The Long-term Natural History of the Weekly
Symptomatic Status of Bipolar I Disorder
Lewis L. Judd, MD Hagop S. Ahishql,MD; PamelaJ . Schettler,PhD;JeanEnd,icott,PhD;Jach Maser, PhD;
David A. Solonton,MD; And.rewC. Leon, PhD;John A. Rice,PhD;Martin B. Keller, MD
Bcckground: To our knowledge, this is the firsr prospectivenatural history study of weekly s1'rnptomaticstatus of patiens with bipolar I disorder(BP-l) during longterm follow-up.
predominatedover manic/hypomanic symptoms (8.995
of weeks) or cycling/mixed symptoms (5.9o/oof weeks).
Subsyndromal,minor depressive,and hypomanic sy'rnptoms combined were nearly 3 times more frequent than
syndromal-levelmajor depressiveand manic symptoms
(29.9o/"vs I L2% of weeks, respectively).Patientswith
BPJ changedsymptom status an averageof 6 times per
year and polarity more than 3 times per year. Longer intake episodesand thosewith depression-onlyor cycling
polaritypredictedgreaterchronicity during long-termfo}Iow-up, as did comorbid drug-use disorder.
filcthods: Analysesare basedon ongoing prospective
follow-up of 146 patientswirh ResearchDiagnosticCriteriaBP-l. who enteredthe NationalInstituteof Mental
Health (Bethesda,Md) CollaborativeDepressionStudy
from 1978 through l98l. Weekly affectivesympromsratus ratings were analyzedby polarity and severity,ranging from as)'mptomatic,to subthresholdlevels, ro fullblou.n major depression and mania. Percentagesof
follovr'-upweeksat each level aswell as number of shifts
in sy'mptom status and polarity during rhe entire follorv-up period were examined.Finally, 2 new measures
of chronicity were evaluatedin relation to previously identified predictorsoI chronicity for BP-1.
Goncluglonr: The longitudinal weekly symptomatic
courseof BP-l is chronic. Overall, rhesyrnptomaticstructure is primarily depressiverather than manic, and subsyndromal and minor affectivesyrnptonrspredominare.
Syntptomseveritylevelsfluctuate,often within the same
patient over time. Bipolar I disorder is expressedas a dimensionalillness featuring the full range (spectrum) of
affectivesymptom severity and polarity.
Rcsults: Patientswith BP-I were symptomaticallyill
47 .3"/oolweeksthroughouta meanof 12.8yearsof followup. Depressivesymptoms (31.9o/o
of total fbllow-up weeks)
Florn thr Deparlntetils
oJ Psy chiatr r-, Ll niv er sit.y
oJ C aliJornia-San Diego,
(Drs ludd, Ahisha|, Schettler,
an<l lvltrser):Dcpartment of
R e s e a r c ha n d T r a i n i n g ,
C ol umbia LIniver nty, N ew
Yorlr, NY (Dr Endicott):
on<lDepartment oJ Pslchiatry,
Brorvn Unirersi ty, Providence,
RI (Drs ,Solonronand Keller),
C o rnell U nit,tr sitl', I thaca, NY
(Dr Le on). ancl W ashington
l l n i v e r s i l y , 5 t L o u i s ,M o
(.Dr Rrct').
Ar ch Gen Psychiatry. 2002;59:53 0-537
RAEPELIN'
uao described
manic-depressiveinsanity
as a cyclical illness. Until
recently,following his lead,
clinical and researchattention concerningmood disorderswas concentratedon the most severesyndromal
manifestations
of thesedisorders.ie. manic
and major depressiveepisodes(MDE).']e
However,recentevidencesuggeststhat the
concept of bipolar I disorder (BP-l) with
episode-freeperiodsof euth)'rniapunctuated by syndromalMDE and mania is inadequate.r0'r2
Analysesof weekly symptomatic statusduring the long-term course
o[anothermooddisorder,unipolarMDD,12
hasshown that, althoughthis illnesshastraditionaily beenexaminedprimarily in terms
of the onset, r'emission,and relapseo[
MDEs, minor and subsyndromaldepressive symptoms dominate its long-term
(REPRINTED) ARCH GEN PSYCHIATRY/VOL 59. IUNE 2002
530
course by nearly a 3;l ratio (43o/ovs l5o/o
of follow-up weeks). Patients with unipolar MDD were lound to be s;'rnptomatic
during 60oloof the follow-up period and to
experiencea changeablecoursein which
major, minor, and subsyndromaldepressivesymptomsalternatedwithin the same
patientover time. tn brief, unipolarMDD
is expressedlongitudinally as a dimensional illness involving the full spectrum
oI depressivesymptom severity.
This new understandingof the longt e r m s y m p t o m a t i cs t r u c t u r e o I u n i p o larity stimulated us to carry oul a similar analysisof the longitudinal symptonl
stnrctureof BP-I,basedon weekly levels
of symptom severity and polaritf in a
largecohort of patientswith BP-l who entered the National Institute of Mental
Health (Bethesda,Md) CollaborativeDepressionStudy (CDS)rr'raduring a major
WWW. ARCHGENPSYCHIATRYCOM
SUBJECTSAND METHODS
SUBJECTS
The analysissample consisredof 146 patients with BP-I
entering the CDS from l97B through l98l at I o[ 5 academic centersduring an alfectiveepisode.r].ro
Patients
experiencedboth depressiveand manic episodesas of
intake or during follow-up, with no evidenceof schizo,
phrenia or schizoaffectivedisorder. Bipolar I diagnosis
was based on the Schedule of Affective Disorders and
j
S c h i z o p h r e n i a r u s i n g R e s e a r c hD i a g n o s t i c C r i t e r i a
(RDC).'o Subjectswere white (genetichypotheseswere
being tested), spoke English, had an IQ score o[ at least
70. and had no evidenceof organicmental disorderor terminal medical illness.All patientsgaveinformed consent
at the 5 academicsites where the data were gathered.
Demographic and clinical characteristics are presented in
loble l.
FOLLOW-UP PROCEDURES
Trained ratersinterviewedpatientsevery 6 months for the
first 5 years of followup and are still continuing to interview them yearly thereafter, using variations of the Longitudinal Interval Follow-up Evaluation (LIFE).t'ZPatient
interviews were the primary information source for LIFE
data, with chronologicalmemory prompts used to obtain
information on changesin weekly symptom severity for
all mood and other mental disorders.Interviews were
supplemented by detailed review of availablemedical
records and all information was integrated into a final rating algorithm score. Weekly symptom ratings were
obtained using LIFE Psychiatric Status Rating (PSR)
scales,which are anchored to diagnostic thresholds for
RDC mental disorders. Collaborative Depression Study
raters routinely undergo rigorous training, resulting in
high intraclass correlation coelficients (lCCs) for rating
changesin symptoms (ICC=0.92), recovery from episodes(ICC=0.95), and subsequentreappearance
ofsl'rnpt o m s( I C C = 0 . 8 8 ) .
Interviewersassigna 5-point rating of the accuracyof
weekly PSRinformation basedon their overall impression
of the subject'srecall, the internal consistencyofin[ormation provided. and evidenceof denial or distortion of illness status. If a subject is severelymanic or depressedat
the scheduledtime of follow-up, the intewiew is rescheduled at a later time. Of the 2516 forms availablefor the analysis sample, 25.8o/owere rated "excellent," 50.4o/a"good,"
20.7o/o"fair ," 2.7o/o"poor ," and 0 .4o/""very poor" in their
accuracv of weekly PSR information. Specific follow-up
weeks were not included in the analysesif accuracyratings were "poor" or "very poor" (9.0o/o
of follow-up weeks
accountingfor 77 forms) or i[there weremissingdata(0.9olo
of weeks). Due to frequent changesin symptom status,it
was inappropriate to impute illnessstatusduring a period
of inaccurateor missing data.
A total of 157 CDS patients met diagnostic criteria
for BP-I and were followed up for up to 20 years. Because
our study focusedon the long-term course,we eliminated
lrom the analysesIl patients (7.0ol.)with less than 2
years of weekly PSR data with "fair" or better accuracy.
Nine of these patients dropped out before 2 years; the
remaining 2 excluded subjects were followed up for
exactly 2 years but had missing data or forms with "poor"
or "very poor" accuracy for some portion of that time.
This left 146 patients with BP,I with at least 2 years of
weekly follow-up data rated "[air" or better accuracyCLASSIFICATION OF WEEKLY SYMPTOM
SEVERITY LEVELS
We have extended the methodology used in our previous
work, describlngthe course of unipolar MDD,r2 to
include symptom severity levels of mania as well as
depression.Each weekly symptom severity level was
assignedas presentedin toble 2, based on the 6-poinr
P S R s c a i e f o r m a j o r d e p r e s s i o na n d m a n i a p l u s t h e
3-point PSRscalefor rating minor depression/dysthl'rnia,
hypomania, DSM-IV atypical depression, DSM-III adjusr
ment disorder with depressedmood, and RDC cyclothymic personality. While affective symptom severity levels
are anchored to the diagnostic thresholds for all depressive and manic conditions, including MDE, mrnor
depressive/dysthymicdisorder, mania, and hypomania,
weekly levels were assignedregardlesso[ whether the
patient was in an RDC-defined episode. Affective symptoms below the thresholds o[ the foregoing RDC conditions were classified as subs;'ndromal depression or subsyndromal mania. Weeks with no affective symptoms
were classified as as)rynptomatic.Weeks with some affective symptoms were then categorizedinto levels o[ pure
depression(no mania/hypomania), pure mania/
hlpomania (no depression), or a combination o[ manic
and depressivesymptoms (cycling/mixed affect). Weeks
with prominent psychotic syrnptoms were counted based
on a PSRscoreof 6 on the 6-point PSRscalefor mania or
MDE.
STATISTICAL ANALYSES
Follow-up weeks spent in the different sl.rnptom status categorieswere computed for each patient as percentagesof
the total number of follow-up weeks with PSRratings of
"fair" or better accuracy.Total and averageyearly numbers of changesin symptom status categoriesand shifts
in symptom polarity were also computed per patient.
Subgroups of patients with BP-I were defined based on
predictors of chronicity previously identified in the BP-I
literature: agei age at onset of first lifetime affective episode;number of lifetime affectiveepisodes;poor social functioning in rhe 5 years prior to intake; family history of
affectivedisorder; alcoholism; and duration, polarity, and
presenceof psychotic featuresin the intake episode.AIthough not previously identified as robust predictors of
chronicity in BP-I, we also examined sex, severity of the
intake episode,drug-use disorder, and comorbid anxiety
disorders.We defined long-term chronicity in 2 ways: (l)
the total percentageof follow-up weeks spent with symptoms at the full-syndromal MDD/manic level, and (2) the
total percentageof follow-up weeks spent with any affective symptoms (at any level other than the asymptomatic
status). Comparisonswere made by analyseso[ variance,
with a 2-tailed ct level o[ .05 defining statisticallysignificant group comparisons.
(REPRINTED) ARCH GEN PSYCHIATRY/VOL 59, JUNE 2OO2
531
WWW.ARCHGENPSYCHIATRY.COM
Table1. DFmographic
andClinicalHistoryCharacleristiss
ot Patients
WithBipolarI Disordar
at CDSIntake*
Age,mean(S0)[range],
y
Female
Education
Highschoolor less
College
or more
Marital
slatus
Married/living
together
Separated/divorced/wid
owed
Never
married
TotalNo.of lifetime
attective
(including
episodes
intake
episode)
1 (intake
episode)
2or3
4-10
>10
Ageat onsetoJtint tifetime
altective
episode,
mean(SD)[rangel,y
Earlyonsetof tirst lifetimeaflective
(age<20 y)
episode
Severity
(worstweekGASscore),
0f intake
episode
mean(SD)lrangel
Inpatient
stalus{intake)
Polarity
priort0 intake
diagnosis
ol affective
episode
Depressive
only
Mania
only
Cycling/mixedt
Polarity
diagnosis
0f entireintake
episode
Depressive
only
Maniconly
Cycling/mixedt
Allfollow-up,
mean(SD)lrangel
Weeks
(median.
884)
(median,
Years
17.0)
Follow-up
withpsychiatric
statusratingsof -taid'0r betteraccuracy,
mean(SD)lrange]
(median,
Weeks
806)
(median,
Years
15.5)
Yearsof lollow-upwithpsychiatric
statusratingsof "fair"or betteraccuracy
15-19
10-14
5-9
2-4
39.2(1s.7)
[17-7s]
81(ss.s)
58(3e.7)
88(60.3)
63(43-2)
35(24.0)
48(32.9)
8 (5.5)
32(21.S)
61(41.8)
45(30.8)
22.e(10.0)
[1-s9]
72(4e.3)
32.9(10.6)
[11-671
132(90.4)
2S(1e.S)
45(30.8)
72(4e.3)
20(13.71
31(21.2)
e5(65.1)
(2S5.1
738.0
040]
) [104"1
14.2(5.7r't2-201
(296.6)
665.0
[104-e88]
(5.7)
12.8
[2-1e]
82(s6.2)
17(11.6)
26(1i.8)
21(14.4)
-Datq3r,e
givenasnumber
(percentage)
of patients
unless
indicated.
otherwise
Patients
intheNational
Institute
of Mental
Health
Depression
Collaborative
_
Study(CDS)
who,asof intake
or anytimeduringfollow-up,
haveResearch
(RDC)
Diagnostic
Criteria
mania
anddepression
butnoschizophrenia
or
schizoaffective
disorder,
(2years)
psychiatric
andwhohaveat least104weeks
of weekly
statusratings
with"verygood,""good,"or "lair';accuracy,
whichwere
thebasisfortheanalyses.
GASindicates
Global
Assessment
Scale.
o{ RDCmajor,
minor,
intermittenl,
plusmania
or dysthymic
depressive
fDiagnosis
disorder,
or hypomania.
affective episode.We hypothesizedthat BP-I would
also be expressedlongitudinally as a dimensionalillness in which patients typically experiencefrequent
changesin polarity and severiryof affectivesymproms
covering the full range of severity of depressionand
rnania.
We also examined 2 new potentiallv useful measures of chronicity in relation to predictors of chronicity previously identified for BP-I, as follows: (1) the
total percentageof follow-up weeks thar parienrsexperienced the full-syndromal level of major depressiveor
m a n i c s y m p t o m s a n d ( 2 ) t h e t o t a l p e r c e n t a g eo f
follow-up weeks they experiencedany affectivesymptoms at any level of severity.We anticipatedgreater
chronicity for BP-l rhan we previouslyfound for unipolar MDD, and we predictedthat our 2 new indices,
characterizingchronicity during the entire follow-up
period, would provide a somewhar different but
complementarypicture than previously reported for
BP-I.
(REPRINTED) ARCH GEN PSYCHIATRY/VOL 59.
JUNE 2OO2
532
SYMPTOM STATUS DURING THE COURSE
OF ILLNESS
Patients were symptomatically ill about half of the
t i m e ( m e a n I S D I , 4 7 . 3 ' / . l 3 4 o / . 1m
: edian,38%) and
asymptomatic for the remainder of follow-up (52.7./.
[34o/.];median,620lo).Fourteenpatients(9.6"/o)of 146
were symptomatic during all of their prospective
follow-up (a finding not artriburable to these parienrs
having a shorter follow-up period). Symptomatically
ill weeks (47.3'/oof follow-up) included a mean ( [SD];
( [ 1 8 . 7 o l ' ] ;m e d i a n , 7 . 5 o / oo) f a l l
median) o[ 14.8o/o
follow-up weeks with subsyndromalsymptoms of
mania or depression; 20.2'/. (l2I.0o/ol; medtan, l2o/o)
of total follow-up with minor depression/dysthymiaor
h y p o m a n i c s y m p t o m s , a n d o n l y 1 2 . 3 o / o( 1 1 4 . 2 ' / o,l
median, 7o/o)of follow-up spent at the syndromal
W\!NV-ARCHGENPSYCHIATRY-COM
Table2. Classification
afAffective
Symptom
Severity
levelsBased
onWeelrly
PSR$caleScores
Across
All4 Groups
ofAlfective
Disorders+
AffectiveSymplom
SeverilyLevel
1. Asymptomatic:
nodepressive
or manic
spectrum
symptoms
whatsoever;
returnto usualself
2. Subsyndromal:
depressive
spectrum
symptoms
belowminordepres$ion
level0r manicspectrum
symptoms
belowhypomania
level
3. Allective
symptoms
attheminordepression
or
hypomania
level
[t0D/tlaniat
'|
1
1
MinorDeprussion/
Hypomaniat
OSif-lIIDepressive
Conditions*$
RDCOyclolhymic
Personalily+
.t
1
1
1
2or3
2or3
a
2
1
lorz
z
,'
:::
rl
J
4
4. Aftective
symptoms
attheMDDor mania
level
o
*Weekly
symptom
severity
levelis assigned
based
oneachweek's
ratlngs
0nallaflective
conditions
regardless
wasin a Research
of whether
thepatient
(RDC)
Diagnostic
Criteria
episode
(MDD),
atthattime.Rated
affective
including
conditions
RDCmajordepressive
disorder
RDCminoror intermittent
depression
personality,
or dysthymia,
RDCmania,
RDChypomania,
RDCcyclothymic
andDSM-Illatypical
(code296.82)
depression
disorder
withdepressed
andadjustment
mood(code309.00).
Weekly
symptom
severity
levels
aremutually
exclusive.
Read
across
thetableforcombinations
Rating
of Psychiatric
Status
Scale(PSR)
values
thatresultin classifying
a particular
weekata givensymptom
level.
Forexample,
severity
wouldbeclassified
a patient
attheminordepression/dysthymia
levelfortheweektheywereratedasPSR3 or 4 onthe6-point
majordepression
scale
minordepression/dyslhymia
or PSR3 onthe3-point
scale
witha PSR
scoreof 1 or 2 onthe6-pointmajordepression
qualifies
scale.
Ellipses
indicate
anyPSRvalueof thisatfective
forthegivensymptom
condition
severity
level,
in
coniunction
withthevalues
shownlor otheraffective
conditions.
Forexample,
a givenweekis classitied
attheMDD/mania
levelbased
ona PSRol 5 or 6 for MDD
regardless
and/ormania,
of PSRvalues
onanyotheraffective
condition(s).
weekly
PSRvalues:
1 = asymptomatic,
returned
to usualself;2 = residual/mild
moderate
aflective
symptoms;
3 = partial
remission,
symptoms
0r
JSix-point
impairment;
4 = marked/major
psychotic
or impairment;
symptoms
5 = definite
withoutprominent
criteria
symptoms
or extreme
impairment;
6 = definite
criteria
psychotic
withprominent
symptoms
impairment.
or extreme
weekly
PSRvalues:
1 = asymptomatic,
returned
(mildsymptoms);
(severe
to usualself;2 = probable
criteria
3 = deJinite
criteria
symptoms).
{Three-point
DSM'Illalypical
depression(code296.82)andadjustmentdisorderwithdepressedmood(c0de309.00).
Slncludes
threshold level of mania and/or MDE. Notably, the 5
CDS centers did not differ in the percentageof weeks
patients with BP-l spent with some affectivesymptoms
o r a s y r n p l o m a r(i F
c *' o ' =I . 0 6 ;P = . 3 8 ) .
As presentedin Toble 3, patientsexperienced3 times
more depressivesyrnptoms(31.9%of total follow-up week)
than manic symptoms (9.3o/oof weeks),and depressive
s)'mptoms were 5 times more frequent than cycling/
mixed syrnptoms(5.9olo
of weeks).Subsyndromaland minor depressive/dysthymicsymptoms were much more
prevalent than MDE-level s1'rnptoms(22.9o/o
vs 8.9oloof
weeks); subsl'ndromal manic and hypomanic s1'rnptoms
were 3 times more common thansymptoms at the threshold for mania (7 .Oo/o
vs 2.3o/o
of weeks).Overall,most of
all symptomatic weeksinvolved subsyndromal,minor depressive,and hlpomanic symptoms(74.0"/.). Only I 2.3olo
of all follow-up weeks were spent with ry.'rnptomsat the
thresholdfor MDE or mania.During RDC-definedMDEs,
patients r,r'ithBP-I had sy'rnptomsat the full syrnptomatic
threshold for only 32.60/oof weeks; during RDC-defined
manic episodes,they experienced the full manic s1,rnptom thresholdfor only 20.5olo
o[weeks.
PERCENTAGE OF WEEKS WITH
PSYCHOTIC SYMPTOMS
Patientswith BP-I spent 2.3% of total follow-up weeks
with psychoticsymptoms-l.3olo of weeksoccurredduring mania and 0.9oloweeksduring MDE. Throughout their
entire course,approximatelyhalf of patients(47.3olo)
had
sonreweekswith psychoticsymptoms-26.0o/ohad psychotic-symptoms during MDEs and 28.l% during manic
eprsoCles.
(REPRINTED) ARCH (;EN PSYCHIATRY/VOL 59. JUNF 2OO2
533
CHANGES IN SYMPTOM STATUS
A changein symptom status was defined as any weekto-week changein symptom severitylevel and,/orpolarity. As presentedin Toble 4, patients experienceda mean
(SD) of 74.3 (Il5.l) changesin symptom statusduring
the entire follow-up, or 5.9 (7.6) times per year. Only
9.60lopatients averagedI or fewer changesin affective
$.rrnptomstatus per year. More than half of the sample
(54.1'/o)changedaffectivesymptom status more than 3
times per year,34.9o/o
more than 5 times per year, ll .6oio
more than l0 timesperyear,and5.5"/omore than 20 times
per year.
CHANGES IN AFFECTIVE SYMPTOM POLARITY
A substantialportion of the symptom statuschangesinvolved shifts in symptom polarity, that is, betweensome
Ievel of depressionand some level of mania,/hlpomania.
This occurreda mean (SD) of 47.2 (l10.8) timesduring
extended follow-up, or 3.5 (7.4) times per year. About
60o/oof patients changedpolarity once per year or less
while 19.2%changedpolarity anaverageof more than j
timesper year,8.2o/o
changedpolarity more than l0 times
per year, and4.l"/o changedpolarity more than 20 times
per yearPATIENT COMBINATIONS OF SYMPTOM
STATUS CATEGORIES
Overall, 90okof patients spent I or more weeks during
follow-up with depressivesymptoms and 86.3% had I
or more weekswith manic/hypomanic symptoms.Only
WWW. ARCHGENPSYCHIATRY.COM
Table
3. Percentage
ofFollow.up
Weeks
Spenl
al Specifio
Allective
Symplom
Categories
Delined
hySymplom
Polarig
andSeuerily
During
Long-term
Follow-up
of146Patienls
WilhBipolar
I Disorder
intheCDS*
Penenlrgeol Follow-up
Wedc$pental [aehLeuel
Mern
AtlecliveSymplom
Se{erilyLevel
(sD)
Weeks
(nodepression 52.7(34.0)
asymptomatic
ormania/lrypomania)
Weeks
withpuredepression
31.e(29.e)
(nomania/hypomania)
Pure
subsyndromal
depression e.4(14.7)
Pure
minor
(17.3)
depression/dysthymia
13.5
threshold
Pure
major
depression
threshold 8.e(12.5)
Weeks
withpuremania/hypomania 9.3(15.6)
(nodepression)
Pure
subsyndromal
2.4(6.8)
mania/hypomania
Pure
hypomania
threshold
4.6(e.e)
Pure
mania
threshold
2.3(4.0)
We€ks
withrycling/mixed
affective
5.e(14.2)
symptomst
Median
(Range)
62{0-99)
23(o-ee)
3 (0-82)
7 (0-82)
5 (0-63)
2.5(0-82)
0 (0-38)
1 (0-81)
1 (0-37)
0 (0-e4)
*Patients
inNational
lnstitute
Health
Collaborative
Deoression
ofMental
(CDS)
Study
who,
asofintake
lollow-up,
liletime
oranytimeduring
have
Research
Diagn0stic
mania
Criteria
anddepression
butnoschizophrenia
or
schizoaffective
(2years)
disorder,
andwhohave
atleast104weeks
ofweekly
"good,"
psychiatric
good,"
status
ratings
with"very
0r"fair"
accuracy.
withcycling/mixed
reached
levels
ol major
affect
depressive
disorder
tweeks
0rmania
anaverage
levels
ol 1.0%
ollollow-up
weeks;
o1minor
depressive
disorder,
dysthymia,
orhypomania
weeks;
anaverage
ol 2.0%ol tollow-up
and
subsyndromal
levels
ot23%offollow-up
oldepression
ormania
anaverage
weeks.
approximately half (48.6'6 had I or more weekswith cycling/rnixedaffectivesymptoms(Table4). ln addition,35
patients (24.0V.) spent I or more weeksduring follow-up
in all 7 possibleswnptom categories(ie,3levelsof depressive s1'mptomseverity,3 levelsof manic/hypomanicseverity, and the asy'rnptomaticstatus).Another 4l patients
(28.1%), during their courseof illness,experienced6 of
the 7 sy'mptomcategories(and of thesepatients,l0o/ohad
(18.5%)spentI or moreweeks
noweeksas1'rnptomatic);27
at 5 slrnptom categories,29 (I9.9o/.)at 4 categories,ll
(7.5'/') at 3 categories,and only 8 Q.l'/o) in 2 s1'rnptom
categories.
Of the l32patienswith I ormoreweekssymptomatic in the depressivespectrum,I05 (79.5'/o)experienced all 3 levels of depressiveseverity.Of the 126 patients with manic sy'rnptoms,6l (48.4o/o)
experiencedall
3 levelsof the manic symptom spectrum.
PREDICTORSOF CHRONICITY DURING
FOLLOW-UP
Greater chronicity, defined in terms of a significantly
higher percentageof follow-up weekswith symptoms at
the full-syndroural MDE/mania level, as well as weeks
with any level of affectivesymptom severity,was significantly associatedwith 4 predictors:poor socialfunctioning in the 5 yearsprior to intake, a longer total duration
of the intake episode,depressive-onlyor cycling/mixed
( v s m a n i c - o n l y ) p o l a r i t y o f t h e i n t a k e e p i s o d e ,a n d
having an RDC diagnosisof drug-usedisorder as of in(REPRTNTED)ARCt't GEN PSYCHTATRITVOL59, JUNE 2002
534
Table
4. Aflecliye
Symplom
Severity
Charaelerisliss
Ouring
Long-term
Follow-uB
of146Patlenls
WithBipolar
I Disorder
in theCDS*
per
N0.ofchanges
insymptom
status
patient,t
mean
nedian
{SD);
lrangel
Duringallof
lotlow-up
Peryear
74.3(115.1);
S.0 [2-273]
5.9{7.6);3.4
[0.2-49.3]
p€rpatient,+
No.0f changes
in polarity
mean(S0);median
[range]
During
allotfollow-up
47.2(110.8)7.sI0-7szl
Peryear
3.5(7.4)
0.6[0.0-48.7]
Patients,
No.(%)
>1 wkasymptomatic
132(90.4)
>1 wkin depression
132(90.4)
spectrum
>1 wk atall3 dopressive
105(79.5)
symptomlevels
>1 wkin manicsp€ctum
126(86.3)
>1 wk atall3 manicsymptom
levels
61(48.4)
>1 wkcycling/mixed
polarity
71(48.6)
*Patients
Institute
Health
Depression
inNational
olMental
Collaborative
(COS)
lollow-up,
haveResearch
Study
who,asoJintake
oranytimeduring
Diagnostic
Criteria
mania
butnoschizophrenia
or
anddepresslon
(2years)
schizoaffective
disorder,
andwhohave
atleast104weeks
0fweekly
"good,"
"faif'
psychiatric
ratings
status
with"verygood,"
or
accuracy.
inlevel
manic/hypomanic
change
oldepressive
and/or
tAnyweek-to-week
theasymptomatic
counts
symptoms,
orchange
from/to
status
as+1.Weeks
withsymptoms
ofboth
depression
andmania/hypomania
add+1.
inpolarity
isdefined
{romsome
level
ofdepression
to
asachange
fChange
weeks
orviceversa
withorwithout
intervening
someleveloi mania/hypomania
Weeks
withsymptoms
and
attheasymptomatic
status.
oJbothdepression
mania/hypomania
add+1t0thecount.
take or during follow-up. Sex,ageat intake, ageofonset
of first affectiveepisode,total number of affective eprsodes, history of affective disorder in first-degree
relatives,severity of intake episode,psychotic features
of the intake episode,and RDC diagnosisof alcoholism
were not significantly associatedwith increasedchron i c i t y ( f o b l e 5 ) . R e s e a r c hD i a g n o s t i c C r i t e r i a diagnosedanxiety disorders(generalizedanxiety disorder, panic disorder, phobic disorder, and obsessivecompulsivedisorder), consideredindividually aswell as
in the aggregate,also did not predict an overall more
chronic course.
Previousrepor6l-e'r8on the long-term picture of BP-l have
largely focused on the course of MDE and manic episodesor have examined it from the perspectiveof patterns of successiveepochs of illness, such as the "kinThese epoch-basedanalysesof major
dling" model.rs-20
affective episodeshave informed us about this illness.
However,we had a different objective: to document the
long-term symptomatic structure of this disorder based
on summary (aggregate)measuresof weekly affective
symptom status.To the bestof our knowledge,this is the
first article describingthe entire long-term weekly naturalistic courseof BP-l in terms of the/ull range of affective symptoms.We believethat the measuresexamined
here provide a more complete picture of the longitudin a l s t r u c t u r e o f B P - I , w h i c h c o m p l e m e n t sp a s t a p proachesfocusing only on major depressive/manicepisodes,and provide valuable new information about the
iong-term course of this iliness.
WWW. ARCHGENPSYCHIATRY.COM
TableS.
Percenlage
ofFollow'up
Weeks.Spent
WithSlmptoms
altheDisoder
lhreshold
forMD0/frlania
0rAnyLeyel0fAlfestiue
Sympt0m
Seveilty
predictors
During
Long-term
Follo*upof1{SPatienls
WithSipolar
t Disordar
in theCoSbyVarious
of Chronicily*
Prediclor
of Chmnicily
% of Follow-up
Wed6ltili Symptoms
at MDD/tlania
Thresholdt
Sex
Male(n = 65)
(n= 81)
Female
(14.3)
12.1
(14.21
12.4
y
Ageatintake,
=40 (n= 88)
>40(n= 58)
Ageat0nsetol firstlifetime
y
affective
episode,
1-20(n = 72r,
21-40(n = 63)
>40(n=11)
TotalNo.of liletime
(including
atfective
episodes
intake
episode)
1-3(n= 40)
4 - 1 0( n= 6 1 )
> 1 0 1 n= 4 5 1
Bestlevel0f socialtunctioning
in 5 y pri0rto intake
(n= 136)
Fairor better
poor/Urossly
Poor/very
(n = 8)
inadequate
Anyaflective
disorder
Dxin tirst-degree
relatives
Yes(n= 47)
No(n= 15)
Totalduration
of intake
episode
<6 mo(n= 50)
6moto<2y(n=60)
-2y(n=35)
o/o0f Follor-upWeolsWithAnytevel
ofAfecliue$ymplomst
f r l l= 0 . 1 0P
; =.92
42.9t33.4)
50.8(34.2)
f r a= 1 . 4 1P;= . 1 6
11.9(12.7)
12.8(16,3)
lrmrt'0.36;P=.72
45.2(33.7)
50.4(34.4)
hu'031;P=.37
13.e(15.6)
11.4(13.3)
6.4(6.2)
= 1.56;P= .21
Fz,r*
48.6i34.6)
46.3{34.2)
43.9{30.8)
=0.13;
Fz,rc
P=.87
9.4(12.6)
12.1(13.5)
15.0(16.1)
Fr,tg= 1.65;P= .20
39.3(34.0)
48.7{32.e)
52.5(34.8)
= 1.70;
Fr,t*
P=.19
11.3(13.4)
28.9(20.0)
lro= 3'51;P<.001
45.9{33.9)
74.4(26.8)
P=.02
\42=2.33:
11.7(12.1'J
6.7(8.8)
P= .15
6o= 1.46;
51.0(33.3)
38.8(33.4)
ts:=1.23;
P=.22
5.5(6.8)
11.9(12.3)
21.9(18.7)
F2,1a2.
16.93;
P<.001(acb<c){
2e.4(30.4)
50.8(29.8)
66.4(34.2)
= 15.22:
Fz.uz
P<.001
{acb<c){
14.8(14.5)
= 5.07;
Fz.rqz
P=.008(b<a,c)*
46.9(29.5)
s0.0(30.5)
52.s(34.5)
=5.56;
(b<c)+
Fa,rc
P=.005
(14.8)
13.8
(14.e)
11.3
(11.0)
11.5
=0.51;
Fz.t€
P=.60
46.3{36.0)
48.6{34.1)
46.1(30.2)
=0.08;
Fz,r€
P=.92
(r6.7)
14.1
(10.4)
10.1
ft.,of= 1.78;
P=.08
48.3(35.0)
46.1(33.0)
t," =0.39;
P- .70
(15.5)
14.3
(13.2)
10.e
frll = 1'44;P= .15
46.8(34.3)
47.6(34.0)
lra=0.13;
P=.90
(16.7)
19.1
(13.0)
10.6
P=.003
4+r=2-99;
63.9(34.0)
43.2(32.8)
tr+=3.02P=.003
Polarity
ol entireintakeepisode
Depressive
Dxonly(n= 20)
Manic
Dxonly(n= 31)
(n= 94)
Cycling/mixed
s.2(5.e)
(15.3)
13.8
Severity
of intakeepisode
GASscorepriorto intake)
{worst-week
11-30(n=55)
31-40(n= 65)
41-67(n= 26)
Psychotic
(based
features
in intakeepisode
onifiakeSADS)
Yes(n= 78)
No(n= 68)
Comorbid
substance
abuse
disorders
EvermetRDCalcoholism
DxS
Yes(n = 58)
No(n= 88)
EvermetRDCdrug-use
disorder
Dxg
Yes(n= 29)
N o( n= 1 1 7 )
*Data
areoivenasmean(SD)patients
unless
otherwise
indicated.
Patients
in National
Institute
ol Mental
Health
Depression
who,asof
Collaborative
Study(CDS)
intake
(RDC)
oranytimeduring
follow-up,
haveResearch
Diagnostic
Criteria
mania
anddepression
butnoschizophrenia
0r schizoaffective
disorder,
andwhohave
at
"good,"
(2years)
psychiatric
ieast104weeks
0fweekly
status
ratings
with"verygood,"
or"fair"accuracy,
whichwerethebasis
fortheanalyses.
MDDindicates
major
depressive
disorder;
Dx,diagnosis;
GAS,
Global
Assessment
Scale;
andSADS,
Schedule
ol Affective
Disorders
andSchizophrenia.
groupvariances.
0f treed0m
adjusted
forunequal
tDegrees
ditferences
based
onposth0cgroupcomparisons.
+Significant
metdiagnosis,
atprobable
level,
or definite
asof intake
follow-up.
or during
$Ever
While BP-I is lesschronic than unipolar MDD, which
did not support our a priori hlpothesesofincreasedqhronicity of BP-I, thesepatientswere nonethelesssymptomatically ill nearly half of their long-term follow-up. Al-
though BP-I is traditionally describedin terms ofepisodes
of MDE and mania, we found that subthreshold,minor
depressive/dysthymic, and hypomanic s)'rnptoms were
the modal expressionsof BP-l during its prospective
(REPRINTED) ARCH GEN PSYCHIATRY/VOL 59. IUNE 2002
535
WwW ARCHGENPSYCHIATRYCOM
course.Symptomsin the depressivespectrumpredominated substantially over manic (3:1) or cycling/mixed
symptoms(5:1). We cannor,however,rule out the possibility that patientswith more distressingdepressive
symptoms may be more likely ro enrer and remain in a
long-term prospectivestudy. Bipolar I is often regarded
as a psychotic disorder,yet slightly more than half of the
patients had no weekswith psychoticsymptoms during
the entire course of illness; psychotic symptoms occurredrelativelymore frequentlyduring manicthan MDD
episodes.Patientsexperiencedfrequentchangesin symptolrr status and polarity in a dynamically fluctuating
course.The full rangeof subslndromal, minor depressivey'
dysthvmic, h;,pomanic, MDE, and manic s1'rnptomswere
observedcommonly within the samepatientsover time.
ln sum, thesedatastrongly support the idea that the longitudinal courseof BP-l is expressedasa dimensionalspectrum involving the completerangeof severityof depressive and manic symptoms.We thereforesubmit that
longitudinal descriptionsof the BP-l coursethat do not
include all levels of affectivesymptom severityand polarity are incomplete.
The definitions of chronicity we have used in this
article, namely. the percenugeof all follow-up weeksspent
at the highestlevel of affectivesymptom severityor with
any affectivesymptoms, are new but provide a complementaryperspective
of the long-termcourseof BP-I.Other
analysesof chronicity in BP-l haveused avariety of definitions basedon specificepochsof tirne,a8 such as time
to recovery from the intake episode,time to first prospectively observedMDE or rnanic episoderelapse,relapse to MDE/manic episode(s)within a specified period oi time, occurrenceof an MDVmanic episodeduring
a particular follow-up interval,5or levelof morbidity during a particular period. Only Turvey et al8analyzedpredictors of the overallpercentageof follow-up spentin ma1oraf{ectiveepisodes.However,their analyses,asall other
studiesof the long-termcourseofBP-I, focusedonly on
MDE and manic episodesrather than the more frequent
periods of minor depression,dysthymia,or hlpomania.
To the best of our knowledge, our study is the first to
characterizeall of long-term follow-up basedon the full
range o[ syndromal and subsyndromal levels of affective symptom severity,Our approachpresentsa definitive picture of the overall chronic nature of BP-l compared rvith other definitions based only on selected
lollow-up intervals, which have produced inconsistent
Iindings. We also found that 2 indices of past chronicity, namely, poor social functioning in the 5 yearsprior
to intake and a longer intake episode,predicted significantly greaters).Tnptomaticchronicity during all of followup. To earlier findings that cycling in the intake episode
predicted greaterchronicity,4'7we now add that a purely
depressiveintake episodealso predicts greaterchronicitv comparedwith purely manic intake episodes.Unlike
Corl'eli et al,5who found that alcoholismpredictedchronicity, defined asbeing in an MDE or manic episodeduring the l5 yearsof follow-up, we found that drug abuse
but not alcoholisrnpredicted greaterchronicity of both
MDE and manic symptoms, and these affective symptorns remain during long-term follow-up. Inconsistent
findings rn chronicity underscoresthe need for reliable
(REPRINTED) ARCII GEN PSYCHIATRY/VOL59, JUNE 2OO2
536
and meaningful definitions o[chronicity, such as rhe ones
we have proposed.
Generalizationto other samplesof BP-l may be limited becausethe CDScohort consistedofseverelyill, tertiary care,whitepatiens.We donotknow theextenttowhich
the history and intake status of our sample are representativeofother patientswirh BP-Iseekingtreatment.Ahhough
interrater agreementfor changesin episodestatushasbeen
shown to be high, there may be some degreeof error in
assigningweekly symptom severitylevels.We may have
underestimatedthe time with subsyndromalsymptoms
and overestimatedthe asymptomatictime sincePSRcoding rules do not allow for subsyndromalsymptoms to be
codedfollowing fullyasymptomaticepisoderecoveryuntil such time assymptoms again reach syndromal levels.
Cycling/mixedexpressionsmay havebeen relativelyuncommonbecausea universallyaccepteddefinition oIthese
forms did not exist when the Scheduleof Affective Disorders and Schizophreniainstrument was developedin
the late 1970s;thus, our analysescannotshedlight on the
question of dysphoric mixed statesusing contemporary
definitions.Nonetheless.the CDSis a unioue databasefor
the perspectivesyrnptomaticstudy of the long-terms1'rnptomatic structure of BP-I. Now that the Zurich Study22l
hasclosed,the CDSis the only available,ongoing prospective naturalistic follow-up study of a large cohort of patients with affectivedisorder of which we are aware.
Algorithms to summarize the doseintensity of mood
stabilizers,antidepressants,and antipsychotic medications have been created and updated over the years to
reflect new treatmentsthat have become availablesince
the study beganin 1978.22However, the CDS is a naturalistic follow-up study of mood disorders, not a controlled treatment investigation. Meaningful analysesof
the adequacy,intensity, and effectofantidepressant,antimanic, and antipsychotic medications on the various
Ievelsof affectivesymptom severitywouid be extremely
complexand arebeyondthe scopeof this article.The predominance of depressiveover manic/hypomanic symptoms should not be interpreted as suggestingthe need
for more aggressive
use of antidepressantmedicationsin
the absenceof a mood stabilizersince there is some evimay induce mania or cycleacdencethat antidepressants
celerationin some bipolar patients.2r
Analysesof within-subject trends over time for particular subgroupsof interest,such as patients with BP-l
with various patterns of cycling or comorbid substance
abuse,are alsobeyond the scopeofour study. The focus
of this article is on characterizingin the aggregatethe
overall long-term symptomatic status of BP-l basedon
the sample as a whole. The relatively large variation
around the means of the long-term outcome measures
we havepresentedsuggeststhat theseindices maybe useful for identifying and characterizingclinically meaningful subgroupsof patientswith BP-I , which we intend to
addressin future manuscripts.
While thesedatasupport the idea that bipolar disorder is best characterizedas a sDectrum of affectivesr.'rnptom severity,2athey do not imply a continuum b.t*..tt
BP-I and BP-ll, which may have rather distinct coursepatNor can we comment on contemporaryimagiterns.25'26
native proposalsto extend the bipolar spectrum to "softer"
WWW. ARCHGENPSYCHIATRY.COM
8, Turvey
WH,ArndtS,Solomon
CL,Coryell
AC,Endicott
DA,Leon
J, Mueller
T,
Keller
M,Akiskal
H.Polarity
sequence,
andchronicity
depressi0n
ofbipolar
I disotder.
J NervMentDls.1999:187:
181-187.
9. Coryell
W,Akiskal
H,Leon
AC,Turvey
D,Endicott
C,Solomon
J. Family
history
levels
andsymptom
duringtreatment
I affective
for bipolar
diso'det.
BiolPsychiatry.2000;47
:1034-1042.
10.JuddLL,AkiskalHS.Delifleatingthel0ngitudinalstructureoldepre
beyond
subtypes
Pharmacopsychiatry.20D0,
clinical
andduration
thresholds.
33:3-7.
11. Akiskal
HS,JuddLL,Gillin
JC,Lemmi
H,Subthreshold
depressions:
clinical
and
polysomnographic
validation
forms.
ofdysthymic,
residual,
andmasked
J.4/fect
Dlsord1997;45:53^63.
12. JuddLL,Akiskal
HS,MaserJD,ZellerPJ,Endicott
J, Coryell
W,Paulus
MP,Kunovac
AC,Mueller
Tl,Rice
M8.A prospective
JL,Leon
JA,Keller
12-year
study
of subsyndromal
andsyndromal
in unipolar
maiordedepressive
sympt0ms
Subnritteel
publication
December
12,
2000;
revifot'
final
pressive
ArchGen Psych
disorders.
iatry.I 998;55:694-700.
sionreceivedAugust6,2001;accepted
September
4,2001.
13.KatzMM,KlermanG.lntroducli0n:overviewollheclinicalstudiesp
FundsfortheconductoJthisstudywereprovidedinpart
J Psychiatry.
1979;136:49-51.
14. KauMM,SecundaSK,
Hirschfeld
RMA,Koslow
Research
Branch
SH.NIMHClinical
by the Roehr Fund of the u niver sityoJCalif ornia, San D i e go.
Pr0gram
Collaborative
onthePsychobiology
/rchGenPsychiaofDepression.
This marutscripthas beenreyiewedby the Publicatry.1979.'36765-772.
tionsCommitteeof theCollaborativeDepression
Studyand
(SADS).
15. SpiEerRL,
EndicotlJ.
Schedule
forAffective
Disorders
andSchizophrenia
has its endorsement.
NY:Biometrics
NewYorkState
Psychiat3rded.NewYork,
Research
Division,
From the National Instituteof Mental Health CollaboricInstitule;
1979.
16.SpitzerRL,EndicottJ,andRobinsE.ResearchDiagnosticCriteriafora
rati,re Program on the Psychobiologyof Depression,CliniDisotders.
R€search
DiGroup
ofFunctional
3tded.NewYork,NY:Eiometrics
cal Studies,conductedwith the participatiln of the t'ollowvision,
NewYorkState
Psychiatric
Institute;1977.
iry investigator.s:
M. B. Keller,MD (Chaitperson,Providence,
17.KellerMB,LavoriPW,FriedmanB,NielsenE,EndicottJ,McDonald
RI); \{. Cory ell, MD (Co-Chei rperson,Iowa City, Iowa) : H.
dreasen
NC.Thelongitudinal
interval
follow-up
acomprehensive
method
evaluation:
forassessing
outcome
in prospective
longitudinal
studies.
ArchGenPsychiaS. Ahishal, MD, L. L.J udd,,MD,J. D. Maser,PhD (San Diego,
try.1987;44:540-548.
Calf) : P.W.Lavoi, PhD,T. I. Mueller,MD, M. T. Shea,PhD
18.Winokur
HS,Akiskal
HS,Endicott
J, Keller
M,Mueller
T.ManicG,C0ryell
(Providence):
MD (ChiJ. Fawcett,MD, W.A. Schet'tner,
(bipolar)
disorder:
thecourse
in lightol a prospective
ten-year
foldepressive
(IowaCity);
cago.Ill); W. Coryell,MD,J. Haley
J. Endicott,
l0w-up
of l31patients.
ActaPsychiat
Scand1994;89:102-110.
PhD,A. C. Leon,PhD,J. Loth,MSW (NewYorh, NY);l. Rice,
gerWP.TheNeurobiology
19. PostRE,Ballan
NewYork,NY:PleofMoodDisorder.
PhD, T. Reich,MD (5t Louis,Mo). Other contributorsinnumPress;
1990.
20.Kessing
PK,Mortensen
PB,Bolwig
TG.Clinical
consequences
LV,Andersen
0t
clrrde:N. C. Andreasen,MD, PhD,P. J. Clayton,MD,l.
study.
J Affect
Disord.1998,
sensitisati0n
inaffective
disorder:
a caseregister
Croughan,MD, G. L. Klermcrn,MDf, R. M. A., Hirschfeld,
-47.
47:41
MD, Nl. M. Katz, PhD, E. Robins,MD, R.W.Shopiro,MD,
A.TheZurich
fromnormal
to
21.Angst
J, D0bler-Mikola
Study,
ll: thecontinuum
R. L. Spitzer, MD, G. Winohur, MDf , andM. A. Young,PhD.
pathological
moodswings.
ClinNeurosci.1984;
depressive
EurArchPsychiatry
234:21-29.
Coresponding author and reprints:LewisL.ludd, MD,
NC,Endicott
W,Faw22. KellerMB,LavoriPW,Klerman
J, Coryell
GL,Andreasen
D epar tment of Psychiatry at the UCSD D epar tment of PsyRMA.Lowlevels
01somatocettJ, RiceJP,Hirschfeld
andlackof predictors
chiatry, 9500Gilman Dr, Lalolla, CA 92093-0603.
patients.
lrch 6enPsychiatherapy
andpsychotherapy
received
bydepressed
lDeceased.
try.1986;43:458-466,
BS,RosoffA,Ackerman
BA.
23, Altshuler
MSW,Mikalauskas
LL,P0stRM,Leverich
revisited.
Antidepressant^induced
mania
andcycle
a controversy
acceleration:
1995;1
AmJ Psychiatry.
52:1130-1138.
2 4 .C a s s a n o B G , D e l l ' 0 s s o L , F r a n k E , M i n i a t i M , F a g i o l i n i A , S h e a r K
reality
0fdiagnostic
criteria
andan
J.Thebipolar
spectrum:
a clinical
in search
E. Manic-Depressive
1. Kra€pelrn
lnsanityandParanoia.
Edinburgh,
Scotland:
assessment
m€thodology.
J Affect
Disord.
1999;54:319-328.
E& S Livingslone;.1921.
W,KellerM,Warshaw
M,Clay25. Akiskal
HS,Maser
JD,ZellerP,Endicott
J, Coryell
2 . A n g s tJ . T h ec o u r s e0 f a f f e c t i vdei s o r d e r sl l,: t y p o l o g o
y l b i p o l am
r anicprospecll:an11-year
tonP,Goodwin
FK.Switching
from"unipolar"
tobipolar
depressiveillness.ArchPsychiatr Nervenkr 1978:226:65-7
3.
predictors
in559patienls.
ArchGenPsytivestudyofclinical
andtemperamental
FK,JamisonKR.Thenatural
3. Goodwin
illness.In:
courseol manic-depressive
14-123.
chiatry.1995:52:1
P0stRM,Ballenger
JC,eds.Neurobiology
of MoodDisorders:
Frontiers
of CliniHS.Long-term
sta26. Coryell
W,Endicott
J, Maser
JD,KellerMD,LeonAC,Akiskal
calNeuroscience.
Vol1. Baltimore,
1984:20-37.
Md:Williams
& Wilkins;
AmJ Psychiatry.1995,
distinctions
in theaffective
bilityof polarity
disorders.
4. KellerMB,LavoriPW,Coryell
W, Andreasen
NC,Endicott
J, Clayton
PJ,Klerman
152:385-390.
RMA.Ditferential
0f puremanic,
mixed/cycling,
GL,Hirsch{eld
outcome
andpure
in classr{ication
anddiagnospectrum:
newconcepts
MS,Thebipolar
depressive
episodes
in patients
withbipolar
illness.
J,4Ml.1986;255:3138-3142. 27.Akiskal
Psychiatric
Associa'
L, ed.Psychiatry
TheAnerican
sis.ln:Grinspoon
Update:
5. WinokurG,Coryell
W, KellerM, Endicott
H.A prospective
J, Akiskal
follow-up
.1983:
Psychiatric
Press;
Vol2.Washington,
DC:American
withbipolarandprimaryunipolar
tionAnnual
Review.
of patients
affective
ArchGenPsydisorder.
-465.
271-292.
chiatry.1993;50.457
05Mdisorders:
beyond
28.Akiskal
HS.Theprevalent
clinical
spectrum
ol bipolar
6. Coryell
W, TurveyC,Endicott
T, Solomon
D, KellerM. 8iJ, LeonAC,Mueller
p0larI affective
predictors
1996;16(suppl
1):45-145.
lV.J ClinPsychophrrmacol
disorder:
of outcome
after15years.
J AffectDis1rd.
KA,Strakowski
SM.
19 9 8 : 5 00: 19 '116 .
29. McEkoy
HGJr,KeckPEJr,Hudson
Jl,Phillips
SL,Pope
Psychiatry.1996:
related
tobipolar
Compr
7 . T u r v e y C L , C o r y e l l W H , S o l o m o n D A , L e o n A C , E n d i c o n J , K e l l e r M , A k i s k aAre
l H impulserontrol
.Longdisorders
disorder?
termpro0nosis
of bipolarI disorder.
ActaPsychiatr
Scand.
l 999;99:1
10-119.
37:229-240.
expressions,such as pharmacologichlpomania, cycloth).'rnic,and impulse-control disorders.2T-2e
Our datamore
properl;- pertain to a dimensionalcontinuum of bipolar
symptom severity from subsyndromal to full-blown syndromal levelswithin the courseof rigorouslydefinedBP-I.
Kraeplin,r rvtro wondered why manic-depressive
episodeserupted periodically,had suggestedthat someday
the origin of the illnesswould be understoodfrom relatively inconspicuoussubs).rynptomatic
foundationsthat persist betweenepisodes.Thesedataprovide support for his
conceptualization.
(REPRIN ED) ARCH GEN PSYCHIATRY/VOL 59, JUNF 2OO2
537
WWW. ARCHGENPSYCHIATRY.COM