You are on page 1of 7

Arti cl e

420 Am JPsychiatry 158:3, March 2001
Axis I Psychiatric Comorbidity and Its Relationship
to Historical Illness Variables in 288 Patients
With Bipolar Disorder
Susan L. McElroy, M.D.
Lori L. Altshuler, M.D.
Trisha Suppes, M.D., Ph.D.
Paul E. Keck, Jr., M.D.
Mark A. Frye, M.D.
Kirk D. Denicoff, M.D.
Willem A. Nolen, M.D., Ph.D.
Ralph W. Kupka, M.D.
Gabriele S. Leverich, L.C.S.W.-C.
Jennifer R. Rochussen
A. John Rush, M.D.
Robert M. Post, M.D.
Objective: Bipolar disorder often co-oc-
curs with other axis I disorders, but little
is known about the relationships be-
tween the clinical features of bipolar ill-
ness and these comorbid conditions.
Therefore, the authors assessed comor-
bid lifetime and current axis I disorders
in 288 patients with bipolar disorder and
the relationships of these comorbid dis-
orders to selected demographic and his-
torical illness variables.
Method: They evaluated 288 outpatients
with bipolar I or II disorder, using struc-
tured diagnostic interviews and clinician-
administered and self-rated question-
naires to determine the diagnosis of bipo-
lar disorder, comorbid axis I disorder di-
agnoses, and demographic and historical
illness characteristics.
Results: One hundred eighty-seven (65%)
of the patients with bipolar disorder also
met DSM-IV criteria for at least one co-
morbid lifetime axis I disorder. More pa-
tients had comorbid anxiety disorders
(N=78, 42%) and substance use disorders
(N=78, 42%) than had eating disorders
(N=9, 5%). There were no differences in
comorbidity between patients with bipo-
lar I and bipolar II disorder. Both lifetime
axis I comorbidity and current axis I co-
morbidity were associated with earlier
age at onset of affective symptoms and
syndromal bipolar disorder. Current axis I
comorbidity was associated with a history
of development of both cycle accelera-
tion and more severe episodes over time.
Conclusions: Patients with bipolar disor-
der often have comorbid anxiety, sub-
stance use, and, to a lesser extent, eating
disorders. Moreover, axis I comorbidity,
especially current comorbidity, may be
associated with an earlier age at onset
and worsening course of bipolar illness.
Further research into the prognostic and
treatment response implications of axis I
comorbidity in bipolar disorder is impor-
tant and is in progress.
(Am J Psychi a t r y 2 0 0 1 ; 1 5 8 : 4 2 0 –4 2 6 )
Epidemiologic data indicate that the rates of sub-
stance use and anxiety disorders in persons with bipolar
disorder are significantly higher than the rates of sub-
stance use and anxiety disorders in the general population
(1–12). Clinical data suggest that bipolar disorder may also
commonly co-occur with eating disorders (6, 13–16). Con-
versely, patients with substance use, anxiety, and eating
disorders often have bipolar disorder (4, 5, 17–22). Al-
though major depressive disorder is also associated with
elevated rates of most of these axis I disorders, epidemio-
logic studies comparing the psychiatric comorbidity of the
depressive and bipolar disorders have often found higher
rates of substance use (3, 9), panic (7), and obsessive-com-
pulsive (8) disorders in bipolar patients than in depressive
Little is known about the relationships between bipolar
disorder and the axis I psychiatric conditions with which it
often co-occurs. For example, few studies have systemati-
cally examined rates of comorbid axis I disorders across
different diagnostic subtypes of bipolar disorder (e.g., bi-
polar I versus bipolar II disorder) or the effects of axis I
psychiatric comorbidity on the phenomenology, course,
outcome, and treatment response of bipolar disorder. Pre-
liminary data have suggested that axis I psychiatric co-
morbidity (e.g., the presence of a substance use or anxiety
disorder) may be associated with more severe subtypes of
bipolar disorder (i.e., bipolar I versus bipolar II disorder)
(3), earlier age at onset of bipolar disorder (22), mixed or
dysphoric rather than pure mania (15, 22, 23), higher rates
of suicidality (7), poorer overall outcome (24–30), and less
favorable response to lithium (22, 28, 30, 31). Other stud-
ies, however, describe findings inconsistent with these (7).
The Stanley Foundation Bipolar Treatment Outcome
Network was established to advance the understanding of
the long-term course and treatment of bipolar disorder
(32). Toward this goal, we used DSM-IV criteria to assess
systematically the comorbid lifetime and current axis I
psychiatric disorders in the first 288 patients with bipolar I
or II disorder who entered the Stanley Foundation Bipolar
Treatment Outcome Network. We hypothesized that non-
bipolar axis I disorders would be highly prevalent in these
Am JPsychiatry 158:3, March 2001 421
patients and that the presence of these disorders would be
associated with negative effects on the presentation and
course of their bipolar disorder, as assessed by historical ill-
ness variables.
The methods of the Stanley Foundation Bipolar Treatment
Outcome Network (32), along with the demographic and clini-
cal features of the first 261 patients (33), are described in greater
detail elsewhere. Patients with any type of bipolar disorder were
recruited from private, academic, and community mental
health clinic outpatient settings by referral and advertisement.
These patients entered into the Stanley Foundation Bipolar
Treatment Outcome Network if they met the following inclu-
sion criteria: 1) age at least 18 years; 2) DSM-IV diagnosis of bi-
polar I disorder, bipolar II disorder, bipolar disorder not other-
wise specified, or schizoaffective disorder, bipolar type; 3)
willingness and ability to perform prospective daily mood
charting and attend monthly evaluation appointments; 4) will-
ingness to be in some form of ongoing treatment with a psychi-
atrist and to consider possible entry into future clinical trials;
and 5) provision of written informed consent after the study
procedures had been fully explained. Patients were not paid for
their participation.
The Stanley Foundation Bipolar Treatment Outcome Network
protocol included an extensive baseline evaluation (32). This
evaluation was typically conducted over several visits and in-
cluded completion of the Structured Clinical Interview for DSM-
IV, Patient Edition (SCID-P) (34) to establish the diagnosis of bi-
polar disorder, illness characteristics (e.g., bipolar disorder sub-
type and age at onset), and comorbid axis I disorder diagnoses.
The evaluation also included completion of structured patient-
rated and clinician-administered questionnaires to determine
demographic (e.g., current annual income, current level of occu-
pational functioning) and historical illness (e.g., age at onset of
affective symptoms; history of dysphoric mania, psychosis, rapid
cycling, and suicide attempts; course of illness; and family his-
tory of psychiatric illness) variables. The SCID-P and question-
naires were administered by highly trained clinical research as-
sistants at the individual sites. All had received SCID-P training
at their own sites as well as extensive training in the SCID-P and
the other instruments at the NIMH site (32). This training was su-
pervised by one of the investigators (G.S.L.). As part of their
SCID-P training, every clinical research assistant viewed and
rated four tapes of SCID-P interviews. Excellent interrater reli-
ability was achieved for the diagnosis of bipolar disorder with an
overall kappa score of 0.92.
In completing the SCID-P and questionnaires, the researchers
obtained information from all sources available in addition to the
patient interview, including medical records and interviews with
family members and treating clinicians. Age at onset of bipolar
and comorbid disorders was defined as the first time the patient
met full DSM-IV criteria for the disorder.
For this study, only patients with bipolar I or II disorder were
included in the analysis. In addition, only “core” DSM-IV sub-
stance use, anxiety, and eating disorders were considered as co-
morbid axis I disorders. DSM-IV not otherwise specified disorders
and impulse control disorders not elsewhere classified, although
assessed, were not included in the analysis.
All statistical analyses were performed with SPSS, version 8.0
(SPSS, Inc., Chicago, 1997). Categorical variables were compared
by using chi-square tests. Continuous variables were compared
by using independent-sample or paired-sample t tests as appro-
priate. A Bonferroni inequality correction was applied to all vari-
ables significant at the p<0.05 level by chi-square or t test. To ex-
amine which historical illness variables were associated with
lifetime and current comorbidity, a stepwise logistic regression
was performed. An alpha of 0.2 was selected as the cutoff for in-
clusion in the regression.
Two hundred eighty-eight patients with bipolar I or bipo-
lar II disorder completed the SCID-P and the patient and cli-
nician questionnaires. All patients were outpatients re-
cruited from the community. One hundred twenty-six (44%)
of the patients were men, the mean current age was 42.8
TABLE 1. Number of Lifetime and Current Comorbid Axis I Disorders Among 288 Patients With Bipolar I and Bipolar II
Lifetime or Current
Comorbid Diagnosis
Number of Comorbid Axis I Disorders
Comparison of Bipolar I
and Bipolar II
All Patients With Bipolar
Diagnosis (N=288) Bipolar I (N=239) Bipolar II (N=49)
Mean SD Mean SD Mean SD t (df=286) p
Lifetime 1.7 2.1 1.8 2.2 1.2 1.4 –1.73 <0.09
Current 0.5 1.0 0.6 1.0 0.4 0.7 –1.39 <0.17
TABLE 2. Lifetime and Current Comorbid Axis I Disorders Among Patients With Bipolar I and Bipolar II Disorder
Number of Current or Lifetime
Comorbid Axis I Disorders
All Patients With Bipolar
Disorder (N=288) Bipolar I (N=239) Bipolar II (N=49)
Comparison of Bipolar I
and Bipolar II
N % N % N % χ
(df=1) p
Lifetime 101 35 80 33 21 43 1.57 0.25
Current 192 67 155 65 37 76 2.08 0.19
One or more
Lifetime 187 65 159 67 28 57 1.57 0.25
Current 96 33 84 35 12 24 2.08 0.19
Two or more
Lifetime 120 42 102 43 18 37 0.59 0.53
Current 37 13 32 13 5 10 0.37 0.65
Three or more
Lifetime 68 24 60 25 8 16 1.73 0.27
Current 16 6 15 6 1 2 1.39 0.33
422 Am JPsychiatry 158:3, March 2001
years (SD=11.3, range=19.1–81.9), the mean age at onset of
illness was 22.3 (SD=10.4, range=1–57), and the mean dura-
tion of illness was 20.6 years (SD=12.2, range=0.1–54.8). One
hundred sixteen (40%) of the patients reported limitation of
occupational functioning by their bipolar illness, 205 (71%)
reported one or more psychiatric hospitalizations, and 160
(56%) had a first-degree relative with a mood disorder. These
features were highly similar to those of the first 261 Stanley
Foundation Bipolar Treatment Outcome Network patients
described in greater detail elsewhere (33).
Table 1 shows that the lifetime DSM-IV axis I psychiatric
comorbidity in this group of bipolar patients was high; pa-
TABLE 3. Lifetime and Current Comorbid Axis I Disorders of Patients With Bipolar I and Bipolar II Disorder
Current or Lifetime Comorbid Axis I
All Patients With Bipolar
Diagnosis (N=288) Bipolar I (N=239) Bipolar II (N=49)
Comparison of Bipolar I
and Bipolar II
N % N % N % χ
(df=1) p
Substance use disorders
Lifetime 122 42 107 45 15 31 3.33 1.00
Current 12 4 9 4 3 6 0.57 1.00
Lifetime 96 33 85 36 11 22 3.14 1.00
Current 7 2 4 2 3 6 3.40 1.00
Lifetime 25 9 22 9 3 6 0.49 1.00
Current 1 0 1 0 0 0 0.21 1.00
Lifetime 24 8 23 10 1 2 3.06 1.00
Current 1 0 1 0 0 0 0.21 1.00
Lifetime 27 9 25 10 2 4 1.95 1.00
Current 2 1 2 1 0 0 0.41 1.00
Lifetime 20 7 20 8 0 0 4.41 0.59
Current 1 0 1 0 0 0 0.21 1.00
Lifetime 45 16 40 17 5 10 1.32 1.00
Current 4 1 4 2 0 0 0.83 1.00
Lifetime 18 6 17 7 1 2 1.78 1.00
Current 1 0 1 0 0 0 0.21 1.00
Anxiety disorders
Lifetime 122 42 100 42 22 45 0.16 1.00
Current 86 30 76 32 10 20 2.52 1.00
Panic disorder/agoraphobia
Lifetime 58 20 48 20 10 20 0.00 1.00
Current 27 9 25 10 2 4 1.95 1.00
Social phobia
Lifetime 47 16 41 17 6 12 0.72 1.00
Current 36 13 33 14 3 6 2.20 1.00
Simple phobia
Lifetime 30 10 26 11 4 8 0.32 1.00
Current 24 8 21 9 3 6 0.38 1.00
Generalized anxiety disorder
Lifetime 8 3 7 3 1 2 0.12 1.00
Current 8 3 7 3 1 2 0.12 1.00
Lifetime 19 7 17 7 2 4 0.61 1.00
Current 12 4 11 5 1 2 0.67 1.00
Lifetime 27 9 22 9 5 10 0.05 1.00
Current 22 8 19 8 3 6 0.19 1.00
Other anxiety disorders
Lifetime 8 3 6 3 2 4 0.37 1.00
Current 6 2 5 2 1 2 0.00 1.00
Eating disorders
Lifetime 17 6 11 5 6 12 4.28 0.59
Current 4 1 3 1 1 2 0.18 1.00
Anorexia nervosa
Lifetime 6 2 4 2 2 4 6.34 0.95
Current 0 0 0 0 0 0
Bulimia nervosa
Lifetime 11 4 8 3 3 6 0.85 1.00
Current 4 1 3 1 1 2 0.18 1.00
There were no significant differences by chi-square between bipolar I comorbidity and bipolar II comorbidity (df=1). Probability multiplied
by 19 to correct for multiple comparisons (Bonferroni’s inequality correction).
Am JPsychiatry 158:3, March 2001 423
tients met criteria for a mean of 1.7 (SD=2.1) lifetime non-
bipolar axis I disorders. Current comorbidity was lower;
patients met criteria for a mean of 0.5 (SD=1.0) current
nonbipolar axis I disorders. Patients with bipolar I disor-
der did not differ from those with bipolar II disorder in
their mean number of comorbid lifetime or current axis I
disorders. As shown in Table 2, 187 (65%) of the patients
met DSM-IV criteria for at least one lifetime comorbid dis-
order, and 96 (33%) met criteria for at least one current co-
morbid disorder. Sixty-eight (24%) patients had three or
more lifetime disorders. Again, bipolar I and bipolar II pa-
tients showed no differences regarding rates of lifetime or
current comorbid disorders.
Table 3 displays the rates of specific comorbid lifetime
and current axis I disorders in the entire group of patients
and by bipolar subtype. Anxiety and substance use disor-
ders were the most common comorbid lifetime disorders,
followed by eating disorders, in all patients as well as in the
bipolar I and II patients as separate groups. In addition,
both as families of disorders and individually, substance
use, anxiety, and eating disorders were equally common in
patients with bipolar I and bipolar II disorders. Alcohol
was the most commonly abused substance, and mari-
juana was the most commonly abused drug, followed by
noncocaine stimulants, cocaine, and sedatives. Panic dis-
order/agoraphobia and social phobia were the two most
common anxiety disorders. Bulimia nervosa was the most
common eating disorder.
Table 4 shows the relationships between the presence
and absence of at least one lifetime or current axis I disor-
der and selected demographic and historical illness vari-
ables. Lifetime axis I comorbidity was associated with
early age at onset of first affective symptoms and bipolar
disorder, development of more severe episodes over time,
and family history of alcoholism and drug abuse. Current
axis I comorbidity was associated with early age at onset of
first affective symptoms and illness, limited occupational
functioning, history of rapid cycling, and history of devel-
opment of both cycle acceleration and more severe epi-
sodes over time. Appendix 1 lists the demographic and
historical illness indicators associated with one or more
lifetime or current comorbid axis I disorders entered into
the regression analysis. As shown in Table 5, significant in-
dicators for lifetime comorbidity by stepwise logistic re-
gression were early onset of bipolar disorder, early onset of
affective symptoms, and family history of drug abuse. Sig-
nificant indicators for current comorbidity were again
early onset of bipolar disorder and affective symptoms, as
well as development of more severe episodes over time
and cycle acceleration.
TABLE 4. Demographic and Historical Illness Measures for 288 Bipolar Patients With and Without Lifetime and Current
Comorbid Axis I Disorders
Number of Lifetime
Comorbid Axis I Disorders
Number of Current
Comorbid Axis I Disorders
One or More
(N=187) Analysis
One or More
(N=96) Analysis
M ea n SD M ea n SD t d f p M ea n SD M ea n SD t d f p
Age (years)
Current 44.8 12.0 41.7 11.2 2.32 286 0.47 43.8 11.6 41.0 10.5 1.92 286 1.00
At onset of bipolar disorder
26.0 10.6 20.3 9.7 4.53 267 0.0002 24.0 10.3 19.0 9.6 3.84 267 0.003
At onset of first affective symptoms 23.8 11.8 17.2 9.2 5.22 282 0.000007 21.3 10.4 16.0 10.0 4.06 282 0.001
At first treatment 30.9 11.3 29.4 11.0 1.16 277 1.00 30.1 11.4 29.5 10.6 0.46 277 1.00
At first hospitalization 30.0 10.9 29.3 10.9 0.43 203 1.00 29.8 10.9 28.9 11.0 0.57 203 1.00
N % N % χ
d f p N % N % χ
d f p
Bipolar disorder diagnosis I or II
bipolar I 80 79 159 85 1.57 1 1.00 155 81 84 88 2.08 1 1.00
Income <$20,000/year 66 65 107 57 2.10 1 1.00 121 63 52 54 2.25 1 1.00
Limited occupational functioning 35 35 81 43 2.05 1 1.00 64 33 52 54 11.55 1 0.03
Suicide attempts (one or more) 23 23 64 34 4.08 1 0.99 53 28 34 35 1.85 1 1.00
Rapid cycling
48 48 105 56 2.10 1 1.00 89 46 64 67 10.34 1 0.05
59 58 113 60 0.13 1 1.00 105 55 67 70 6.37 1 0.29
Dysphoric mania
49 49 123 66 8.12 1 0.14 106 55 66 69 4.88 1 0.69
Psychiatric hospitalization 76 75 129 69 1.25 1 1.00 140 73 65 68 0.85 1 1.00
Cycle acceleration 38 38 98 52 5.75 1 0.42 77 40 59 61 11.71 1 0.03
More severe episodes 51 50 134 72 12.78 1 0.02 108 56 77 80 16.00 1 0.001
Psychiatric history in first-degree
Unipolar disorder 54 53 106 57 0.27 1 1.00 106 55 54 56 0.03 1 1.00
Bipolar disorder 34 34 89 48 5.20 1 0.55 81 42 42 44 0.06 1 1.00
Any mood disorder 62 61 128 68 1.46 1 1.00 127 66 63 66 0.01 1 1.00
Suicide 16 16 39 21 1.07 1 1.00 33 17 22 23 1.36 1 1.00
Alcoholism 26 26 88 47 12.46 1 0.009 68 35 46 48 4.18 1 1.00
Drug abuse 18 18 70 37 11.89 1 0.02 55 29 33 34 0.99 1 1.00
Probability multiplied by 22 to correct for multiple comparisons (Bonferroni’s inequality correction).
Clinician report (Clinician Questionnaire or Structured Clinical Interview for DSM-IV, Patient Edition); all others based on patient report.
Rapid cycling—four or more episodes per year.
424 Am JPsychiatry 158:3, March 2001
Our findings of high rates of lifetime and current sub-
stance use, anxiety, and, possibly, eating disorders in out-
patients with bipolar disorders are consistent with existing
data from community (1–12) and clinical (4–6, 13–16, 22,
23, 28–30, 35–37) samples, suggesting that these disorders
co-occur with bipolar disorder more often than expected
by chance alone. Indeed, in our group, bipolar disorder
was twice as likely to be accompanied by another lifetime
axis I psychiatric disorder (N=187, 65%) than to exist by it-
self (N=101, 35%). Moreover, consistent with other find-
ings, axis I comorbidity was associated with a more ad-
verse onset and course of bipolar disorder (with earlier age
at onset of affective symptoms and the bipolar disorder
syndrome and higher rates of development of cycle accel-
eration and more severe episodes over time) (7, 14, 22–31).
These findings, however, must be considered in view of
several methodological limitations. First, this study in-
cluded only patients with bipolar disorder and lacked di-
rect comparisons with a normal control group, another
psychiatrically ill group, or an epidemiologic sample. Also,
interviewers were not blind to the patients’ established di-
agnoses of bipolar disorder. The rates of comorbid axis I
disorders found in our group might therefore be falsely el-
evated because of Berkson’s bias (38) or interviewer bias.
In addition, historical illness variables reported here were
obtained retrospectively, sometimes by self-report, and it
is possible that patients with comorbid disorders were bi-
ased toward overreporting earlier age at onset and more
severe course of their bipolar symptoms.
A second limitation is that our group of bipolar patients
may not be representative of the disorder because of the
inclusion criteria of the Stanley Foundation Bipolar Treat-
ment Outcome Network. Specifically, only outpatients
willing to participate in research and to be compliant with
treatment were included. Our results, therefore, might not
be generalizable to other bipolar populations, including
those who are hospitalized or those who are not in treat-
ment. However, comparison of the Stanley Foundation Bi-
polar Treatment Outcome Network bipolar patients with
other clinical populations, including those of Winokur
(39), the NIMH Collaborative Study (40), and the National
Depressive and Manic-Depressive Association (41), re-
veals many similarities (33).
Another potential limitation is that there is considerable
phenomenologic overlap among the manic, depressive,
and mixed symptoms and episodes of bipolar disorder
and other axis I disorders. Rates of certain axis I disorders
might be falsely elevated because of misattribution of af-
fective symptoms to these other disorders. Indeed, an im-
portant complication of studies of comorbidity in general
is that as the number of recognized psychiatric diagnoses
increases, rates of comorbidity increase as well (6).
Despite these limitations, our study has several strengths.
First, to our knowledge, this is the largest cohort of indi-
viduals with bipolar disorder to have their comorbid axis I
disorders systematically assessed with structured inter-
views. Epidemiologic studies assessing axis I comorbidity
of patients with bipolar disorder have evaluated cohorts of
168 (3, 7, 8), 29 (10), and 99 (11) subjects, respectively. The
largest clinical study assessing axis I comorbidity in bipo-
lar patients (29) assessed comorbid substance abuse in
134 hospitalized bipolar I patients with a manic or mixed
episode and found that 33% met criteria for an alcohol use
disorder and 34% met criteria for a drug use disorder. Sec-
ond, this is the first study to use DSM-IV criteria to diag-
nose both bipolar and comorbid axis I disorders. Third,
patients were drawn from five different sites that were
geographically diverse, including four sites in the United
States and one site in Europe. Fourth, relationships among
comorbidity and demographic and historical illness vari-
ables were systematically evaluated.
Our findings of high rates of substance use, anxiety, and,
to a lesser degree, eating disorders in patients with bipolar
disorder have several important clinical implications. The
first is that a comprehensive evaluation of patients with
bipolar disorder should include a systematic search for
and assessment of these and other comorbid disorders.
Also, patients with uncomplicated bipolar disorder, espe-
cially children, adolescents, and young adults, need to be
observed carefully for the development of other axis I dis-
orders and educated about their risk for developing them.
Conversely, patients presenting with multiple axis I disor-
ders should be carefully evaluated for the development of
mood disorders, especially bipolar disorder. Second, in
young people with a family history of bipolar disorder, the
emergence of an early-onset anxiety, substance use, or
eating disorder should trigger consideration of a prodro-
mal mood disorder, including a bipolar disorder. Third,
our findings that axis I psychiatric comorbidity may be as-
sociated with earlier age at onset and development of cy-
cle acceleration and more severe episodes over time—
clinical features of bipolar disorder that themselves have
been associated with more severe illness or poorer out-
come—suggest that bipolar disorder with a comorbid axis
I disorder should be carefully and ultimately perhaps dif-
ferentially managed.
TABLE 5. Significant Indicators of Lifetime and Current Co-
morbid Axis I Disorders in Patients With Bipolar Disorders
According to Logistic Regression
Indicator p
Patients with one or more lifetime axis I disorders
Early onset of bipolar disorder 0.002
Early onset of affective symptoms 0.03
Family history of drug abuse 0.002
Patients with one or more current axis I disorders
Early onset of bipolar disorder 0.03
Early onset of affective symptoms 0.0002
More severe episodes 0.005
Cycle acceleration 0.02
Am JPsychiatry 158:3, March 2001 425
Our data, however, do not permit conclusions regarding
the comparative treatment of bipolar patients with and
without various comorbid axis I disorders. To our knowl-
edge, no large, prospective, controlled studies have as-
sessed the effects of treating different comorbid disorders
on the outcome of bipolar disorder or compared the effi-
cacy of different mood-stabilizing agents in bipolar pa-
tients with selected comorbid disorders. Such prospective
data are necessary to determine whether differential treat-
ment based on comorbidity improves the outcome of bi-
polar disorder. Nonetheless, consideration of axis I co-
morbidity in bipolar disorder might represent yet another
clinically useful way of characterizing or subtyping the ill-
ness. Furthermore, greater attention should be focused on
the efficacy of agents with antimanic or mood-stabilizing
properties in the treatment of pathological substance use,
anxiety, and eating syndromes.
How might the extensive overlap of bipolar disorder with
these other axis I disorders be explained? One possibility is
that these disorders are in fact distinct but unrelated dis-
ease entities, representing either risk factors for each other
or the sharing of similar end states from different etiologic
mechanisms. This possibility is supported in part by the
findings of high rates of substance use disorders in the
families of patients with comorbid substance use disorder,
suggesting the inheritance of two (or more) illnesses in
some patients. Another possibility, however, is that bipolar
disorder is related to substance use, anxiety, and eating dis-
orders (or at least some forms of these disorders) and, by
extension, that all of these disorders may be related to one
another and share a common underlying pathophysiologic
etiology (8). This possibility is supported by three large-
scale epidemiologic studies indicating that mood, anxiety,
and substance use disorders are highly comorbid with one
another (3, 9, 11, 42). It is also supported by a growing body
of research suggesting that substance use, anxiety, and eat-
ing disorders may be associated with elevated familial rates
of mood disorders and may respond to agents with thymo-
leptic (antidepressant and/or mood-stabilizing) properties
(21, 43). Indeed, preliminary research indicates that dys-
regulation in several common neurotransmitter systems—
including the dopaminergic, serotonergic, noradrenergic,
and γ-aminobutyric-acid-ergic—may be important in the
pathophysiology of all of these disorders (4, 44, 45). Further
research examining the overlap of these disorders would
therefore appear to be just as important as further research
into their differences.
In sum, our findings are consistent with others suggest-
ing that bipolar disorder is more often than not accompa-
nied by other axis I disorders. Moreover, axis I comorbidity
may be associated with clinical features that themselves
have been associated with greater severity and/or poorer
outcome of bipolar disorder. In subsequent work, we will
extend these observations to the impact of axis I comor-
bidity on the prospectively evaluated course of illness and
differential response to treatment of bipolar disorder.
Received June 15, 1998; revisions received Dec. 23, 1999, and
March 2, 2000; accepted March 24, 2000. From the Stanley Founda-
tion Bipolar Treatment Outcome Network, including the Biological
Psychiatry Program, Department of Psychiatry, University of Cincin-
nati College of Medicine; the University of California, Los Angeles,
Neuropsychiatric Institute and the West Los Angeles Veterans Affairs
Medical Center, Los Angeles; the Southwest Medical Center at Dallas;
the Biological Psychiatry Branch, NIMH, Bethesda, Md.; and the HC
Rümke Group, Willem Arntsz Huis and Academic Hospital, Utrecht,
the Netherlands. Address reprint requests to Dr. McElroy, Biological
Psychiatry Program, University of Cincinnati College of Medicine,
P.O. Box 670559, 231 Bethesda Ave., Cincinnati, OH 45267-0559; (e-mail).
Supported in part by the Theodore and Vada Stanley Foundation.
1. Boyd JH, Burke JD, Gruenberg E, Holzer LE III, Rae DS, George
LK, Karno M, Stoltzman T, McEvoy L, Nestadt G: Exclusion crite-
ria of DSM-III: a study of co-occurrence of hierarchy free syn-
dromes. Arch Gen Psychiatry 1984; 41:983–989
2. Helzer J, Pryzbeck T: The co-occurrence of alcoholism with
other psychiatric disorders in the general population and its
impact on treatment. J Stud Alcohol 1988; 49:219–224
3. Regier DA, Farmer ME, Rae DS, Locke BZ, Keith SJ, Judd LL,
Goodwin FK: Comorbidity of mental disorders with alcohol
and other drug abuse: results from the Epidemiologic Catch-
ment Area (ECA) study. JAMA 1990; 264:2511–2518
4. Goodwin FK, Jamison KR: Manic-Depressive Illness. New York,
Oxford University Press, 1990
5. Brady KT, Lydiard RB: Bipolar affective disorder and substance
abuse. J Clin Psychopharmacol 1992; 12(suppl 1):17–22
APPENDIX 1. Demographic and Historical Illness Indicators
Associated With One or More Lifetime or Current Comor-
bid Axis I Disorders in 288 Bipolar Patients: Logistic Re-
gression Analysis
The following indicators of demographic and illness variables were
tested. Chi-square tests were completed on each variable and
Bonferroni’s inequality correction was applied. An alpha of 0.2
was chosen as the cutoff for inclusion in the logistic regression.
Current age
Early onset of bipolar disorder (14 years or younger)
Early symptoms (10 years or younger)

Bipolar disorder diagnosis (bipolar I versus bipolar II)
Income (<$20,000)
Limited occupational functioning
Suicide attempts (one or more)
History of rapid cycling
History of dysphoric mania
History of psychosis
History of psychiatric hospitalization
Cycle acceleration over time
More severe episodes over time
Family history of a depressive disorder
Family history of bipolar disorder
Family history of any mood disorder
Family history of schizophrenia
Family history of suicide
Family history of alcoholism
Family history of drug abuse
This variable was included in the logistic regression for zero versus
one or more lifetime and current axis I disorders.
This variable was included in the logistic regression for zero versus
one or more lifetime axis I disorders.
This variable was included in the logistic regression for zero versus
one or more current axis I disorders.
426 Am JPsychiatry 158:3, March 2001
6. Strakowski SM, McElroy SL, Keck PE Jr, West SA: The co-occur-
rence of mania with medical and other psychiatric disorders.
Int J Psychiatry Med 1994; 24:305–328
7. Chen Y-W, Dilsaver SC: Comorbidity of panic disorder in bipolar
illness: evidence from the Epidemiologic Catchment Area sur-
vey. Am J Psychiatry 1995; 152:280–282
8. Chen Y-W, Dilsaver SC: Comorbidity for obsessive-compulsive
disorder in bipolar and unipolar disorders. Psychiatry Res
1995; 59:57–64
9. Kessler RC, Nelson CB, McGonagle KA, Edlund MJ, Frank RG,
Leaf PJ: The epidemiology of co-occurring addictive and men-
tal disorders: implications for prevention and service utiliza-
tion. Am J Orthopsychiatry 1996; 66:17–31
10. Kessler RC, Rubinow DR, Holmes C, Abelson JM, Zhao S: The ep-
idemiology of DSM-III-R bipolar I disorder in a general popula-
tion survey. Psychol Med 1997; 27:1079–1089
11. Ravelli A, Bijl RV, van Zessen G: Comorbiditeit van psychia-
trische stoornissen in de Nederlandse bevolking: resultaten
van de Netherlands Mental Health Survey and Incidence
Study (NEMESIS). Tijdschrift voor Psychiatrie 1988; 40:531–
12. Szádóczky E, Papp ZS, Vitrai J, Ríhmer Z, Füredi J: The preva-
lence of major depressive and bipolar disorders in Hungary:
results from a national epidemiologic survey. J Affect Disord
1998; 50:153–162
13. Strakowski SM, Tohen M, Stoll AL, Faedda GL, Goodwin DC: Co-
morbidity in mania at first hospitalization. Am J Psychiatry
1992; 149:554–556
14. Strakowski SM, Tohen M, Stoll AL, Faedda GL, Mayer PV, Kolbr-
ener ML, Goodwin DC: Comorbidity in psychosis at first hospi-
talization. Am J Psychiatry 1993; 150:752–757
15. McElroy SL, Strakowski SM, Keck PE Jr, Tugrul KL, West SA, Lon-
czak HS: Differences and similarities in mixed and pure mania.
Compr Psychiatry 1995; 36:187–194
16. Kruger S, Shugar G, Cooke RG: Comorbidity of binge eating dis-
order and the partial binge-eating syndrome with bipolar dis-
order. Int J Eat Disord 1996; 19:45–52
17. Savino M, Perugi G, Simonini E, Soriani A, Cassano GB, Akiskal
HS: Affective comorbidity in panic disorder: is there a bipolar
connection? J Affect Disord 1993; 28:155–163
18. Bowen R, South M, Hawkes J: Mood swings in patients with
panic disorder. Can J Psychiatry 1995; 39:91–94
19. Hudson JI, Pope HG Jr, Yurgelun-Todd D, Jonas JM, Frankenburg
FR: A controlled study of lifetime prevalence of affective and
other psychiatric disorders in bulimic outpatients. Am J Psychi-
atry 1987; 144:1283–1287
20. Halmi KA, Eckert E, Marchi P, Sampugnaro V, Apple R, Cohen J:
Comorbidity of psychiatric diagnoses in anorexia nervosa. Arch
Gen Psychiatry 1991; 48:712–718
21. Simpson SG, Al-Mufti R, Andersen AE, De Paulo JR: Bipolar II af-
fective disorder in eating disorder inpatients. J Nerv Ment Dis
1992; 180:719–722
22. Sonne SC, Brady KT, Morton WA: Substance abuse and bipolar
disorder. J Nerv Ment Dis 1994; 182:349–352
23. Himmelhoch JM, Garfinkel ME: Sources of lithium resistance in
mixed mania. Psychopharmacol Bull 1986; 22:613–620
24. Black DW, Winokur G, Hulbert J, Nasrallah A: Predictors of im-
mediate response in the treatment of mania: the importance
of comorbidity. Biol Psychiatry 1988; 24:191–198
25. Black DW, Winokur G, Bell S, Nasrallah A, Hulbert J: Compli-
cated mania: comorbidity and immediate outcome in the
treatment of mania. Arch Gen Psychiatry 1988; 45:232–236
26. Black DW, Hulbert J, Nasrallah A: The effect of somatic treat-
ment and comorbidity on immediate outcome in manic pa-
tients. Compr Psychiatry 1989; 30:74–79
27. Tohen M, Waternaux CM, Tsuang MT, Hunt AT: Four-year fol-
low-up of twenty-four first-episode manic patients. J Affect Dis-
ord 1990; 19:79–86
28. Young LT, Cooke RG, Robb JC, Leavitt AJ, Joffe RT: Anxious and
non-anxious bipolar disorder. J Affect Disord 1993; 29:49–52
29. Keck PE Jr, McElroy SL, Strakowski SM, West SA, Sax KW, Hawkins
JM, Bourne ML, Haggard P: Twelve-month outcome of patients
with bipolar disorder following hospitalization for a manic or
mixed episode. Am J Psychiatry 1998; 155:646–652
30. Goldberg JF, Garno JL, Leon AC, Kocsis JH, Portera L: A history of
substance abuse complicates remission from acute mania in
bipolar disorder. J Clin Psychiatry 1999; 60:733–740
31. O’Connell RA, Mayo JA, Flatow L, Cuthbertson B, O’Brien BE:
Outcome of bipolar disorder on long-term treatment with lith-
ium. Br J Psychiatry 1991; 159:123–129
32. Leverich GS, Nolen W, Rush AJ, McElroy SL, Keck PE Jr, Denicoff
RD, Suppes T, Altshuler LL, Kupka R, Frye MA, Kramlinger KG,
Post RM: The Stanley Foundation Bipolar Treatment Outcome
Network, I: longitudinal methodology. J Affect Disord (in
33. Suppes T, Leverich GS, Keck PE Jr, Nolen WA, Denicoff KD, Alt-
shuler LL, McElroy SL, Rush AJ, Kupka R, Frye MA, Bickel M, Post
RM: The Stanley Foundation Bipolar Treatment Outcome Net-
work, II: demographics and illness characteristics of the first
261 patients. J Affect Disord (in press)
34. First MB, Spitzer RL, Gibbon M, Williams JBW: Structured Clini-
cal Interview for DSM-IV Axis I Disorders, Patient Edition (SCID-
P), version 2. New York, New York State Psychiatric Institute, Bi-
ometrics Research, 1995
35. Winokur G, Coryell W, Akiskal HS, Maser JD, Keller MB, Endicott J,
Mueller T: Alcoholism in manic-depressive (bipolar) illness: famil-
ial illness, course of illness, and the primary-secondary distinc-
tion. Am J Psychiatry 1995; 152:365–372; correction, 152:1106
36. Krüger S, Cooke RG, Hasey GM, Jorna T, Persad E: Comorbidity
of obsessive-compulsive disorder in bipolar disorder. J Affect
Disord 1995; 34:117–120
37. Pini S, Cassano GB, Simirno M, Russo A, Montgomery SA: Prev-
alence of anxiety disorders comorbidity in bipolar depression,
unipolar depression and dysthymia. J Affect Disord 1997; 42:
38. Berkson J: Limitations of the application of fourfold table anal-
ysis to hospital data. Biometric Bulletin 1946; 2:47–53
39. Winokur G: The Iowa 500: heterogeneity and course in manic-
depressive illness (bipolar). Compr Psychiatry 1975; 16:125–131
40. Katz MM, Secunda SK, Hirschfeld RMA, Koslow SH: NIMH Clini-
cal Research Branch Collaborative Program on the Psychobiol-
ogy of Depression. Arch Gen Psychiatry 1979; 36:765–771
41. Lish JD, Dime-Meenan S, Whybrow PC, Price RA, Hirschfeld
RMA: The National Depressive and Manic-Depressive Associa-
tion (DMDA) survey of bipolar members. J Affect Disord 1994;
42. Kessler RC, McGonagle KA, Zhao S, Nelson CB, Hughes M, Eshle-
man S, Wittchen H-U, Kendler KS: Lifetime and 12-month prev-
alence of DSM-III-R psychiatric disorders in the United States:
results from the National Comorbidity Survey. Arch Gen Psychi-
atry 1994; 51:8–19
43. Hudson JI, Pope HG Jr: Affective spectrum disorder: does antide-
pressant response identify a family of disorders with a common
pathophysiology? Am J Psychiatry 1990; 147:552–564
44. Musselman DL, DeBattista C, Nathan KI, Kilts CD, Schatzberg
AF, Nemeroff CB: Biology of mood disorders, in The American
Psychiatric Press Textbook of Psychopharmacology, 2nd ed.
Edited by Schatzberg AF, Nemeroff CB. Washington, DC, Amer-
ican Psychiatric Press, 1998, pp 549–588
45. Dilsaver SC: The pathophysiologies of substance abuse and af-
fective disorders: an integrative model? J Clin Psychopharma-
col 1987; 7:1–10