Anda di halaman 1dari 5

B R I T I S H J O U R N A L O F P S YC H I AT RY ( 2 0 0 1 ) , 1 7 8 , 1 7 2 ^ 1 7 6

Cerebral white matter lesions in bipolar affective criteria for inclusion in either the good or
the poor outcome groups were eliminated
from the study and did not undergo MRI.
disorder: relationship to outcome
P. BRIAN MOORE, DEBRA
DEBR A J. SHEPHERD, DONALD ECCLESTON, Subjects
IAIN C. MACMILLAN, UPTAL GOSWAMI, VICTOR L. McALLISTER Patients with a DSM±IV (American Psy-
and I. NICOL FERRIER chiatric Association, 1994) diagnosis of bi-
polar affective disorder aged 20±65 years
were identified from the in-patient and
out-patient services of the Newcastle
Mental Health Trust. To be included, all
Background Twenty per cent of Chronic affective disorders are a major patients had been diagnosed as suffering
patients with bipolar affective disorder cause of psychiatric morbidity (Scott et al,
al, from bipolar disorder for at least 3 years
1988), affecting up to 1% of the popu- and had had a minimum of two episodes
suffer anillness that respondsinadequately
lation. Studies of bipolar patients have of illness. After identification, their case
to treatment and has a poor outcome. identified a group whose illnesses respond notes were reviewed. Those meeting the cri-
Many patients, but not all, with bipolar poorly to treatment and have a poor out- teria for bipolar disorder (DSM±IV) were
disorder show white matter abnormalities come, with 20% remaining unwell for up subjected to a full clinical evaluation by
on T2-weighted magnetic resonance to a year (Cole et al,al, 1993). Treatment an experienced psychiatrist (I.C.M. or
resistance might be anticipated in individ- P.B.M.), who ensured that they met the cri-
imaging (MRI).
uals whose illness had specific neurobio- teria for inclusion but not the exclusion
logical underpinnings. In bipolar subjects, criteria for the study. A full history, includ-
Aims To explore the hypothesis that
T2-weighted magnetic resonance imaging ing family history of mental illness,
white matter abnormalities on MRI are (MRI) has revealed deep subcortical white smoking history and obstetric history, was
seen more frequently in subjects whose matter lesions (DWML) and periventricular compiled and old case notes were reviewed.
illness has a poor outcome compared with white matter lesions (PVWML), which To screen out subjects with considerable
those with a good outcome or controls. have differing neuropathologies (Fazekas cognitive decline (early dementia or
et al,
al, 1993). Although not all MRI studies pseudodementia), a Mini-Mental State
Method Two groups of age- and find an excess of white matter lesions Examination (MMSE; Folstein et al, al, 1975)
gender-matched patients with bipolar (Brown et al, al, 1992; Strakowski et al, al, was completed. This instrument is insuffi-
1993), the majority do (Dupont et al, al, ciently sensitive to detect subtle cognitive
disorder (14 with a good outcome and15
1987, 1990, 1995a
1995a,b; Swayze et al,
al, 1990; changes, for example those reported in
with a poor outcome) and15 controls, Figiel et al,
al, 1991; McDonald et al,al, 1991; patients with depression, thereby prevent-
aged 20^65 years, were studied. Axial T2- Puzynski et al,
al, 1995; Woods et al,
al, 1995), ing inappropriate exclusion of patients.
weighted MRI scans were examined for and meta-analysis indicates that more After a complete description of the study
bipolar patients than controls exhibit of the patients, written informed consent
the presence and severity of white matter
WMLs (Altshuler et al, al, 1995). Specific, was obtained. Thereafter, subjects were allo-
abnormalities. but as yet unidentified subgroups, may ex- cated to either a good or poor outcome
hibit WMLs. Dupont et al (1990) reported group (see below) or eliminated because
Results Significantly more poor
that patients showing WMLs were more of an intermediate outcome. Immediately
outcome group members had deep seriously unwell and performed poorly on prior to MRI scanning, the patient was in-
subcortical punctate, but not neuropsychological testing. terviewed to assess current mental state
periventricular, white matter and a Beck Depression Inventory (BDI;
hyperintensities than the good outcome Beck et al,
al, 1961) was completed by the
Hypothesis
subject.
group (P
(Pˆ0.035)
0.035) or controls (P
(Pˆ0.003)
0.003) These observations lead us to predict that
and these abnormalities were of greater poor outcome bipolar patients have under-
severity (P
(Pˆ0.030
0.030 and P50.014, lying structural brain abnormalities. It is Patient selection criteria
our hypothesis that poor outcome bipolar These are set out in the Appendix.
respectively).
subjects have more white matter abnor-
Conclusions Subcortical white matter malities on MRI than either good outcome
patients or controls. Poor outcome group (B)
lesions are associated with poor outcome
At the time of study, patients in this group
bipolar disorder.
METHOD had been unwell for 2 years or more,
Declaration of interest This despite adequate therapy. Any periods of
To explore our hypothesis we have com- remission lasted 8 weeks or less, during
research was supported financially by
pared T2-weighted MRI scans of poor and which time they still had significant func-
Newcastle City HealthTrust. good outcome bipolar patients with healthy tional impairment. All patients had made
controls. Patients who did not meet the a poor response to lithium. Almost all were

17 2
W H I T E M AT T E R L E S I ON S IN
I N B I P OL 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 the Table 1 Characteristics of patient groups and controls
treatment of affective disorder.
Good outcome group Poor outcome group Controls

Good outcome group (A) n (m, f) 14 (8, 6) 15 (7, 8) 15 (7, 8)


At the time of study, these patients had Age, years (s.d.) 47.4 (10.10) 42.1 (13.9) 41.9 (12.6)
been clinically euthymic for at least 8 eight Length of education, years (s.d.) 12.7 (3.0) 10.9 (3.1) 14.0 (1.1)
weeks. After any episode of illness, they % Right-handed 92% 80% 80%
had shown full symptomatic recovery and Non-Caucasian 1 1 0
had returned to normal premorbid func- Rapid cycling (4
(4 4 episodes/year) 0 7 ^
tioning. The good and poor outcome Family history of bipolar affective disorder 4 4 0
groups were age and gender matched. Age of illness onset, years (s.d.) 31.4 (3.5) 26.3 (10.0) ^
Illness duration, years (s.d.) 16.0 (7.9) 15.8 (10.8) ^
No. of episodes of illness, mean (s.d.) 8.6 (6.0) 8.9 (4.3) ^
Controls
No. of hospitalisations, mean (s.d.) 2.4 (1.9) 5.0 (4.7)1 ^
Controls were recruited from the families MMSE scores, mean (s.d.) 29.0 (1.0) 28.0 (2.3) 30 (0)
of patients and the staff of Newcastle Uni-
BDI scores, mean (s.d.) 9.0 (6.7) 16.0 (7.0)1 ^
versity and its associated hospitals. Con-
trols were selected to match, as closely as MMSE, Mini-Mental State Examination; BDI, Beck Depression Inventory.
1. P50.01, Student's t-test v. good outcome.
possible, the ages and genders of subjects
in the poor outcome group. No controls
met the exclusion criteria applied to the Magnetic resonance imaging or non-parametric methods: for parametric
patient groups. All subjects were examined using a General data, ANOVA, ANCOVA, Student's t-test
Electrics (Slough, UK) MR max plus 0.5 and Pearson product moment coefficient
tesla scanner. Sagittal, coronal and axial were calculated; for non-parametric data,
Patient characteristics T1-weighted scan sequences, axial T2- including white matter hyperintensity
weighted scan sequences and coronal inver- gradings, w2, Mann±Whitney and Spearman
In total, 17 patients were recruited into the
sion recovery scan sequences were recorded rank correlations were calculated using the
poor outcome group and 16 into the good
on all subjects. The T1-weighted and inver- computerised statistical package Minitab
outcome group. One patient was excluded
sion recovery scans were examined by a 10.2 for Windows (Minitab Inc., PA).
because he experienced a panic attack
during MRI scanning and the scanning consultant neuroradiologist (V.L.M.) for
sequence could not be completed. Three the presence of structural abnormalities.
patients were eliminated after the scans Up to 15 axially oriented 7-mm slices sep- RESULTS
were examined because of evidence of de- arated from each other by 1 mm were
myelination (n(nˆ1),
1), an Arnold±Chiari mal- recorded during the T2-weighted sequence. The principal findings of the study are
formation with hydrocephalus (n (nˆ1)
1) and Axial T2-weighted scan sequences with shown in Table 2. Comparing the numbers
cerebrovascular infarcts (n
(nˆ1).
1). A total of a relaxation time of 2300 ms and echo of subjects with DWMLs, significantly
29 patients completed the study, 15 in the delays of 25 and 100 ms were examined more (odds ratioˆ11.4)
ratio 11.4) patients (7/15,
poor outcome and 14 in the good outcome for the presence of white matter hyper- 47%) in the poor outcome group had
group. intensities. To be counted, WMLs had to DWMLs than in the good group (1/14,
The patient characteristics of the good be present on both the proton density and 7%, Fisher's exact test, Pˆ0.035)
0.035) or con-
(A) and poor (B) outcome groups and con- T2 scans. Two investigators (D.J.S. and trols (0/15, 0%, Pˆ0.003).
0.003). The grades
trols are compared in Table 1. Good out- V.L.M.) examined scans independently (severity) of DWMLs in the poor outcome
come patients tend to be older, better and blindly and their results were in com- group significantly exceeded (Mann±Whitney
educated and less likely to be left handed, plete agreement. White matter lesions were U-test) the good outcome patients
but the differences do not reach signifi- classified as either deep subcortical or peri- (Pˆ0.030)
0.030) or controls (P(P50.014). In con-
cance. The illness duration, gender ratios, ventricular and images were graded 0±3 trast, the grade of DWMLs in the good
ages and prevalence of a family history of using the scheme proposed by Fazekas patient group and controls did not differ
bipolar disorder are similar in both groups. (Fazekas et al,
al, 1987, 1993). significantly (P
(Pˆ0.96).
0.96).
The major difference between the groups In contrast, PVWMLs were not present
arises in the severity of illness at the time in strikingly different numbers in the good
of investigation. All good outcome patients Statistical analysis (8/14, 57%), poor (10/15, 67%) or control
were euthymic on the day of scanning, Prior to analysis, all data were checked (7/15, 47%) groups. Fisher's exact test
reflected in mean BDI scores of 9.0, for deviation from normality using the failed to reveal any differences between
s.d.ˆ6.7.
s.d. 6.7. In contrast, the poor outcome Anderson±Darling method (Stephens, the numbers in each group with PVWMLs.
group had significantly higher (P (P50.05, 1980). The MMSE scores, age of onset of Although more patients in the poor out-
unpaired Student's t-test) BDI scores of illness and duration of education differed come group had more severe abnormalities,
16.0, s.d.ˆ7.0,
s.d. 7.0, and several of this group significantly from normality. All data were Mann±Whitney tests failed to reveal any
showed clinical mood disturbances. analysed using the appropriate parametric significant differences between the groups

17 3
MOO
M OO R E E T A L

T
Table
able 2 White matter abnormalities on magnetic resonance imaging in bipolar subjects and controls subcortical hyperintensities varied from
55% to 18%, whereas the percentages in
Poor outcome Good outcome Controls normal controls varied from 17% to 0%.
This variation could have resulted from sev-
group group
eral factors. In particular, risk factors for
n 15 14 15 the formation of hyperintensities, especially
Deep punctate white matter lesions advancing ageing and hypertension, may
Fazekas grade: 0 8 13 15 not have been controlled for sufficiently
rigorously. However, our study suggests
1 4 0 0
that a major cause of the variation is likely
2 3 1 0
to be the inclusion of poor prognosis pa-
3 0 0 0
tients in the study. Studies based on hospi-
P (Mann^Whitney) compared with controls 50.014 0.96 tal populations tend to overrepresent the
Periventricular white matter lesions more chronic patients with long illness
Fazekas grade: 0 5 6 8 histories who suffer frequent relapse. It is
1 6 8 6 notable that the study of Strakowski et al
2 4 0 1 (1993), which failed to find significant dif-
3 0 0 0 ferences between controls and patients,
P (Mann^Whitney) compared with controls 0.16 0.74 considered patients only in their first epi-
sode of manic illness. This factor alone is
likely to reduce significantly the propor-
(good v. poor, Pˆ0.22;
0.22; good v. controls, Hyperintense ventricular caps and smooth tions of poor outcome patients included in
Pˆ0.74;
0.74; poor v. controls, Pˆ0.16).
0.16). periventricular halos may result from their study because only approximately
It is concluded that DWMLs but not myelin pallor and are associated with the 10±20% of all bipolar patients go on to
PVWMLs are linked significantly to poor loss of ependymal lining of the ventricles develop a treatment-resistant disorder (Cole
outcome in bipolar disorder. and local increase in tissue fluid. In con- et al,
al, 1993). Furthermore, Strakowski's
Obstetric, smoking and family histories trast, extensive irregular PVWMLs have a study excluded patients with a family his-
of mental disorder were collected. A de- similar aetiology to punctate lesions tory of bipolar disorder, thereby elimin-
tailed statistical analysis of the data is (Fazekas et al,
al, 1993). Hence, analysis of ating those with a genetic predisposition
not warranted because of the small sample white matter abnormalities in diseased that could be associated with the develop-
size. However, inspection of the data states needs to differentiate clearly between ment of organic lesions. Dupont et al
failed to suggest that, for an individual, these lesions. Although our study shows (1990) noted that bipolar patients with
there was a clear link between WMLs that there is an association between deep WMLs had significantly more hospitalisa-
and these histories. A larger study would punctate and periventricular lesions in our tions, higher Hamilton Rating Scale for
be required to explore correlations between patient group (Spearman: Pˆ0.39, 0.39, Pˆ Depression scores and received more neuro-
these variables. 0.04), only the deep punctate lesions are leptics than patients without WMLs. All of
associated with a poor outcome bipolar dis- these measures of poor outcome may con-
DISCUSSION order. A similar finding has been reported firm the current finding that WMLs are as-
in elderly patients with depression ± in this sociated with a poorer prognosis. This is
Deep white matter lesions group, DWMLs but not PVWMLs are asso- further suggested by the observation of a
The principal finding of this study is that ciated with depression (O'Brien et al, al, follow-up 1 year later, that approximately
DWMLs are seen most frequently in bi- 1996). This difference probably reflects half of Dupont's patients with WMLs failed
polar patients with a poor outcome when the differing neuropathological bases of to recover. Further support for this hypoth-
compared with those who have a good out- these abnormalities. The present results esis is provided by studies in elderly, uni-
come or with normal controls. In contrast, suggest that a factor associated with poor polar subjects that have linked DWMLs to
the frequencies of PVWML hyperintensities outcome in bipolar disorder is the presence outcome. Individuals with DWMLs are re-
do not differ significantly between these of microvascular abnormalities in the ported to respond less well to in-patient
groups. Although the pathology of WMLs cerebral white matter. treatment, including electroconvulsive ther-
in bipolar disorder is not known, they A number of studies have reported apy (Hickie et al,al, 1995), and to relapse
may be direct consequences of the disease WMLs in bipolar patients but many fail more rapidly (O'Brien et al,
al, 1998).
process, they may result from independent to differentiate periventricular from sub- Although we have classified outcome
processes that then predispose to treatment cortical punctate lesions. White matter on the basis of a clinical assessment and
resistance or they may simply be coinciden- lesions have been reported more frequently level of functioning, others, for example
tal findings. Clinicopathological studies in in young bipolar patients (aged 560 years) McGlashan (1984) in his outcome studies
elderly subjects indicate that DWMLs and than controls by Figiel et al (1991), Swayze at Chestnut Lodge, have described out-
PVWMLs have different aetiologies. Discrete et al (1990) and Dupont et al (1987, 1990), come on a more systematic basis using five
and confluent subcortical punctate WMLs whereas Strakowski et al (1993) failed to dimensions: hospitalisation, employment,
probably arise from perivascular abnormal- find any difference for newly diagnosed social activity, symptoms and global func-
ities and in severe cases may indicate the patients with mania. In these studies, the tioning, rated 0±4. With the exception of
presence of areas of microcystic infarct. percentages of bipolar patients with social activity, these were recorded for

17 4
W H I T E M AT T E R L E S I ON S IN
I N B I P OL 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- drugs and yet had markedly different appears to be multi-factorial. Preliminary
come patients would score 0 or 1, 0, no DWML frequencies, the good group being investigations suggest that cerebral dys-
data, 0 and 0, respectively, on each of similar to the controls, who were drug free. rhythmias (Cole et al,al, 1993) are associated
these dimensions. In contrast, the corre- Only patients received medication, yet with treatment resistance. These observa-
sponding scores of our good outcome PVWMLs were found in considerable tions may link to the growing evidence that
patients would be 4, 3±4 (including house- numbers in the controls. poor outcome patients respond much better
work as employment), no data, 3±4 and 4, With the exception of sodium valproate to anticonvulsants than to lithium. Some of
respectively. This reflects the study design usage, the most striking differences between the poor outcome treatment-resistant
to maximise the contrast between good the groups was the number of patients in patients were clinically unwell at the time
and poor outcome patient groups. the poor outcome compared to the good of scanning and it is possible that differ-
outcome group receiving neuroleptics (6 v. ences between the good and poor outcome
Periventricular white matter 2), antidepressants (6 v. 2) or anticholiner- groups reflect differences in current mental
lesions gics (4 v. 0). There are insufficient patients state. However, there is evidence (Dupont
receiving different drugs to warrant a statis- et al,
al, 1990) that WMLs observed in bipolar
The frequency of PVWMLs in our study did
tical analysis of the data, especially in view patients are stable over time. Nevertheless,
not differ significantly between the good
of the pharmacological heterogeneity of the it would seem prudent to re-examine
and the poor outcome groups and normal
different neuroleptics or antidepressants. patients with DWMLs in the future when
controls. The frequency of all PVWMLs
There is, however, little evidence in the they are clinically euthymic to establish
was high, at 57%, 67% and 47%, respec-
literature that either neuroleptic or anti- whether the lesions reflect the clinical state
tively, for the good and poor prognosis
depressant usage results in WMLs. of the patient or more fundamental abnorm-
groups and controls. These rates are com-
alities of cerebral architecture.
parable with other studies (Altshuler et al,al,
1995). However, interpretation of the Age and white matter lesions
presence of PVWMLs is problematic. APPENDIX
It has been emphasised that age is an
Studies (Sze et al,
al, 1986; Leifer et al,
al, 1990) Patient selection criteria
important determinant of the prevalence
suggest that the hyperintense caps seen at
of WMLs. Woods et al (1990) reported that The inclusion criteria are:
the end of lateral ventricles on T2-weighted
the prevalence of both DWMLs and (a) DSM ^ IV bipolar affective disorder;
scans are a normal finding, as is the
DVWMLs in bipolar subjects but not in
presence of a thin white hyperintense line (b) age 20^65 years;
controls increased after the age of 30 years.
around the ventricle margins. When we (c) two episodes minimum and a history of illness for
In contrast, Altshuler et al (1995) reported
rated our scans with these appearances as at least 3 years.
that the prevalence of only PVWMLs in bi-
normal, the differences between the groups
polar subjects but not in controls increased Patients were then divided into two groups:
increased but nevertheless still failed to
after the age of 30 years. This difference
reach statistical significance at the 5% (a) Good outcome group (A):
may result from the lack of control for
level. Failure to find significantly different (i) return to premorbid level of functioning
cardiac risk factors in Woods' study, espe-
numbers of good and poor outcome between illness;
cially in view of the likelihood that sub-
patients and controls with PVWMLs may
cortical lesions are vascular in origin. In (ii) currently euthymic for at least 8 weeks;
be a consequence of the low power of a
our bipolar subjects, we found no signifi- (iii) if prescribed, good response to lithium.
study based on 44 subjects. We are un-
cant effect of age upon the prevalence of
certain whether some PVWMLs are related (b) Poor outcome group (B):
DWMLs (ANCOVA: Fˆ0.02, 0.02, Pˆ0.88),
0.88),
to a poor outcome, and further studies are (i) symptomatic non-recovery for at least the
whereas the PVWMLs increased with age
indicated. past 2 years;
(ANCOVA: Fˆ5.57, 5.57, Pˆ0.007).
0.007). The mean
age of our poor outcome group (42.1, (ii) any period of well-being in the past 2 years
Medication s.d.ˆ13.9
s.d. 13.9 years) was comparable to that of for 8 weeks or less;
The medication taken by good and poor the controls (41.9, s.d.ˆ12.6
s.d. 12.6 years) but less (iii) failure to regain premorbid functioning
outcome groups was similar and reflected than that of the good outcome group (47.4, during periods of well-being.
the recent trend in bipolar management to s.d.ˆ10.1
s.d. 10.1 years). Hence, it is improbable
The exclusion criteria are:
focus upon treatment with mood stabili- that differences in the ages of groups could
sers. Thus, 10, 6 and 0 of the 14 good contribute significantly to the increased pre- (a) Psychiatric: evidence of cognitive decline; other
Axis I comorbid condition; bipolar disorder
outcome patients received lithium, car- valence of DWMLs in the poor outcome
other than type I or type II; learning disabilities.
bamazepine and/or valproate, respectively, groups. Age similarities rather than group
whereas the corresponding numbers were differences (F
(Fˆ2.73,
2.73, Pˆ0.08)
0.08) account for (b) Neurological: cerebrovascular disease; neuro-
degenerative disorders; head injury with concus-
12, 8 and 5 for the 15 poor outcome the similarities in the prevalence of
sion; epilepsy; idiopathic parkinsonism; systemic
patients. Four patients in the good outcome PVWMLs in patient groups and controls. illness with cerebral consequences; focal neuro-
group received lithium/anticonvulsant The presence of DWMLs distinguishes logical signs on examination.
combinations whereas nine poor outcome poor outcome from good outcome patient
(c) Medical: hepatic disorder; cardiovascular
patients received this combination. groups and controls. Nevertheless, these
disorder; renal failure; hypertension (blood
It is improbable that lithium will give abnormalities were found in only 47% of pressure 4150/100 untreated or any treated
rise to WMLs. Good and poor outcome the poor prognosis patient group. The aeti- hypertension); endocrine disorder (excluding
groups had broadly similar usage of these ology of treatment resistance therefore corrected hypothyroidism).

17 5
MOO
M OO R E E T A L

(d) Pharmacological: medication (corticosteroids,


antihypertensives); alcohol dependence or
misuse; illicit drug use or solvent misuse. CLINICAL IMPLICATIONS

(e) Radiological: metal implants.


& The presence of deep subcortical white matter lesions (DWMLs) on MRI in bipolar

patients indicates a poorer clinical outcome.


ACKNOWLEDGEMENTS
& We would suggest that any patient found to have DWMLs early in the course of
We wish to thank Mrs M. Cheek for patiently typing
illness might benefit from intense therapeutic input to optimise outcome.
this manuscript, all the staff at the Department of
Neuroradiology, Newcastle General Hospital, for & When examining a research paper on bipolar disorder, the clinical features of the
their time and courtesy when scanning the subjects
patient group require careful scrutiny.
and, finally, Newcastle City Health Trust
Trust for a grant
to cover the costs of MRI.
LIMITATIONS

REFERENCES & The difficulties of identifying and recruiting poor outcome bipolar patients means
that the study is necessarily of limited size and should be replicated by a larger multi-
American Psychiatric Association (1994) Diagnostic
and Statistical Manual of Mental Disorders (4th edn) centre study.
(DSM ^ IV).Washington, DC: APA.
& There is no consensus on reliable and meaningful definitions of outcome in bipolar
Altshuler, L. L., Curran, J. G., Hauser, P., et al (1995)
T2 hyperintensities in bipolar disorder: magnetic disorder.
resonance imaging comparison and literature meta-
analysis. American Journal of Psychiatry,
Psychiatry, 152,
152, 1139^1194. & The study requires replication.
Beck, A. T.,Ward,
T., Ward, C. H., Mendelson, M., et al (1961)
An inventory for measuring depression. Archives of
General Psychiatry,
Psychiatry, 4, 561^585.
Brown, F.W., Lewine, R. J., Hudgins, P. A., et al (1992)
White matter hyperintensity signals in psychiatric and
P. BRIAN MOORE PhD, DEBRA J. SHEPHERD, MB, DONALD ECCLESTON, MD, IAIN C. MACMILLAN,
nonpsychiatric subjects. American Journal of Psychiatry,
Psychiatry,
149,
149, 520^525. MRCPsych, UPTALGOSWAMI, MD,VICTOR L. MCALLISTER, FRCR, I. NICOL FERRIER, MD, Queen Elizabeth
Hospital,Gateshead, UK
Cole, A. J., Scott, J., Ferrier, I. N., et al (1993)
Patterns of treatment resistance in bipolar affective
Correspondence: P. B. Moore,Tranwell Unit,Queen Elizabeth Hospital,Windy Nook Road,Gateshead
disorder. Acta Psychiatrica Scandinavica,
Scandinavica, 88,
88, 121^123.
NE9 6SX,UK
Dupont, R. M., Jernigan,T. L., Gillin, J. C., et al (1987)
Subcortical signal hyperintensities in bipolar patients (First received 1 February 2000, final revision 27 June 2000, accepted 28 June 2000)
detected by MRI. Psychiatric Research,
Research, 21,
21, 357^358.
_ , _ , Butters, N., et al (1990) Subcortical
abnormalities detected in bipolar affective disorders Hickie, I., Scott, E., Mitchell, P., et al (1995) Subcortical Scott, J., Barker,W. A. & Eccleston, D. (1988) The
using magnetic resonance imaging: clinical and hyperintensities on magnetic resonance imaging: clinical Newcastle Chronic Depression Study. Patient
neuropsychological significance. Archives of General correlates and prognostic significance in patients with characteristics and factors associated with chronicity.
Psychiatry,
Psychiatry, 47,
47, 55^59. severe depression. Biological Psychiatry,
Psychiatry, 37,
37, 151^160. British Journal of Psychiatry,
Psychiatry, 152,
152, 28^33.

_ (1995a) Diagnostic
, Butters, N., Schafter, K., et al (1995a Leifer, D., Buonanno, F. S. & Richardson, E. P. (1990) Stephens, M. (1980) Tests based on EDF statistics. In
specificity of focal white matter abnormalities in bipolar Clinicopathologic correlations of cranial magnetic Goodness-of-fit Techniques (eds R. B. D'Agostino & M. A.
and unipolar mood disorder. Biological Psychiatry,
Psychiatry, 38,
38, resonance imaging of periventricular white matter. Stephens), pp. 97^105. New York: Marcel Dekker.
482^486. Neurology,
Neurology, 40,
4 0, 911^918.
Strakowski, S. M.,Woods, B. T., Tohen, M., et al
_ (1995b) Magnetic
, Jernigan,T. L., Heindel,W., et al (1995b McDonald,
McDonald,W. W. M., Krishnan, K. R. R., Doraiswamy, (1993) Subcortical signal hyperintensities in mania at first
resonance imaging and mood disorders. Localisation of P. M., et al (1991) Occurrence of subcortical hospitalization. Biological Psychiatry,
Psychiatry, 33,
33, 204^206.
white matter and other subcortical abnormalities. hyperintensities in elderly subjects with mania.
Archives of General Psychiatry,
Psychiatry, 52,
52, 747^755. Psychiatric Research: Neuroimaging,
Neuroimaging, 40,
40, 211^220. Swayze, I. I.V. W., Andreasen, N. C., Alliger, R. J., et
al (1990) Structural brain abnormalities in bipolar
Fazekas, F., Chawluk, J. B., Alavi, A., et al (1987) MR
McGlashan, T. H. (1984) The Chestnut Lodge Follow- affective disorder.Ventricular enlargement and focal
signal abnormalities at 1.5 T in Alzheimer's dementia and
up Study I, Follow up methodology and study samples. signal hyperintensities. Archives of General Psychiatry,
Psychiatry, 47,
47,
normal ageing. American Journal of Neuroradiology,
Neuroradiology, 8,
Archives of General Psychiatry,
Psychiatry, 41,
41, 573^585. 1054^1059.
421^426.
O'Brien, J., Desmond, P., Ames, D., et al (1996) A Sze, G., De Armond, S. J., Brant-Zawadzki, M., et al
_ , Kleinert, R., Offenbacher, H., et al (1993)
Pathologic correlates of incidental MRI white matter magnetic resonance imaging study of white matter (1986) Foci of MRI signal (pseudo lesions) anterior to the
signal hyperintensities. Neurology,
Neurology, 43,
43, 1683^1689. lesions in depression and Alzheimer's disease. British frontal horns: histologic correlations of a normal finding.
Journal of Psychiatry,
Psychiatry, 168,
168, 477^485. American Journal of Neuroradiology,
Neuroradiology, 7, 381^387.
Figiel, G. S., Krishnan, K. R. R., Rao,V. P., et al (1991)
Subcortical hyperintensities on brain magnetic _ , Ames, D., Chiu, E., et al (1998) Severe deep white Woods, B. T.,Yurgelun-Todd, D., Benes, F. M., et al
resonance imaging: a comparison of normal and bipolar matter lesions and outcome in elderly patients with (1990) Progressive ventricular enlargement in
subjects. Journal of Neuropsychiatry and Clinical major depressive disorder: follow-up study. British schizophrenia: comparison to bipolar affective disorder
Neuroscience,
Neuroscience, 3, 18^22. Medical Journal,
Journal, 317,
317, 982^984. and correlation with clinical course. Biological Psychiatry,
Psychiatry,
27,
27, 341^352.
Folstein, M. F., Folstein, S. E. & McHugh, P. R. (1975) Puzynski, S., Beresewicz, M., Koszewska, I., et al
A practical method for grading the cognitive state of (1995) Cortical and subcortical brain changes in bipolar _ , _ , Mikulis, D., et al (1995) Age-related MRI

patients for the clinician. Journal of Psychiatric Research,


Research, affective disorders: a magnetic resonance imaging study. abnormalities in bipolar illness: a clinical study. Biological
12,
12, 189^198. European Neuropsychopharmacology,
Neuropsychopharmacology, 5, 270. Psychiatry,
Psychiatry, 38,
38, 846^847.

17 6

Anda mungkin juga menyukai