bipolar disorder (nice)

8/8/2019 Bipolar Disorder (Nice)

http://slidepdf.com/reader/full/bipolar-disorder-nice 1/5

BackgroundBackground Twenty per centof Twenty per cent of patientswith bipolar affective disorderpatientswith bipolar affective disordersuffer anillnessthatrespondsinadequatelysuffer anillnessthatrespondsinadequatelytotreatmentandhas a pooroutcome.totreatment andhas a pooroutcome.Manypatients, but not all, withbipolarManypatients, but not all, withbipolardisorder show whitematter abnormalitiesdisorder show white matter abnormalitieson Ton T 22-weighted magnetic resonance-weighted magnetic resonanceimaging (MRI).imaging (MRI).

AimsAims To explore thehypothesis thatTo explorethehypothesis thatwhitematter abnormalities on MRI arewhitematter abnormalities on MRI areseenmore frequently insubjects whoseseenmore frequently insubjects whoseillnesshas a poor outcome comparedwithillnesshas a poor outcome compared withthose with a good outcome orcontrols.those with a good outcome orcontrols.

MethodMethod Two groups of age- andTwo groupsof age- andgender-matched patients with bipolargender-matched patientswith bipolar

disorder (14 with a good outcome and15disorder (14 with a good outcome and15with a poor outcome) and 15 controls,with a poor outcome) and15 controls,aged 20^65 years, werestudied. Axial Taged 20^65 years, werestudied. Axial T 22--weighted MRI scans were examined forweighted MRI scanswere examinedforthepresence and severity of white matterthepresence and severity of white matterabnormalities.abnormalities.

ResultsResults SignificantlymorepoorSignificantlymorepooroutcome groupmembers had deepoutcome groupmembershad deepsubcortical punctate, butnotsubcorticalpunctate, but notperiventricular, white matterperiventricular, white matterhyperintensities thanthegood outcomehyperintensities thanthegood outcomegroup (group ( PP 0.035) orcontrols (0.035) orcontrols ( PP 0.003)0.003)and these abnormalitieswere ofgreaterand these abnormalitieswere ofgreaterseverity (severity ( PP 0.030 and0.030 and PP55 0.014,0.014,respectively).respectively).

ConclusionsConclusions SubcorticalwhitematterSubcorticalwhite matterlesions areassociated withpoor outcomelesions are associatedwith poor outcomebipolardisorder.bipolardisorder.

Declaration of interestDeclaration of interest ThisThisresearchwas supported financiallybyresearchwas supported financiallyby

Newcastle City HealthTrust.Newcastle City HealthTrust.

Chronic affective disorders are a majorChronic affective disorders are a majorcause of psychiatric morbidity (Scottcause of psychiatric morbidity (Scott et al et al ,,1988), affecting up to 1% of the popu-1988), affecting up to 1% of the popu-lation. Studies of bipolar patients havelation. Studies of bipolar patients haveidentified a group whose illnesses respondidentified a group whose illnesses respondpoorly to treatment and have a poor out-poorly to treatment and have a poor out-come, with 20% remaining unwell for upcome, with 20% remaining unwell for up

to a year (Coleto a year (Cole et al et al , 1993). Treatment, 1993). Treatmentresistance might be anticipated in individ-resistance might be anticipated in individ-uals whose illness had specific neurobio-uals whose illness had specific neurobio-logical underpinnings. In bipolar subjects,logical underpinnings. In bipolar subjects,TT22-weighted magnetic resonance imaging-weighted magnetic resonance imaging(MRI) has revealed deep subcortical white(MRI) has revealed deep subcortical whitematter lesions (DWML) and periventricularmatter lesions (DWML) and periventricularwhite matter lesions (PVWML), whichwhite matter lesions (PVWML), whichhave differing neuropathologies (Fazekashave differing neuropathologies (Fazekaset al et al , 1993). Although not all MRI studies, 1993). Although not all MRI studiesfind an excess of white matter lesionsfind an excess of white matter lesions(Brown(Brown et al et al , 1992; Strakowski, 1992; Strakowski et al et al ,,1993), the majority do (Dupont1993), the majority do (Dupont et al et al ,,

1987, 1990, 19951987, 1990, 1995 aa ,,bb ; Swayze; Swayze et al et al , 1990;, 1990;FigielFigiel et al et al , 1991; McDonald, 1991; McDonald et al et al , 1991;, 1991;PuzynskiPuzynski et al et al , 1995; Woods, 1995; Woods et al et al , 1995),, 1995),and meta-analysis indicates that moreand meta-analysis indicates that morebipolar patients than controls exhibitbipolar patients than controls exhibitWMLs (AltshulerWMLs (Altshuler et a l e t a l , 1995). Specific,, 1995). Specific,but as yet unidentified subgroups, may ex-but as yet unidentified subgroups, may ex-hibit WMLs. Duponthibit WMLs. Dupont et al et al (1990) reported(1990) reportedthat patients showing WMLs were morethat patients showing WMLs were moreseriously unwell and performed poorly onseriously unwell and performed poorly onneuropsychological testing.neuropsychological testing.

HypothesisHypothesisThese observations lead us to predict thatThese observations lead us to predict thatpoor outcome bipolar patients have under-poor outcome bipolar patients have under-lying structural brain abnormalities. It islying structural brain abnormalities. It isour hypothesis that poor outcome bipolarour hypothesis that poor outcome bipolarsubjects have more white matter abnor-subjects have more white matter abnor-malities on MRI than either good outcomemalities on MRI than either good outcomepatients or controls.patients or controls.

METHODMETHOD

To explore our hypothesis we have com-To explore our hypothesis we have com-pared Tpared T 22-weighted MRI scans of poor and-weighted MRI scans of poor and

good outcome bipolar patients with healthygood outcome bipolar patients with healthycontrols. Patients who did not meet thecontrols. Patients who did not meet the

criteria for inclusion in either the good orcriteria for inclusion in either the good orthe poor outcome groups were eliminatedthe poor outcome groups were eliminatedfrom the study and did not undergo MRI.from the study and did not undergo MRI.

SubjectsSubjects

Patients with a DSM±IV (American Psy-Patients with a DSM±IV (American Psy-chiatric Association, 1994) diagnosis of bi-chiatric Association, 1994) diagnosis of bi-polar affective disorder aged 20±65 yearspolar affective disorder aged 20±65 yearswere identified from the in-patient andwere identified from the in-patient andout-patient services of the Newcastleout-patient services of the NewcastleMental Health Trust. To be included, allMental Health Trust. To be included, allpatients had been diagnosed as sufferingpatients had been diagnosed as sufferingfrom bipolar disorder for at least 3 yearsfrom bipolar disorder for at least 3 yearsand had had a minimum of two episodesand had had a minimum of two episodesof illness. After identification, their caseof illness. After identification, their casenotes were reviewed. Those meeting the cri-notes were reviewed. Those meeting the cri-teria for bipolar disorder (DSM±IV) wereteria for bipolar disorder (DSM±IV) weresubjected to a full clinical evaluation bysubjected to a full clinical evaluation by

an experienced psychiatrist (I.C.M. oran experienced psychiatrist (I.C.M. orP.B.M.), who ensured that they met the cri-P.B.M.), who ensured that they met the cri-teria for inclusion but not the exclusionteria for inclusion but not the exclusioncriteria for the study. A full history, includ-criteria for the study. A full history, includ-ing family history of mental illness,ing family history of mental illness,smoking history and obstetric history, wassmoking history and obstetric history, wascompiled and old case notes were reviewed.compiled and old case notes were reviewed.To screen out subjects with considerableTo screen out subjects with considerablecognitive decline (early dementia orcognitive decline (early dementia orpseudodementia), a Mini-Mental Statepseudodementia), a Mini-Mental StateExamination (MMSE; FolsteinExamination (MMSE; Folstein et al et al , 1975), 1975)was completed. This instrument is insuffi-was completed. This instrument is insuffi-ciently sensitive to detect subtle cognitiveciently sensitive to detect subtle cognitive

changes, for example those reported inchanges, for example those reported inpatients with depression, thereby prevent-patients with depression, thereby prevent-ing inappropriate exclusion of patients.ing inappropriate exclusion of patients.After a complete description of the studyAfter a complete description of the studyof the patients, written informed consentof the patients, written informed consentwaswas obtained. Thereafter, subjects wereallo-obtained. Thereafter, subjects wereallo-catedcated to either a good or poor outcometo either a good or poor outcomegroup (see below) or eliminated becausegroup (see below) or eliminated becauseof an intermediate outcome. Immediatelyof an intermediate outcome. Immediatelyprior to MRI scanning, the patient was in-prior to MRI scanning, the patient was in-terviewed to assess current mental stateterviewed to assess current mental stateand a Beck Depression Inventory (BDI;and a Beck Depression Inventory (BDI;BeckBeck et al et al , 1961) was completed by the, 1961) was completed by the

subject.subject.

Patient selection criteriaPatient selection criteriaThese are set out in the Appendix.These are set out in the Appendix.

Poor outcome group (B)Poor outcome group (B)At the time of study, patients in this groupAt the time of study, patients in this grouphad been unwell for 2 years or more,had been unwell for 2 years or more,despite adequate therapy. Any periods of despite adequate therapy. Any periods of remission lasted 8 weeks or less, duringremission lasted 8 weeks or less, duringwhich time they still had significant func-which time they still had significant func-

tional impairment. All patients had madetional impairment. All patients had madea poor response to lithium. Almost all werea poor response to lithium. Almost all were

17 217 2

BRITIS H JO URNAL OF P SYCHIATRYBRITIS H JOURN AL OF P SYCHIATRY ( 2 0 0 1 ) , 1 7 8 , 1 7 2 ̂ 1 7 6( 2 0 0 1 ) , 1 7 8 , 1 7 2 ^ 1 7 6

Cerebral white matter lesions in bipolar affectiveCerebral white matter lesions in bipolar affective

disorder: relationship to outcomedisorder: relationship to outcome

P. BRIAN MOORE, DEBRA J. SHEPHERD, DONALD ECCLESTON,P. BRIAN MOORE, DEBRA J. SHEPHERD, DONALD ECCLESTON,

IAIN C. MACMILLAN, UPTAL GOSWAMI, VICTOR L. McALLISTERIAIN C. MACMILLAN, UPTAL GOSWAMI, VICTOR L. McALLISTERand I. NICOL FERRIERand I. NICOL FERRIER



W H I T E M AT T E R L E S I O N S I N B I P O L A R A F F E C T I V E D I S O R D E RW H I T E M AT T E R L E S I O N S I N B I P O L A R A F F E C T I V E D I S O R D E R

patients of a regional unit specialising in thepatients of a regional unit specialising in thetreatment of affective disorder.treatment of affective disorder.

Good outcome group (A)Good outcome group (A)At the time of study, these patients hadAt the time of study, these patients hadbeen clinically euthymic for at least 8 eightbeen clinically euthymic for at least 8 eightweeks. After any episode of illness, theyweeks. After any episode of illness, theyhad shown full symptomatic recovery andhad shown full symptomatic recovery andhad returned to normal premorbid func-had returned to normal premorbid func-tioning. The good and poor outcometioning. The good and poor outcomegroups were age and gender matched.groups were age and gender matched.

ControlsControlsControls were recruited from the familiesControls were recruited from the familiesof patients and the staff of Newcastle Uni-of patients and the staff of Newcastle Uni-versity and its associated hospitals. Con-versity and its associated hospitals. Con-

trols were selected to match, as closely astrols were selected to match, as closely aspossible, the ages and genders of subjectspossible, the ages and genders of subjectsin the poor outcome group. No controlsin the poor outcome group. No controlsmet the exclusion criteria applied to themet the exclusion criteria applied to thepatient groups.patient groups.

Patient characteristicsPatient characteristicsIn total, 17 patients were recruited into theIn total, 17 patients were recruited into thepoor outcome group and 16 into the goodpoor outcome group and 16 into the goodoutcome group. One patient was excludedoutcome group. One patient was excludedbecause he experienced a panic attackbecause he experienced a panic attackduring MRI scanning and the scanningduring MRI scanning and the scanning

sequence could not be completed. Threesequence could not be completed. Threepatients were eliminated after the scanspatients were eliminated after the scanswere examined because of evidence of de-were examined because of evidence of de-myelination (myelination ( nn 1), an Arnold±Chiari mal-1), an Arnold±Chiari mal-formation with hydrocephalus (formation with hydrocephalus ( nn 1) and1) andcerebrovascular infarcts (cerebrovascular infarcts ( nn 1). A total of 1). A total of 29 patients completed the study, 15 in the29 patients completed the study, 15 in thepoor outcome and 14 in the good outcomepoor outcome and 14 in the good outcomegroup.group.

The patient characteristics of the goodThe patient characteristics of the good(A) and poor (B) outcome groups and con-(A) and poor (B) outcome groups and con-trols are compared in Table 1. Good out-trols are compared in Table 1. Good out-come patients tend to be older, bettercome patients tend to be older, better

educated and less likely to be left handed,educated and less likely to be left handed,but the differences do not reach signifi-but the differences do not reach signifi-cance. The illness duration, gender ratios,cance. The illness duration, gender ratios,ages and prevalence of a family history of ages and prevalence of a family history of bipolar disorder are similar in both groups.bipolar disorder are similar in both groups.The major difference between the groupsThe major difference between the groupsarises in the severity of illness at the timearises in the severity of illness at the timeof investigation. All good outcome patientsof investigation. All good outcome patientswere euthymic on the day of scanning,were euthymic on the day of scanning,reflected in mean BDI scores of 9.0,reflected in mean BDI scores of 9.0,s.d.s.d. 6.7. In contrast, the poor outcome6.7. In contrast, the poor outcomegroup had significantly higher (group had significantly higher ( PP 55 0.05,0.05,unpaired Student'sunpaired Student's t t -test) BDI scores of -test) BDI scores of

16.0, s.d.16.0, s.d. 7.0, and several of this group7.0, and several of this groupshowed clinical mood disturbances.showed clinical mood disturbances.

Magnetic resonance imagingMagnetic resonance imagingAll subjects were examined using a GeneralAll subjects were examined using a GeneralElectrics (Slough, UK) MR max plus 0.5Electrics (Slough, UK) MR max plus 0.5tesla scanner. Sagittal, coronal and axialtesla scanner. Sagittal, coronal and axialTT11-weighted scan sequences, axial T-weighted scan sequences, axial T 22--weighted scan sequences and coronal inver-weighted scan sequences and coronal inver-sion recovery scan sequences were recordedsion recovery scan sequences were recordedon all subjects. The Ton all subjects. The T 11-weighted and inver--weighted and inver-sion recovery scans were examined by asion recovery scans were examined by aconsultant neuroradiologist (V.L.M.) forconsultant neuroradiologist (V.L.M.) for

the presence of structural abnormalities.the presence of structural abnormalities.Up to 15 axially oriented 7-mm slices sep-Up to 15 axially oriented 7-mm slices sep-arated from each other by 1 mm werearated from each other by 1 mm wererecorded during the Trecorded during the T 22-weighted sequence.-weighted sequence.

Axial TAxial T 22-weighted scan sequences with-weighted scan sequences witha relaxation time of 2300 ms and echoa relaxation time of 2300 ms and echodelays of 25 and 100 ms were examineddelays of 25 and 100 ms were examinedfor the presence of white matter hyper-for the presence of white matter hyper-intensities. To be counted, WMLs had tointensities. To be counted, WMLs had tobe present on both the proton density andbe present on both the proton density andTT22 scans. Two investigators (D.J.S. andscans. Two investigators (D.J.S. andV.L.M.) examined scans independentlyV.L.M.) examined scans independentlyand blindly and their results were in com-and blindly and their results were in com-

plete agreement. White matter lesions wereplete agreement. White matter lesions wereclassified as either deep subcortical or peri-classified as either deep subcortical or peri-ventricular and images were graded 0±3ventricular and images were graded 0±3using the scheme proposed by Fazekasusing the scheme proposed by Fazekas(Fazekas(Fazekas et al et al , 1987, 1993)., 1987, 1993).

Statistical analysisStatistical analysisPrior to analysis, all data were checkedPrior to analysis, all data were checkedfor deviation from normality using thefor deviation from normality using theAnderson±Darling method (Stephens,Anderson±Darling method (Stephens,1980). The MMSE scores, age of onset of 1980). The MMSE scores, age of onset of illness and duration of education differedillness and duration of education differed

significantly from normality. All data weresignificantly from normality. All data wereanalysed using the appropriate parametricanalysed using the appropriate parametric

or non-parametric methods: for parametricor non-parametric methods: for parametricdata, ANOVA, ANCOVA, Student'sdata, ANOVA, ANCOVA, Student's t t -test-testand Pearson product moment coefficientand Pearson product moment coefficientwere calculated; for non-parametric data,were calculated; for non-parametric data,including white matter hyperintensityincluding white matter hyperintensitygradings,gradings, ww22, Mann±Whitney and Spearman, Mann±Whitney and Spearmanrank correlations were calculated using therank correlations were calculated using thecomputerised statistical package Minitabcomputerised statistical package Minitab10.2 for Windows (Minitab Inc., PA).10.2 for Windows (Minitab Inc., PA).

RESULTSRESULTS

The principal findings of the study areThe principal findings of the study areshown in Table 2. Comparing the numbersshown in Table 2. Comparing the numbersof subjects with DWMLs, significantlyof subjects with DWMLs, significantlymore (odds ratiomore (odds ratio 11.4) patients (7/15,11.4) patients (7/15,47%) in the poor outcome group had47%) in the poor outcome group hadDWMLs than in the good group (1/14,DWMLs than in the good group (1/14,7%, Fisher's exact test,7%, Fisher's exact test, PP 0.035) or con-0.035) or con-trols (0/15, 0%,trols (0/15, 0%, PP 0.003). The grades0.003). The grades(severity) of DWMLs in the poor outcome(severity) of DWMLs in the poor outcomegroup significantly exceeded(Mann±Whitneygroup significantlyexceeded(Mann±Whitney

U-test) the good outcome patientsU-test) the good outcome patients((PP 0.030) or controls (0.030) or controls ( PP 55 0.014). In con-0.014). In con-trast, the grade of DWMLs in the goodtrast, the grade of DWMLs in the goodpatient group and controls did not differpatient group and controls did not differsignificantly (significantly ( PP 0.96).0.96).

In contrast, PVWMLs were not presentIn contrast, PVWMLs were not presentin strikingly different numbers in the goodin strikingly different numbers in the good(8/14, 57%), poor (10/15, 67%) or control(8/14, 57%), poor (10/15, 67%) or control(7/15, 47%) groups. Fisher's exact test(7/15, 47%) groups. Fisher's exact testfailed to reveal any differences betweenfailed to reveal any differences betweenthe numbers in each group with PVWMLs.the numbers in each group with PVWMLs.Although more patients in the poor out-Although more patients in the poor out-come group had more severe abnormalities,come group had more severe abnormalities,

Mann±Whitney tests failed to reveal anyMann±Whitney tests failed to reveal anysignificant differences between the groupssignificant differences between the groups

17 317 3

Table 1Table 1 Characteristics of patient groups and controlsCharacteristics of patient groups andcontrols

Good outcome groupGoodoutcomegroup Poor outcome groupPoor outcome group ControlsControls

nn (m, f)(m, f) 14 (8, 6)14 (8, 6) 15 (7, 8)15 (7, 8) 15 (7, 8)15 (7, 8)

Age, years (s.d.)Age, years (s.d.) 47.4 (10.10)47.4 (10.10) 42.1 (13.9)42.1 (13.9) 41.9 (12.6)41.9 (12.6)

Length of education, years (s.d.)Length of education, years(s.d.) 12.7 (3.0)12.7 (3.0) 10.9 (3.1)10.9 (3.1) 14.0 (1.1)14.0 (1.1)

% Right-handed% Right-handed 92%92% 80%80% 80%80%

Non-CaucasianNon-Caucasian 1 1 11 00

Rapid cycling (Rapid cycling ( 44 4 episodes/year)4 episodes/year) 0 0 77 ^ ^

Family history of bipolar affective disorderFamilyhistory of bipolar affective disorder 4 4 44 00

Age of illness onset, years (s.d.)Age of illnessonset, years (s.d.) 31.4 (3.5)31.4 (3.5) 26.3 (10.0)26.3 (10.0) ^ ^

Illness duration, years (s.d.)Illness duration, years ( s.d.) 16.0 (7.9)16.0 (7.9) 15.8 (10.8)15.8 (10.8) ^ ^

No. of episodes of illness, mean (s.d.)No. of episodes of illness, mean (s.d.) 8.6 (6.0)8.6 (6.0) 8.9 (4.3)8.9 (4.3) ^ ^

No. of hospitalisations, mean (s.d.)No. of hospitalisations, mean (s.d.) 2.4 (1.9)2.4 (1.9) 5.0 (4.7)5.0 (4.7) 11 ^ ^

MMSE scores, mean (s.d.)MMSE scores, mean (s.d.) 29.0 (1.0)29.0 (1.0) 28.0 (2.3)28.0 (2.3) 30 (0)30 (0)

BDI scores, mean (s.d.)BDI scores, mean (s.d.) 9.0 (6.7)9.0 (6.7) 16.0 (7.0)16.0 (7.0) 11 ^ ^

MMSE,Mini-Mental State Examination; BDI, Beck Depression Inventory.MMSE,Mini-Mental State Examination; BDI, Beck Depression Inventory.1.1. PP 55 0.01, Student's0.01, Student's tt-test-test v.v. good outcome.good outcome.



M O OR E E T A LM O OR E E T A L

(good(good v.v. poor,poor, PP 0.22; good0.22; good v.v. controls,controls,PP 0.74; poor0.74; poor v.v. controls,controls, PP 0.16).0.16).

It is concluded that DWMLs but notIt is concluded that DWMLs but notPVWMLs are linked significantly to poorPVWMLs are linked significantly to pooroutcome in bipolar disorder.outcome in bipolar disorder.

Obstetric, smoking and family historiesObstetric, smoking and family historiesof mental disorder were collected. A de-of mental disorder were collected. A de-tailed statistical analysis of the data istailed statistical analysis of the data isnot warranted because of the small samplenot warranted because of the small samplesize. However, inspection of the datasize. However, inspection of the datafailed to suggest that, for an individual,failed to suggest that, for an individual,

there was a clear link between WMLsthere was a clear link between WMLsand these histories. A larger study wouldand these histories. A larger study wouldbe required to explore correlations betweenbe required to explore correlations betweenthese variables.these variables.

DISCUSSIONDISCUSSION

Deep white matter lesionsDeep white matter lesionsThe principal finding of this study is thatThe principal finding of this study is thatDWMLs are seen most frequently in bi-DWMLs are seen most frequently in bi-polar patients with a poor outcome whenpolar patients with a poor outcome whencompared with those who have a good out-compared with those who have a good out-come or with normal controls. In contrast,come or with normal controls. In contrast,

the frequencies of PVWML hyperintensitiesthe frequencies of PVWML hyperintensitiesdo not differ significantly between thesedo not differ significantly between thesegroups. Although the pathology of WMLsgroups. Although the pathology of WMLsin bipolar disorder is not known, theyin bipolar disorder is not known, theymay be direct consequences of the diseasemay be direct consequences of the diseaseprocess, they may result from independentprocess, they may result from independentprocesses that then predispose to treatmentprocesses that then predispose to treatmentresistance or they may simply be coinciden-resistance or they may simply be coinciden-tal findings. Clinicopathological studies intal findings. Clinicopathological studies inelderly subjects indicate that DWMLs andelderly subjects indicate that DWMLs andPVWMLshave different aetiologies.DiscretePVWMLshave different aetiologies. Discreteand confluent subcortical punctate WMLsand confluent subcortical punctate WMLsprobably arise from perivascular abnormal-probably arise from perivascular abnormal-

ities and in severe cases may indicate theities and in severe cases may indicate thepresence of areas of microcystic infarct.presence of areas of microcystic infarct.

Hyperintense ventricular caps and smoothHyperintense ventricular caps and smoothperiventricular halos may result fromperiventricular halos may result frommyelin pallor and are associated with themyelin pallor and are associated with theloss of ependymal lining of the ventriclesloss of ependymal lining of the ventriclesand local increase in tissue fluid. In con-and local increase in tissue fluid. In con-trast, extensive irregular PVWMLs have atrast, extensive irregular PVWMLs have asimilar aetiology to punctate lesionssimilar aetiology to punctate lesions(Fazekas(Fazekas et al et al , 1993). Hence, analysis of , 1993). Hence, analysis of white matter abnormalities in diseasedwhite matter abnormalities in diseasedstates needs to differentiate clearly betweenstates needs to differentiate clearly betweenthese lesions. Although our study showsthese lesions. Although our study shows

that there is an association between deepthat there is an association between deeppunctate and periventricular lesions in ourpunctate and periventricular lesions in ourpatient group (Spearman:patient group (Spearman: PP 0.39,0.39, PP

0.04), only the deep punctate lesions are0.04), only the deep punctate lesions areassociated with a poor outcome bipolar dis-associated with a poor outcome bipolar dis-order. A similar finding has been reportedorder. A similar finding has been reportedin elderly patients with depression ± in thisin elderly patients with depression ± in thisgroup, DWMLs but not PVWMLs are asso-group, DWMLs but not PVWMLs are asso-ciated with depression (O'Brienciated with depression (O'Brien et al et al ,,1996). This difference probably reflects1996). This difference probably reflectsthe differing neuropathological bases of the differing neuropathological bases of these abnormalities. The present resultsthese abnormalities. The present resultssuggest that a factor associated with poorsuggest that a factor associated with poor

outcome in bipolar disorder is the presenceoutcome in bipolar disorder is the presenceof microvascular abnormalities in theof microvascular abnormalities in thecerebral white matter.cerebral white matter.

A number of studies have reportedA number of studies have reportedWMLs in bipolar patients but many failWMLs in bipolar patients but many failto differentiate periventricular from sub-to differentiate periventricular from sub-cortical punctate lesions. White mattercortical punctate lesions. White matterlesions have been reported more frequentlylesions have been reported more frequentlyin young bipolar patients (agedin young bipolar patients (aged 55 60 years)60 years)than controls by Figielthan controls by Figiel et al et al (1991), Swayze(1991), Swayzeet al et al (1990) and Dupont(1990) and Dupont et al et al (1987, 1990),(1987, 1990),whereas Strakowskiwhereas Strakowski et al et al (1993) failed to(1993) failed tofind any difference for newly diagnosedfind any difference for newly diagnosed

patients with mania. In these studies, thepatients with mania. In these studies, thepercentages of bipolar patients withpercentages of bipolar patients with

subcortical hyperintensities varied fromsubcortical hyperintensities varied from55% to 18%, whereas the percentages in55% to 18%, whereas the percentages innormal controls varied from 17% to 0%.normal controls varied from 17% to 0%.This variation could have resulted from sev-This variation could have resulted from sev-eral factors. In particular, risk factors foreral factors. In particular, risk factors forthe formation of hyperintensities, especiallythe formation of hyperintensities, especially

advancing ageing and hypertension, mayadvancing ageing and hypertension, maynot have been controlled for sufficientlynot have been controlled for sufficientlyrigorously. However, our study suggestsrigorously. However, our study suggeststhat a major cause of the variation is likelythat a major cause of the variation is likelyto be the inclusion of poor prognosis pa-to be the inclusion of poor prognosis pa-tients in the study. Studies based on hospi-tients in the study. Studies based on hospi-tal populations tend to overrepresent thetal populations tend to overrepresent themore chronic patients with long illnessmore chronic patients with long illnesshistories who suffer frequent relapse. It ishistories who suffer frequent relapse. It isnotable that the study of Strakowskinotable that the study of Strakowski et al et al

(1993), which failed to find significant dif-(1993), which failed to find significant dif-ferences between controls and patients,ferences between controls and patients,considered patients only in their first epi-considered patients only in their first epi-

sode of manic illness. This factor alone issode of manic illness. This factor alone islikely to reduce significantly the propor-likely to reduce significantly the propor-tions of poor outcome patients included intions of poor outcome patients included intheir study because only approximatelytheir study because only approximately10±20% of all bipolar patients go on to10±20% of all bipolar patients go on todevelop a treatment-resistant disorder (Coledevelop a treatment-resistant disorder (Coleet al et al , 1993). Furthermore, Strakowski's, 1993). Furthermore, Strakowski'sstudy excluded patients with a family his-study excluded patients with a family his-tory of bipolar disorder, thereby elimin-tory of bipolar disorder, thereby elimin-ating those with a genetic predispositionating those with a genetic predispositionthat could be associated with the develop-that could be associated with the develop-ment of organic lesions. Dupontment of organic lesions. Dupont et al et al

(1990) noted that bipolar patients with(1990) noted that bipolar patients with

WMLs had significantly more hospitalisa-WMLs had significantly more hospitalisa-tions, higher Hamilton Rating Scale fortions, higher Hamilton Rating Scale forDepression scores and received more neuro-Depression scores and received more neuro-leptics than patients without WMLs. All of leptics than patients without WMLs. All of these measures of poor outcome may con-these measures of poor outcome may con-firm the current finding that WMLs are as-firm the current finding that WMLs are as-sociated with a poorer prognosis. This issociated with a poorer prognosis. This isfurther suggested by the observation of afurther suggested by the observation of afollow-up 1 year later, that approximatelyfollow-up 1 year later, that approximatelyhalf of Dupont's patients with WMLs failedhalf of Dupont's patients with WMLs failedto recover. Further support for this hypoth-to recover. Further support for this hypoth-esis is provided by studies in elderly, uni-esis is provided by studies in elderly, uni-polar subjects that have linked DWMLs topolar subjects that have linked DWMLs to

outcome. Individuals with DWMLs are re-outcome. Individuals with DWMLs are re-ported to respond less well to in-patientported to respond less well to in-patienttreatment, including electroconvulsive ther-treatment, including electroconvulsive ther-apy (Hickieapy (Hickie et al et al , 1995), and to relapse, 1995), and to relapsemore rapidly (O'Brienmore rapidly (O'Brien et al et al , 1998)., 1998).

Although we have classified outcomeAlthough we have classified outcomeon the basis of a clinical assessment andon the basis of a clinical assessment andlevel of functioning, others, for examplelevel of functioning, others, for exampleMcGlashan (1984) in his outcome studiesMcGlashan (1984) in his outcome studiesat Chestnut Lodge, have described out-at Chestnut Lodge, have described out-come on a more systematic basis using fivecome on a more systematic basis using fivedimensions: hospitalisation, employment,dimensions: hospitalisation, employment,social activity, symptoms and global func-social activity, symptoms and global func-

tioning, rated 0±4. With the exception of tioning, rated 0±4. With the exception of social activity, these were recorded forsocial activity, these were recorded for

17 417 4

Table 2Table 2 Whitematter abnormalities on magnetic resonance imaging in bipolar subjects and controls White matter abnormalities on magnetic resonance imaging in bipolar subjects and controls

Poor outcomePoor outcome

groupgroup

Good outcomeGood outcome

groupgroup

ControlsControls

nn 1515 1414 1515

Deep punctate white matter lesionsDeep punctate white matter lesions

Fazekas grade:Fazekas grade: 0 0 88 1313 1515

11 44 00 00

22 33 11 00

33 00 00 00PP (Mann^Whitney) comparedwith controls(Mann^Whitney) compared with controls 55 0.0140.014 0.960.96

Periventricular white matter lesionsPeriventricular white matter lesions

Fazekas grade:Fazekas grade: 0 0 55 66 88

11 66 88 66

22 44 00 11

33 00 00 00PP (Mann^Whitney) comparedwith controls(Mann^Whitney) comparedwith controls 0.16 0.16 0.740.74



W H I T E M AT T E R L E S I O N S I N B I P O L A R A F F E C T I V E D I S O R D E RW H I T E M AT T E R L E S I O N S I N B I P O L A R A F F E C T I V E D I S O R D E R

our patients. In retrospect, our poor out-our patients. In retrospect, our poor out-come patients would score 0 or 1, 0, nocome patients would score 0 or 1, 0, nodata, 0 and 0, respectively, on each of data, 0 and 0, respectively, on each of these dimensions. In contrast, the corre-these dimensions. In contrast, the corre-sponding scores of our good outcomesponding scores of our good outcomepatients would be 4, 3±4 (including house-patients would be 4, 3±4 (including house-

work as employment), no data, 3±4 and 4,work as employment), no data, 3±4 and 4,respectively. This reflects the study designrespectively. This reflects the study designto maximise the contrast between goodto maximise the contrast between goodand poor outcome patient groups.and poor outcome patient groups.

Periventricular white matterPeriventricular white matterlesionslesionsThe frequency of PVWMLs in our study didThe frequency of PVWMLs in our study didnot differ significantly between the goodnot differ significantly between the goodand the poor outcome groups and normaland the poor outcome groups and normalcontrols. The frequency of all PVWMLscontrols. The frequency of all PVWMLswas high, at 57%, 67% and 47%, respec-was high, at 57%, 67% and 47%, respec-tively, for the good and poor prognosistively, for the good and poor prognosisgroups and controls. These rates are com-groups and controls. These rates are com-parable with other studies (Altshulerparable with other studies (Altshuler et al et al ,,1995). However, interpretation of the1995). However, interpretation of thepresence of PVWMLs is problematic.presence of PVWMLs is problematic.Studies (SzeStudies (Sze et al et al , 1986; Leifer, 1986; Leifer et al et al , 1990), 1990)suggest that the hyperintense caps seen atsuggest that the hyperintense caps seen atthe end of lateral ventricles on Tthe end of lateral ventricles on T 22-weighted-weightedscans are a normal finding, as is thescans are a normal finding, as is thepresence of a thin white hyperintense linepresence of a thin white hyperintense linearound the ventricle margins. When wearound the ventricle margins. When werated our scans with these appearances asrated our scans with these appearances asnormal, the differences between the groupsnormal, the differences between the groupsincreased but nevertheless still failed toincreased but nevertheless still failed toreach statistical significance at the 5%reach statistical significance at the 5%level. Failure to find significantly differentlevel. Failure to find significantly differentnumbers of good and poor outcomenumbers of good and poor outcomepatients and controls with PVWMLs maypatients and controls with PVWMLs maybe a consequence of the low power of abe a consequence of the low power of astudy based on 44 subjects. We are un-study based on 44 subjects. We are un-certain whether some PVWMLs are relatedcertain whether some PVWMLs are relatedto a poor outcome, and further studies areto a poor outcome, and further studies areindicated.indicated.

MedicationMedicationThe medication taken by good and poorThe medication taken by good and poor

outcome groups was similar and reflectedoutcome groups was similar and reflectedthe recent trend in bipolar management tothe recent trend in bipolar management tofocus upon treatment with mood stabili-focus upon treatment with mood stabili-sers. Thus, 10, 6 and 0 of the 14 goodsers. Thus, 10, 6 and 0 of the 14 goodoutcome patients received lithium, car-outcome patients received lithium, car-bamazepine and/or valproate, respectively,bamazepine and/or valproate, respectively,whereas the corresponding numbers werewhereas the corresponding numbers were12, 8 and 5 for the 15 poor outcome12, 8 and 5 for the 15 poor outcomepatients. Four patients in the good outcomepatients. Four patients in the good outcomegroup received lithium/anticonvulsantgroup received lithium/anticonvulsantcombinations whereas nine poor outcomecombinations whereas nine poor outcomepatients received this combination.patients received this combination.

It is improbable that lithium will giveIt is improbable that lithium will give

rise to WMLs. Good and poor outcomerise to WMLs. Good and poor outcomegroups had broadly similar usage of thesegroups had broadly similar usage of these

drugs and yet had markedly differentdrugs and yet had markedly differentDWML frequencies, the good group beingDWML frequencies, the good group beingsimilar to the controls, who were drug free.similar to the controls, who were drug free.Only patients received medication, yetOnly patients received medication, yetPVWMLs were found in considerablePVWMLs were found in considerablenumbers in the controls.numbers in the controls.

With the exception of sodium valproateWith the exception of sodium valproateusage, the most striking differences betweenusage, the most striking differences betweenthe groups was the number of patients inthe groups was the number of patients inthe poor outcome compared to the goodthe poor outcome compared to the goodoutcome group receiving neuroleptics (6outcome group receiving neuroleptics (6 v.v.

2), antidepressants (62), antidepressants (6 v.v. 2) or anticholiner-2) or anticholiner-gics (4gics (4 v.v. 0). There are insufficient patients0). There are insufficient patientsreceiving different drugs to warrant a statis-receiving different drugs to warrant a statis-tical analysis of the data, especially in viewtical analysis of the data, especially in viewof the pharmacological heterogeneity of theof the pharmacological heterogeneity of thedifferent neuroleptics or antidepressants.different neuroleptics or antidepressants.There is, however, little evidence in theThere is, however, little evidence in theliterature that either neuroleptic or anti-literature that either neuroleptic or anti-

depressant usage results in WMLs.depressant usage results in WMLs.

Age and white matter lesionsAge and white matter lesionsIt has been emphasised that age is anIt has been emphasised that age is animportant determinant of the prevalenceimportant determinant of the prevalenceof WMLs. Woodsof WMLs. Woods et al et al (1990) reported that(1990) reported thatthe prevalence of both DWMLs andthe prevalence of both DWMLs andDVWMLs in bipolar subjects but not inDVWMLs in bipolar subjects but not incontrols increased after the age of 30 years.controls increased after the age of 30 years.In contrast, AltshulerIn contrast, Altshuler et al et al (1995) reported(1995) reportedthat the prevalence of only PVWMLs in bi-that the prevalence of only PVWMLs in bi-polar subjects but not in controls increasedpolar subjects but not in controls increased

after the age of 30 years. This differenceafter the age of 30 years. This differencemay result from the lack of control formay result from the lack of control forcardiac risk factors in Woods' study, espe-cardiac risk factors in Woods' study, espe-cially in view of the likelihood that sub-cially in view of the likelihood that sub-cortical lesions are vascular in origin. Incortical lesions are vascular in origin. Inour bipolar subjects, we found no signifi-our bipolar subjects, we found no signifi-cant effect of age upon the prevalence of cant effect of age upon the prevalence of DWMLs (ANCOVA:DWMLs (ANCOVA: F F 0.02,0.02, PP 0.88),0.88),whereas the PVWMLs increased with agewhereas the PVWMLs increased with age(ANCOVA:(ANCOVA: F F 5.57,5.57, PP 0.007). The mean0.007). The meanage of our poor outcome group (42.1,age of our poor outcome group (42.1,s.d.s.d. 13.9 years) was comparable to that of 13.9 years) was comparable to that of the controls (41.9, s.d.the controls (41.9, s.d. 12.6 years) but less12.6 years) but less

than that of the good outcome group (47.4,than that of the good outcome group (47.4,s.d.s.d. 10.1 years). Hence, it is improbable10.1 years). Hence, it is improbablethat differences in the ages of groups couldthat differences in the ages of groups couldcontribute significantly to the increased pre-contribute significantly to the increased pre-valence of DWMLs in the poor outcomevalence of DWMLs in the poor outcomegroups. Age similarities rather than groupgroups. Age similarities rather than groupdifferences (differences ( F F 2.73,2.73, PP 0.08) account for0.08) account forthe similarities in the prevalence of the similarities in the prevalence of PVWMLs in patient groups and controls.PVWMLs in patient groups and controls.

The presence of DWMLs distinguishesThe presence of DWMLs distinguishespoor outcome from good outcome patientpoor outcome from good outcome patientgroups and controls. Nevertheless, thesegroups and controls. Nevertheless, theseabnormalities were found in only 47% of abnormalities were found in only 47% of

the poor prognosis patient group. The aeti-the poor prognosis patient group. The aeti-ology of treatment resistance thereforeology of treatment resistance therefore

appears to be multi-factorial. Preliminaryappears to be multi-factorial. Preliminaryinvestigations suggest that cerebral dys-investigations suggest that cerebral dys-rhythmias (Colerhythmias (Cole et al et al , 1993) are associated, 1993) are associatedwith treatment resistance. These observa-with treatment resistance. These observa-tions may link to the growing evidence thattions may link to the growing evidence thatpoor outcome patients respond much betterpoor outcome patients respond much better

to anticonvulsants than to lithium. Some of to anticonvulsants than to lithium. Some of the poor outcome treatment-resistantthe poor outcome treatment-resistantpatients were clinically unwell at the timepatients were clinically unwell at the timeof scanning and it is possible that differ-of scanning and it is possible that differ-encesences between the good and poor outcomebetween the good and poor outcomegroups reflect differences in current mentalgroups reflect differences in current mentalstate. However, there is evidence (Dupontstate. However, there is evidence (Dupontet al et al , 1990) that WMLs observed in bipolar, 1990) that WMLs observed in bipolarpatients are stable over time. Nevertheless,patients are stable over time. Nevertheless,it would seem prudent to re-examineit would seem prudent to re-examinepatients with DWMLs in the future whenpatients with DWMLs in the future whenthey are clinically euthymic to establishthey are clinically euthymic to establishwhether the lesions reflect the clinical statewhether the lesions reflect the clinical state

of thepatient or more fundamental abnorm-of the patient or more fundamental abnorm-alities of cerebral architecture.alities of cerebral architecture.

APPENDIXAPPENDIX

Patient selection criteriaPatient selection criteriaThe inclusion criteria are:The inclusion criteria are:

(a)(a) DSM^ IV bipolar affective disorder;DSM^ IV bipolar affectivedisorder;

(b)(b) age 20^65 years;age 20^65 years;

(c)(c) two episodesminimumand a history of illness fortwo episodesminimum and a history of illness forat least 3 years.atleast 3 years.

Patients were then divided into two groups:Patients were then divided into two groups:

(a)(a) Good outcomegroup (A):Good outcome group (A):

(i)(i) return topremorbidlevelof functioningreturn to premorbid levelof functioningbetween illness;between illness;

(ii)(ii) currentlyeuthymic for at least 8 weeks;currentlyeuthymic forat least 8 weeks;

(iii)(iii) if prescribed, goodresponse tolithium.if prescribed, good response tolithium.

(b)(b) Pooroutcome group (B):Pooroutcome group (B):

(i)(i) symptomatic non-recovery for at leastthesymptomaticnon-recovery for at least thepast 2 years;past 2 years;

(ii)(ii) anyperiod of well-beingin the past 2 yearsanyperiodof well-beinginthe past 2 yearsfor 8 weeks or less;for 8 weeks or less;

(iii)(iii) failure toregain premorbid functioningfailure to regain premorbid functioningduringperiods of well-being.duringperiods of well-being.

The exclusion criteria are:The exclusion criteria are:

(a)(a) Psychiatric: evidence of cognitive decline; otherPsychiatric: evidence of cognitive decline; otherAxis I comorbid condition; bipolar disorderAxis I comorbid condition; bipolar disorderother thantype I or type II; learning disabilities.other thantype I or type II; learning disabilities.

(b)(b) Neurological: cerebrovascular disease; neuro-Neurological: cerebrovascular disease; neuro-degenerative disorders; head injury with concus-degenerative disorders; head injury with concus-sion; epilepsy; idiopathic parkinsonism; systemicsion; epilepsy; idiopathic parkinsonism; systemicillness with cerebral consequences; focal neuro-illness with cerebral consequences; focal neuro-logical signs on examination.logical signs on examination.

(c)(c) Medical: hepatic disorder; cardiovascularMedical: hepatic disorder; cardiovasculardisorder; renal failure; hypertension (blooddisorder; renal failure; hypertension (bloodpressurepressure 44 150/100 untreated or any treated150/100 untreated or any treatedhypertension); endocrine disorder (excludinghypertension); endocrine disorder (excludingcorrected hypothyroidism).corrected hypothyroidism).

17 517 5



M O OR E E T A LM O OR E E T A L

(d)(d) Pharmacological: medication (corticosteroids,Pharmacological: medication (corticosteroids,antihypertensives); alcohol dependence orantihypertensives); alcohol dependence ormisuse; illicit druguse or solventmisuse.misuse; illicitdrug use or solvent misuse.

(e)(e) Radiological: metalimplants.Radiological: metal implants.

ACKNOWLEDGEMENTSACKNOWLEDGEMENTS

We wish to thank Mrs M.Cheek for patiently typingWe wish to thank Mrs M.Cheek for patiently typingthis manuscript, all the staff at the Department of this manuscript, all the staff at the Department of Neuroradiology, Newcastle General Hospital, forNeuroradiology, Newcastle General Hospital, fortheir time and courtesy when scanning the subjectstheir time and courtesy when scanning the subjectsand, finally, Newcastle City HealthTrust for a grantand, finally, Newcastle City HealthTrust for a grantto cover the costs of MRI.to cover the costs of MRI.

REFERENCESREFERENCES

American Psychiatric Association (1994)American Psychiatric Association (1994) DiagnosticDiagnosticand Statistical Manual of Mental Disordersand Statistical Manual of Mental Disorders (4th edn)(4th edn)(DSM^ IV).Washington, DC: APA.(DSM^ IV).Washington, DC: APA.

Altshuler, L. L., Curran, J. G., Hauser, P.,Altshuler, L. L., Curran, J. G., Hauser, P., et al et al (1995)(1995)TT22 hyperintensities in bipolar disorder: magnetichyperintensities in bipolar disorder: magneticresonance imaging comparison and literature meta-resonance imaging comparison and literature meta-analysis.analysis. American Journal of PsychiatryAmerican Journal of Psychiatry ,, 152152 , 1139^1194.,1139^1194.

Beck, A. T.,Ward, C. H., Mendelson, M.,Beck, A. T.,Ward, C. H., Mendelson, M., et al et al (1961)(1961)An inventory for measuring depression.An inventory for measuring depression. Archives of Archives of General PsychiatryGeneral Psychiatry ,, 44, 561^585., 561^585.

Brown, F.W., Lewine, R. J., Hudgins, P. A.,Brown, F.W., Lewine, R. J., Hudgins, P. A., et al et al (1992)(1992)White matter hyperintensity signals in psychiatric andWhite matter hyperintensity signals in psychiatric andnonpsychiatric subjects.nonpsychiatric subjects. American Journal of PsychiatryAmerican Journal of Psychiatry ,,149149, 520^525., 520^525.

Cole, A. J., Scott, J., Ferrier, I. N.,Cole, A. J., Scott, J., Ferrier, I. N., et al et al (1993)(1993)Patterns of treatment resistance in bipolar affectivePatterns of treatment resistance in bipolar affectivedisorder.disorder. Acta Psychiatrica ScandinavicaActa Psychiatrica Scandinavica ,, 8888 , 121^123., 121^123.

Dupont, R. M., Jernigan,T.L., Gillin, J.C .,Dupont, R. M., Jernigan,T.L., Gillin, J. C., et al et al (1987)(1987)Subcortical signal hyperintensities in bipolar patientsSubcortical signalhyperintensities in bipolar patientsdetected by MRI.detected by MRI. Psychiatric ResearchPsychiatric Research ,, 2121, 357^358., 357^358.

_ _ ,, _ _ , Butters, N.,, Butters, N., et al et al (1990)(1990) SubcorticalSubcorticalabnormalities detected in bipolar affective disordersabnormalities detected in bipolar affective disordersusing magnetic resonance imaging: clinical andusing magnetic resonance imaging: clinical andneuropsychological significance.neuropsychological significance. Archives of GeneralArchives of GeneralPsychiatryPsychiatry ,, 4747, 55^59., 55^59._ _

, Butters, N., Schafter, K.,, Butters, N., Schafter, K., et al et al (1995(1995 aa )) DiagnosticDiagnosticspecif icity of focal white matter abnormalities in bipolarspecificity of focal white matter abnormalities in bipolarand unipolar mood disorder.and unipolar mood disorder. Biological PsychiatryBiological Psychiatry ,, 3838 ,,482^486.482^486._ _

, Jernigan,T.L. , Heindel,W.,, Jernigan,T. L., Heindel,W., et al et al (1995(1995 bb )) MagneticMagneticresonance imaging and mood disorders. Localisation of resonance imaging and mood disorders. Localisation of white matter and other subcortical abnormalities.white matter and other subcortical abnormalities.Archives of General PsychiatryArchives of General Psychiatry ,, 5252 , 747^755., 747^755.

Fazekas, F., Chawluk, J. B., Alavi, A.,Fazekas, F., Chawluk, J. B., Alavi, A., et al et al (1987)(1987) MRMRsignal abnormalities at 1.5 T in Alzheimer's dementia andsignal abnormalities at 1.5 T in Alzheimer's dementia andnormal ageing.normal ageing. American Journal of NeuroradiologyAmerican Journal of Neuroradiology ,, 88,,421^426.421^426._ _

, Kleinert, R., Offenbacher, H.,, Kleinert, R., Offenbacher, H., et al et al (1993)(1993)Pathologic correlates of incidental MRI white matterPathologic correlates of incidental MRI white mattersignal hyperintensities.signal hyperintensities. NeurologyNeurology,, 4343, 1683^1689., 1683^1689.

Figiel, G. S. , Krishnan, K. R. R., Rao,V. P.,Figiel,G. S., Krishnan, K. R. R., Rao, V. P., et al et al (1991)(1991)Subcortical hyperintensities on brain magneticSubcortical hyperintensities on brain magneticresonance imaging: a comparison of normal and bipolarresonance imaging: a comparison of normal and bipolarsubjects.subjects. Journal of Neuropsychiatry and Clinical Journal of Neuropsychiatry and ClinicalNeuroscienceNeuroscience ,, 33,18^22.,18^22.

Folstein, M.F., Folstein, S. E. & McHugh, P. R. (1975)Folstein, M. F., Folstein, S. E. & McHugh, P. R. (1975)A practical method for grading the cognitive state of A practical method for grading the cognitive state of patients for the clinician.patients for the clinician. Journal of Psychiatric Research Journal of Psychiatric Research ,,1212, 189^198., 189^198.

Hickie, I., Scott, E., Mitchell, P.,Hickie, I., Scott, E., Mitchell, P., et al et al (1995)(1995) SubcorticalSubcorticalhyperintensities on magnetic resonance imaging: clinicalhyperintensities on magnetic resonance imaging: clinicalcorrelates and prognostic significance in patients withcorrelates and prognostic significance in patients withsevere depression.severe depression. Biological PsychiatryBiological Psychiatry ,, 3737, 151^160.,151^160.

Leifer, D., Buonanno, F. S. & Richardson, E. P. (1990)Leifer, D., Buonanno, F. S. & Richardson, E. P. (1990)Clinicopathologic correlations of cranial magneticClinicopathologic correlations of cranial magneticresonance imaging of periventricular white matter.resonance imaging of periventricular white matter.NeurologyNeurology ,, 4040, 911^918., 911^918.

McDonald,W. M., Krishnan, K. R. R., Doraiswamy,McDonald,W. M., Krishnan, K. R. R., Doraiswamy,P. M.,P. M., et al et al (1991)(1991) Occurrence of subcorticalOccurrence of subcorticalhyperintensities in elderly subjects with mania.hyperintensities in elderly subjects with mania.Psychiatric Research: NeuroimagingPsychiatric Research: Neuroimaging ,, 4040 , 211^220., 211^220.

McGlashan,T.H. (1984)McGlashan,T.H. (1984) The Chestnut Lodge Follow-The Chestnut Lodge Follow-

up Study I, Follow up methodology and study samples.up Study I,Follow up methodology and study samples.Archives of General PsychiatryArchives of General Psychiatry ,, 4141, 573^585., 573^585.

O'Brien, J., Desmond, P., Ames, D.,O'Brien, J., Desmond, P., Ames, D., et al et al (1996)(1996) AAmagnetic resonance imaging study of white mattermagnetic resonance imaging study of white matterlesions in depression and Alzheimer's disease.lesions in depression and Alzheimer's disease. BritishBritish Journal of Psychiatry Journal of Psychiatry ,, 168168 , 477^485., 477^485._ _

, Ames,D.,Chiu, E.,, Ames,D.,Chiu, E., et al et al (1998)(1998) Severe deep whiteSevere deep whitematter lesions and outcome in elderly patients withmatter lesions and outcome in elderly patients withmajor depressive disorder: follow-up study.major depressive disorder: follow-up study. BritishBritishMedical JournalMedical Journal ,, 317317, 982^984., 982^984.

Puzynski, S., Beresewicz, M., Koszewska, I.,Puzynski, S., Beresewicz, M., Koszewska, I., et al et al

(1995)(1995) Cortical and subcortical brain changes in bipolarCortical and subcortical brain changes in bipolaraffective disorders: a magnetic resonance imaging study.affective disorders: a magnetic resonance imaging study.European NeuropsychopharmacologyEuropean Neuropsychopharmacology ,, 55, 270., 270.

Scott, J., Barker,W. A. & Eccleston, D. (1988)Scott, J., Barker,W. A. & Eccleston, D. (1988) TheTheNewcastle Chronic Depression Study. PatientNewcastle Chronic Depression Study. Patientcharacteristics and factors associated with chronicity.characteristics and factors associated with chronicity.British Journal of PsychiatryBritish Journal of Psychiatry ,, 152152 , 28^33., 28^33.

Stephens, M. (1980)Stephens, M. (1980) Tests based on EDF statistics.InTests based on EDF statistics.InGoodness-of-fit TechniquesGoodness-of-fit Techniques (eds R. B.D'Agostino & M. A.(eds R. B. D'Agostino & M. A.Stephens), pp. 97^105. NewYork: Marcel Dekker.Stephens), pp. 97^105. NewYork: Marcel Dekker.

Strakowski, S. M.,Woods, B. T.,Tohen, M.,Strakowski, S. M.,Woods, B. T.,Tohen, M., et al et al

(1993)(1993) Subcortical signalhyperintensities in mania at firstSubcortical signal hyperintensities in mania at firsthospitalization.hospitalization. Biological PsychiatryBiological Psychiatry ,, 3333, 204^206., 204^206.

Swayze, I. I. V.W., Andreasen, N.C. , Alliger, R. J.,Swayze, I. I. V.W., Andreasen, N.C. , Alliger, R. J., etet

al al (1990)(1990) Structural brain abnormalities in bipolarStructural brain abnormalities in bipolaraffective disorder.Ventricular enlargement and focalaffective disorder.Ventricular enlargement and focal

signal hyperintensities.signalhyperintensities. Archives of General PsychiatryArchives of General Psychiatry ,, 4747,,1054^1059.1054^1059.

Sze, G., De Armond, S. J., Brant-Zawadzki, M.,Sze, G., De Armond, S. J., Brant-Zawadzki, M., et al et al

(1986)(1986) Fociof MRI signal (pseudo lesions) anterior to theFoci of MRIsignal (pseudo lesions) anterior to thefrontal horns: histologic correlations of a normal finding.frontal horns: histologic correlations of a normal f inding.American Journal of NeuroradiologyAmerican Journal of Neuroradiology ,, 77, 381^387., 381^387.

Woods, B. T., Yurgelun-Todd, D., Benes, F. M.,Woods, B. T., Yurgelun-Todd, D., Benes, F. M., et al et al

(1990)(1990) Progressive ventricular enlargement inProgressive ventricular enlargement inschizophrenia: comparison to bipolar affective disorderschizophrenia: comparison to bipolar affective disorderand correlation with clinical course.and correlation with clinical course. Biological PsychiatryBiological Psychiatry ,,2727, 341^352., 341^352._ _

,,_ _

, Mikulis, D.,, Mikulis, D., et al et al (1995)(1995) Age-related MRIAge-related MRIabnormalities in bipolar illness: a clinical study.abnormalities in bipolar illness: a clinical study. BiologicalBiologicalPsychiatryPsychiatry ,, 3838 , 846^847., 846^847.

17 617 6

CLINICAL IMPLICATIONSCLINICAL IMPLICATIONS

&& Thepresence of deep subcortical whitematter lesions (DWMLs) on MRIinbipolarThepresence of deep subcortical whitematter lesions (DWMLs) on MRIinbipolarpatients indicates a poorer clinical outcome.patients indicates a poorer clinical outcome.

&& Wewould suggest that any patient found to have DWMLs early in the course of Wewould suggest that any patient found to have DWMLs early in the course of illness might benefit fromintense therapeutic input to optimise outcome.illness might benefit fromintense therapeutic input to optimise outcome.

&& When examining a research paper on bipolar disorder, the clinical features of the When examining a research paper on bipolar disorder, the clinical features of thepatient group require careful scrutiny.patient group require careful scrutiny.

LIMITATIONSLIMITATIONS

&& The difficulties of identifying and recruiting poor outcomebipolar patients meansThe difficulties of identifying and recruiting poor outcomebipolar patients means that the study is necessarily of limited size and should be replicated by a larger multi- that the study is necessarily of limited size and should be replicatedby a larger multi-centre study.centre study.

&& There is no consensus on reliable and meaningful definitions of outcome in bipolarThere is no consensus on reliable and meaningful definitions of outcome in bipolardisorder.disorder.

&& The study requires replication.The study requires replication.

P. BRIAN MOORE PhD, DEBRA J. SHEPHERD,MB,DONALD ECCLESTON,MD, IAIN C. MACMILLAN,P. BRIAN MOORE PhD,DEBRA J. SHEPHERD,MB,DONALD ECCLESTON, MD, IAIN C. MACMILLAN,MRCPsych,UPTALGOSWAMI, MD,VICTOR L. MMRCPsych,UPTALGOSWAMI, MD,VICTOR L. M CC ALLISTER, FRCR, I.NICOL FERRIER, MD, Queen ElizabethALLISTER,FRCR, I.NICOL FERRIER, MD,Queen ElizabethHospital,Gateshead,UKHospital,Gateshead,UK

Correspondence:P.B. Moore,Tranwell Unit,Queen Elizabeth Hospital,Windy Nook Road,GatesheadCorrespondence:P.B. Moore,Tranwell Unit,Queen Elizabeth Hospital,Windy Nook Road,GatesheadNE9 6SX,UKNE9 6SX,UK

(First received 1 February 2000, f inal revision 27 June 200 0, accepted 28 June 200 0)(First received 1 February 2000, f inal revision 27 June 200 0, accepted 28 June 200 0)

bipolar disorder (nice)

Documents