Eye(2004)18,1188–1206 &2004NaturePublishingGroupAllrightsreserved0950-222X/04$30.00 www.nature.com/eye Posterior ischaemic SSHayreh C A optic neuropathy: M B R clinical features, I D pathogenesis, and G E management O P H T H A L M Abstract nonarteritic PIONproduced significant O L improvementofvisualacuityaswellasvisual Purpose To investigateandpresent a O fields,butnotsoinarteriticorsurgicalPION. G comprehensiveaccountoftheclinicalfeatures, However,some spontaneousvisual IC pathogenesis,andmanagement ofposterior improvementalso occurred insome untreated A ischaemicoptic neuropathy (PION). L nonarteritic PIONcases. Methods Thisretrospectivestudyisbasedon S Conclusions PIONisadistinctclinicalentity Y 53consecutive eyesof42patients withPION but should bediagnosedonlyafter exclusion M seeninmyclinicsince1973,whofulfilledthe P ofallothercausesofvisualloss.Inallpatients inclusion criteria.They weresystematically O olderthan50,GCAmustberuledout.Thereis S evaluated,treated,andfollowed byme.All IU patients hadinitiallydetailed ophthalmic usually markedvisualloss, withcentralfield M evaluation ofthe anterior andposterior defectbeingthe most common. Thestudy suggests thathigh-dose steroid therapyin segments,includingvisual field with nonarteritic PION, soonafter theonset of Departmentof Goldmannperimeterandfluorescein fundus visualloss, resulted insignificant visual OphthalmologyandVisual angiography.All patients aged50years and Sciences,Collegeof improvementcomparedtotheuntreatedcases, olderwere also investigatedfor giant cell Medicine,Universityof but not inarteritic andsurgicalPION. arteritis(GCA). Everyattemptwas madeto Iowa,IowaCity,IA,USA Eye (2004)18, 1188–1206. ruleoutothercausesofvisualloss.Follow-up doi:10.1038/sj.eye.6701562 Correspondence:SSHayreh evaluationwassimilartotheinitialevaluation Departmentof exceptangiography. Aetiologically, PIONcan Keywords: giant cellarteritis; ischaemic optic OphthalmologyandVisual bedivided intothree types:arteritic dueto Sciences neuropathy;posterior ischaemic optic GCA,nonarteritic not dueto GCA,and UniversityHospitals& neuropathy;radical neck dissection; spinal surgicalfollowinga surgicalprocedure. Clinics surgery 200HawkinsDrive Steroidtherapywas givento onlythose IowaCity nonarteritic PIONpatients whooptedto try IA52242-1091,USA that,but was givento allarteriticPION Tel: þ13193562947 Introduction patients. Fax: þ13193537996 Results PIONwasnonarteriticin28patients In mystudieson theblood supplyofthe optic E-mail:sohan-hayreh@ uiowa.edu (35eyes),arteriticin 12(14 eyes),andsurgical nerve,it wasfound that, unlike the rest ofthe inthree (four eyes).Visualacuity varied optic nerve,the anterior part ofthe optic nerve Received:4September between20/20andnolightperceptionFitwas (opticnervehead)hasitsowndistinctsourceof 2003 countfingers orlessin19of35eyeswith bloodsupply,thatis,posteriorciliary arteries Accepted:4September nonarteritic PION,four of14in arteritic, and (PCAs),which do notsupplythe restofthe 2003 allfourwithsurgicalPION.Themostcommon optic nerve.1,2Fromclinical andexperimental SupportedinpartbyGrants visualfield defectwas centralvisualloss, studies, Ifound in1974thatischaemic optic EY-1151andRR-59from aloneorin combination withother types of neuropathy involvingthe optic nerve headisa theNationalInstitutesof visualfield defects. Initially,opticdisc and distinct clinical entity,dueto interferencewith Health,Bethesda,Maryland, fundus showed noabnormality but thedisc the PCAcirculation,andI namedit anterior andinpartbyunrestricted usuallydeveloped pallor inabout6–8weeks. ischaemic optic neuropathy (AION).3An grantfromResearchto PreventBlindness,Inc.,New Aggressivetreatmentwithhigh-dosesystemic accompanying editorial, commenting onmy York. steroidduringthe veryearly stagesof paper,stated: ‘TheabbreviationAIONadopted Posteriorischaemicopticneuropathy SSHayreh 1189 byHayrehalmostofnecessitypresupposestheexistence Inclusion criteria ofaposteriorcounterpart‘PION’whichmight,perhaps, beappliedtothepost-traumaticopticatrophyassociated IstressedinmyoriginalpaperonPION:‘Therearesome with damagetothe optic nerve withinits bonycanal.’4 diagnoseswhich should not bemade until other Later,in1978,havingseenmorepatientswithischaemic possibilities have been carefully ruled out.I thinkPION opticneuropathy,Iclassifiedischaemicopticneuropathy isoneofthese.’7Iadheredtothisruleverystrictlyinthe into twotypes:AION, involving theanterior part ofthe presentstudy.Therefore,Imadeallpossibleattemptsto optic nerve,and posterior ischaemic optic neuropathy excludefromthis studypatients whose diagnosiswas (PION) involvingthe restofthe optic nerve.5,6In 1981, evenslightly indoubt, evenif theywere referredto me Isubmitted apaperdescribing PIONindetail tothe withadiagnosisofPION.Myinclusioncriteriawere:(1) Archives ofOphthalmologyfor publication; it was suddenonsetofvisualdeterioration;(2)presenceofoptic rejected with the remarkthat‘thereis noreasonto nerve relatedvisual fielddefect(s);(3) initially, normal believethat anysuch clinical entityexists’. It was opticdiscandrestofthefundusonophthalmoscopyand published elsewhere.7 Innumerable reports have since fluorescein angiography;and(4) exclusionof anyother establishedPION asa clinical entity,but allofthem ocular,optic nerve, or neurologic disorders,including except twoareanecdotalcase reports,basedon oneor compressive, inflammatory,orother mechanisms,as twopatientsonly.Theexceptionsare:(1)byIsayamaetal8 causeofvisual loss. Forpatients with GCA,Iused the in14PIONpatients and(2) bySaddaet al9 in72PION criterion ofpositivetemporal artery biopsyforGCA, in patients. allbutthreewhereforvariouslogisticreasonsIcouldnot Thus,basedonitsbloodsupply,theopticnervecanbe getatemporalarterybiopsy;thesethreepatientshadall dividedintotwodistinctregions:(a)theopticnervehead theclassicalclinicalfindingsofGCA,includinganorexia, almostentirelysuppliedbythePCAcirculation1,2and(b) weight loss,anaemia, malaise,tender temporalarteries, the rest ofthe opticnerve (ieposterior part ofthe optic scalp tenderness, headache,neck pain andjaw nerve) supplied from multiple other sources2,10–12(see claudication, highly elevatederythrocytesedimentation below). Pathogenetically, aswellas clinically,acute rate(ESR), and/or C-reactive protein (CRP)and ischaemiaofthe optic nerve resultsintwo verydistinct thrombocytosis,sothattherewasnodoubt oftheir typesofischaemicopticneuropathy:(a)AIONinvolving diagnosisofGCA. theopticnervehead,and(b)PIONinvolvingasegment PIONwasnonarteriticin28(35eyes)patients,arteritic ofthe posterior part ofthe optic nerve. in12(14eyes)patients,andsurgicalinthree(foureyes) Aetiologically, PIONcan beclassified into three patients. types:(1) arteriticPION duetogiant cellarteritis Attheinitialvisit,allpatientshadadetailedsystemic (GCA),(2) nonarteritic PIONduetocausesother than andophthalmichistory,includingspecificquestioning,in GCA,and(3)surgicalPIONattributable toa surgical detail,ofpatientsaged50yearsandolderfortheocular procedure.9,13 andsystemic signsand symptomsofGCA.All patients The objectiveofthe presentstudyisto presenta hadophthalmic evaluation byme;this included Snellen comprehensive account ofthe clinical features, visual acuity andvisual fieldtesting with a Goldmann pathogenesisand managementofPIONbased ona perimeter usingI2e, I4e, andV4eisopters (unlessthe retrospective studyof53consecutive PIONeyesof43 visual losswas soseverethatthe fields could notbe patients, systematicallyevaluated, treated,and plotted), andmost ofthem alsohad anAmslergrid followed duringthe past 30yearsinmyOcular evaluation ofthe sizeofthe centralscotoma. All had Vascular Clinic at the University ofIowa Hospitals external andslit-lamp examinationofthe anterior andClinics. segment, lensandvitreous,intraocular pressure measurement,ophthalmoscopicexaminationand,almost invariably, fluorescein fundus angiography.Allpatients aged50yearsandolderor suspected tohave GCAalso Materials andmethods hadESR (Westergren)and, from1985,CRP to ruleout Thisretrospectivestudyisbasedon53consecutiveeyes GCA;ifGCAwassuspected,theywerepromptlystarted of42Caucasianpatients andone Blackpatientwith on high-dosesteroid therapy,similar tothe steroid PIONseenintheOcularVascularClinicattheUniversity therapyregimendescribed indetail elsewhere.14–16 ofIowa Hospitals andClinics since1973, whowere Temporalartery biopsy wascarried outas soon as systematically evaluated,treated,and followedby me convenientto confirmthe diagnosis. To ruleout other personally, except forthreepatients,investigated causesofvisual loss,includinginflammatory, similarlybyDrRandyKardon,myneuro-ophthalmology compressive, infiltrative orother forms ofoptic colleague. neuropathy (including Leber’s optic neuropathy), Eye Posteriorischaemicopticneuropathy SSHayreh 1190 appropriateinvestigations werecarried out, including mildinFigures2dand3c,moderateinFigures2aand3b, orbital echographic,magnetic resonanceimaging, and/ marked inFigures1c,d, 2b, 3aand 4a,b,andseverein orcomputertomography,andneurologicevaluation.All Figure1a,b. nonarteritic patients had asystemic evaluation. Ateach follow-up visit, allpatients hadthe sameevaluation as on the initialvisit except forfluoresceinfundus Statistical analyses angiography. The linear mixedmodelanalysis for repeatedmeasures wasused tocompareLogMAR visual acuityandvisual Steroidtherapy field grade(0–4Fsee above)changes innonarteritic PIONeyes thathadsteroid treatmentandthose not Arteritic PION To prevent further visual loss,all treated.Thefactorsinthemodelweretreatment(steroid patients weretreatedwith a steroid therapyregimen or none), time (baselineandfinal follow-up), andthe similar tothatI haveused inmyother studies inGCA treatment–timeinteractioneffect.Specificmeancontrasts patients; it isdescribed elsewhere.14–16 thatweretested included (1)between-treatmentgroup Nonarteritic PION Since thereisno known comparison ofmeans atbaseline andat last follow-up, treatmentforthisandnoinformationwasavailableasto and(2)testofmeanchangefrombaselinetofinalfollow- whethersteroidtherapyisbeneficialtothesepatientsor up withineach treatmentgroup. The P-valuesfor these not,allthesepatientsweregiventheoptiontotrysteroid twosets ofcomparisons were adjustedusing therapyiftheywishedafterafullexplanationofthepros Bonferroni’smethod. ABonferroniadjusted andconsofthetherapy.Of28patientswith nonarteritic P-valueo0.05 wasconsidered statistically significant. PION, 14opted totry the steroid therapy,andit was The observed prevalenceofsevenmajor systemic started the dayofthe initial visitto theclinic or soon diseasesin nonarteritic PIONpatients waseach after.Theinitialdosewasusually80mgoralPrednisone comparedwith those expected inthe age-matched daily,except for threewhoweregivena single controlpopulation from estimatesreportedfor US intravenous mega-doseofcorticosteroids initiallyFthis Caucasianpopulation for1990–1992 bythe National wascompletelyrandom,withoutanyparticularreasonto Centerfor Health Statistics.17 Statistical significance was do so.Usually,80mgPrednisonewasgiven for2 weeks assessed usingexact binomialprobabilities. andthengradually taperedoff,with the wholesteroid therapy lastingforabout 2–21months. 2 SurgicalPION Two ofthethreehadintravenous Results mega-dose steroid therapyfor 2 dayson the daythe visual loss wasdiscovered,followed byrapidtapering Demographic characteristics within afewdays. This retrospective studyhad43patients (53 eyes):28 (35 eyes) withnonarteritic PION,12(14 eyes)with Visual statusevaluation arteriticPIONandthree(foureyes)withsurgicalPION. Nonarteritic PIONwasseenin17womenand11 men The visual outcome wasevaluated ina maskedmanner (nineright,12left,andsevenbotheyes),arteriticPIONin withoutknowledgeofthetypeofPION,treatmentgiven, 10womenandtwomen(threeright,sevenleft,andtwo orother factors influencingit. The methods used for botheyes),andsurgicalPIONin2womenandoneman evaluationofvisualacuityandvisualfieldswereexactly (oneright, oneleft, andone both eyes).Agerange was the sameasdescribed indetail elsewhere.14In addition, 20–90(median 61.5,interquartile 52–70) yearsin visual field losswasoriginally gradedsubjectively into nonarteritic PION, 62–83(median 73.4,interquartile 71– arbitrary gradesfrom 0 (normal)to 4 (severe loss)in 79)yearsinarteriticPION,and50–82(median77.3)years stepsof0.5;the findings werethencondensed for insurgicalPION.Ofthe12patientswitharteriticPION, descriptivepurposesintomild(grades0.5–1.0),moderate threehadoccult GCA18 with no systemic symptomsor (1.5–2.0),marked(2.5–3.0),andsevere(3.5–4.0)loss.The signs ofGCA. grade wasjudged byqualitatively assessingon computing clinically the amountofisopteric visual field loss andfactoring inthe functional disability produced Follow-up bythatdefect;forexample,inferiorand/orcentralvisual fielddefect,producingfarmorefunctionaldisability,was This wasevaluatedseparately forthe threetypes of assignedamuchhighergradethanacorrespondingloss PION. inthe upper fieldorelsewhere.These gradesarebest Nonarteritic PION Ofthe35eyes, threeeyeswere described by thefollowing figures: normalinFigure2c, followed less than2 weeksafterthe onset ofPION. Eye Posteriorischaemicopticneuropathy SSHayreh 1191 The follow-upfor the remaining32eyesafteronset Table2 VisualfielddefectsinarteriticandnonarteriticPION variedbetween 1.3and214(median 16.2)months. eyes Arteritic PION The follow-upfor allthe 14eyeswith Typeofvisualfielddefect Nonarteritic Arteritic arteriticPIONafteronsetvariedfrom1.3to55.6(median PION PION 11.4)months. Totaleyesa 32 13 SurgicalPIONThefoureyeswerefollowedfor4.6,9.3, Superioraltitudinaldefect 3 0 9.3,and56.6months after the onset ofPION. Inferioraltitudinaldefect 0 1 Centralscotomaalone 6 3 Superiornasalparacentral 1 2 Visualsymptoms scotoma Inferiornasalparacentral 1 0 Apartfromthe symptom ofsuddenvisual loss of scotoma variable degree, fiveeyeshadalsoexperienced Centralscotomawithother 9 2 amaurosisfugaxbeforethat(threeofthemwitharteritic fielddefects PION). Centrocecalscotomaalone 5 0 Markedgeneralized 1 2 constrictiononlywithno centralscotoma Visualacuity Nasalperipheralloss 1 0 Table1 summarizes theinitial visual acuity inthethree Temporalperipheralloss 2 0 Onlyperipheralislandfield 5 2 types ofPION. remaining Inferiornasaldefect 4 2 Superiortemporaldefect 1 1 Visualfields Superiornasaldefect 2 0 Peripheralconstrictionwith 3 0 Table2 summarizes thevarious typesof visual field centraldefect defects seeninarteritic andnonarteritic PIONon initial examination.Centralvisualloss,aloneorincombination aVisualacuitynolightperceptionortoopoortorecordvisualfieldsin threenonarteriticPIONeyesandonearteriticPIONeye. with other types ofvisual fielddefects, wasthe most common visual fielddefect seeninthe presentstudyin PION(Figures1,2a,band3).Forexample,ineyeswhere nasalperiphery intwo; inthe twoarteritic PIONeyesit thevisualfieldscouldbeplotted,27ofthe32eyes(84%) waslocatedintheinferiorperipheryinoneandinferior with nonarteritic PIONandnineof13eyes(69%) with temporalperiphery inone. arteritic PIONhada centralfield defect.However,in Asdiscussedabove,theseverityofthevisualfieldloss contrast tothis,a marked generalizedperipheral wasgraded into mild,moderate, marked,andsevere. constriction, with onlya smallnormal centralresidual Figure5gives the information innonarteritic PION. In field,waspresentinoneeyewithnonarteriticPIONand arteriticPION,thevisualfieldcouldnotbeplottedinone twoeyeswith arteritic PION(Figure4).Whenonlya eyewith light perception visiononlyand inthe peripheral island visual fieldwaspresent,in thefive remaining13eyesit wasmildin five,moderate inone, nonarteriticPIONeyes,itwasinthetemporalperiphery markedinfour,andsevereinthree.Amongthefoureyes intwo, inferiortemporal periphery inoneand inferior withsurgicalPION,intheonlyseeingeyeitwassevere. Table1 InitialvisualacuityindifferenttypesofPION Opticdisc changes Visualacuity Nonarteritic Arteritic Surgical The optic discwasinitially normal inalleyes.The disc PION PION PION developed pallorofvariabledegreeafter a variable 20/20–20/25 6 4 0 lengthoftime inallexcept seveneyes(five with 20/30–20/40 1 2 0 nonarteritic PIONandtwo arteriticPION)with a final 20/50–20/70 3 1 0 visualacuityof20/20or20/25andnocentralvisualfield 20/80 1 0 0 defect except intwowith a tinyparacentral scotoma. 20/200–20/400 5 3 0 Thiswouldindicatethatthesesevennervessufferedlittle Countingfingers 14 2 1 Handmotion 2 0 0 or nopermanent ischaemic optic nerve damage.In the Lightperception 2 1 0 eyesthatdeveloped optic discpallor,it wasusually not Nolightperception 1 1 3 possibleto determine the precise timeof itsonset becauselogisticallyitwasnotpossibletoexamineevery Totaleyes 35 14 4 patienteveryweekorso.Thetimeintervalbetweenthese Eye Posteriorischaemicopticneuropathy SSHayreh 1192 Figure1 Fourvisualfieldsshowingvaryingsizesanddensitiesofcentralscotomawithnormalperipheralvisualfieldsinnonarteritic PION.Righteyefieldsof(a)a74-year-oldwomanand(b)an87-year-oldwoman;and(c)and(d)fieldsintheleftandrighteyes, respectively,ofa52-year-oldman.Noteinallfigures:scotomawithI4eisthesolidblackareaandwithV4eisthedottedarea. patients’visitstotheclinicwasusually1monthoreven generalized pallor, it wasusually more markedinthe longerforvariousgeographicandclimaticreasons.(Our temporal partFthis agreeswith centralscotoma being TertiaryCareUniversityofIowaHospitaldrawspatients themostcommonvisualfielddefectinPION.Therewas fromanareawitharadiusofabout300–400miles.)Also, no evident difference inthe opticdisc pallor amongthe sincethe development ofpallor isa gradual andsubtle differenttypesofPION.Inthepresentseries,intwoeyes process,it is notalways possibletopinpoint itsexact with nonarteritic PION, the discofthe eyewith PION onset.As farasIcould judge,itwasusually 6–8weeks, developedanincreaseincupsizecomparedtothefellow although sometimes it wasas shortas only3–4weeks, normal eye,inaddition totemporal pallor. while inothercases it wasover 8weeks. Whetherthis depends upon the siteoflesioninthe optic nerve or Rest of the ophthalmicevaluation severity ishardtocomment. Inthe eyesofthe present study,mild-to-moderatetemporalpalloroftheopticdisc External and slit-lampevaluation ofthe anterior wasthe most common finding, andwhen a dischad segment, intraocular pressuremeasurements,and Eye Posteriorischaemicopticneuropathy SSHayreh 1193 Figure 2 Visual fields of both eyes of a 74-year-old woman with nonarteritic PION, showing marked visual improvement with high-dosesteroidtherapy.(a)and(b)arevisualfieldsofleftandrighteyes,respectively,2daysaftertheonsetofvisualloss,witha visualacuityof20/300inbotheyes.Thepatientwasstartedon80mgPrednisonedailythatday.(c)Fieldoflefteye13daysafter(a). (d)Fieldofrighteye34daysafter(b).Totaldurationofsteroidtherapywas21months.Finalvisualacuitywas20/20inthelefteyeand 2 20/40intherighteye.After6months’follow-upthevisualfieldsandacuityremainedstable. ophthalmoscopyrevealednoabnormalityinanyofthese artery disease (stenosis and/or plaqueinnine), eyes, except forage-relatedlenschanges. peripheral vasculardisease(one), vasospasticdiseases (migraine infourandRaynaud’s diseaseinone), gastrointestinal ulcer (four),smoking (in nineF34715 Systemic diseases associated with nonarteriticPION packyears) andhypercholesterolaemia(11).Magnetic Thefollowingsystemicdiseaseswereseenin28patients resonance imaging and/orcomputer tomography was with nonarteritic PION: arterial hypertension performedin16patients to rule out anyothercause of (13patients), diabetesmellitus (six), ischaemic heart visual loss.Temporalartery biopsywasperformedin disease (seven), othercardiacdiseases(including eightto ruleout GCA. valvulardisease,atrialfibrillation,patentforamenovale, Whenthe prevalence ofmajorsystemic diseasesin congestive heartfailure,andechocardiographic nonarteritic PIONpatients ofthe presentstudywas abnormalities, insix), arterial hypotension comparedwiththose expected inthe age-matched (threeForthostatic inone), thyroid disease (three), control population inthe US Caucasianpopulation, it rheumatologic diseases(threeFtwoofthem had showed a significantly higher prevalenceofarterial systemiclupus),cerebrovascularaccidents(five),carotid hypertension (P¼0.022), ischaemic heart disease Eye Posteriorischaemicopticneuropathy SSHayreh 1194 Figure3 Threevisualfieldsoflefteyeofan81-year-oldmanwithnonarteriticPION:(a)onthedayofthevisualloss,withvisual acuityonlycountfingers.Patientwasstartedon80mgPrednisonedailythatday.(b)Visualfieldsthenextdaywithvisualacuity 20/30,and(c)15dayslaterwithvisualacuity20/25.Totaldurationofsteroidtherapywas2months.Onafollow-upof19monthsthe visualfieldsandacuityremainedstable. (P¼0.026),cerebrovasculardisease (P¼0.006),carotid ESR was90.5(range 13–130)mm/h,andmedian CRP artery and peripheral vasculardisease (Po0.0001), 4.75 (range1.5–12.2)mg/dl. diabetes mellitus (P¼0.014), migraine (P¼0.039), and gastrointestinalulcers (P¼0.011)inthe nonarteritic SurgicalPION PION patients. Inthepresentstudy,therewereonlythreepatientswith surgicalPION. This isa newly emerging type ofPION, with medicolegal implications.Therefore,Ifeelthata ESR and CRP summary oftheircase reports isinstructive. InpatientswithnonarteriticPION,themedianESRwas 25(range1–89)mm/h,andCRP o0.5mg/dlinall 1. A55-year-oldwoman,witharterialhypertensionand patients except 2 (1.1mginoneand 10.6mginanother ischaemic heart disease,hadlumbarfusion spinal with myocardialinfarction,and congestiveheart surgeryinthepronepositionlastingforabout6h.At failureFinboth the temporalartery biopsywas theendoftheprocedure,shehadbilateralgrossfacial negative).In patients with arteritic PION, the median andorbital oedema. Duringsurgery shelost about Eye Posteriorischaemicopticneuropathy SSHayreh 1195 Figure 4 Visual fields of (a) right and (b) left eyes of a 79-year-old woman with arteritic PION, showing remaining, markedly constrictedcentralvisualfields,withcompletelossofperipheralfieldsinbotheyes. Figure5 Twographsshowingchangefrominitialtofinalvisualfieldsin15eyesofpatientstreatedwithsteroidtherapy(leftgraph) and15eyesofthosenottreated(rightgraph).Normal ¼ normalvisualfields. 400ml ofblood,and hersystolic bloodpressure was rapidtapering,without anybenefit. Theeyesdidnot around90mmHganddiastolicaround60mmHg,and recover anyvision ona follow-up for9 monthsand later,during recovery fromanaesthesia, diastolic developedbilateral optic atrophy. bloodpressure wasaslowas 45mmHg.When she 2. A17-year-oldmandevelopedorbitalfloorfractureon becamealert aftersurgery,shereportednolight theleftsidefollowingfootballinjury.Thelefteyehad perception inboth eyes. Bothfundi were normal. normalvisualacuityonexaminationthatdayandalso Magnetic resonance imagingofthe brainandorbits the nextdayafter recovery fromanaesthesia for showed noabnormality. Shehadintravenous surgerytorepairthefracture.However,laterthatday megadose steroid therapyfor 2days followed by hedevelopedmarkedorbitaloedema.Whenseennext Eye Posteriorischaemicopticneuropathy SSHayreh 1196 morning he hadno lightperception inthateye. The andthereforecould not haveshown anysignificant optic discand funduswerenormal, and improvement;theywereexcludedfromdataanalysison electroretinography wasnormal. Hehadintravenous visual acuity improvement. Ofthe remaining14eyes megadose steroid therapyfor 2 days followedby withaninitialvisualacuityof20/40orworse,nineeyes rapidtapering, without any benefit. The eyedidnot (64%) showed visual improvement. Among the16eyes recoverany visionon afollow-up for 16months and ofuntreatedpatients,fourshowed visual acuity developed optic atrophy. improvement(Figure6).Again,fourofthe16eyes,with 3. An 82-year-old womanhad cataract surgeryin her aninitial visual acuityof 20/20–20/25,wereexcluded right eye with retrobulbaranaesthesia.At herfirst fromdataanalysisforvisualacuityimprovement.Ofthe chancetoseewiththateyethenextday,shereported remaining12eyeswithavisualacuityof20/70orworse, onlycountfingersvision.Thevisualfieldexamination four(33%) improved. showed a large absolutecentral scotomawith a Oftheeyesthathadvisualacuityatbaseline20/40or normal peripheral field.The optic disc andfundus worse,themeanvisualacuity(logMAR)atbaselinedid wereperfectlynormal. At 5weeksafter surgery,the not differ significantly between thesteroidtreatedand optic discdeveloped pallor.The visual acuityand untreatedeyes(P¼0.165),indicatingthatthetwogroups fieldsdidnotchangeonfollow-upforalmost5years. hadsimilarvisualacuitytobeginwith.Thetreatedeyes In hermedical history,shehad arterial hypertension, showed a significant improvement frombaseline acuity hyperlipidaemia, mitral annularcalcification, (P¼0.031),withthemeanacuityatfinalfollow-upbeing transient ischaemic attacks,cortical atrophyon significantly better thanthe untreated eyes(P¼0.023). magnetic resonanceimaging,erythrocyte Thus, the findings ofthisstudy showedthat eyesof sedimentation rate27mm/h,and C-reactiveprotein patientstreatedwithhigh-dosesystemicsteroidtherapy 0.1mg/dl. showedsignificantvisualacuityimprovementcompared to untreated eyes. Effectofsteroidtherapy onvisualfunction The following criteriawereused for visual acuity Visual acuitystable or worse innonarteritic PION improvement.(a)Visualacuityimprovementmustbeat Ofthe16eyesinthetreatedgroup,fiveremainedstable least X2 lines ofSnellenchart; (b)Theremust be (20/20–20/25intwo, 20/80 inone, handmotioninone concomitant improvement in thecentral 51ofvisual andno light perception inone),and twodeteriorated fields, torule out apparentvisual acuityimprovement (from20/40to 20/400,and20/60 to20/400).Ofthe 16 fromsimple eccentric fixation orlearning experience; eyesinuntreatedgroup, nineremainedstable(20/ and(c) The visual improvementmust have been 25–20/25infour,andcount fingers infive),and two maintained onfollow-up andnot be justfluctuation or deteriorated (from20/200 tocount fingers). transient. InnonarteriticPIONThiswasevaluatedin32ofthe35 eyeswithnonarteriticPION,becausetheremainingthree Time intervalbetween theonset ofvisual loss andstart of eyesoftwopatientshadafollow-upoflessthan2weeks steroidtherapy andwereexcluded fromthis evaluation.Ofthe 32eyes, This wasasfollows. (i)In nineeyeswith improvement: 16wereofpatients treatedwith high-dosesteroid the therapywasstarted the sameday5, and2,7 and9 therapy (seeabovefor therapyregimen)andthe dayslaterin2,1,and1,respectively.(ii)Intwoeyeswith remaining16hadnotreatment.Visualimprovementwas deterioration:itwasstarted27daysaftertheonsetwhen evaluatedonfollow-upbythechangeinvisualacuity,as the patient wasfirstseen. (iii)In fiveeyeswith stable well ascentraland peripheral visual fields. visualacuityof20/20,20/25,20/80,handmotionandno light perception,it wasstarted 1, 27,7,5, and24days, Visual acuity respectively, afteronset. Figure6 givesthe initialpretreatmentand finalvisual acuity inthe treatedgroup,and initialandfinal visual acuity inthe untreated group. Time intervalbetween start ofsteroidtherapyand onsetof visual improvement Visual acuityimprovementinnonarteritic PION Itwasthedayafterthestartoftherapyinfour,andafter In the treated groupof16eyes,nine showed 7,9,13, and16days,with no datainoneeye.In the four improvementbyX2linesofSnellenchart(Figure6).Two untreatedeyeswith visual acuityimprovement,it was ofthe16eyeshadaninitialvisualacuityof20/20–20/25, seen4,6,43, and53days after theonset ofPION. Eye Posteriorischaemicopticneuropathy SSHayreh 1197 Figure6 Twographsshowingchangefrominitialtofinalvisualacuityin16eyesofpatientstreatedwithsteroidtherapy(upper graph)and16eyesofthosenottreated(lowergraph). Visualfields that, theeyesof thetreatedgroup showeda significant improvementfrom baseline(Po0.001)andit was Figure5shows the changesinthe visual fields ineyes significantly greater(P¼0.030)thaninthe untreated with nonarteritic PION(both treatedand untreated eyes. Thedataalso indicated a bettermean visual groups) during follow-up. Themean visual field grade fieldgrade at finalfollow-up for thetreatedeyes at baselinedid notdiffersignificantly betweenthe comparedtothe untreated eyes(P¼0.064).Thus, the steroidtreatedanduntreatedeyes(P¼1.000),indicating findings ofthisstudy showed thateyesofpatients thatthe twogroups weresimilarto beginwith. The treatedwith high-dosesystemicsteroid therapyshowed eyesofthe untreated groupdidnot show asignificant significant visual fieldimprovementcomparedto improvementfrom baseline(P¼0.465). Incontrast to untreated eyes. Eye