Operational criteria for the classification of chronic ... · mental state (disorientation in two of three fields, confused, an abnormal digit span, or comatose); hepatic encephalopathy

Jounal ofNeurology, Neurosurgery, and Psychiatry 1997;62:51-60

Operational criteria for the classification ofchronic alcoholics: identification of Wemicke'sencephalopathy

D Caine, G M Halliday, J J Kril, C G Harper

AbstractObjectives-To establish better opera-tional criteria for the diagnosis ofWernicke's encephalopathy. Currentcriteria for diagnosing Wernicke'sencephalopathy require the presence ofthree clinical signs (oculomotor abnor-malities, cerebellar dysfunction, and analtered mental state), although it hasoften been reported that most patients donot filfil all these criteria.Methods-The clinical histories of 28alcoholics with neurological and neu-ropsychological assessments and defini-tive neuropathological diagnoses wereexamined to determine clinical signs foruse in a screening schedule. Operationalcriteria were then proposed for differenti-ating patients with Wernicke'sencephalopathy alone or in combinationwith Korsakoff's psychosis or hepaticencephalopathy. The new criteria forWernicke's encephalopathy require two ofthe following four signs; (1) dietary defi-ciencies, (2) oculomotor abnormalities,(3) cerebellar dysfunction, and (4) eitheran altered mental state or mild memoryimpairment. Reproducibility and validitytesting of these criteria were performedon 106 alcoholics screened from a largenecropsy sample.Results-Despite rater variability withregard to specific symptoms, within andbetween rater reliability for diagnosticclassification using the criteria retrospec-tively on patient records was 100% forthree independent raters. Validity testingshowed that Wernicke's encephalopathywas underrecognised only when occurringwith hepatic encephalopathy (50% sensi-tivity).Conclusions-By contrast with currentcriteria, the proposed operational criteriashow that the antemortem identificationof Wernicke's encephalopathy can beachieved with a high degree of specificity.

( Neurol Neurosurg Psychiatry 1997;62:51-60)

Keywords: alcoholism; diagnostic criteria; neurologicalimpairnent; Wernicke-Korsakoff syndrome

It is well known that the pathology ofWernicke's encephalopathy may not be associ-ated with the full clinical triad and that in theacute phase clinical symptoms may resolvewith thiamine treatment.'-" These findings

suggest that there can be a complete resolutionof the underlying brain abnormality inWernicke's encephalopathy, similar to the res-olution of neuropathology in hepaticencephalopathy. However, unlike hepaticencephalopathy, in which the clinical signs area precursor of the pathology, many patientswith the neuropathology of Wernicke'sencephalopathy do not have recorded signs ofthe classic triad.' 2 This raises the question ofwhether the clinical correlates of the patholog-ical lesions are being missed during life orwhether the pathology represents clinicallysilent lesions. A similar situation is seen inAlzheimer's disease, in which many non-demented aged patients show a similar typeand distribution of neuropathology to theirdemented counterparts. 12 Recently researchinto the pathophysiology of aging andAlzheimer's disease has been successfullyadvanced by the use of operational criteriawhich can be applied across clinics and labora-tories.'3 14 The aim of the present study is topropose operational criteria for the diagnosisof Wernicke's encephalopathy.Most patients world wide with Wernicke's

encephalopathy are alcoholics,'5 with 30%-80% of chronic alcoholics having the clinicalor biochemical signs of thiamine deficiency.'6Two extensive clinicopathological studies' 2found that the incidence of oculomotor abnor-malities was low and therefore reliance on theclassic triad would result in significant under-recognition of these patients during life.Because an altered mental state is also a com-mon neurological sign associated with hepaticencephalopathy,'7 the overlap between the twodiseases poses substantial and often unrecog-nised difficulties for diagnosis. Operational cri-teria for recognising and separating thesecommon conditions are needed. Therefore theaim of the present study was to devise and testsuch operational criteria for classifying chronicalcoholics for both treatment and research. Wehave assessed only alcoholics in order toanalyse a large and relatively homogeneousgroup of patients with Wemicke's encephal-opathy. The criteria account for the diagnosesof other common medical complications asso-ciated with this patient population.4 Key signsfor operational criteria for diagnosis are pro-posed and the reliability and predictive validity(specificity and sensitivity) of using these crite-ria to classify chronic alcoholics tested.

MethodsTo determine characteristic signs and

Neuropsychology Unit,Royal Prince AlfredHospital,Camperdown, 2050,AustraliaD CainePrince ofWalesMedical ResearchInstitute, High Street,Randwick, 2031,AustraliaGM HallidayNeuropathologyDivision, DepartmentsofPathology,University of Sydney,2006 and Royal PrinceAlfred Hospital,Camperdown, 2050,AustraliaGM HallidayJ J KrilC G HarperCorrespondence to:Dr G Halliday, Prince ofWales Medical ResearchInstitute, Villa 2, HighStreet, Randwick. 2031,Australia.Received 25 March 1996and in final revised form30 August 1996Accepted 2 September 1996

51

on January 17, 2020 by guest. Protected by copyright.

http://jnnp.bmj.com

/J N

eurol Neurosurg P

sychiatry: first published as 10.1136/jnnp.62.1.51 on 1 January 1997. Dow

nloaded from

http://jnnp.bmj.com/

Caine, Halliday, Kril, Harper

symptoms for alcoholics with and withoutWernicke's encephalopathy, the histories of 28patients, well known to the neuropsychologyunit, Royal Prince Alfred Hospital, who haddied between 1989 and 1993 and hadnecropsy confirmation of their clinical diag-noses were examined. These had a variety ofneurological presentations, including eightpatients with Wernicke's encephalopathy(classic triad but with no evidence of a perma-nent amnesic syndrome), six patients with theWernicke-Korsakoff syndrome (alcoholicswith Wernicke's encephalopathy fulfilling cur-rent criteria for Korsakoffs psychosis'8 20)),nine patients with hepatic encephalopathy(neurological abnormalities associated withdecompensated liver disease'7), and eightpatients with alcohol related brain dysfunction(neurological or neuropsychological impair-ment, although not fulfilling the criteria forWernicke's encephalopathy, Wernicke-Korsakoff syndrome, or hepatic encephalopa-thy). Three patients had both Wernicke'sencephalopathy and hepatic encephalopathy.In all patients the clinical diagnosis matchedthe neuropathological diagnosis, patients withalcohol related brain dysfunction having noneuropathological abnormality.

STANDARDISED CLINICAL EXAMINATIONPatients in the neuropsychology unit under-went a 90 minute clinical assessment by a con-sultant neurologist or neurology registrar,comprising a detailed history following a stan-dard protocol, a complete general medicalexamination, and a detailed neurologicalexamination including an examination of men-tal status. They also underwent clinical neu-ropsychological assessment which included

Table 1 Schedule of clinical signs and symptoms with domains highlighted (see textfordefinitions)

Previous historyor at first Developedpresentation subsequently At any time

Wernick's encephalopathyWE diagnosed clinically

Dietary deficiencyUndernutritionVitamin deficiency

Eye signsOphthalmoplegiaNystagmusGaze palsy

Gerebellar signsAtaxiaUnsteadyCerebellar dysfunction

Seresi=mFrontal lobe ysfunction

Planning deficitAbstraction deficitMental rigidityBehavioural disturbance

AmneSiaMild memory impairmentMild-moderate memory problemsConfabulating

Altered mental stateDisorientedComatoseConfusedDigit span abnormal

Hepatic encephalopathyDiagnosed clinically

Liver diseaseJaundiceHaematemesisMelaenaAscitesAsterixis

standardised administration and scoring of thefollowing neuropsychological tests: Wechslermemory scale, Rey auditory verbal learningtest, Rey-Osterrieth complex figure test, trailmaking test, and Stroop test.2' The diagnosisof Wernicke's encephalopathy, Wernicke-Korsakoff syndrome, hepatic encephalopathy,or alcohol related brain dysfunction was madeusing the history obtained, information fromreferring agencies, the results of the clinicaland neuropsychological examinations, andthose of any additional tests which may havebeen ordered.

PATHOLOGICAL DIAGNOSISA full necropsy was performed for all patientsincluding macroscopic and microscopicalexamination of all organs. A detailed neu-ropathological examination was performed.The external brain features were examinedbefore and after fixation in neutral bufferedformalin for 14 days. The brain was thenembedded in 3% agar and sectioned at 3 mmintervals in the coronal plane with a rotaryslicer. The cut surfaces of each slice wereexamined for neuropathological abnormalities.Standardised blocks of the superior frontal,superior parietal, inferior temporal, andparahippocampal cortices, hippocampus,amygdala, anterior and posterior basal gangliaincluding the basal forebrain, thalamus,mamillary bodies of the hypothalamus, mid-brain, pons, medulla oblongata, and cerebel-lum (vermis, lateral cortices and dentatenuclei) were sampled for paraffin embedding.Sections were cut at 10 mm and stained withhaematoxylin and eosin, luxol fast blue, andsilver stains before microscopical examinationand diagnosis (by CGH). Other specialisedstains were performed if required to excludecerebral infarction, head injury, or degenera-tive conditions such as Alzheimer's disease.Serious liver disease was diagnosed if liver cir-rhosis or hepatitis was confirmed at necropsy.Hepatic encephalopathy was diagnosed if thepatient had serious liver disease and Alzheimertype II astrocytes in the brain (common in thebasal ganglia, cerebral cortex, or pons).Wernicke's encephalopathy was diagnosedwhen mamillary body and periventricularlesions were evident.2223 In acute Wernicke'sencephalopathy the pathology was largely con-fined to abnormalities of blood vessels withpetechial haemorrhages commonly occurring.In chronic Wernicke's encephalopathy neu-ronal loss and gliosis were evident in themamillary bodies. A diagnosis of normal brainwas made if no neuropathological abnormalitywas detected.

EVALUATION OF CLINICAL SIGNS AND

SYMPTOMSA range of clinical signs and symptoms repre-senting deficits in the major clinical domainsknown to be affected in alcoholics4 was evalu-ated for establishing operational criteria.Related clinical symptoms and alternativeterms used to describe a sign or symptom weregrouped together under each of these domainsand a standardised schedule of the clinical

52


http://jnnp.bmj.com

/J N

eurol Neurosurg P


nloaded from


Operational criteria for the classification of chronic alcoholics: identification of Wernicke's encephalopathy

signs designed (table 1). For example, cerebel-lar damage resulting in ataxia is commonlypresent in alcoholics, but in the absence of for-mal testing it is often recorded as an unsteadi-ness. Therefore both ataxia and unsteadinesswere used as indicators of cerebellar dysfunc-tion. The presence of either of these signs wassufficient to result in a positive score for cere-bellar abnormalities.

Evaluation of the clinical profiles of the 28patients disclosed 11 commonly encounteredclinical domains. These were Wemicke'sencephalopathy (diagnosed during life usingthe classic triad), dietary deficiencies (a bodymass index lower than 2 SD below normal asevidence of undernutrition, a history of grosslyimpaired dietary intake, or an abnormal thi-amine status), eye signs (oculomotor abnor-malities such as ophthalmoplegia, nystagmus,or gaze palsy), cerebellar signs (see above, orabnormalities of past pointing, dysdiadokoki-nesia, or impaired heel-shin testing), seizures(either as part of a withdrawal syndrome or inisolation, or a longstanding history of anticon-vulsant medication), frontal lobe dysfunction(abnormalities in planning, insight, or abstrac-tion with formal neuropsychological testing orwhen neurological examination elicited thesecharacteristics), amnesia (a stable and persist-ing inability to form new memories), mildmemory impairment (failure to remember twoor more words in the four item memory test,24or impairment on more elaborate neuropsy-chological tests of memory function), alteredmental state (disorientation in two of threefields, confused, an abnormal digit span, orcomatose); hepatic encephalopathy (grades 3or 4, in which signs of impaired mental staterange from confusion to coma17), and liver dis-ease (signs of systemic abnormalities sec-ondary to chronic liver disease includingmelaena, haematemesis, jaundice, ascites, orasterixis). Abnormal liver function tests in iso-lation were not considered sufficient to have apositive score for liver disease. In addition,poor performance on memory tests on onlyone occasion was only considered sufficient fora positive score for amnesia if the patient wasunable to form any new memories.

ESTABLISHING OPERATIONAL CRITERIARetrospective patient note analysis was con-

Table 2 Operational criteria for the clinical diagnosis of chronic alcoholics. See textfordefinition of clinical signs. Multiple diagnoses are possible

Clinical diagnoses

No clinical Liver diseaseClinical domains signs only ARBD HE WE WKSWE Absent Absent Absent Absent Optional OptionalDietary deficiency Absent Absent t tEye signs Absent Absent * 1tCerebellar signs Absent Absent * t tSeizures Absent Absent *Frontal lobe

dysfunction Absent Absent *Amnesia Absent Absent * Absent Absent PresentMild memory

impairment Absent Absent * t- t-Altered mental state Absent Absent Absent Present f .HE Absent Absent Absent Optional Absent AbsentLiver disease Absent Present Present

*Any one of these signs; tany two of these signs, but not both of those marked.ARBD = alcohol related brain dysfunction; HE = hepatic encephalopathy; WE = Wernicke'sencephalopathy; WKS = Wernicke-Korsakoff syndrome.

ducted blind to the neuropathological diagnosisusing the standardised schedule to record thepresence or absence of clinical features at twotime points: symptoms present at or before ini-tial neurological examination and symptomsoccurring after initial presentation. The clini-cal domains were then categorised accordingto pathological diagnosis. Operational criteriafor different diagnoses were devised on thebasis of this information (table 2).

VALIDATING THE OPERATIONAL CRITERIATest populationAll adult patients necropsied between 1986and 1994 from the Royal Prince AlfredHospital, as well as 10% of necropsiedpatients from the New South Wales Instituteof Forensic Medicine (referred for neu-ropathological examination), constituted thebase from which the test samples were drawn(4600 patients). Of these 303 had a history ofalcoholism (7% of sample) based on informa-tion from telephone interviews to general prac-titioners, written questionnaires to relatives ofpatients, and clinical details from hospitalrecords. Alcoholism was established if greaterthan 80 g ethanol was consumed each day formost of their adult life (usually > 30 years) asrecorded from these multiple sources.Informant histories have been shown to pro-vide accurate assessment of alcohol intake.25Exclusion criteria were an average daily alco-hol consumption of less than 80 g ethanol,inadequate clinical history (no neurologicalexamination or single admission only), or neu-ropathological evidence of a cerebral infarc-tion, head injury or neurodegenerativecondition (for example, Alzheimer's disease).A total of 197 of the 303 alcoholics selectedwere excluded, leaving 106 alcoholics for thestudy (35% of alcoholics, 2-3% of the totalsample, 16 women and 90 men, including the28 patients used for establishing the opera-tional criteria). The age range was 33 to 79years (mean 58 (SD 10)).

INCLUSION CRITERIA AND DEFINITION OFPATIENTSInclusion criteria were consumption of greaterthan 80 g ethanol per day for most of theiradult life, neurological assessment, clinicalassessment at more than one time point, fullnecropsy, interview with general practitioneror relatives, and an extensive and systematicneuropathological examination that excludedconditions other than those directly attribut-able to chronic alcohol consumption.

TESTING THE OPERATIONAL CRITERIAThe standardised schedule and operationalcriteria were applied to the longitudinalpatient records of all 106 patients for clinicalpatient classification. About 30% of eachpatient type were randomly selected for analy-ses of between and within rater reliability (32patients in total). These patients were classi-fied independently by three researchers blindto neuropathological diagnosis (DC, GMH,JJK). Symptoms and signs occurring at anystage were used for classification. The predic-

53


http://jnnp.bmj.com

/J N

eurol Neurosurg P


nloaded from



tive validity of the clinical classification againstthe criterion of clinicopathological diagnosis isexpressed in terms of sensitivity and speci-ficity,26 where sensitivity = (true positives/(true positives + false negatives)) x 100% andspecificity = (true negatives/(false positives +true negatives)) x 100%. The microscopicalpathology for all patients with either a falseclinical but positive pathological diagnosis or atrue clinical but negative pathological diagno-sis (17% of all patients) was carefullyrechecked blind to the clinical classification(by CGH). If discrepencies between diagnosesstill existed (10 patients), the neuropathologywas rechecked by two independent researchers(GMH, JJK) and neuropathological diagnosisdetermined by consensus agreement of allthree researchers. The original neuropatholog-ical diagnosis was confirmed in all of thesepatients.

ResultsDIAGNOSTIC GROUPSApplication of the operational criteria to the106 patients confirmed the major diagnosesidentified in the 28 detailed patients and estab-lished a pattern for six different diagnosticgroups. The clinical characteristics and pro-gression of symptoms for each group follow.

Alcoholics with compensated liver diseaseonly had clinical evidence of compensated liverdisease in concert with the pathology for seriousliver disease and a normal brain. patients wereexcluded if neurological impairment or dietarydeficiency was evident. These patients had oneor more of the symptoms of liver disease, butdid not progress to develop all symptoms orhepatic coma.

Alcoholics with hepatic encephalopathy hadclinical evidence of decompensated liver disease(altered mental state) in concert with thepathology for serious liver disease and hepaticencephalopathy. Patients were excluded ifanother encephalopathy was diagnosed clini-cally or if dietary deficiencies, eye signs, cere-bellar signs, mild memory impairment, oramnesia were evident, unless the patient hadalso been given a definitive clinical diagnosis ofhepatic encephalopathy.The presence of cirrhosis of the liver was

often not apparent until the patient presentedwith symptoms of decompensated liver disease.Therefore these patients represent the moresevere end of the range of patients with liverdisease. The most frequent symptoms includedthe combination of jaundice, ascites, melaena,and haematemesis, whereas asterixis was lessfrequent (patients 1 and 2, appendix). Thelongest interval between initial presentationwith liver disease and death was 15 months(patient 2) and was independent of whether thepatient ceased drinking during this time.Therefore, although these patients may havehad cirrhosis of the liver for an extended period,mortality seemed inevitable after the develop-ment of hepatic encephalopathy and was mostcommonly due to sepsis, gastrointestinal haem-orrhage, or raised intracranial pressure.

Alcoholics with Wernicke's encephalopathy

alone had a definitive clinical diagnosis ofWernicke's encephalopathy (classic triad) orclinical evidence of at least two of the four fol-lowing signs; (1) dietary deficiencies, (2) eyesigns, (3) cerebellar signs, and (4) either analtered mental state or mild memory impair-ment; and Wernicke's encephalopathy atnecropsy. Amnesic patients and patients witha diagnosis of hepatic encephalopathy wereexcluded from this group. All patients had aprogressive and usually short clinical course(average of three years between first presenta-tion and death). At least two of the four signs ofWernicke's encephalopathy were present at, orsoon after, first presentation (patients 3 and 4,appendix). In many patients only a single signof the Wernicke's encephalopathy triad wasnoted initially, whereas at subsequent presen-tations there were usually additional neurolog-ical signs. Most were still drinking at the time ofdeath. It should be noted that clinically, manyof these patients did not fulfill current diag-nostic criteria using the full clinical triad (eyesigns, cerebellar signs, and an altered mentalstate). We found that patients with Wernicke'sencephalopathy pathology had a minimum oftwo clinical signs which may have includeddietary deficiencies. Thus some patients wereincluded with only one neurological abnormal-ity in the presence of dietary deficiencies. Thiscriterion differs substantially from the classicdefinition.

Alcoholics with the Wernicke-Korsakoffsyndrome fulfilled the above criteria for thediagnosis of Wernicke's encephalopathy andhad documented evidence of amnesia and dis-orientation in the absence of an acute confu-sional state (patients 5 and 6, appendix).These patients were often institutionalised dueto their profound memory dysfunction andtherefore tended to have a longer history ofneurological signs and symptoms (> 10 years)documented either in patient notes or inferredfrom nursing home institutionalisation. Initialpresentation was usually associated with theeffects of intoxication or withdrawal (includingseizures). There was progressive deteriorationin both the range of neurological symptomsand the severity of cognitive symptoms(patient 6). In particular, stable amnesia wasusually noted after detoxification at either theinitial or second presentation (within oneyear). At this time institutionalisation usuallyoccurred and was associated with a reductionor cessation of drinking and improved nutri-tional status. This is likely to contribute to therelatively long survival time of these patients.

Alcoholics with alcohol related brain dys-function had clinical evidence of neurologicalimpairment in at least one domain (cerebellarsigns, seizures, frontal lobe dysfunction, mem-ory impairment, altered mental state), but didnot reach clinical criteria for any of the diag-nostic groups above (patients 7 and 8, appen-dix). These patients had no significantneuropathology and usually presented with thefull range of withdrawal symptoms (includingseizures, patient 7) or with complex cognitivedeficits (dissimilar to the Wernicke-Korsakoffsyndrome in that memory impairment was

54


http://jnnp.bmj.com

/J N

eurol Neurosurg P


nloaded from



either mild (patient 8), or, if amnesic, associ-ated with a longstanding history of seizures).Neither eye signs nor dietary deficiencies wereever seen in these patients (patients 7 and 8).Most patients had longstanding histories ofseizures (most had epilepsy diagnosed in earlyadulthood) or neuropsychiatric abnormalitieswith significant frontal lobe features (mostlycoincident with their history of alcoholism).

Alcoholics without neurological complica-tions were defined by exclusion criteria.Patients were excluded if they fulfilled any ofthe criteria above or if dietary deficiencies,liver disease, or neurological, psychological, orneuropathological abnormalities were evident.These alcoholic patients usually presentedwith intoxication, unsteadiness, blackouts, orwithdrawal symptoms but without any perma-nent sequelae; in particular, no neurologicalsigns or cognitive dysfunction were evident(patient 9, appendix).

Alcoholics with multiple diagnoses werealso found. Some patients with neurologicalimpairment also had evidence of liver disease.Compensated liver disease did not contributeto the neurological signs in any patient.However, significant overlap in clinical signswas evident in patients with hepaticencephalopathy and Wernicke's encephalopa-thy. In all such patients, the course of hepaticencephalopathy followed that described abovefor hepatic encephalopathy alone, although itwas somewhat more prolonged. In patientswith Wemicke's encephalopathy and hepaticencephalopathy diagnosed during life, the pre-sentation of Wemicke's encephalopathy andhepatic encephalopathy occurred concurrently(patient 10, appendix), possibly indicating amore classic presentation of the Wernicke'sencephalopathy triad in these patients. Onepatient had a history of the Wernicke'sencephalopathy triad and Korsakoffs psy-chosis which preceded the complication ofhepatic encephalopathy (patient 11, appen-dix).

RELIABILITY AND PREDICTIVE VALIDITY OFOPERATIONAL CRITERIAIn the 30% of patients randomly chosen forreanalysis of patient classification, rater retest-ing gave the same diagnosis in all patients(100% accuracy) even with test-retest intervalsof up to four months. Further, despitebetween rater agreement for individual signsand symptoms on the standardised schedule ofonly 83%, it was 100% for diagnostic classifi-cation in all instances. This indicates thatgrouping symptoms and signs into broaddomains for patient classification is reliableand reproducible with regard to overall diag-nosis while allowing for interrater differenceswith regard to specific signs and symptoms.

Table 3 shows the sensitivity of the opera-tional criteria and the percentage of patientswith different clinical signs in each alcoholicgroup. Only the diagnosis of Wernicke'sencephalopathy in combination with amnesiaor hepatic encephalopathy proved problematicfor retrospective classification. Out of 40patients with Wemicke's encephalopathy

pathology, six were not clinically diagnosedusing the operational criteria, whereas onlynine patients fulfilled the classic clinical triad(22 5%). Using the operational criteria, twopatients with Korsakoffs psychosis and a his-tory of alcoholism were institutionalised for anextended time and had no clinical record ofthe signs of Wernicke's encephalopathy. Theremaining patients had additional hepaticencephalopathy clinically and pathologically,and the clinical signs of Wernicke'sencephalopathy may have been less obvious.

Altered mental state was found at somestage in all encephalopathic patients and mildmemory impairment was seen in most groups,but particularly those classified with alcoholrelated brain dysfunction or Wemicke'sencephalopathy (either alone or in combina-tion; table 3). Seizures were most prevalent inpatients with alcohol related brain dysfunc-tion, although the Wernicke-Korsakoff syn-drome group also had an above averagenumber of such patients (table 3). Of particularimportance is the finding that cerebellar signsconcentrated in all patients classified withWernicke's encephalopathy, although patientswith the Wernicke-Korsakoff syndrome hadeye signs more often than those withWemicke's encephalopathy or Wemicke'sencephalopathy and hepatic encephalopathy(table 3). Conversely, there was a high preva-lence of dietary deficiencies in patients withWemicke's encephalopathy and Wernicke'sencephalopathy plus hepatic encephalopathycompared with the Wemicke-Korsakoff syn-drome. This is most likely due to the institu-tionalisation of many in this group. Othersigns concentrated in particular alcoholicgroups because of the inclusion/exclusion cri-teria. Dietary deficiencies and eye signsoccurred only in partients with Wernicke'sencephalopathy (with or without amnesia orhepatic encephalopathy), amnesia was mostcommon in patients with the Wernicke-Korsakoff syndrome, and liver disease andhepatic encephalopathy were common inpatients with these pathological diagnoses(table 3).

Table 3 also shows the specificity of theoperational criteria and the percentage of over-lapping clinical signs in each alcoholic group.The false positive rate for clinical classificationcan be obtained by subtracting the specificityvalue from 100. Interestingly, only in patientsclassified with hepatic encephalopathy or liverdisease could the diagnosis not be substanti-ated on pathological grounds. Of the 18patients classified with hepatic encephalopathyusing the operational criteria, no Alzheimertype II astrocytes could be found in the brainsof two patients. Both patients had serious liverpathology, and examination of the clinicalrecords showed evidence of decompensatedliver disease. Yet despite this, hepaticencephalopathy could not be confirmed neu-ropathologically, although death occurredwithout longstanding coma in both patients(within three hours of coma onset). Of the 11patients fulfilling the operational criteria forcompensated liver disease only, four had no

55


http://jnnp.bmj.com

/J N

eurol Neurosurg P


nloaded from



Table 3 Specificity and sensitivity of the clinical diagnosis based on pathological criteria and the proportion ofpatients with different clinical signs.Clinical signs which concentrate in particular alcoholic groups are in bold

Clinical diagnoses

No clinical Liver diseaseAll 106 signs only ARBD HE WE WKS WE+HEpatients (n = 22) (n = 11) (n = 15) (n = 18) (n = 16) (n = 16) (n = 8)

Sensitivity 94 100 100 100 100 100 88 50Specificity 99 100 96 100 98 100 100 100Mean number of signs (/11) 2 0 1 2 3 3 *5 *5Clinical domains:WE 8 0 0 0 0 *31 13 *25Dietary deficiency 9 0 0 0 0 *25 13 *38Eye signs 19 0 0 0 0 31 *56 25Cerebellar signs 36 0 0 36 17 *81 *81 50Epilepsy 1 6 0 0 *50 11 19 *32 0Frontal lobe dysfunction 14 0 0 14 0 *31 *44 13Amnesia 17 0 0 14 0 0 *100 0Mild memory impairment 29 0 0 50 6 *63 50 50Altered mental state 29 0 0 0 56 31 56 *63HE 22 0 0 0 *89 0 0 *88Liver disease 47 0 *100 21 *100 44 31 75

For abbreviations see table 2. * > Twice the proportion seen in all 106 patients. Values apart from mean number are %.

serious liver pathology. Seven of 16 patientsclassified with Wernicke's encephalopathy alsofulfilled the criteria for liver disease, althoughthree of these patients had no serious liverpathology. All seven patients with a false posi-tive diagnosis for liver disease using the opera-tional criteria had clinical symptoms and signsthat could be attributed to concomitant gas-trointestinal disease (for example, gastriculcer) producing melaena or haematemesis.

DiscussionThe operational criteria reported here can beused to classify accurately alcoholics with neu-rological deficits. The criteria were designed todifferentiate alcoholics with and without theWemicke-Korsakoff syndrome, and to differ-entiate Wemicke's encephalopathy fromhepatic encephalopathy as these representlongstanding diagnostic difficulties in alcoholresearch.17 28 31 Other medical complicationsoften encountered in alcoholics, but which donot result in CNS deficits, were not incorpo-rated. By grouping together signs and symp-toms into clinical domains and redefiningcurrent diagnostic criteria for Wernicke'sencephalopathy, the overall accuracy of theproposed criteria is high and the need for pointto point agreement on individual signs is min-imised.The existing clinical triad for Wernicke's

encephalopathy was modified so that the diag-nosis only required any two of the followingsigns; eye signs, cerebellar dysfunction, alteredmental state, and dietary deficiencies. Thismodification produced highly specific andnon-overlapping diagnoses for Wemicke'sencephalopathy, the Wemicke-Korsakoff syn-drome, hepatic encephalopathy, andWernicke's encephalopathy plus hepaticencephalopathy (98%-100% specificity).Furthermore, sensitivity for the diagnosis ofWemicke's encephalopathy was improvedfrom 31% using the classic triad to 100% withthe proposed criteria. The use of the proposedoperational criteria for the classification ofneurological impairment in chronic alcoholicswill allow Wemicke's encephalopathy to bediagnosed with a high degree of confidence.

The validation of the criteria relies on ade-quate testing procedures. The test populationstudied was derived from both hospital andforensic sources, representing a broader crosssection of patients than those from a singlesource. The use of forensic material minimisedthe bias of patient selection towards those withovert neurological complications, as all suddenor accidental deaths are subject to forensicnecropsy in Australia. In addition, patientswith a wide variety of neurological diseaseswere initially selected for neuropathologicalevaluation and from these, alcoholics werechosen for study. Pathological diagnoses wererevalidated for all patients that were clinicallymismatched. Because of the retrospectiveanalysis and inclusion criteria, a proportion ofpatients at the lower end of hazardous alcoholconsumption may not have been included inour alcoholic population. However, thesepatients are least likely to have medical com-plications due to their alcoholism and mayonly have neurological deficits disclosed bydetailed neuropsychological assessment.32 Weare therefore confident of the validity of oursample and test populations and that ourmethodology has not produced substantialbias towards alcoholics with neurologicalimpairment.

DIFFERENTIATION OF DIAGNOSTIC GROUPSUsing the operational criteria, we have beenable to clinically define alcoholics withWemicke's encephalopathy neuropathology.Only two institutionalised (> 15 years) alco-holics with Korsakoffs psychosis hadWemicke's encephalopathy neuropathologywithout enough clinical evidence of the disease(false positive rate of 6% excluding patientswith hepatic encephalopathy ). This substan-tially reduces the current rate (> 80%) of neg-ative clinical diagnosis,' 2 and the use of theoperational criteria prospectively is likely toeliminate this error altogether. The high rateof negative clinical diagnosis of Wernicke'sencephalopathy in patients with positive neu-ropathology in past studies strongly suggeststhat most patients with clinical Wemicke'sencephalopathy go undetected during life,rather than the alternative explanation of

56


http://jnnp.bmj.com

/J N

eurol Neurosurg P


nloaded from



patients with a clinical episode of Wernicke'sencephalopathy having resolving neuropathol-ogy.The question of whether Wernicke's

encephalopathy is both clinically and neu-ropathologically resolvable can be assessed inthose alcoholics with neurological symptomsbut no neuropathology. In the present study48 out of the 106 patients examined had noneuropathological abnormality. Most of these(69%) also had no signs of Wernicke'sencephalopathy in their medical records orfrom informant interviews with general practi-tioners or family members. Of the remainingpatients (alcohol rated brain dysfunction)most were epileptic from young adulthood orhad longstanding neuropsychiatric syndromes.Although it cannot be discounted that some ofthese neurologically impaired patients mayhave had a past episode of clinical Wernicke'sencephalopathy without any permanent braindamage (true negatives), none had any evi-dence or history of dietary deficiency, eyesigns, or an altered mental state, suggestingthat overt encephalopathy is unlikely to havebeen missed. Mild memory impairment wascommon in these patients and cerebellar dys-function was seen in 36%, although coexistentepilepsy and longstanding treatment with anti-convulsants may account for these symp-toms.33 Therefore, it seems unlikely that asignificant number of alcoholics have hadWernicke's encephalopathy clinically withoutevidence of permanent brain damage.The finding that Wernicke's encephalopa-

thy and hepatic encephalopathy occur togetherin some alcoholics has important implicationsfor the acute management and treatment ofpatients with suspected hepatic encephalopa-thy. It is well established that the most com-mon and often isolated sign of Wernicke'sencephalopathy is a disturbed mental state' 2and that some patients present in coma34 35making differential diagnosis difficult. Theprospective use of the operational criteria out-lined should improve diagnostic classificationof patients with these life threateningencephalopathies. In particular, previous cere-bellar signs or mild memory impairment aremore common in Wernicke's encephalopathythan hepatic encephalopathy, with dietarydeficiencies and eye signs distinguishingpatients with Wernicke's encephalopathy. Inaddition, alcoholics with both ence-phalopathies were notable for having physicalsigns in many domains. Our results indicatethat in patients suspected of having alcoholichepatic encephalopathy in the presence of orwith a history of neurological abnormalities, astrong suspicion of Wernicke's encephalopa-thy should be entertained and appropriatetreatment measures taken.

Using the operational criteria descibed,alcoholics with Wernicke's encephalopathycan be distinguished from those with theWernicke-Korsakoff syndrome with a highdegree of sensitivity and specificity. In the pre-sent study, two factors distinguished betweenthe diseases: an amnesic syndrome with (1) aclear sensorium and (2) which does not remit

with treatment. That is, alcoholics with theWernicke-Korsakoff syndrome can be differ-entiated from alcoholics with Wernicke'sencephalopathy but without Korsakoffs psy-chosis by the severity and stability of memoryloss regardless of other cognitive deficits. It isimportant to distinguish permanent memoryloss in alcoholics who are sober and thiaminereplete from memory loss in intoxicated, thi-amine deficient alcoholics. One important testof this distinction may be memory recoverywith thiamine supplementation. In agreementwith our definition, past studies define theWernicke-Korsakoff syndrome clinically as anamnesic syndrome in the presence of pre-served intelligence and learned behav-iour. 18-203637 In particular, alcoholics withWernicke-Korsakoff syndrome have a consis-tent loss of anterograde episodic memory.'8 36 38whereas many past studies have not empha-sised the underlying signs of Wernicke'sencephalopathy in patients with the Wernicke-Korsakoff syndrome, most support the sugges-tion that the Wernicke-Korsakoff syndromeis the inevitable outcome of Wernicke'sencephalopathy. 18 29 39 Interestingly, in our testpopulation this occurred only in the propor-tion of alcoholics whose survival time wasextended by institutional intervention, as pre-viously suggested.40 The inclusion ofWernicke's encephalopathy as a prerequisite inour criteria for the Wernicke-Korsakoff syn-drome was useful in excluding patients withmemory deficits from other causes and did notsignificantly decrease diagnostic sensitivity.However, the use of Wernicke's encephalopa-thy as an essential inclusion criterion forWernicke-Korsakoff syndrome was a source oferror for retrospective analysis in a small pro-portion of patients.We were unable to identify a subset of alco-

holic patients with a dementia-like processwhen degenerative neuropathology and headinjury were excluded. This supports otherpathological studies which have not been ableto define pathological characteristics for alco-holic dementia.284' The absence of alcoholicdementia in the present sample suggests thatcoincident neurodegenerative conditions con-tribute to this clinical profile. Alcoholicdementia has been used to describe those whohave a gradual cognitive decline,29 whereas theterm Korsakoffs psychosis has been used forpatients whose onset of memory loss is rapidlyacquired. The Wernicke-Korsakoff syndromeis generally diagnosed in younger patients thanthose with alcoholic dementia.29 42 43 Afterexcluding degenerative conditions, the alco-holics studied had a mean age (SD) of 58 (10)years and none had gradual cognitive decline.Careful clinicopathological studies have yet tosubstantiate the existence and aetiology of thisproposed condition.

IMPLICATIONS FOR PATHOGENESISThe association between the reversibleencephalopathic symptoms of Wernicke'sencephalopathy and permanent brain damageis not well understood. The neuropathologyunderlying Wernicke's encephalopathy can

57


http://jnnp.bmj.com

/J N

eurol Neurosurg P


nloaded from



vary in type (acute v chronic) and extent,23 andsome of these differences may account for thereversible nature of some symptoms and signs,in particular the eye signs. Lesions in brain-stem periventricular regions are likely toaccount for the eye signs with the type andextent of damage determining their clinicalresolution."' By contrast, cerebellar dysfunc-tion is seen as a permanent neurological signin alcoholics.' 4'10 The coexistence of cerebellarabnormalities and poor nutritional state inmany patients with Wernicke's encephalopa-thy supports the suggestion that thiamine defi-ciency is important in the pathogenesis ofcerebellar damage.3 44 Although often termedalcoholic cerebellar degeneration, a direct neu-rotoxic effect of alcohol has not been shown.Furthermore, malnourished non-alcoholicshave also been reported to have cerebellardegeneration.44 In addition, a proportion ofalcoholics with a history of seizures have per-manent cerebellar degeneration which may bedue to either repeated hypoxic episodescaused by seizures or to longstanding medica-tion with anticonvulsants.33

Patients with Wernicke's encephalopathyhad a progressive and relentless course with amaximum survival time of three years. By con-trast, patients with the Wemicke-Korsakoffsyndrome had considerably longer diseasedurations due to substantial institutionalisedcare. The importance of rehabilitation in thesuccessful long term management of patientswith the Wernicke-Korsakoff syndrome hasbeen noted previously.4046 The increased sur-vival time of institutionalised patients with theWemicke-Korsakoff syndrome suggests thatthe brain damage underlying the disorder maybe related to cessation of alcohol or correctionof thiamine deficiency. Interestingly, in somepatients with Wemicke's encephalopathyseizures precede the amnesia necessary for adiagnosis of the Wernicke-Korsakoff syn-drome (all patients with the Wemicke-Korsakoff syndrome with seizures also hadWemicke's encephalopathy neuropathology).This suggests that the Wernicke-Korsakoffsyndrome may occur by different pathogenicmechanisms: brain damage in association withsevere thiamine deficiency alone (the major-ity), or the combined brain damage due to thi-amine deficiency and seizures.

SummaryWe have devised operational criteria to signifi-cantly improve the identification of patientswith Wemicke's encephalopathy. The use ofthese proposed criteria for the differentialdiagnosis of alcoholics may improve ourunderstanding of the underlying aetiologicalfactors contributing to their neurologicalimpairment. The existing classic triad wasmodified to include the presence of dietarydeficiencies and required only two rather thanthree signs for a clinical diagnosis. Using thiscriterion, the diagnosis of Wernicke'sencephalopathy either alone or with amnesia(Wemicke-Korsakoff syndrome) or hepaticencephalopathy improved from 22% to 85%.

Notably, Wernicke's encephalopathy neu-ropathology was found in a significant numberof patients with hepatic encephalopathy, sug-gesting that these have a high risk of additionalWemicke's encephalopathy and should betreated with parenteral thiamine.

This study was funded by the National Health and MedicalResearch Council of Australia. We are grateful for the assis-tance of the staff of the Histopathology Laboratory,Department of Pathology, University of Sydney, and toProfessors RF Butterworth and J Willoughby and Dr JDGWatson for their comments on the manuscnrpt.

AppendixBRIEF HISTORIES OF REPRESENTATIVE PATIENTS

Alcoholics with hepatic encephalopathyPatient 1 was a 51 year old male alcoholic who had hadthe signs and symptoms of hepatic encephalopathy forthe preceding two months with two hospital admis-sions. He initially presented with bleeding oesophagealvarices and was discharged one month later when sta-bilised. He was readmitted one week later drowsy,jaundiced, and dehydrated. For the next three weeks hehad recurrent variceal bleeds and resistant ascites withincreasing encephalopathy. Six days before death hehad a large haematemesis and became hypotensive. Hewas resuscitated, but his condition remained graveuntil death.

Patient 2 was a 65 year old male alcoholic who hadhad three major hospital admissions over the preceding15 months for the management of his alcoholic liverdisease. He was abstinent for this time period. Aboutone year before death he presented with several dayshistory of weakness, confusion, and disorientation andwas admitted to hospital for management of his hepaticencephalopathy. During the next nine months he hadfurther short hospital stays for increased ascites. Onfinal admission he presented semiconscious andhypothermic with evidence of aspirant in the right lungand ascites. He was rousable and responded to namebut had low blood pressure, was acidotic, and hadascites with a very high white cell count. Antibioticswere started and CT ordered to discount intracranialhaemorrhage; however, on rewarming he deterioratedand died despite intubatation.

Alcoholics with Wernicke's encephalopathy alonePatient 327 was a 51 year old male alcoholic who hadhad increasing periods of binge drinking accompaniedby poor diet over the preceding three years. Three daysbefore death he was admitted to hospital, heavily intox-icated, with delirium, confusion, and agitation. Heshowed ophthalmoplegia, dysarthria, and strabismustremor. He was given thiamine and became alert andoriented briefly before vomiting large amounts of bloodand bile and aspirating.

Patient 4 was a 59 year old male alcoholic who hadno major alcohol related diseases besides Wernicke'sencephalopathy. When first assessed in 1985 for anoticeable Spigelian hernia, he was able to give adetailed, informative family and personal history. Hewas immediately admitted to hospital and underwentsuccessful surgery but became confused and disori-ented in time and place but not person over the nextweek and experienced some hallucinations indicative ofalcohol withdrawal. He was discharged one monthlater. Two years before death he was admitted to hospi-tal for detoxification, showing signs of nystagmus, oph-thalmoplegia, and confusion. He was seen at theneuropsychology unit and tested after seven days sobri-ety. At this time he showed poor planning and con-struction and scored 26/38 on the short mental statusexamination despite scoring only 1/4 for simple calcula-tions, 1/3 for similarities, and 0/4 for construction. Hewas able to learn four objects after two trials and couldremember two of four objects after a five minute delay.Because of defective vision, all memory tests were given

58


http://jnnp.bmj.com

/J N

eurol Neurosurg P


nloaded from



in the auditory mode. One month before death, he wasadmitted to hospital during an acute Wernicke'sencephalopathy episode with ocular signs, confusion(disoriented in time and place, with no recall of theprevious nights events), and severe ataxia. He did notimprove with thiamine therapy (100 mg daily). Formaltesting disclosed global impairment with the examinersconcluding that Wernicke's encephalopathy had par-tially resolved with thiamine therapy. However,reassessment 10 days later showed substantial improve-ment in attention span, alertness, and ability to followsimple commands. He was discharged three weekslater. The next week he was readmitted with back painand was again confused and disoriented in time andplace. Just before death, his memory skills were poorand he was grossly confused, although the memorylapses were not continuous nor of sufficient degree ornature to be consistent with a diagnosis of Korsakoff3spsychosis. This history indicates a persisting confu-sional state related to heavy ethanol intoxication andthiamine deficiency.

Alcoholics with the Wemnicke-Korsakoff syndromePatient 5 was a 78 year old female alcoholic who hadbeen under constant institutionalised care for eightyears before death and had been diagnosed withKorsakoff3s psychosis before admission. Estimates ofalcohol intake were in excess of 100 g per day from theage of 20 without adequate diet. The first report ofmemory impairment was 20 years before death. Shewas persistently disoriented in time and place and hadnystagmus, cerebellar ataxia, and severely impaireddeclarative memory. She was unable to remember herdoctor from week to week despite being in his care forsix years. Disorientation and amnesia persisted untildeath despite documentation of abstinence from alco-hol for the final eight years. This neuropsychologicalprofile indicates the persisting, stable memory impair-ment diagnostic of Korsakoffs psychosis.

Patient 6 was a 60 year old chronic alcoholic whowas admitted to hospital in 1980 for severe headaches,when it was noted that he had problems with his mem-ory. In 1983, he was admitted to hospital with chestpains and during investigations was noted to have nys-tagmus and ataxia. His memory, however, wasregarded as normal. He was readmitted to hospital in1990 because of suicidal intention with bilateral nystag-mus and a broad based ataxic gait. Neuropsychologicaltesting showed him to be a man of average to aboveaverage intelligence, with a normal immediate memoryspan (short mental state 30/38). However, he wasunable to learn any new material (structured orunstructured) beyond his immediate memory span. Herecalled none of four words after five minutes, andthere was nearly total loss of information after interfer-ence. He was unable to recall personal events from theprevious 10 years. He was institutionalised withoutimprovement in memory function for almost threeyears until his death by suicide. This neuropsychologicalprofile indicates the persisting, stable memory impair-ment of Korsakoffs psychosis.

Alcoholics with alcohol related brain dysfunctionPatient 7 was a 58 year old male alcoholic who devel-oped epilepsy a year before death with no CT abnor-mality. Three months before death he was diagnosedwith carcinoma of the lung and declined treatment. Hehad several grand mal seizures in the two days beforedeath.

Patient 8 was a 66 year old male alcoholic who had-drank heavily since the age of 18. Examination in theneuropsychological unit disclosed a mild memorydeficit. He was able to learn four unrelated words afterone attempt, but his recall was very poor. It was notuntil he had repeated the test that he was able toremember two of the four words. All other tests wereadequate. Peripheral neuropathy was noted.

Alcoholics without neurological complicationsPatient 9 was a 41 year old man with a long history ofalcohol misuse. By the age of 29 he had had 34 admis-

sions to detoxification units. On neuropsychologicalassessment 18 months before death no impairment wasnoted. He scored 37/38 on the short mental stateexamination and had insight into his alcohol problem.He remained cognitively stable until death, which wasdue to acute alcohol intoxication.

Akoholics with multiple diagnosesPatient 10 was a 46 year old woman who presentedwith hepatic encephalopathy after a 27 year history ofconsuming about one bottle of scotch a day. Two and ahalf years before death she was admitted for one monthafter a short history of weight loss, vomiting, anorexia,and increasing jaundice with hepatomegaly and raisedserum transaminases. Three months before death, shewas admitted with hepatomegaly and jaundice, but noevidence of oesophageal varices or peptic ulceration.Peripheral neuropathy and cerebellar dysfunction wereevident. Clinical evidence of encephalopathy returnedafter reduction of dietary protein content. She had ashort history of auditory and visual hallucinations. Shehad asterixis, spider naevae, pronounced jaundice,clubbing of fingers and toes, and hepatosplenomegaly.Investigations disclosed anaemia, hyponatraemia, highserum, and urinary osmolarity, and pulmonaryoedema. She was assessed by the neuropsychology unitdenying any symptoms of cognitive impairment or neu-rological dysfunction. Bilateral nystagmus was presenton lateral gaze and there was generalised muscle wast-ing. She had an abnormal tandem gait and abnormalheel-shin test although she was not clinicallyencephalopathic. On short mental state examinationshe scored 20 out of 38 with problems in digit span,learning, calculations, information, similarities, con-struction, and recall. On full neuropsychological assess-ment she showed severe memory impairment andpronounced impairment of higher order cognitive func-tions. Confusion developed and hepatic failure withmalaena became evident, although renal failure super-vened.

Case 11 was a 46 year old male alcoholic with a 10year history of alcohol-related diseases including bleed-ing gastric ulcers, chronic liver disease, nystagmus,tremor, memory blanks and a five year history of ataxiaand frequent vomiting. He was first noted to havememory problems in early 1990 when he returned tohis doctor's surgery having forgotten his prior atten-dance that morning. In the final year of his life, he wasstill living independently although he was often disori-ented in time and place and showed gross memory lossand unsteady gait. He was admitted to hospital withepigastric pain one month before death, was confused,disoriented in time and place, and had poor short termmemory. His family reported an increase in mentalimpairment over the previous fortnight. He was for-mally tested two weeks later and was found to have ageneral amnesic syndrome. Although alert and coopera-tive, he showed perserveration and confabulation. Hecould do no calculations and could not remember factsfrom personal or historical events. He was last admittedto hospital in hepatic coma, surviving for three days.This history indicates an amnesic deficit of at least 12months duration with intermittent episodes ofWernicke's encephalopathy related to heavy ethanolintoxication and thiamine deficiency followed by arapidly decompensating liver disease.

1 Harper CG, Giles M, Finlay-Jones R. Clinical signs in theWemicke-Korsakoff complex: a retrospective analysis of131 patients diagnosed at necropsy. J Neurol NeurosurgPsychiatry 1986;49:341-5.

2 Naidoo DP, Bramdev A, Cooper K. Wernicke'sencephalopathy and alcohol-related disease. PostgradMedJ 1991;67:978-81.

3 Kril JJ. Neuropathology of thiamine deficiency disorders.Metab Brain Dis 1996;11:9-17.

4 Watson JDG, Kril JJ, Lennane KJ. Neurological, neu-ropathological and psychiatric aspects of alcoholism.Medicine International 1995;23:61-5.

5 Shogry MEC, Curnes JT. Mamillary body enhancement onMR as the only sign of acute Wemicke encephalopathy.Am J Neuroradiol 1994;15:172-4.

6 Vortmeyer AO. Wernicke's encephalopathy, Korsakoffsamnesic state and thiamine deficient encephalopathy.Alcohol 1993;28:199-200.

59 on January 17, 2020 by guest. P

rotected by copyright.http://jnnp.bm

j.com/

J Neurol N

eurosurg Psychiatry: first published as 10.1136/jnnp.62.1.51 on 1 January 1997. D

ownloaded from



7 Butterworth RF. Pathophysiologic mechanisms responsiblefor the reversible (thiamine-responsive) and irreversible(thiamine non-responsive) neurological symptoms ofWernicke's encephalopathy. Drug Alcohol Rev 1993;12:315-22.

8 Charness ME. Brain lesions in alcoholics. Alcohol Clin ExpRes 1993;17:2-11.

9 Price J, Kerr R. The Wernicke-Korsakoff syndrome: clini-cal corelates and dilemmas. Aust Drug Alcohol Rev 1988;7:57-60.

10 Victor M, Adams RD, Collins GH. The Wernicke-KorsakoffSyndrome. 2nd ed. Philadelphia: FA Davis Company,1989.

11 D'Aprile P, Gentile MA, Carella A. Enhanced MR in theacute phase of Wernicke encephalopathy. Am _7Neuroradiol 1994;15:591-3.

12 Arriagada PV, Marzloff K, Hyman BT. Distribution ofAlzheimer-type pathologic changes in non-dementedelderly individuals matches the pattern in Alzheimer'sdisease. Neurology 1992;42:1681-8.

13 McKhann G, Drachman D, Folstein M, Katzman R, PriceD, Stadlan EM. Clinical diagnosis of Alzheimer's dis-ease: Report of the NINCDS-ADRDA work group underthe auspices of the Department of Health and HumanServices task-force on Alzheimer's disease. Neurology1984;34:939-44.

14 Morris JC, Heyman A, Mohs RC, et al. The consortium toestablish a registry for Alzheimer's disease (CERAD). I.Clinical and neuropsychological assessment ofAlzheimer's disease. Neurology 1989;39:1159-65.

15 Harper C, Fornes P, Duyckaerts C, Lecomte D, Hauw JJ.An international perspective on the prevalence of theWernicke-Korsakoff syndrome. Metab Brain Dis 1995;10:17-24.

16 Darnton-Hill I, Truswell AS. Thiamin status of a sample ofhomeless clinic attenders in Sydney. Med _7 Aust 1990;152:5-9.

17 Butterworth RF. Hepatic encephalopathy. The Neurologist1995;1:95-104.

18 Butters N, Stuss DT. Diencephalic amnesia. In: Boller F,Grafman J, eds. Handbook of neuropsychology. ElsevierScience Publishers, 1989;3:107-48.

19 Kopelman MD. The Korsakoff syndrome. Br Psychiatry1995;166: 154-73.

20 O'Connor M, Verfaellie M, Cermak LS. Clinical differenti-ation of amnesic subtypes. In: Baddeley AD, Wilson BA,Watts FN, eds. Handbook of memory disorders. New York:John Wiley and Sons, 1995;53-80.

21 Spreen 0, Strauss E. A compendium of neuropsychologicaltests. Oxford: Oxford University Press, 1991.

22 Harper C. Wernicke's encephalopathy: a more commondisease than realised. _7 Neurol Neurosurg Psychiatry 1979;42:226-31.

23 Torvik A. Topographic distribution and severity of brainlesions in Wernicke's encephalopathy. Clin Neuropathol1987;6:25-9.

24 Strub RL, Black FW. The mental status examination in neu-rology. 2nd ed. Philadelphia: FA Davis Co, 1985.

25 Rice JP, Reich T, Bucholz KK, et al. Comparison of directinterview and family history diagnosis of alcohol depen-dence. Alcohol Clin Exp Res 1995;19:1018-23.

26 McKeith IG, Fairbairn AF, Perry RH, Thompson P. Theclinical diagnosis and misdiagnosis of senile dementia of

Lewy body type (SDLT). Br 7 Psychiatry 1994;165:324-32.

27 Byrne C, Halliday G, Ellis J, Harper C. Thalamic vacuola-tion in acute Wernicke's encephalopathy. Metab BrainDis 1993;8:107-14.

28 Victor M. Alcoholic dementia. Can _7 Neurol Sci 1994;21:88-99.

29 Cutting J. The relationship between Korsakov's syndromeand "alcoholic dementia". Br _7 Psychiatry 1978;132:240-51.

30 Butters N. Alcoholic Korsakoffs syndrome: someunresolved issues concerning etiology, neuropathology,and cognitive deficits. _7 Clin Exp Neuropsychol 1985;7:181-210.

31 Bowden SC. Separating cognitive impairment in neurologi-cally asymptomatic alcoholism from Wernicke-KorsakoffSyndrome: is the neuropsychological distinction justified?Psychol Bull 1990;107:355-66.

32 Waugh M, Jackson M, Fox GA, Hawke SH, Tuck RR.Effects of social drinking on neuropsychological perfor-mance. Br.7Addict 1989;84:659-67.

33 Sobaniec-Lotowska ME, Sobaniec W. Morphological fea-tures of encephalopathy after chronic administration ofthe antiepileptic drug valproate to rats. A transmissionelectron microscopic study of capillaries in the cerebellarcortex. Exp Toxic Pathol 1996;48:65-75.

34 Wallis WE, Willoughby E, Baker P. Coma in the Wernicke-Korsakoff syndrome. Lancet 1978;2:400-1.

35 Torvik A, Lindboe CF, Rogde S. Brain lesions in alco-holics. A neuropathological study with clinical correla-tions. .7Neurol Sci 1982;56:223-48.

36 Squire LR, Knowlton B, Musen G. The structure and orga-nization of memory. Ann Rev Psychol 1993;44:453-95.

37 Zola-Morgan S, Squire LR. Neuroanatomy of memory.Ann Rev Neurosci 1993;16:547-63.

38 Moss MB, Albert MS, Butters N, Payne M. Differentialpatterns of memory loss among patients with Alzheimer'sdisease, Huntington's disease, and alcoholic Korsakoffssyndrome. Arch Neurol 1986;43:239-46.

39 Ryan C, Butters N. Further evidence for a continuum-of-impairment encompassing male alcoholic Korsakoffpatients and chronic alcoholic men. Alcohol Clin Exp Res1980;4: 190-8.

40 Price J, Kerr R, Hicks M, Nixon PF. The Wernicke-Korsakoff syndrome: a reappraisal in Queensland withspecial reference to prevention. Med .7 Aust 1987;147:561-5.

41 Joyce EM. Aetiology of alcoholic brain damage: alcoholicneurotoxicity or thiamine malnutrition? Br Med Bull1994;50:99-114.

42 Carlen PL, McAndrews MP, Weiss RT, et al. Alcohol-related dementia in the institutionalized elderly. AlcoholClin Exp Res 1994;18:1330-4.

43 Whelan G. Alcohol-related health problems in the elderly.Med JAust 1995;162:325-7.

44 Adams RD. Nutritional cerebellar degeneration. Amsterdam:Elsevier, 1976.

45 Mancall EL, McEntee WJ. Alterations of the cerebellarcortex in nutritional encephalopathy. Neurology 1965;15:303-13.

46 Lennane KJ. Management of moderate to severe alcohol-related brain damage (Korsakoffs syndrome). Med3 Aust1986;145:136-43.

60


http://jnnp.bmj.com

/J N

eurol Neurosurg P


nloaded from


Documents

Operational criteria for the classification of chronic ... · mental state (disorientation in two of three fields, confused, an abnormal digit span, or comatose); hepatic encephalopathy