatient’s CTS for 3 years without complica-tion.

D E BarananoDepartments of Neuroscience and Medicine, JohnsHopkins School of Medicine, Baltimore, MD, USA

N R MillerWilmer Eye Institute, Division of Neuro-

ophthalmology, Departments of Ophthalmology andNeurology, Johns Hopkins School of Medicine,

Baltimore, MD USA

Correspondence to: Dr Neil R. Miller, MD, MaumeneeB-109, Johns Hopkins Hospital, 600 North WolfeStreet, Baltimore, 21287, USA; [email protected]


References

1 Naumann M. Evidence-based medicine:botulinum toxin in focal hyperhidrosis. J Neurol2001;248:31–3.

588 PostScript


on Novem


http://bjo.bmj.com

/B

r J Ophthalm


ownloaded from

http://bjo.bmj.com/

2 Boroojerdi B, Ferbert A, Schwarz M, et al.Botulinum toxin treatment of synkinesia andhyperlacrimation after facial palsy. J NeurolNeurosurg Psychiatry 1998;65:111–4.

3 Riemann R, Pfennigsdorf S, Riemann E, et al.Successful treatment of crocodile tears by injectionof botulinum toxin into the lacrimal gland: a casereport. Ophthalmology 1999;106:2322–4.

4 Hofmann RJ. Treatment of Frey’s syndrome(gustatory sweating) and ‘crocodile tears’(gustatory epiphora) with purified botulinumtoxin. Ophthal Plast Reconstr Surg2000;16:289–91.

5 Keegan DJ, Geerling G, Lee JP, et al. Botulinumtoxin treatment for hyperlacrimation secondary toaberrant regenerated seventh nerve palsy orsalivary gland transplantation. Br J Ophthalmol2002;86:43–6.

6 Yavuzer R, Basterzi Y, Akata F. Botulinum toxin Afor the treatment of crocodile tears. Plast ReconstrSurg 2002;110:369–70.

7 Montoya FJ, Riddell CE, Caesar R, et al.Treatment of gustatory hyperlacrimation(crocodile tears) with injection of botulinum toxininto the lacrimal gland. Eye 2002;16:705–9.

Retinal arterial collapse pressurein eyes with retinal arterialocclusive diseasesRetinal arterial occlusions may be primarilyor secondarily associated with low retinalarterial pressure. Based on previous ophthal-modynamometric studies1–6 the purpose ofthe present study is to estimate the retinalvessel pressure in patients with central retinalartery or branch retinal artery occlusions andpatients with amaurosis fugax.

Case reportThis prospective clinical non-interventionalcomparative study included nine eyes ofseven patients (mean age 68.8 (SD 13.7)years) with central retinal artery occlusion(n = 1 eye), branch retinal artery occlusion(n = 2), ischaemic ophthalmopathy (n = 2),or amaurosis fugax (n = 4). An age matchedcontrol group consisted of 27 eyes of 21subjects attending the hospital because ofcataract or refractive problems. After medicalpupil dilatation, a conventional Goldmanncontact lens fitted with a pressure sensor

mounted into the holding ring was put ontothe cornea. By slightly pressing the contactlens, pressure was applied onto the globe, andthe pressure when the central retinal vein orartery started to pulsate was noted. Themethods applied in the study adhered to thetenets of the declaration of Helsinki. Themethod has already been described in detail.6

In the study group, central retinal arterycollapse pressure measured 43.9 (SD 33.4)arbitrary units (AU) and was significantly(p = 0.004) lower than in the control group(78.0 (SD 19.2) AU) (fig 1). Within the studygroup, central retinal artery collapse pressurewas lowest in the eye with central retinalartery occlusion, showing a pulse synchronicmovement of the erythrocyte column in thevessel without applying any pressure onto theglobe. In the two eyes with branch retinalartery occlusion, collapse pressure in thearterial branch lying in the oedematous partof the fundus measured 36.7 AU and 0 AUrespectively. These values were significantly(p = 0.005) lower than the values obtained inthe control group (fig 1). In the eyes withbranch retinal artery occlusion, collapsepressure in the arterial branch in the non-oedematous part of the fundus was in thenormal range (93.1 AU and 93.3 AU, respec-tively). In the patient suffering from ischae-mic ophthalmopathy, central retinal arterycollapse pressure was lower in the eye moreseverely affected than in the contralateral eye(14.7 AU v 51.7 AU). Both values weresignificantly (p = 0.02) lower than the valuesof the control group. In the eyes withamaurosis fugax, mean central retinal arterycollapse pressure measured 73.0 (SD 15.4)AU which was not significantly (p = 0.55)different from central retinal artery collapsepressure in the control group (fig 1). Centralretinal vein collapse pressure did not varysignificantly between the study groups andthe control group (8.8 (SD 12.2) AU v 6.1 (SD8.4) AU; p = 0.54).

CommentCentral retinal artery collapse pressure asdetermined by the new ophthalmodynamo-metric technique was significantly lower ineyes with retinal artery occlusive diseasesthan in normal eyes (fig 1). Correspondingly,in the eyes with branch retinal artery occlu-sions, measurements were lower in thearterial branch affected by the occlusion thanin the retinal artery branch with intactperfusion. As a corollary, in the patientsuffering from ischaemic ophthalmopathy,the central retinal artery collapse pressurewas lower in the eye more severely affectedbecause of a complete stenosis of the carotidartery than in the contralateral eye.Interestingly, the eyes with amaurosis fugaxdid not show significantly lower measure-ments than normal eyes. This agrees withprevious studies using other ophthalmo-dynamometric techniques for evaluation ofcarotid artery perfusion.4 In conclusion, usinga new ophthalmodynamometer with biomi-croscopic observation of central retinal ves-sels during the examination, central retinalartery collapse pressure measurements weresignificantly lower in eyes with retinalarterial occlusive diseases than in normaleyes. Future studies may show whetherdetermination of the central retinal arterycollapse pressure in patients with increasedrisk for retinal arterial occlusions maybe suitable to predict which patients havea higher risk for eventual retinal artery

occlusion compared with other patients witha similar risk profile.

J B Jonas

Correspondence to: J B Jonas, Department ofOphthalmology, Faculty of Clinical Medicine

Mannheim, University of Heidelberg, Theodor-Kutzer-Ufer 1-3, 68167 Mannheim, Germany;[email protected]


References

1 Rios-Montenegro EN, Anderson DR, David NJ.Intracranial pressure and ocular hemodynamics.Arch Ophthalmol 1973;89:52–8.

2 Yablonski M. A new fundus lensophthalmodynamometer. Am J Ophthalmol1978;86:644–7.

3 Krieglstein GK, da Silva FA. Comparativemeasurements of the ophthalmic arterial pressureusing the Mikuni dynamometer and the Stepanik-arteriotonograph. Albrecht Von Graefes Arch KlinExp Ophthalmol 1979;212:77–91.

4 Zaret CR, Sacks JG, Holm PW. Suctionophthalmodynamometry in the diagnosis ofcarotid stenosis. Ophthalmology1979;86:1510–12.

5 Entenmann B, Robert YC, Pirani P, et al. Contactlens tonometry—application in humans. InvestOphthalmol Vis Sci 1997;38:2447–51.

6 Jonas JB. Reproducibility ofophthalmodynamometric measurements of thecentral retinal artery and vein collapse pressure.Br J Ophthalmol 2003;87:577–9.

Modified self sealing sclerotomyfor drainage of subretinal fluidduring scleral buckling surgeryDrainage of subretinal fluid is probably themost dangerous step in scleral bucklingsurgery for uncomplicated retinal detach-ment. The most common complicationsinclude subretinal haemorrhage, retinal per-foration, and vitreoretinal incarceration.1 2

Sclerotomy to drain subretinal fluid is tradi-tionally made with a sharp blade, diathermyto the sclera and choroid is performed,followed by perforation of the choroid toallow drainage of subretinal fluid. Suture ofthe sclerotomy at the end of the procedurehas been recommended to avoid retinalincarceration.

The purpose of this study was to determinethe effectiveness and safety of a modified selfsealing sclerotomy technique for drainage ofsubretinal fluid during scleral buckling sur-gery.

Patients and methodsTwenty consecutive patients undergoingscleral buckling for primary rhegmatogenousretinal detachment from two vitreoretinalsurgery centres were enrolled in this prospec-tive study. A scleral buckling procedure wasperformed using a circumferential scleralband (Mira 240, Mira, Waltham, MA, USA)sutured with the posterior border located12 mm posterior to the limbus, and addingany necessary segmental sponges (Mira).Cryoretinopexy was performed using a CTUOphthalmic Cryo Unit (Keeler, London, UK)to seal retinal tears. After surgery, sulfurhexafluoride (SF6) gas was used in allpatients. The drainage site was chosen basedon retinal elevation, as shown by intraopera-tative retinal examination with indirect

Figure 1 Box plots showing the retinal arterycollapse pressure in the study group of patientswith central retinal artery occlusion (CRAO),branch retinal artery occlusion, ischaemicophthalmopathy, amaurosis fugax, and in thecontrol group.

Proprietary interest: none.

PostScript 589


on Novem


http://bjo.bmj.com

/B

r J Ophthalm


ownloaded from

http://bjo.bmj.com/

ophthalmoscopy. A 3–4 mm half depthscleral incision was created perpendicular tothe limbus using an angled bevel up blade(Alcon Laboratories, Fort Worth, TX, USA)with its sharp advancing edge directedperperdicular to the scleral surface (fig 1A).With a crescent knife, a 3 mm tunnel incisionwas then made to create a scleral flap parallelto the limbus (fig 1B). The scleral flap wasretracted, and a 27 gauge needle was used toperforate the scleral bed and choroid (fig 1C).Subretinal fluid was expressed (fig 1D) anddried with a cotton swab. In all cases, thesurgical wound was inspected for adequateclosure at the end of the operation. Drainageof subretinal fluid and complications asso-ciated with this technique were assessed byintraoperative binocular indirect ophthalmo-scopy, and recorded on surgical reports andrecords from postoperative visits.

ResultsAll 20 patients had self sealing sclerotomythat did not require suturing at the end ofsurgery. Treating the choroid with diathermyor cautery was not necesary in any case, andthe planned needle drainage procedure wassuccessful in all patients. Drainage was slowand gradual. No serious complications werelinked to the drainage of subretinal fluid withthis technique. Spontaneous drainage wasseen in one patient with thin sclera, and twolimited subretinal haemorrhage that did notmigrate to the posterior pole were observed.

CommentsWe describe the results and complications of20 scleral buckling procedures for primaryrhegmatogenous retinal detachments inwhich subretinal fluid was drained using asutureless sclerotomy technique. A self seal-ing incision with a stepped wound construc-tion is not new to ophthalmology. Cataract

and vitreoretinal surgeons pioneered thistechnique as part of phacoemulsificationcataract extraction and vitreous surgeries.3–5

Possible advantages to this type of incisioninclude shortened operating time andreduced incidence of postoperative woundleak.

This new way of constructing the scler-otomy for drainage has many advan-tages, and we propose it as an alternativeto standard sclerotomy incision to drainsubretinal fluid during scleral bucklingsurgery for uncomplicated retinaldetachments.

J B Yepez, J Cedeno de Yepez, A ValeroThe Retina and Vitreous Service, Clinica de Ojos de

Maracaibo, Maracaibo, Venezuela

J F ArevaloRetina and Vitreous Service, Clinica Oftalmologica

Centro Caracas, Caracas, Venezuela

Correspondence to: Dr J Fernando Arevalo, ClinicaOftalmologica Centro Caracas, Centro Caracas PH-1,

Av Panteon, San Bernardino, Caracas 1010,Venezuela; [email protected]

References

1 Hilton GF, Grizzard WS, Avins LR, et al. Thedrainage of subretinal fluid: a randomizedcontrolled clinical trial. Retina 1981;1:271–80.

2 Wilkinson CP, Bradford RH Jr. Complicationsof draining subretinal fluid. Retina1984;4:1–4.

3 Fine IH, Fichman RA, Grabow HB. Clear CorneaCataract Surgery and Topical Anesthesia.Thorofare, NJ: Slack Inc, 1993.

4 Chen JC. Sutureless pars plana vitrectomy throughself-sealing sclerotomies. Arch Ophthalmol1996;114:1273–5.

5 Kwok AK, Tham CC, Lam DS, Li M, et al.Modified sutureless sclerotomies in parsplana vitrectomy. Am J Ophthalmol1999;127:731–3.

Conjunctival dendrite in a caseof primary herpes simplexinfectionOcular involvement in primary herpes sim-plex infection is usually in the form offollicular conjunctivitis, blepharitis, andsometimes corneal involvement in the formof superficial punctuate keratitis, dendrite, or(rarely) geographical ulcer.1

We report a case of dendritiform lesion inthe conjunctiva in a young girl with primaryherpes simplex infection. To the best of ourknowledge, conjunctival dendritiform lesionhas not been reported before in primaryherpes simplex infection.

Case reportA 20 year old girl presented to our outpatientdepartment with complaints of redness anddiscomfort in her right eye of two days’duration. She gave a history of fever of oneweek’s duration followed by appearance ofvesicles at the right side angle of the mouthand on the right upper lid. Past ocular andsystemic history was unremarkable.

Visual acuity was 6/6 unaided in both theeyes. There were vesicles at the angle of themouth (fig 1A) and on the right upper lid.Slit lamp examination of the right eye withfluorescein staining revealed a dendritiformpattern of staining in the lower bulbarconjunctiva (fig 1). Cornea was clear andrest of the anterior segment was unremark-able. Left eye examination was unremark-able. Fundus examination in both the eyeswas within normal limits. The patient wasadvised to use topical acyclovir 3% eyeointment five times a day and tablet acyclovir400 mg five times a day.

On follow up after two days, there wassuperficial punctuate keratitis in the inferiorhalf of the cornea in the right eye. The patientwas asked to continue the same medication.One week later, the vesicles were absent andthe conjunctiva and cornea were clear. Themedication was discontinued.

Darouger et al, in a study of primary herpessimplex ocular infection, found 64% of thepatients to be over fifteen years of age.2

Follicular conjunctivitis (7%), blepharocon-junctivitis (16%), and corneal dendritic ulcers(15%) were some of the lesions reported.Appearance of a dendritic ulcer on theconjunctiva, to the best of our knowledge,has not been reported in primary herpessimplex infection.

Dendritic lesions on histopathologicalstudy show that they are composed of roundepithelial cells and variable sized syncytiacontaining bizarre shaped nuclei. The epithe-lial cells contain viral DNA.3 Recurrent infec-tion with the virus in the form of epithelialkeratitis commonly produces dendriticlesions on the cornea.1

Figure 1 (A) A 3–4 mm half depth scleral incision is created perpendicular to limbus using anangled bevel up blade with its sharp advancing edge directed perpendicular to the sclera surface.(B) With a crescent knife, a 3 mm tunnel incision is then made to create a scleral flap parallel to thelimbus. (C) The scleral flap is retracted, and a 27 gauge needle is then used to perforate the scleralbed and choroid. (D) Subretinal fluid is expressed and dried with a cotton swab.

Presented in part at the Vitreous Society AnnualMeeting, Fajardo, Puerto Rico, November 2001.

doi: 10.1136/bjo.2003.029272


Supported in part by the Fundacion Arevalo-Coutinhopara la Investigacion en Oftalmologia (FACO),Caracas, Venezuela.

The authors have no proprietary or financial interest inany products or techniques described in this article.

590 PostScript


on Novem


http://bjo.bmj.com

/B

r J Ophthalm


ownloaded from

http://bjo.bmj.com/

Conjunctival dendrite is an interesting andapparently rarely reported lesion.

U SridharCornea Services, ICARE Eye Hospital, Noida, UP,

India

Y BansalDept of Optometry, ICARE Eye Hospital, Noida, UP,

India

S ChoudhuryTraining Centre, ICARE Eye Hospital, Noida, UP, India

A K GuptaAcademic Director ICARE Eye Hospital, Noida, UP,

India

Correspondence to: Dr U Sridhar, ICARE Eye Hospital,E-3A Sector 26, Noida, UP, India; u_sridhar@

yahoo.com


References

1 Leisegang TJ, Melton J J III, Daly PJ, et al.Epidemiology of ocular herpes simplex. ArchOphthalmol 1989;107:1155–9.

2 Darouger S, Wishart MS, Viswalingam ND.Epidemiological and clinical features of primaryherpes simplex ocular infection. Br J Ophthalmol1985;69:2–6.

3 Tabery HM. Early epithelial changes in recurrentherpes simplex viral keratitis. Acta OphthalmolScand 1998;76:349–52.

Severe ocular trauma caused byan ostrichThe ostrich is a strange and harmless lookingbird; however, in Africa attacks by ostricheson humans are not uncommon and some-times result in death. We recently treatedsuch a patient with an eye injury.

Case reportA 35 year old male patient presented with aninjury sustained from being kicked in the faceby an ostrich (Struthio camelus).

On examination the right eye was found tobe normal but he had vision of bare lightperception on the left with proptosis of theglobe and severe chemosis of the conjunctiva.Both upper and lower lids were avulsedmedially. There was limitation in all positionsof gaze which was more noticeable onattempted abduction (fig 1). The eye was

soft and he had an oedematous cornea and afull hyphema. The posterior segment couldnot be visualised.

Computed tomography showed a blowoutfracture of the left orbit (fig 2). There was afracture of the left nasal bone with acomminuted lateral wall fracture. The medialorbital wall was also fractured with opacifi-cation of the left ethmoid sinus and hernia-tion of the medial rectus into the sinus.

An intraocular haemorrhage as well ashaemorrhage in the retrobulbar space wasnoted.

Under general anaesthesia, both the upperand lower lids were repaired and thehyphema was washed out. A posteriorrupture1 was suspected clinically but the siteof rupture could not be identified. The eyewas subsequently eviscerated.

CommentOstriches usually inflict injury in one of twoways: the more serious injury is that of aslash or laceration, usually to the lowerabdomen or limbs, caused by the ostrichkicking in a forward and downward motionwith its powerful foot. The toenail of theostrich is sharp and is used by the ostrich forprotection against predators.2 The secondtype of injury is seen more commonly. Thisoccurs when the ostrich uses its bony breastplate as a ram to knock the person to theground. The ostrich then jumps upon thevictim and, because an ostrich weighs 75–150 kg, this may cause contusion of the torsowith rib fractures.

Our patient was bending while repairing afence when he was kicked by an ostrich. Hewas struck in the face and sustained exten-sive facial trauma extending from his nasalbones to his orbital walls and ethmoidalsinus. The trauma also resulted in irreparableblunt trauma to the eye.

The injury caused was severe with nopossibility of repair of the globe, and is theonly documented case of an eye being lostdue to injury by an ostrich

L M Levitz, T R Carmichael, M NissenbaumUniversity of the Witwatersrand, Department of

Ophthalmology, Johannesburg, South Africa

Correspondence to: Professor T R Carmichael, PO Box752062, Garden View, 2047, South Africa;

[email protected]


References

1 Kuhn F, Morris R, Witherspoon CD, et al. Astandardized classification of ocular trauma.Ophthalmology 1996;103:240–3.

2 Salvat SA. Hunters and Hunted of the Savannah.London: Orbis, 1971:155.

Swollen optic discs in a patientwith the chromosome 22q11.2deletion syndromeThe chromosome 22q11.2 deletion syndrome(22q11DS) encompasses velocardiofacial syn-drome (VCFS), DiGeorge syndrome (DGS),and conotruncal anomaly face syndrome(CTFS) and is the result of a microdeletionof chromosome band 22q11.2. It is a rela-tively common genetic anomaly estimated tooccur in approximately one in 4000 livebirths. The 22q11.2 deletion can arise denovo or can have an autosomal dominant

The authors have no proprietory interests in theproducts mentioned in the article.

Figure 1 Dendritiform pattern of staining in the lower bulbar conjunctiva. (A) Vesicles at the angleof the mouth.

Figure 1 Blunt trauma to left eye with absenceof abduction.

Figure 2 Computed tomography showingblowout fracture into the ethmoid sinus andcomminuted lateral wall fracture.

PostScript 591


on Novem


http://bjo.bmj.com

/B

r J Ophthalm


ownloaded from

http://bjo.bmj.com/

inheritance. The condition is thought to bedue in part to abnormal development of thepharyngeal arch structures. Clinical findingsare extensive and highly variable betweenpatients. Prominent features include cardiacdefects, cleft palate, dysmorphic facies, mal-descent of the thymus, hypoparathyroidism,immune deficiency and developmental delay.Ocular findings include hypertelorism,1 ret-inal vascular tortuousity, narrow palpebralfissures, small optic nerves, iris nodules,cataracts,2 and iris coloboma.3 We present acase of a boy who was found to have bilateraldisc swelling that led to a diagnosis of22q11DS.

Case reportA 14 year old boy presented to the accidentand emergency department after having ageneralised seizure. He had been admitted toanother hospital, 2 days before this, with asudden onset of collapse and subsequentrespiratory arrest. At that time he was notedto have swollen optic discs and a headcomputed tomography scan done there wasreported as normal. He had further seizuresafter admission to our hospital. Blood testingrevealed low plasma calcium and highplasma phosphate levels. The patient hadbeen complaining of back pain in recentmonths and his mother said that he hadshrunk by a couple of centimetres over thepast year. She also said that he had alwaysbeen clumsy and he had been diagnosed asdyslexic at the age of 7. He had a history ofbehavioural problems which his family saidoften settled during holidays in the sun. Athoracic spine x ray revealed wedge-shapedfractures of three vertebral bodies. Furtherblood testing showed a relatively lowparathormone level confirming the diagnosisof hypoparathyroidism. He was then referred

to the ophthalmology department for assess-ment of his disc swelling.

He had been seen in the eye clinic a yearbefore this presentation complaining ofcoloured lines in his field of vision. Noabnormality was found and his discs on thatoccasion were noted to be normal. Onexamination this time he was noticed tohave dysmorphic features, notably palpebralfissures slanting medially upwards, andabnormally formed low set ears. Visualacuities were 6/5 in both eyes, colour visiontesting was unremarkable, and there was norelative afferent pupillary defect. There wasno sign of cataract formation or any otherabnormality of the anterior segments.Funduscopy revealed bilateral disc swellingwith extensive peripapillary haemorrhages inboth eyes (fig 1). The diagnosis of hypopara-thyroidism combined with the patient’sabnormal facies raised the suspicion of agenetic disorder and blood was then sent forchromosomal analysis. His hypoparathroid-ism was treated with vitamin D and calciumsupplements and he responded well with hiscalcium reaching normal levels within a fewdays. Examination 6 weeks after his calciumhad normalised showed most of the haemor-rhages and disc swelling had cleared (fig 2).Results of chromosomal testing revealed aregion 11.2 microdeletion of the long arm (q)of chromosome 22. Other members of hisfamily are now also being tested.

CommentChromosome 22q11.2 deletion syndrome isone of the more common causes of congenitaland childhood hypoparathyroidism whichcan lead to hypocalcaemia. Hypocalcaemiais a known cause of disc swelling,4 themechanism of which is not known. Somepatients with hypocalcaemic disc swellinghave a loss of acuity and field typical of an

optic neuropathy, while in others the featuresresemble papilloedema, with no significantvisual loss.5 The swelling usually resolvesafter the calcium level returns to normal.6

22q11DS is probably underdiagnosed. Thiscase illustrates the importance of a correctand early diagnosis of this relatively commongenetic disorder so that treatment can beginin an effort to prevent further medical anddevelopmental complications. The highlyvariable clinical features require a high levelof awareness of the condition across severaldifferent disciplines. Patients, especially chil-dren, presenting with swollen optic discs andwho have normal imaging studies of thebrain should have a calcium level checked. Ifabnormal and it is found to be due tohypoparathryoidism then chromosomal ana-lysis should be considered, especially if otherparts of the history or examination raise thesuspicion of a genetic disorder.

R A M Girgis, H D R McKee, J R InnesHull and East Yorkshire Eye Hospital, Fountain Street,

Hull, Yorkshire HU3 2JZ, UK

Correspondence to: Mr Girgis, Hull and East YorkshireEye Hospital, Fountain Street, Hull, Yorkshire HU3

2JZ, UK; [email protected]


References

1 Kitano I, Park S, Kato K, et al. Craniofacialmorphology of conotruncal anomaly facesyndrome. Cleft Palate Craniofac J1997;34:425–9.

2 Mansour AM, Goldberg RB, Wang FM, et al.Ocular findings in the velo-cardio-facialsyndrome. J Pediatr Ophthalmol Strabismus1987;24:263–6.

3 Morrison DA, Fitzpatrick DR, Fleck BW. Iriscoloboma and a microdeletion of chromosome22: del(22)(q11.2). Br J Ophthalmol2002;86:1316.

4 Walsh FB, Murray R. Ocular manifestations ofdisturbances in calcium metabolism.Am J Ophthalmol 1953;36:1657–76.

5 McLean C, Lobo R, Brazier JD. Optic discinvolvement in hypocalcaemia withhypoparathyroidism: papilloedema or opticneuropathy? Neuro-Ophthalmology1998;20:117–24.

6 Kent CJ, Jakobiec FA. Ophthalmic manifestationsof some metabolic and endocrine disorders. In:Albert DM, Jakobiec FA, eds. Principles andpractice of ophthalmology, 2nd ed. Philadelphia:Saunders, 2000:4730–42.

The correlation of phenylephrine1% with hydroxyamphetamine1% in Horner’s syndromePharmacological testing in Horner’s syn-drome involves the use of cocaine to confirmthe diagnosis and hydroxyamphetamine tolocalise the lesion to the post-ganglionic(third order) or non-postganglionic neuron.However, hydroxyamphetamine bromide 1%(Paredrine) is not always readily available tothe ophthalmologist. An alternative drug forlocalising the site of the lesion is phenylephr-ine 1% which can easily be prepared bydilution of stronger concentrations (2.5% or10%) and which is almost universally avail-able in most ophthalmologists’ offices.Because of the principle of denervation super-sensitivity, a Horner’s syndrome produced bya lesion interrupting the postganglionic fibresshould dilate the pupil when phenylephrine1% is placed in the conjunctival sac. The pupil

Figure 1 Shows bilateral disc swelling with extensive peripapillary haemorrhages in both eyes.

Figure 2 Examination 6 weeks later showed most of the haemorrhages and disc swelling hadcleared.

592 PostScript


on Novem


http://bjo.bmj.com

/B

r J Ophthalm


ownloaded from

http://bjo.bmj.com/

of a patient with central (first order) Horner’ssyndrome should not dilate, while a pre-ganglionic (second order) pupil may dilateminimally1; a normal pupil may, at best,dilate minimally. The purpose of this studywas to compare the pupillary response ofpatients with Horner’s syndrome to pheny-lephrine 1% and hydroxyamphetamine 1%.

Fourteen consecutive patients withHorner’s syndrome were prospectively testedwith cocaine 10%, hydroxyamphetamine 1%,and phenylephrine 1% on separate days, atleast 3 days apart. All pupils were measuredin the same room lighting with a standardruler 1 hour after instillation of two drops.Phenylephrine 1% was prepared on eachexamination day by diluting phenylephrine10% hydrochloride (0.1 ml) with 0.9 ml ofpreservative free natural tear drops. Theclinician interpreting each pharmacologicaltest was masked to the results of the otherpharmacological tests and the cause ofHorner’s syndrome.

Nine of the 14 patients were men. Theaverage age was 59 years (range 34–74 years).All patients underwent magnetic resonanceimaging of head and neck, magnetic reso-nance angiography of the neck, and com-puted tomography of the chest. Horner’ssyndrome was considered to be central intwo patients and pre-ganglionic in one, basedon neuroimaging. All 11 patients diagnosedwith postganglionic Horner’s syndrome hadnormal neuroimaging, an isolated Horner’ssyndrome, positive cocaine test and positivehydroxyamphetamine test. Table 1 shows thebaseline pupil size and the change with eachof the three pharmacological tests. Afterpharmacological testing with phenylephrine1%, the mean increase in pupil size inpatients with postganglionic Horner’s syn-drome was 2.3 mm (SD 1.1 mm) and in thecontralateral normal pupil was 0.2 mm (SD0.2 mm) (paired t test: p,0.0001). Thesensitivity of phenylephrine 1% was 81%and the specificity 100%. The mean increasein pupil diameter on the affected side inpatients with postganglionic Horner’s syn-drome after hydroxyamphetamine 1% was0.27 mm (SD 0.3 mm) and was 2.65 mm (SD0.3) in the contralateral normal pupil.

CommentThe law of denervation supersensitivitystates that an organ deprived of its normalinnervation becomes more sensitive to thechemical transmitter normally released fromthose nerves. Thompson and Mensher docu-mented supersensitivity of the iris dilator tophenylephrine 1% in one patient.1 Theyfurther tested 13 patients but used a 10%

concentration, which dilates the normalpupil. They determined that the affectedpupil of the three patients with post-gang-lionic lesions dilated sooner and more vigor-ously than the unaffected pupil. Ramsay2

tested 14 patients with phenylephrine 1%and found in patients with Horner’s syn-drome that 71% of pupils were supersensi-tive. However, the responses of the post-ganglionic and non-post-ganglionic Horner’ssyndrome were not reported separately.Other studies have reported on the use ofphenylephrine 10%1 and epinephrine 1%,3

and adrenaline (epinephrine) 0.1%1 in thepharmacological testing of Horner’s, but nonehas investigated the efficacy of phenylephrine1% in identifying post-ganglionic Horner’slesion. Our study shows that phenylephrinedilates the post-ganglionic Horner’s pupil,but not the non-post-ganglionic or normalpupil. We found the sensitivity of 81% and aspecificity of 100%. Hydroxyamphetamine 1%has been shown to have a sensitivity of 93%and specificity of 83%.4 Only one out of 11patients with Horner’s syndrome in our serieshad a lesion, which localised to the post-ganglionic neuron with hydroxyampheta-mine 1%, but not with phenylephrine 1%.

Phenylephrine 1% has some limitations:

(1) It does not dilate the normal pupil. Ifneither pupil dilates with phenylephrine 1%it could be either because the lesion is non-post-ganglionic or the drops are ineffective.

(2) The degree of supersensitivity is deter-mined by the extent of denervation. A partialpost-ganglionic lesion may be difficult todistinguished from a preganglionic lesion; inboth conditions the iris may dilate minimally.In our study there was only one patient witha preganglionic lesion.

(3) Tests for supersensitivity may also besubject to false positive errors owing to thevariations in penetration of the drug. Thedrug should be placed strictly on intactcorneas so that the same dose reaches eachiris.

(4) Supersensitivity of the iris dilatorincreases with age. Phenylephrine sensitivityof the iris increases by 0.23 mm per decadeafter age 20.

In summary, we here report the first seriesof patients with Horner’s syndrome, whichcompared the pupillary response of phenyl-ephrine 1% to hydroxyamphetamine 1%.Phenylephrine 1% correlates well with theresults of hydroxyamphetamine 1% in localis-ing the lesion to the post-ganglionic neuronand is a reliable alternative to hydroxyamphe-tamine 1% should pharmacological testing be

desired and hydroxyamphetamine 1% not beavailable.

H V Danesh-MeyerDepartment of Ophthalmology, University of

Auckland, Auckland, New Zealand

H V Danesh-Meyer, P Savino, R SergottNeuro-ophthalmic Services, Wills Eye Hospital,

Thomas Jefferson Medical School, Philadelphia, PA,USA

Correspondence to:Associate professor Helen V Danesh-Meyer,Department of Ophthalmology, University of

Auckland, Private Bag 92019, Auckland, NewZealand; [email protected]


References

1 Thompson HS, Mensher JH. Adrenergic mydriasisin Horner’s syndrome. Am J Ophthalmol1971;72:472–80.

2 Ramsay DA. Dilute solutions of phenylephrineand pilocarpine in the diagnosis of disorderedautonomic innervation of the iris. J Neurol Sci1986;73:125–34.

3 Maloney WF, Younge BR, Moyer NJ. Evaluationof the causes and accuracy of pharmacologiclocalization Horner’s syndrome. Am J Ophthalmol1980;90:394–402.

4 Cremer SA, Thompson S, Digree KB, et al.Hydroxyamphetamine mydriasis in Horner’ssyndrome. Am J Ophthalmol 1990;110:71–6.

Tetraspanin protein KAI1 expres-sion in retinoblastomaKAI1/CD82 is a metastasis suppressor genelocated on human chromosome 11p11.2. It isa member of the structurally distinct familyof cell surface glycoprotein, transmembrane4-protein superfamily.1 KAI1 was initiallyisolated as a gene that suppressed metastasisof rat prostate tumour cells.2 KAI1 is down-regulated in several types of human malig-nancies.3–5 The purpose of this study was toinvestigate the expression of KAI1 in retino-blastoma and to correlate clinicopathologi-cally.

MethodsThere were 30 archival specimens of retino-blastomas from 2000 to 2002. There were 12tumours with no invasion of choroid, or opticnerve and 18 tumours with invasion ofchoroid/optic nerve, and one tumour whichhad metastasised to the submandibularregion. There were six well differentiatedtumours, six moderately differentiatedtumours, and 18 poorly differentiatedtumours (table 1). Immunohistochemical

Table 1 Mean change in pupillary diameter in patients with Horner’s syndrome

Baseline pupil

Change in pupil diameterfollowing



Cocaine Hyrdoxyamphetamine Phenylephrine

Size (mm) 10% (mm) 1% (mm) 1% (mm)

Normal pupil Horner’s pupil Normal pupil Horner’s pupil Normal pupil Horner’s pupil Normal pupil Horner’s pupil

Central (n = 2) 3.25 2 1.5 0.25 2.1 2.25 0.3 0.25Pre-ganglionic(n = 1)

4 2.5 3.0 0.5 2.5 2.9 0.2 0.5

Post-ganglionic(n = 11)

4.0 2.9 1.8 0.2 1.8 0.3 0.2 1.9

PostScript 593


on Novem


http://bjo.bmj.com

/B

r J Ophthalm


ownloaded from

http://bjo.bmj.com/

staining was performed using a sensitivelabelled streptavidin biotin (LSAB kit, Dako)on tumours using monoclonal antibodies fortetraspanin KAI1 (C33, Novacastra) and forproliferation index Ki-67 (Clone MIB-1,Dako, Denmark) after antigen retrieval.

The immunoanalysis for KAI1 was donebased on the percentage of cells and thestaining intensity. KAI1 was scored afteranalysing 15 tumour fields and on a fourtiered scale: 0, negative (0%–5% positive cells);1+, weak to moderate (6%–30%); 2+, intense(31%–50%); and 3+, very intense (.50%).Labelling index for Ki 67 positive cells wasexpressed by counting a minimum of 500 cellsin the staining area. Tumours were dividedinto two groups; group I .50% of cellsshowing positive Ki 67 expression and groupII ,50% of cells showing positive for Ki 67.Data were analysed for statistical significance.

ResultsThe immunohistochemical details are givenin table 1. KAI1 expression was seen in the

control lymphoid follicle of the tonsil (fig 1A).Intense KAI1 positivity with more than 80%positivity was seen in all 12/12 tumours withno invasion (fig 1B). Among the 18/18tumours with invasion, KAI1 was decreasedin all 18. The invading front of the tumourhad less KAI1 than the tumour at the centralportion. Retinoblastomas with focal anddiffuse invasion of choroid had negativeKAI1 immunoreactivity. Tumours with pre-laminar optic nerve invasion had weak KAI1immunoreactivity (fig 1C). Tumours at thepost-laminar and surgical end of the opticnerve (inset, fig 1C) and at the metastatic site(fig 1D) had negative KAI1 immunoreactiv-ity. Negative KAi1 reactivity was seen in 50%(9/18) of poorly differentiated retinoblasto-mas. Statistically significant correlation wasobserved between KAi1 expression and Ki 67labelling index in the whole study group(p,0.001). No statistically significant corre-lation was observed between KAI1 expressionand differentiation. Significant statistical cor-relation was observed when KAI1 expression

was compared with tumour invasion(p,0.001).

CommentRetinoblastoma joins a growing list of can-cers in which downregulation of KAI1 isassociated with tumour progression. In ourstudy KAI1 was identified by the monoclonalantibody CD33.6 It was originally shown asinhibitory to syncytium formation induced byhuman T cell leukaemia virus type I, and thisspecific inhibition to syncytium formationinduced by some human T cell line by thisantibody was strongly associated with alteredglycosylation of cell surface antigen, suggest-ing that the C33 antigen—that is, KAI1,might have a possible role in the cell to celladhesion mechanism.7

Thus, KAI1 may link to the cell surfacemolecules, such as integrins, E-cadherin, andother TM4SF members, and loss of KAI1function may have a significant role in theprogression of retinoblastoma.7 The mechan-ism of KAI1 downregulation is not known.The 5’ promoter region of the gene contains aCpG island,8 raising the possibility of genesilencing by promoter methylation. Thus,biologically, our findings suggest a potentialimplication of KAI1 in tumour progressionand these molecules may provide novelinsights into tumour progression in retino-blastoma.

AcknowledgementFinancial support: Indian Council Of MedicalResearch, India.

S AmirthaLakshmi, V Pushparaj,V Krishnamurthy, J Biswas, S Krishnakumar

Department of Ocular Pathology, Medical and VisionResearch Foundation, Sankara Nethralaya, Chennai,

India

M P ShanmugamDepartment of Ocular Oncology, Medical and VisionResearch Foundation, Sankara Nethralaya, Chennai,

India

Correspondence to: Dr Subramanian Krishnakumar,Vision Research Foundation, Sankara Nethralaya, 18,College Road, Chennai 600 006, Tamil Nadu, India;

[email protected]

Accepted for publication 1 August 2003

References

1 Maecker HT, Todd SC, Levy S. The tetraspaninsuperfamily: molecular facilitators. Faseb J1997;11:428–42.

2 Dong JT, Lamb PW, Rinker-Schaeffer CW, et al.Kai1, a metastasis suppressor gene for prostatecancer on human chromosome 11p11. 2. Science(Washington DC) 1995;268:884–6.

3 Yu Y, Yang J-L, Markovic B, et al. Loss of KAI1messenger RNA expression in both high-gradeand invasive human bladder cancers. Clin CancerRes 1997;3:1045–9.

4 Huang C I, Kohno N, Ogawa E, et al.Correlation of reduction in mrp-1/cd9and KAI1/cd82 expression with recurrencesin breast cancer patients. Am J Pathol1998;153:973–83.

5 Guo X, Friess H, Graber HU, et al. KAI1expression is up-regulated in early pancreaticcancer and decreased in the presence ofmetastases. Cancer Res 1996;56:4876–80.

6 Imai T, Fukudome K, Takagi S, et al. C33 antigenrecognized by monoclonal antibodies inhibitoryto human T cell leukemia virus type 1-induced

Table 1 Clinicopathological and immunohistochemical features of theretinoblastomas used in the study

No

Age(years)/sex

Clinicopathologicalparameters Chemotherapy

KAI1immunoreactivity Ki67

% Cellsstained

Stainintensity

% cellsstained

1 2/M Both eyes, (RE) well diff No 90 3+ 402 5/M LE, poor diff No 90 3+ 373 1/F RE, well diff No 90 3+ 634 2/M Both eyes, (RE), well diff No 50 2+ 405 8/M LE, poor diff No 90 3+ 406 4/F LE, poor diff No 80 3+ 607 1/F LE, poor diff No 95 3+ 498 5/F RE, poor diff No 45 2+ 509 1/F Both eyes, (RE), mod diff No 80 3+ 4010 2/F Both eyes, RE, mod diff No 85 3+ 7011 2/F Both eyes, LE, mod diff No 40 2+ 6612 1.5/M RE, well diff No 85 3+ 5013 7/F LE, mod diff, No choroidal

invasion, surgical end of theoptic nerve invaded

Postop chemo 3cycles

1 0 56

14 1/M RE, poor diff, Pre laminar No 20 1+ 6015 2/F RE, poor diff, diffuse choroidal

invasion, surgical end of opticnerve invaded


1 0 70

16 4/M Both eyes, LE, poor diff Postop chemo 4cycles

1 0 70

17 4/F RE, mod diff, lamina cribrosa No 40 2+ 6418 2/M RE, poor diff, lamina cribrosa No 20 1+ 6619 1.1 M/F RE, poor diff, diffuse choroidal

invasion and post laminar opticnerve invasion


1 0 80

20 3/M LE, poor diff, post laminarinvasion


10 1+ 60

21 4/F RE, mod diff, focal choroidalinvasion

No 10 1+ 69

22 2/F RE, poor diff, post laminar No 0 0 7023 2.5/F RE, poor diff, diffuse No 10 1+ 6024 3/M RE, poor diff, post laminar No 1 0 7025 5/M RE, poor diff, diffuse choroidal

invasion, surgical end of ONinvaded


1 0 60

26 5/F Both eyes,(RE) well diff, prelaminar, focal choroidalinvasion

No 10 1+ 75

27 2/F Both eyes, (LE) poor diff, diffuse No 10 1+ 6028 2/M RE, poor diff, post laminar No 1 0 8029 1/F Both eyes, (RE) well diff, lamina

cribrosaPreop and postopchemo

10 1+ 61

30 2/F Poor diff, diffuse choroidal,post laminar, metastasis tosubmandibular area

Postop chemo 1 0 60

diff = differentiated; Postop = postoperative; chemo = chemotherapy.Proprietary interest: The authors have no financialinterest in any of the materials used in the study.

594 PostScript


on Novem


http://bjo.bmj.com

/B

r J Ophthalm


ownloaded from

http://bjo.bmj.com/

syncytium formation is a member of a new familyof transmembrane proteins including cd9, cd37,cd53, and cd63. J Immunol 1992;149:2879–86.

7 Rubinstein E, Boucheix C. Tetraspans. In: Kreis T,Vale R, eds. Guidebook to the extracellular matrix

and adhesion proteins. Oxford: Oxford UniversityPress, 1999:321–4.

8 Dong JT, Isaacs WB, Barrett JC, et al. Genomicorganization of the human KAI1 metastasis-suppressor gene. Genomics 1997;41:25–32.

Bilateral ocular surfacesquamous neoplasia: aclinicopathological case reportOcular surface squamous neoplasia (OSSN)was first described by Lee and Hirst1 as anumbrella term that encompasses intraepithe-lial and invasive squamous cell carcinoma ofthe conjunctiva and cornea. The incidence ofOSSN ranges from 0.02 to 3.5 per 100 000population and varies geographically, withgreater frequency near the equator.Generally, it is a slow growing tumour thatrarely metastasises, but is capable of causingextensive local tissue destruction. BilateralOSSN is extremely rare2–7 and offers a uniqueopportunity to study the biological character-istics of bilateral OSSN of the conjunctiva.The following case report describes theclinical presentation, histopathology, andimmunohistochemical evaluation of tumourproliferation markers of a patient diagnosedwith bilateral OSSN.

Case reportAn 86 year old white woman was referredto the Doheny Eye Institute because ofredness in her right eye that had deve-loped over a period of several months.She had undergone a mastectomy in1954. She had no history of oculartrauma, toxin exposure, or tobacco use. Herbrother and sister died from liver cancer.An ophthalmic examination revealed avisual acuity of 20/100 in each eye.Ectropion and indurated lower eyelid mar-gins were present bilaterally with no lossof cilia. A closer examination revealed athickened epithelium that lined the palpebralconjunctiva and cul del sac of the righteye (fig 1A). The left lower palpebral con-junctiva showed similar changes. However,there was a focal nodule on the inferiorbulbar conjunctiva (fig 1B).

The patient underwent a biopsy of the rightpalpebral conjunctiva. Histopathologicalexamination of the specimen revealed

Figure 1 (A) Microphotograph showing the KAI1 expression in control lymphoid follicle of thetonsil (magnification6200). (B) Microphotograph showing the intense membrane immunoreactivityof KAI1 in retinoblastoma with no invasion (magnification6200). (C) Microphotograph showingthe weak KAI1 immunoreactivity of tumour cells invading the lamina cribrosa of the optic nerve(magnification 6100). Inset shows negative KAI1 in tumour cells at the surgical end of optic nerve.(D) Microphotograph showing the negative KAI1 immunoreactivity of tumour cells at the metastaticsite in the submandibular region (magnification6200).

Figure 1 (A) Clinical photograph ofthe right eye showing diffusethickened epithelium involving thepalpebral conjunctiva and cul de sac.(B) Clinical photograph of the left eyeshowing irregular thickenedepithelium and a focal mass noted onthe inferior bulbar conjunctiva. (C)Photomicrograph from the leftconjunctival biopsy showingsquamous cell carcinoma in situ.Note that multiple abnormal mitoticfigures (inset) are present(haematoxylin and eosin; originalmagnification6250; inset6325). (D)Immunohistochemistry for humanpapillomavirus (HPV) showingpositive staining in majority of thetumour cells (original magnification6250).

PostScript 595


on Novem


http://bjo.bmj.com

/B

r J Ophthalm


ownloaded from

http://bjo.bmj.com/

acanthotic epithelium with squamous meta-plasia, occasional dyskeratotic cells, paraker-atosis, and hyperkeratosis. Multipleabnormal mitotic figures were present. Thebasement membrane was intact, and adiagnosis of conjunctival squamous cellcarcinoma in situ was diagnosed. A subse-quent biopsy of the left conjunctiva revealedfull thickness squamous metaplasia of theepithelium with acanthotic and markeddysplastic changes. Multiple abnormal mito-tic figures were seen and the basementmembrane was also intact, with an extensivechronic inflammatory cell infiltrate in thestroma (fig 1C). These findings were alsoconsistent with a diagnosis of squamous cellcarcinoma in situ of the conjunctiva.

Immunohistochemical analysis (Dako,Carpinteria, CA, USA) revealed that neoplas-tic cells were positive for pankeratin, humanpapillomavirus (HPV) (fig 1D), and Ki-67(MIB-1) in both specimens. Moreover, bothbiopsies indicated the presence of a few bcl-2positive cells. The right eye biopsy was p53positive and the left was p53 negative. Theimmunohistochemically positive cells werecounted by methods described previously.8

Table 1 summarises the immunohistochem-ical findings of both biopsies. Because of thepatient’s fragile health, surgical interventionwas deferred and she was treated with topicalmitomycin C in both eyes. At the follow upexamination 13 months after the biopsy, amass was found in the right lower palpebralconjunctiva, but there was no evidence ofsuch lesion in the left conjunctiva ormetastasis.

CommentThe aetiology of bilateral OSSN remainsunclear. Causative factors that are believedto contribute to the development of unilateralOSSN include ultraviolet light exposure,ocular trauma, predisposing genetic factors,and infection with HPV. Previous reportshave provided convincing evidence of anassociation with HPV type 16 in some casesof bilateral conjunctival dysplasia.3 4 It hasbeen postulated that the interaction betweenHPV and ultraviolet light may have a role inthe development of HPV related tumours inpatients who are exposed to the sun.9

However, both conjunctival lesions in thepresent case were located in areas that werenot exposed to sunlight, suggesting a possi-bility that HPV infection in both eyes mayhave led to the development of bilateralOSSN.

Conjunctival OSSN has been described as aslow growing, well differentiated tumour oflow grade malignancy that responds to localexcision and rarely metastasises. Immu-nohistochemical studies of bilateral lesionsallow comparison of the proliferative potencyin both eyes, and offer a unique opportunity

to study some biological aspects of bilateraltumours under the same environmentalconditions at the same point in time. Inrecent years p53, bcl-2, and MIB-1 have beenused as markers of proliferative potency.8 10

The p53 gene is a common cellular target inhuman carcinogenesis and is thought to havean important role in cellular proliferation. Incontrast with the wild type p53, mutants ofthe p53 gene produce an abnormal proteinwith a long half life and are thus immuno-histochemically detectable. Also, p53 hasbeen reported to be a prognostic marker inseveral tumours.8 10 Bcl-2 is a proto-oncogenethat is thought to have a role in oncogenesisby inhibiting programmed cell death andpreserving cells from p53 induced apoptosis.8

However, the mutant p53 protein alsoinduces apoptosis and decreases the expres-sion of bcl-2 proteins. Mahomed et al8

suggested that the interplay between theeffects of the increased mutant p53 proteinsand the absence of bcl-2 expression intumorigenesis may promote clonal expan-sion, leading to progressively increasedgenomic instability. The synergy of thepresence of mutant p53 and absence of bcl-2 in the present case might have allowed theprogression of the tumour in the rightconjunctiva.

Ki-67 is a nuclear antigen expressed in allstages of the cell cycle except the restingstage. MIB-1 is a monoclonal antibody thatrecognises the Ki-67 antigen, which is amarker of cellular proliferation and reportedto be a prognostic factor for various cancers.The high immunoreactivity of MIB-1 inconjunctival OSSN is usually associated withhighly aggressive tumour growth.8 Ourresults demonstrate a higher immunoreactiv-ity of MIB-1 in the right conjunctival speci-men. These findings indicate that the rightconjunctival specimen is more aggressivethan the left, and is consistent with thispatient’s clinical course.

In conclusion, this case represents a rareexample of conjunctival pathology: OSSN asa bilateral tumour. To our knowledge, this isalso the first report that compares the rightand left biopsies of conjunctival OSSN byimmunohistochemical analysis of potentialoncogenic factors. Enhanced expression ofMIB-1 and presence of mutant p53 protein inthe absence of bcl-2 may contribute to theaggressive biology of OSSN.

AcknowledgementsThis work was supported in part by NEI core grantEy03040 and by an unrestricted grant fromResearch to Prevent Blindness Inc, New York, USA.

Y UsuiDepartment of Ophthalmology, Tokyo Medical

University, Tokyo, Japan

Y Usui, G O Waring, R F See, N A RaoA Ray Irvine Ocular Pathology Laboratory, KeckSchool of Medicine of the University of Southern

California, Los Angeles, CA, USA

R F See, N A RaoDoheny Eye Institute, Keck School of Medicine of the

University of Southern California, Los Angeles, CA,USA

N A RaoDepartment of Ophthalmology, Keck School of

Medicine of the University of Southern California, LosAngeles, CA, USA

A C MarroneDepartment of Ophthalmology, Torrance Memorial

Medical Center, Torrance, CA, USA

Correspondence to: Yoshihiko Usui, MD, Departmentof Ophthalmology, Tokyo Medical University, 6-7-1

Nishi-shinjuku, Shinjuku-ku, Tokyo 160-0023,Japan; [email protected]


References

1 Lee GA, Hirst LW. Ocular surface squamousneoplasia. Surv Ophthalmol 1995;39:429–50.

2 Cervantes G, Rodriguez AA Jr, Leal AG.Squamous cell carcinoma of the conjunctiva:clinicopathological features in 287 cases.Can J Ophthalmol 2002;37:14–19.

3 Odrich MG, Jakobiec FA, Lancaster WD, et al. Aspectrum of bilateral squamous conjunctivaltumors associated with human papillomavirustype 16. Ophthalmology 1991;98:628–35.

4 Tritten JJ, Beati D, Sahli R, et al. [Bilateralconjunctivo-palpebral tumor in animmunocompetent man caused by humanpapilloma virus.] Klin Monatsbl Augenheilkd1994;204:453–5.

5 Olurin O. Bilateral conjunctival epithelioma in anadolescent. Ann Ophthalmol 1971;3:633–6.

6 Lam A, Borzeix A, Seck CM, et al. [A familialform of epidermoid carcinoma of the conjunctiva.]J Fr Ophtalmol 1996;19:143–8.

7 Tabin G, Levin S, Snibson G, et al. Late recurrencesand the necessity for long-term follow-up in cornealand conjunctival intraepithelial neoplasia.Ophthalmology 1997;104:485–92.

8 Mahomed A, Chetty R. Human immunodeficiencyvirus infection, Bcl-2, p53 protein, and Ki-67analysis in ocular surface squamous neoplasia.Arch Ophthalmol 2002;120:554–8.

9 Arends MJ, Wyllie AH, Bird CC. Papillomavirusesand human cancer. Hum Pathol 1990;21:686–98.

10 Aoki S, Kubo E, Nakamura S, et al. Possibleprognostic markers in conjunctival dysplasia andsquamous cell carcinoma. Jpn J Ophthalmol1998;42:256–61.

Employing endoscopic guidancefor placement of a blackdiaphragm aniridia intraocularlens following destructiveAcanthamoeba sclerokeratitisAnterior segment reconstruction can beparticularly challenging when anatomic land-marks are lost. We describe a case ofdestructive Acanthamoeba sclerokeratitisresulting in aniridia, aphakia, loss of limbalarchitecture, and corneal opacification thatwas approached surgically with penetratingkeratoplasty and placement of a black dia-phragm aniridia intraocular lens under endo-scopic guidance.

Case reportOur patient, a 48 year old female contact lenswearer, was diagnosed with Acanthamoebakeratitis in June 2000. Before our evaluation,she had been treated with tobramycin and

Table 1 Comparative immunohistochemical findings

Right conjunctival biopsy Left conjunctival biopsy

Immunohistochemical markerPankeratin + + + + + + + +HPV + + + + + + + +P53 + + + + 2

Bcl-2 2 2

MIB-1 + + + + +

2 = 5% or less positive cells; + = 6% to 25%; + + = 26% to 50%; + + + = 51% to 75%; + + + + = morethan 75%.

596 PostScript


on Novem


http://bjo.bmj.com

/B

r J Ophthalm


ownloaded from

http://bjo.bmj.com/

dexamethasone ointment, topical trifluridine,oral acyclovir, oral prednisone, and topicalprednisolone acetate 1%. We diagnosedAcanthamoeba keratitis and began aggressivetreatment with polyhexamethyl biguanide,chlorhexidine, and oral clotrimazole. ByJanuary 2001, she was culture negative, buthad developed necrotising sclerokeratitiswith limbal involvement, dense corneal opa-cification, and descemetocoele formation.Urgent penetrating limbal keratoplasty wasperformed. Upon placement of the lid spec-ulum, spontaneous perforation of the corneaoccurred with prolapse of the lens andnecrotic iris. We performed a 12.5 mmdiameter keratolimbal resection, removedresidual lens and necrotic iris, and performedanterior vitrectomy. A 13.0 mm keratolimbalgraft was placed and covered with anamniotic membrane graft. (fig 1)

Eighteen months later, the patient hadnegative cultures, a quiet eye, an opaquecorneal graft, controlled intraocular pressure,and counting fingers vision with projection toall four quadrants. However, the patientcomplained of severe glare and light sensi-tivity. Soft contact lens wear was unsuccess-ful because of irregular postsurgicaltopography. After extensive discussion, thishighly motivated patient elected to pursuefurther anterior segment reconstructive sur-gery to address the aniridia, aphakia in theabsence of capsular support, and cornealopacity.

Penetrating keratoplasty and implantationof a sulcus fixated Morcher 67F blackdiaphragm polymethylmethacrylate lens wasplanned. External landmarks for transscleralsuture fixation had been lost due to infec-tious necrosis and the large keratolimbalgraft. Instead, suture placement was guidedby an ocular endoscope (URAM E2MicroProbe Laser System, EndoOptiks, NewJersey, USA). Following excision of an8.0 mm diameter corneal button, a 10–0prolene suture on an STC-6 needle (EthiconInc, New Jersey, USA) was passed externallyunder a scleral flap and viewed internally viathe endoscope as it entered the ciliary sulcus.The suture was passed through the lensfixation loop. A 25 gauge needle was passedexternally into the ciliary sulcus under endo-scopic visualisation, the STC-6 needle waspassed into its bore, and the complex guidedout of the eye. This process was repeated forthe opposing haptic, the sutures were tied,and an 8.0 mm donor button was placed. Inthe early postoperative period, the intraocularlens was positioned without obvious decen-tration or tilt, and the patient reportedsubstantial improvement in her glare symp-toms.

CommentA black diaphragm intraocular lens designallows simultaneous treatment of aniridiaand aphakia.1 The Morcher 67F has a13.5 mm length, 10 mm diameter optic, anda 5 mm central clear zone. Precise hapticcapture in the ciliary sulcus is necessary tominimise risks of haptic-optic crowding,mechanical irritation, and tilt or decentrationof a small optic zone. Unfortunately, lensdecentration and tilt is commonly observedfollowing transscleral fixation of lenses.2 3

This can be attributed to suboptimal hapticposition following blind passage of fixationsutures.3 Althaus and Sundmacher havedescribed the usefulness of direct endoscopicvisualisation in eyes undergoing transscleralsuture lens fixation.4 In our patient, accuratelens position was critical, and the risk of lensmalposition high, given her unfavourableanatomy. Our experience confirms that endo-scopic visualisation is valuable for the place-ment of transscleral lens fixation sutures,particularly when surgical landmarks are lostand when mild lens malposition mightadversely affect the surgical outcome.

A M Poothullil, S D McLeod, S LinDepartment of Ophthalmology, University of

California San Francisco, San Francisco, CA, USA

S D McLeodThe Francis I Proctor Foundation University of

California San Francisco, San Francisco, CA, USA

Correspondence to: Dr S D McLeod, Department ofOphthalmology, 10 Koret Way, K-301, San

Francisco, CA 94143, USA; [email protected]


References

1 Thompson CG, Fawzy K, Gryce IG, et al.Implantation of a black diaphragm intraocularlens for traumatic aniridia. J Cataract Refract Surg1999;25:808–13.

2 Schein, et al. A randomized trial of intraocularlens fixation techniques with penetratingkeratoplasty. Ophthalmology1993;100:1437–43.

3 Hayashi K, Hayashi H, Nakao F, et al. Intraocularlens tilt and decentration, anterior chamber depth,and refractive error after trans-scleral suturefixation surgery. Ophthalmology1999;106:878–82.

4 Althaus C, Sundmacher R. Endoscopicallycontrolled optimization of trans-scleral suturefixation of posterior chamber lenses in the ciliarysulcus. Ophthalmologe 1993;90:317–24.

Figure 1 Endoscopic view of the STC-6 needleas it enters the ciliary sulcus.

A video clip of thisprocedure is availableon the BJO website(www.bjophthalmol.com)

MAILBOX

Variant CJD and tonometryWe read with interest the paper by Lim et al.1

The authors interestingly showed the impor-tance of tonometer head wiping in reducingcorneal epithelial cell count and the signifi-cance of damaged prism surfaces in trappingdebris. They hypothesised the possible risk ofvariant Creutzfeld-Jakob disease (vCJD)transmission from cornea epithelial cells

present on the tonometer surface. The ques-tion is really one of what constitutes theinfectious dose of vCJD for this mode oftransmission, and this is currently unknown.

There has been one definite, one probable,and two possible cases of CJD throughcorneal transplantation but one can hardlycompare the prion load in a full thicknesscorneal graft with a mean epithelial cellcount of 9/10 (after wiping or wipe/Milton).Two of the four cases of transmission hadhad multiple graft procedures.

The evidence from animal studies on CJDinfected corneal transmission is also variable.Herzberg transplanted CJD infected corneasonto two Capuchin monkeys; both remaineddisease free for up to 4 years.2 Manuelidis et alshowed transmission of CJD when infectedcorneas were placed directly into the anteriorchamber of uninfected guinea pigs3 but didnot show transmission of CJD after penetrat-ing keratoplasty. Tateishi injected emulsifiedCJD infected cornea into the brains of sixmice and only one developed characteristicchanges after 2.8 years.4

These studies certainly suggest that anintraocular/intracerebral delivery must beneeded for transmission but even then, theinocula that produced disease after intracer-ebral inoculation only irregularly transmitdisease after peripheral (intraocular) inocu-lation.5 In addition, host genetic factors(homozygosity at codon 129) will also deter-mine the risk of susceptibility.6 There is alsoconvincing evidence that the agent strain andhost genotype will determine whether ocularinvolvement even occurs in experimentalrodent models of scrapie.7

More recently, western blot analysis oneyes of patients that had been infected withsporadic CJD and vCJD8 confirmed earlierresults (also in human eyes) that PrPSc couldonly be detected in the retina and not incornea or sclera. From the evidence, the riskof transmission from tonometry would sug-gest this to be more theoretical than practical,and may be reduced further—as the authorssuggest—by adequate wiping or regularreplacement of tonometer prisms. The ques-tion of risk can only be truly resolved bytransmission studies using primary humandiseased eye tissue.

J S Mehta, R J Osborne, P A BloomThe Western Eye Hospital, London, UK

Correspondence to: Dr J S Mehta, 9 SandringhamCourt, King & Queen Wharf, Rotherhithe Street,London SE16 5SQ, UK; [email protected]


References

1 Lim R, Dhillon B, Kurian KM, et al. Retention ofcorneal epithelial cells following Goldmanntonometry: implications for CJD risk.Br J Ophthalmol 2003;87:583–6.

2 Herzberg L. Creutzfeldt-Jakob disease andcorneal grafts. Med J Aust 1979;1:248.

3 Manuelidis EE, Angelo JN, Gorgacz EJ, et al.Experimental Creutzfeldt-Jakob diseasetransmitted via the eye with infected cornea.N Engl J Med 1977;296:1334–6.

4 Tateishi J. Transmission of Creutzfeldt-Jakobdisease from human blood and urine into mice.Lancet 1985;2:1074.

5 Brown P, Gibbs DJ, Rodgera-Johnson P, et al.Human spongiform encephalopathy: the NationalInstitutes of Health series of 300 cases ofexperimental transmitted disease. Ann Neurol1994;35:513–29.

PostScript 597


on Novem


http://bjo.bmj.com

/B

r J Ophthalm


ownloaded from

http://bjo.bmj.com/

6 Mehta JS, Franks WA. The sclera, the prion, andthe ophthalmologist. Br J Ophthalmol2002;86:587–92.

7 Foster JD, Fraser H, Bruce ME. Retinopathy inmice with experimental scrapie. NeuropatholAppl Neurobiol 1986;12:185–96.

8 Head MW, Northcott V, Rennison K, et al. Prionprotein accumulation in eyes of patients withsporadic and variant CJD. Invest Ophthalmol VisSci 2003;44:342–6.

Intravitreal triamcinoloneacetonide for exudative agerelated macular degenerationWe read the article by Jonas et al onintravitreal triamcinolone injections for exu-dative age related macular degeneration withinterest.1 The paper stated that visual acuityincreased significantly (p,0.001) from 0.16(SD 0.11) to a mean maximum of 0.23 (0.17).The authors therefore picked the best fromone of up to 10 postoperative visual acuitymeasurements and compared it with a singlepreoperative visual acuity measurement. Thisis misleading the reader regarding the trueeffectiveness of the treatment.

The Macular Photocoagulation StudyGroup found that the differences in betweentwo repeated tests were one line or more in13% of cases and the differences were great-est in patients with visual acuity of 20/100 orworse.2 By taking up to 10 postoperativemeasurements, Jonas et al greatly increasedthe chances of a positive result. The differ-ence between mean pre-injection 0.16 (20/125 or 6/36) and best mean postoperative0.23 (20/87 or 6/26) was less than one line onthe Snellen chart.

Their table 1 gave the mean visual acuitypre-injection and at various time intervalspost-injection. At 1 and 2 months, the pvalues were 0.04. It was unclear whether thep values were one or two tailed but both weredescribed as not significant (NS) in table 1.Multiple significance testing at each of anumber of time points is generally notrecommended—if it is done, some kind ofadjustment to the p values is needed.3 4

Looking at the results presented in table 1,readers might conclude that triamcinolonehad a transient and doubtful beneficial effecton the visual acuity.

The authors go on to further analyse theresults into improvements of three and six ormore lines. The vision was tested on a Snellenchart which has irregular steps. Three or sixlines do not therefore represent a constantchange in visual angle (as in a logMAR chart)and therefore the analysis was confusing.

Variations in intraocular pressure of 5 or6 mm Hg occur diurnally in normal indivi-duals as well as glaucomatous patients.5–7

While there is little doubt that triamcinolonemay affect the intraocular pressure, thecomparison of the baseline with the highestvalue (p,0.001) was misleading as was thecomparison of the highest value with that at7 months (p,0.001). Of more interest mightbe the number of patients who had very highlevels (the range extended to 64 mm Hg) andwhether these very high intraocular pressuresresponded to treatment.

The authors’ experience in using triamci-nolone is well recognised. We congratulatethem on an otherwise excellent piece of work.

D Wong, I Campbell, C Groenwald,E Lancaster

Royal Liverpool University Hospital, Prescot Street,Liverpool L7 8XP, UK

Correspondence to: Mr David Wong, Royal LiverpoolUniversity Hospital, Prescot Street, Liverpool L7 8XP,

UK; [email protected]


References

1 Jonas JB, Kreissig I, Degenring R. Intraocularpressure after intravitreal injection oftriamcinolone acetonide. Br J Ophthalmol2003;87:24–7.

2 Blackhurst DW, Maguire MG. Reproducibility ofrefraction and visual acuity measurement under astandard protocol. The Macular PhotocoagulationStudy Group. Retina 1989;9:163–9.

3 Altman DG. Practical statistics for medicalresearch. London: Chapman and Hall, 1991.

4 Matthews R. The numbers don’t add up. NewScientist 2003;177:28.

5 Pointer JS. The diurnal variation of intraocularpressure in non-glaucomatous subjects: relevancein a clinical context. Ophthalmic Physiol Opt1997;17:456–65.

6 De Vivero C, O Brien C, Lanigan L, et al. Diurnalintraocular pressure variation in low-tensionglaucoma. Eye 1994;8(Pt 5):521–3.

7 Smith J. Diurnal intraocular pressure. Correlationto automated perimetry. Ophthalmology1985;92:858–61.

Authors’ replyWe thank Wong and colleagues very muchfor their interest in our study.1 We agree withthem that comparing a single preoperativemeasurement with the best out of a series ofpostoperative measurements gives a tendencytowards a falsely high increase in visualacuity after the triamcinolone acetonideinjection. That there was an increase in visualacuity after the injection in some patients,however, may have been demonstrated intable 1 giving the visual acuity measurementsbefore and after the application of triamci-nolone acetonide. At 1 and 2 months, thedifference from the preoperative values wassignificant with a p value of 0.04.Unfortunately, the values are described asnon-significant in what is a typographicalerror in the manuscript. We regret this error.The authors agree with Wong and colleagues,that the effect of triamcinolone acetonide istemporary, and that repeated injections maybe necessary. In some patients, repeatedintravitreal injections of triamcinolone aceto-nide were associated with repeated re-increase in visual acuity.2

We also agree with Wong and coworkersthat comparing the preoperative intraocularpressure measurement with the highest valueduring the follow up again falsely increasedthe number of ocular hypertensive subjectsafter the injection. In this matter, however, itwas not in favour of the treatment since itartificially increased the number of patientswith unwanted side effects. Concerning thenumber of patients with very high intraocularpressures after the injection, a preliminaryanalysis shows that about 1% of the eyesinjected will need filtering surgery (owndata).

The authors would like to thank Wong andcolleagues for constructively commenting onthe study and pointing out, as did theauthors, that the intravitreal injection oftriamcinolone acetonide as treatment forexudative macular degeneration is still inclinical evaluation and that its therapeuticeffect has not been proved so far. This may beeven more important as a recent randomisedstudy using 4 mg triamcinolone acetonidedid not demonstrate an effect on the risk of

loss of visual acuity during the first year ofthe study in eyes with classic choroidalneovascularisation.3 A significant anti-angio-genic effect, however, was found 3 monthsafter treatment. In view of the anti-angio-genic, anti-inflammatory, and anti-oedema-tous effects of intravitreal triamcinolone,further randomised trials using differentdoses of triamcinolone acetonide in eyes withdifferent types of subfoveal neovascularisa-tion may be warranted.

J B Jonas, I Kreissig, R DegenringDepartment of Ophthalmology, Theodor-Kutzer-Ufer,

1-3 Mannheim, Germany; [email protected]


References

1 Jonas JB, Kreissig I, Degenring R. Intraocularpressure after intravitreal injection oftriamcinolone acetonide. Br J Ophthalmol2003;87:24–7.

2 Jonas JB, Kreissig I, Degenring RF. Repeatedintravitreal injections of triamcinolone acetonideas treatment of progressive exudative age-relatedmacular degeneration. Graefes Arch Clin ExpOphthalmol 2002;240:873–4.

3 Gillies MC, Simpson JM, Luo W, et al. Arandomised clinical trial of a single dose ofintravitreal triamcinolone acetonide forneovascular age-related macular degeneration:one-year results. Arch Ophthalmol2003;121:667–73.

Viscosurgery in diabeticvitrectomyGrigorian and colleagues recently recountedtheir experience of using viscosurgery toremove epiretinal membranes (ERMs) fromeyes with proliferative diabetic retinopathy(PDR).1 They concluded that ‘‘viscodissec-tion’’ (injection of Healon between thefibrovascular proliferations and the retina)is safe and is equally as effective as its non-use. On the contrary, their study shows thatviscodissection is not cost effective (becauseof the costs of both the viscoelastic and theextra operating time), and confirms that thetechnique is inherently unsafe in PDR.2 3

The use of Healon to aid dissection offibroglial and fibrovascular ERMs duringvitrectomy was introduced in the 1980s4 butwas not widely adopted. Viscoelastic materialsquirting out from under the ERMs was‘‘messy’’ and led to the formulation (in the1990s) of yellow tinted Healon to aid itsvisualisation and simplify its removal5 andthen Healon GV (viscosity 10 times that ofHealon) for adherent diabetic ERMs.6 In1984, we began undertaking ‘‘viscodelamina-tion’’ in diabetic vitrectomy.2 7 This techniquewas primarily directed at stripping the poster-ior hyaloid membrane (PHM) from detached,ischaemic and atrophic peripheral retina.Viscodelamination was especially useful incombined tractional and rhegmatogenousretinal detachments (CTRD) with very lim-ited non-rhegmatogenous posterior vitreousdetachment (PVD) present. Because of theprohibitive cost of Healon, methylcellulose1% was injected in the majority of eyes.

To summarise our experience, stripping ofthe PHM usually proceeded uneventfullyduring slow pressurisation of the closedretrohyaloid compartment by viscoelastic, asdid separation of any loosely adherent,sparsely vascularised ERMs that were con-tained within the peripheral vitreous cortex.2

In well photocoagulated eyes, the separation

i

u

vl

w

iv

ppi

p

mh

r

vWbm

rm

598 PostScript


on Novem


http://bjo.bmj.com

/B

r J Ophthalm


ownloaded from

http://bjo.bmj.com/

sometimes continued posteriorly, culminat-ing in a complete PVD. In the case of moreadherent fibrovascular ERMs, their ‘‘visco-elevation’’ sometimes occurred throughstretching of the vascular and glial tent pegsconnecting the ERMs to the retina. The PHMand ERMs could then be removed en blocusing the suction cutter. However, instead ofstretching, the vascular connections betweenthe ERM and the retina tended to bedisrupted. Avulsion generally occurred atthe point of greatest weakness at the originsof neovascular outgrowths from the retinalveins. Although correlating with ERM vascu-larity and with the density of neovascularoutgrowths from the retina, ERM retinaladherence was unpredictable, and bleedingwas ultimately an inevitable consequence ofthe perpendicular hydraulic forces necessaryto effect peeling of more adherent ERMs.Fortunately, the bleeding from side puncturesin the retinal veins was constrained by theviscoelastic (so called ‘‘haemorrhagic con-finement’’)2 8 and a high ambient intraocularpressure during the surgery. However, as waspredictable in theory, but again unpredictablein practice, the hydraulic tension sometimesdisrupted the retina ahead of, and instead of,peeling the ERMs. Furthermore, recurrentfibroglial membranes were sometimesobserved later even in eyes where viscodela-mination had proceeded uneventfully. Thishas been attributed to the difficulty incompletely removing viscoelastic from theretinal surface, with preretinal retention ofgrowth factors.2 3 Not for nothing is oneviscoelastic mixture marketed as Viscoat.We had discontinued viscosurgery in PDRby 1988 in favour of purely mechanicalmethods that minimise ERM elevation.3

Fifteen years on and Grigorian and collea-gues have clearly come to a very differentconclusion from ours despite reporting aconsiderable excess of iatrogenic posteriorretinal breaks during, and recurrent detach-ment after, viscosurgery.1 By back calculationfrom their assiduously collected data, itappears that 20 posterior breaks wereinduced in 65 eyes undergoing viscodissec-tion compared with four in 89 eyes havingconventional surgery. This trend was con-firmed in groups of eyes with pathology ofsimilar (‘‘relatively high’’) complexity. Thus,there were 10 iatrogenic posterior breaks in34 viscodissections in eyes in the range C4-6compared with three in 26 conventionaloperations. (It is acknowledged thatGrigorian et al state that the complexity score‘‘does not account for the degree of adhe-sion,’’ neither was it ‘‘a good predictor of theamount of traction necessary to dissect amembrane.’’) The intraoperative problemsappear to have been reflected in the ultimateoutcomes. After 6 months of follow up, forexample, a detached retina was evident inseven of 43 eyes (16%) in the viscosurgerygroup compared with three of 58 eyes(5%) undergoing conventional surgery.Furthermore, although eyes with CTRDseemed to fare well whether or not Healonwas used, this was not the case in eyes withtractional detachments (with or withoutvitreous haemorrhage). Six of 30 eyes (20%)had a detached retina 6 months after

viscodissection compared with only two of37 eyes (5%) after conventional surgery.Indeed, most of the data favoured conven-tional surgery. Lower viscosity Healon wasproposed as a future means of reducingthe frequency of iatrogenic breaks, but thisis unlikely to be helpful in their attempt toachieve the impossible—that is, a worth-while increase in the ease and speed ofERM removal without an unacceptableadded risk of retinal haemorrhage, tears,and scarring. Better by far would be to avoidviscoelastics altogether.

D McLeodUniversity of Manchester, Manchester Royal Eye

Hospital, Oxford Road, Manchester M13 9WH, UK;[email protected]


References

1 Grigorian RA, Castellarin A, Fegan R, et al.Epiretinal membrane removal in diabetic eyes:comparison of viscodissection with conventionalmethods of membrane peeling. Br J Ophthalmol2003;87:737–41.

2 McLeod D, James CR. Viscodelamination at thevitreoretinal juncture in severe diabetic eyedisease. Br J Ophthalmol 1988;72:413–19.

3 Charles S. Techniques and tools for dissection ofepiretinal membranes. Graefes Arch Clin ExpOphthalmol 2003;241:347–52.

4 Stenkula S, Ivert I, Gislason I, et al. The use ofsodium-hyaluronate (Healon) in the treatment ofretinal detachment. Ophthalmic Surg1981;12:435–43.

5 Stenkula S, Ivert I, Berglin L, et al. Healon yellowas a surgical tool in maneuvering intraoculartissues. Ophthalmic Surg 1992;23:708–710.

6 Crafoord S, Stenkula S. Healon GV in posteriorsegment surgery. Acta Ophthalmol (Copenh)1993;71:560–1.

7 Barry PJ, Hiscott PS, Grierson I, et al. Reparativeepiretinal fibrosis after diabetic vitrectomy. TransOphthalmol Soc UK 1985;104:285–96.

8 Folk JC, Packer AJ, Weingeist TA, et al. Sodiumhyaluronate (Healon) in closed vitrectomy.Ophthalmic Surg 1986;17:299–306.

NOTICES

Cataract surgeryThe latest issue of Community Eye Health(No 48) discusses a solution to reduce

worldwide cataract blindness, includingsutureless non-phaco cataract surgery. Forfurther information please contact: Journalof Community Eye Health, InternationalResource Centre, International Centre forEye Health, Department of Infectious andTropical Diseases, London School ofHygiene and Tropical Medicine, KeppelStreet, London WC1E 7HT, UK (tel: +44(0)20 7612 7964; email: [email protected]; website: www.jceh.co.uk). Annualsubscription (4 issues) UK£28/US$45. Freeto developing country applicants.

Elimination of avoidable blindnessThe 56th World Health Assembly (WHA)considered the report on the eliminationof avoidable blindness (doc A56/26) andurged Member States to: (1) Commit them-selves to supporting the Global Initiativefor the Elimination of Avoidable Blindnessby setting up a national Vision 2020 plan by2005; (2) Establish a national coordinat-ing committee for Vision 2020, or a nationalblindness prevention committee to helpimplement the plan; (3) Implement theplan by 2007; (4) Include effective moni-toring and evaluation of the plan with theaim of showing a reduction in the magni-tude of avoidable blindness by 2010; (5)To support the mobilisation of resourcesfor eliminating avoidable blindness. TheWHA also urged the Director-General tomaintain and strengthen WHO’s collabora-tion with Member States and the partnersof the Global Initiative for the Eliminationof Avoidable Blindness as well as aid inthe coordination and support of nationalcapability.

XVth Meeting of the InternationalNeuro-Ophthalmology SocietyThe XVth Meeting of the InternationalNeuro-Ophthalmology Society will takeplace 18–22 July 2004, in Geneva,Switzerland. Further details: Prof. ASafran, University Hospital Geneva, c/oSYMPORG SA, Geneva (fax: +4122 8398484; email: [email protected]; website:www.symporg.ch).

4th International Congress onAutoimmunityThe 4th International Congress on Auto-immunity will take place 3–7 November2004 in Budapest, Hungary. The deadlinefor the receipt of abstracts is 20 June2004. Further details: Kenes InternationalGlobal Congress Organisers and Associa-tion Management Services, 17 Rue duCendrier, PO Box 1726, CH-1211 Geneva 1,Switzerland (tel: +41 22 908 0488; fax:+41 22 732 2850; email: [email protected]; website: www.kenes.com/autoim2004).

XVI International Congress for EyeResearchThe XVI International Congress for EyeResearch will be held on 29 August–3 Sep-tember 2004 in Sydney, Australia. For fur-ther information, please contact: [email protected] (website: www.tourhosts.com.au/icer2004).

CORRECTION

In the article by Brodsky et al in the Februaryissue (Br J Opthalmol 2004;88:268–272), aportion of the text within fig 2 was incor-rectly labelled. The label under ‘‘+ Superiorrectus contracture’’ is currently printed as‘‘Compensatory head tilt away from side offixing eye.’’ It should have been printed as‘‘Compensatory head tilt toward the side ofthe fixing eye.’’

PostScript 599


on Novem


http://bjo.bmj.com

/B

r J Ophthalm


ownloaded from

http://bjo.bmj.com/

Documents

Br J Ophthalmol 88 PostScript · 2004-03-10 · the Agence Franc¸aise de Se´curite´ Sanitaire des Produits de Sante´. They are very similar in composition, both being based on