Differences in incidence of suicide attempts
during phases of bipolar I and II disorders
With a lifetime risk of a non-fatal suicide attempt
ranging from 25% to 56% (1–3), patients with
bipolar disorder (BD) are at higher risk for suicide
attempts than are patients with any other Axis I
disorder (4). While information on risk factors for
suicidal behavior is accumulating (5), a major
problem for research in this area is the lack of
studies relating suicidal behavior to the most
pathognomonic feature of the disorder: the recur-
rent and pleomorphic course.
The long-term course of BD is chronic and
dominated by depressive symptoms (6–8). In two
recent prospective studies, suicidal behavior was
related to depressive aspects of the illness (9, 10).
Marangell et al. (9) found that history of suicide
attempts and percentage of days spent depressed in
the year prior to the participants? entry into the
Valtonen HM, Suominen K, Haukka J, Mantere O, Leppa ¨ ma ¨ ki S,
Arvilommi P, Isometsa ¨ ET. Differences in incidence of suicide attempts
during phases of bipolar I and II disorders.
Bipolar Disord 2008: 10: 588–596. ª 2008 The Authors Journal
compilation ª 2008 Blackwell Munksgaard
Background: Differences in the incidence of suicide attempts during
various phases of bipolar disorder (BD), or the relative importance of
static versus time-varying risk factors for overall risk for suicide
attempts, are unknown.
Methods: We investigated the incidence of suicide attempts in different
phases of BD as a part of the Jorvi Bipolar Study (JoBS), a naturalistic,
prospective, 18-month study representing psychiatric in- and outpatients
with DSM-IV BD in three Finnish cities. Life charts were used to
classify time spent in follow-up in the different phases of illness among
the 81 BD I and 95 BD II patients.
Results: Compared to the other phases of the illness, the incidence of
suicide attempts was 37-fold higher [95% confidence interval (CI) for
relative risk (RR): 11.8–120.3] during combined mixed and depressive
mixed states, and 18-fold higher (95% CI: 6.5–50.8) during major
depressive phases. In Cox?s proportional hazards regression models,
combined mixed (mixed or depressive mixed) or major depressive phases
and prior suicide attempts independently predicted suicide attempts.
No other factor significantly modified the risks related to these time-
varying risk factors; their population-attributable fraction was 86%.
Conclusions: The incidence of suicide attempts varies remarkably
between illness phases, with mixed and depressive phases involving the
highest risk by time. Time spent in high-risk illness phases is likely the
major determinant of overall risk for suicide attempts among BD
patients. Studies of suicidal behavior should investigate the role of
both static and time-varying risk factors in overall risk; clinically,
management of mixed and depressive phases may be crucial in
Hanna M Valtonena,b, Kirsi
Suominena,b, Jari Haukkaa, Outi
Manterea, Sami Leppa ¨ma ¨kia,c, Petri
Arvilommia,band Erkki T Isometsa ¨a,c
aDepartment of Mental Health and Alcohol
Research, National Public Health Institute, Helsinki,
bDepartment of Psychiatry, Jorvi Hospital, Helsinki
University Central Hospital, Espoo,cDepartment of
Psychiatry, Helsinki University Central Hospital,
Key words: bipolar disorder – bipolar I disorder –
bipolar II disorder – depressive mixed phase –
depressive phase – mixed phase – prospective
study – suicide attempt
Received 23 March 2007, revised and accepted for
publication 28 September 2007
Corresponding author: Erkki T. Isometsa ¨, MD, PhD,
Department of Psychiatry, Institute of Clinical
Medicine, University of Helsinki, PO Box 22,
Helsinki, 00014 Finland. Fax: + 358 9 4716 3735;
The authors of this paper do not have any commercial associations
that might pose a conflict of interest in connection with this manu-
Bipolar Disorders 2008: 10: 588–596
ª 2008 The Authors
Journal compilation ª 2008 Blackwell Munksgaard
study were significantly associated with suicide
attempts or completions. Likewise, we have previ-
ously reported that a depressive phase at the index
episode independently predicted suicide attempters
(10). Furthermore, mixed states are also related
to suicidal behavior in retrospective and cross-
sectional studies (11–20).
Overall, few prospective studies have focused on
suicidal behavior in BD. In addition, there is a lack
of studies on risk factors for attempted suicide that
take into account the time spent in the risk phases
during follow-up. To our knowledge, no previous
studies have examined the incidence of suicide
attempts, or differences in incidence during various
phases of BD. Moreover, previous studies have
been unable to estimate the relative importance of
static versus time-varying risk factors for overall
risk for suicide attempts or the importance of time
spent at risk for overall risk.
We previously reported in our cross-sectional
study that the highest level of suicidal ideation was
related to the mixed phases of the illness (20), and
in our preceding prospective study, that depressive
aspects of the illness at baseline independently
predicted future suicide attempts (10). In this
prospective study, our aim was to investigate
whether the incidence of attempted suicide is
higher during depressive and mixed phases than
in other phases, and to examine if common risk
factors associated with suicidal behavior in BD
(comorbid anxiety disorders, comorbid personality
disorders, comorbid substance dependence ⁄abuse,
hopelessness, severity of depression and lifetime
psychotic features) modify this risk during depres-
sive and mixed phases. Furthermore, we estimated
the population-attributable fraction (PAF) for
these putative high-risk phases.
The background and methodology of the Jorvi
Bipolar Study (JoBS) have been described in
detail elsewhere (21). In brief, the JoBS is a
collaborative research project between the Depart-
ment of Mental Health and Alcohol Research of
the National Public Health Institute, Helsinki,
and the Department of Psychiatry, Jorvi Hospital,
Helsinki University Central Hospital (HUCH),
Espoo, Finland. The Department of Psychiatry at
Jorvi Hospital provides secondary care psychiatric
services to all residents of Espoo, Kauniainen and
Kirkkonummi (261,116 inhabitants in 2002). The
HUCH Ethics Committee approved the study
Screening, diagnostic evaluation and baseline
In the first phase, both psychiatric in- and outpa-
tients at the Department of Psychiatry at Jorvi
Hospital were screened for a possible new episode
of DSM-IV BD from 1 January 2002 to 28
February 2003. Attending mental health profes-
sionals in the department screened every patient
between 18 and 59 years of age who was: (i)
seeking treatment; (ii) being referred; or (iii)
already receiving care and now showing signs of
deteriorating clinical state for the presence of BD
using the Mood Disorder Questionnaire (MDQ)
(22). After a positive MDQ screen or suspicion of
BD (n = 28), the patient was fully informed of the
study protocol and written informed consent was
requested. Of the 1,630 patients screened, 546
scored positive (a positive MDQ screen or suspi-
cion of BD); 49 of these refused a face-to-face
interview, and 7 could not be contacted because
they neither appeared for their appointment nor
answered repeated telephone requests.
In the second phase of sampling, experienced
research psychiatrists (OM, HMV, PA, KS, SL,
Marita Pippingsko ¨ ld) interviewed 490 patients
face-to-face using the Structured Clinical Interview
for DSM-IV Disorders, research version with
psychotic screen (SCID-I ⁄P) (23). Diagnosis was
made using all available information from the face-
to-face interview and psychiatric records; if the
diagnosis was uncertain, the attending personnel,
Of the 546 patients with positive screens, 490
were interviewed and 201 patients were assigned a
diagnosis of BD after the SCID-I/P interview and a
current episode of the disorder. Ten patients
refused to participate, leaving 191 patients in the
bipolar cohort study. In the 20 randomly selected,
videotaped diagnostic interviews blindly assessed
by another diagnostician, we found complete
agreement (for BD overall, j = 1.0; for BD I
j = 1.0;for
j = 1.0) (21). The Structured Clinical Interview
for DSM-IV personality disorders (SCID-II) (24)
was used to assess diagnoses on Axis II. The cohort
baseline measurements included the following
observer scales: the Young Mania Rating Scale
(YMRS) (25), the 17-item Hamilton Depression
Rating Scale (HAM-D) (26), the Scale for Suicidal
Ideation (SSI) (27) and the Social and Occupa-
tional Functioning Assessment Scale of DSM-IV
(SOFAS) (28). The self-reported scales included the
21-item Beck Depression Inventory (BDI) (29), the
Beck Anxiety Inventory (BAI) (30), the Beck
Incidence of suicide attempts in BD
Hopelessness Scale (BHS) (31) and the Perceived
Social Support Scale-Revised (PSSS-R) (32). A
bipolar II depressive mixed state was defined
according to Benazzi and Akiskal (three or more
simultaneous intra-episode hypomanic symptoms
present for at least 50% of the time during a major
depressive episode) (33).
Details of the follow-up methodology (e.g., the life
chart methodology) have been described in detail
elsewhere (8, 34). The life chart methodology was
usedtointegrateall availableinformation aboutthe
nature and duration of different phases during the
18-month follow-up. The JoBS initially included
191 patients with DSM-IV BD, 15 (8%) of whom
neither continued in treatment nor participated in
the 6-month follow-up. Of the 176 included in the
6-month follow-up, 3 patients (2%) died between
the 6- and 18-month follow-up visits [2 (1%) by
suicide], and 10 patients dropped out. Overall, this
study included 176 patients; the mean length of
follow-up time was 1.6 person-years. Of the 176
patients,85 (48%) were males, 91 (52%) females, 81
(46%) BD I patients, and 95 (54%) BD II patients.
Of the 176 patients, 89 (51%) had attempted suicide
before follow-up [40 ⁄81 (49%) of BD I and 49 ⁄95
(52%) of BD II patients], 88 (50%) had lifetime
comorbid substancedependence or abuse, 95 (54%)
any lifetime anxiety comorbidity, 79 (45%) comor-
bid personality disorder and 87 (49%) had lifetime
psychotic features (10).
Integration of information into a life chart
After baseline assessments, patients were prospec-
investigated at 6 and 18 months by repeated SCID-
I ⁄P interviews. In addition, all observer- and self-
reported scales were included in both follow-up
assessments. All medical and psychiatric records
up were examined by gathering all available data,
which were then integrated into the form of a
graphic life chart based on DSM-IV criteria, anal-
Study (34). This was created after reviewing all the
information from the follow-up period with the
patient at the 6- and 18-month interviews, which
on symptom ratings and visits to attending person-
nel, we also inquired about change points in the
patients? psychopathologic states using probes
related to important life events to improve the
accuracy of the assessment. Time after the baseline
euthymia, manic, hypomanic, major depressive,
mixed, depressive mixed, cyclothymic, substance-
induced mood phase, and depressive and hypo-
manic symptoms (8). However, we deviated from
the DSM-IV in two ways: by accepting hypomanias
lasting between 2 and 3 days as hypomanic epi-
sodes, and by including depressive mixed episodes
for both BD I and BD II patients. These were
defined according to Benazzi and Akiskal (33) to
comprise major depressive episodes with three or
more concurrent intra-episode hypomanic symp-
toms present for at least 50% of the time (8). Mixed
states were defined as mixed affective episodes (full
syndromal criteria of mania and major depression
concurrently) in the DSM-IV (35).
Information about the occurrence of a suicide
attempt during the follow-up was based on both the
interview and the psychiatric and medical records.
A suicide attempt was defined as self-injurious
behavior with a non-fatal outcome accompanied by
evidence (either explicit or implicit) that the person
had intended to die (36); self-harm with no suicidal
intention was excluded. As non-fatal suicidal
behavior is among the main research foci of the
JoBS, we carefully evaluated suicidal behavior (e.g.,
suicide attempts with intention to die) not commu-
nicated to professional staff, suicide attempts not
leading to an emergency room visit, the timing
of suicide attempts, and the lethality of suicide
attempts. The BHS and the HAM-D were used at
baseline and at the 6- and 18-month follow-ups; the
highest value obtained was used in the analysis. The
incidence of suicide attempts during combined
mixed and depressive mixed states was first com-
pared to those in the other phases (including mania,
depressive symptoms, cyclothymia, and substance-
induced phases, but excluding major depressive
phases); thereafter the incidence of suicide attempts
during major depressive phases was compared to
those in other phases (excluding the combined
mixed phase). As there were no suicide attempts
during mania, hypomania, hypomanic symptoms,
euthymia, cyclothymia or a substance-induced
phase, and only four suicide attempts during
depressive symptoms, these phases were combined
and classified as the ?other phases? in the analyses.
The Poisson regression model was used to investi-
gate univariate relative risks (RRs). Cox?s propor-
tional hazards regression model with individual
frailty was used to investigate the relative impor-
tance ofindividual risk factorsof attemptedsuicide.
Valtonen et al.
In the model predicting suicide attempts during
follow-up, gender, age, subtype of BD, lifetime
comorbid substance dependence ⁄abuse, lifetime
comorbid anxiety disorder, suicide attempts before
follow-up, comorbid personality disorders, lifetime
psychoticfeatures,and the highest valueof the BHS
and HAM-D were used as fixed covariates. The
current phase in follow-up period was used as a
time-varying covariate. For each individual, the
follow-up time was cut into time periods, where the
value of the time-dependent variable (i.e., phase)
was constant. Thus, the follow-up time for each
individual consisted of several contiguous time
periods, each defined by specific entry and exit
times. For continuous variables (age, and HAM-D
and BHS scores), hazard ratios (HRs) were calcu-
lated for increments of 10 units. The interactions
between the time-varying phase and other risk
factors were tested using Cox?s model with the
likelihood ratio test. The proportional hazards
assumption was checked using the Grambsch and
Therneau (37) method and we detected no devia-
tion from proportionality. The PAF was defined as
pEðk ? 1Þ=½1þpEðk ? 1Þ?, where pEis the probabil-
ity of exposure and k is the RR (exposed versus
unexposed). We calculated PAF for the time-
varying phase, comparing combined mixed and
major depressive phases to other phases.
Incidence of suicide attempts during different phases of
During the follow-up, 20% of patients (35 ⁄176)
reported at least one suicide attempt; altogether
there were 53 suicide attempts, all of which could be
timed. No suicide attempts occurred during either
manic ⁄hypomanic or euthymic phases. Of the 53
suicide attempts, 39 (73.6%) occurred during a
major depressive phase, 7 (13.2%) during a depres-
sive mixed phase, 3 (5.7%) during a mixed phase
and 4 (7.5%) during depressive symptoms (Fig. 1).
The incidence of suicide attempts (Table 1) was
high in depressive phases (0.42) and highest in
combined mixed (mixed and depressive mixed)
phases (0.88), but low in other phases (0.02, other
Factors associated with high incidence of suicide
attempts during follow-up
The incidence of suicide attempts was higher
among females than among males and among
BD II than among BD I patients (Table 1). When
other risk factors were taken into account, how-
ever, a subtype of BD or sex did not independently
predict suicide attempts. Lifetime comorbid anxi-
ety disorders and comorbid personality disorders
were related to a higher incidence of suicide
attempts, whereas lifetime comorbid substance
abuse ⁄dependence showed no association with
the higher incidence of suicide attempts. The higher
the HAM-D or BHS scores, the higher were the
RRs of incidence of suicide attempts.
While many of the static variables raised the
unadjusted RR to 2- to 3-fold, and high hopeless-
ness even to 8-fold, the order of magnitude in the
increase of RR was higher for time-varying risk
factors. The incidence of suicide attempts increased
37-fold during combined mixed and depressive
mixed states compared to the other phases (mania,
depressive symptoms, cyclothymia, and substance-
induced phase) and 18-fold during the major
depressive phase compared to these other phases.
In Cox?s proportional hazard model, which
predicted suicide attempts during follow-up, the
predetermined covariates included gender, age,
ence ⁄abuse, any lifetime comorbid anxiety disor-
der, suicide attempts before follow-up, comorbid
personality disorders, lifetime psychotic features,
Incidence per person−year
0.0 0.5 1.0 1.5 2.0 2.5
Fig. 1. Incidence of suicide attempts per person-year with 95%
confidence intervals, with respect to phases of bipolar disorder.
Incidence of suicide attempts in BD
the highest BHS score, the highest HAM-D score,
and the time spent in risk phases, both combined
mixed phases (mixed and depressive mixed) and
depressive phases. In this model, suicide attempts
before follow-up increased the risk almost 4-fold,
whereas the major depressive phase increased the
depressive mixed) increased the risk almost 28-fold
(Fig. 2). None of the comorbid disorders or the
static cross-sectional measures of depression or
hopelessness independently predicted suicide at-
tempts. We detected no significant interaction
between either of the high-risk phases and the static
risk factors or completed suicide of first-degree
no evidence indicated that the static risk factors
modified the effect of the high-risk phases).
The attributable fractions of the time-varying risk factors
Besides the hazard, time spent in high-risk phases
also markedly influences PAFs. Patients spent 38%
of their time in major depressive phases or in
combined mixed phases. Bipolar II patients spent
more time in risk phases than did BD I patients (in
the depressive phase: 56.5 BD II versus 36.1 BD I
person-years; in combined mixed phases: 6.7 BD II
versus 4.8 BD I person-years).
Table 1. Risk factors for suicide attempts
attemptsPerson-years IncidenceRR 95% CIHR 95% CI
Age at baselinea
Subtype of bipolar disorder
Suicide attempt before follow-up
Lifetime comorbid substance dependence ⁄abuse
Any lifetime comorbid anxiety disorder
Comorbid personality disorder
Lifetime psychotic features
Time spent in other phasesd
Time spent in mixed phasese
Time spent in major depressive phases
1.38–4.55 1.27 0.57–2.84
1.13–3.67 1.32 0.58–3.02
2.49–10.4444 3.98 1.60–9.90
Relative risk (RR) from univariate Poisson model, and hazard ratio (HR) from multivariate Cox?s model.
aFor continuous variables (age, HAM-D and BHS) HRs are calculated for increments of 10 units.
bHAM-D max is defined as the maximum score on the HAM-D and BHS max is defined as the maximum score on the BHS. Both scales
were divided into tertiles.
cOne BHS score was missing.
dOther phases include manic and hypomanic phases, hypomanic symptoms, euthymic, cyclothymic, and substance-induced mood
eMixed phases include both mixed and depressive mixed phases.
RR = relative risk; 95% CI = 95% confidence interval; HR = hazard ratio; BD = bipolar disorder; HAM-D = Hamilton Depression Rating
Scale; BHS = Beck Hopelessness Scale.
Valtonen et al.
The total PAF of the two high-risk phases was
86% (the pooled estimate of HR was 17.7). Due to
the much higher proportion of time spent in major
depressive than in combined mixed phases, the
attributable fraction of the major depressive phase
is higher, even though RR may be higher in the
combined mixed phase.
To our knowledge, no previous study has investi-
gated variation in the incidence of suicide attempts
during different phases of BD. The incidence of
suicide attempts varies remarkably depending on
the type of illness phase: it increased 37-fold during
mixed phases and 18-fold during major depressive
phases. In Cox?s proportional hazards regression
models, combined mixed (mixed or depressive
mixed) or depressive phases, and prior suicide
attempts independently and strongly predicted
suicide attempts. Given a PAF of 86% and a lack
of significant interactions, the risk factors that best
predicted suicide attempts were obviously the high-
risk phases, not the individual characteristics of the
Strengths and limitations
The present study has some major methodological
strengths, as it involved a relatively large (n = 176)
sample of both in- and outpatients in three Finnish
The patients were carefully diagnosed using struc-
tured interviews with excellent reliability for BD I
and II (j = 1.0 and 1.0, respectively), plus infor-
mation on all comorbid Axis I and II disorders at
was the life chart methodology. The graphic life
chart used in this study is similar, but not identical,
to the Longitudinal Interval Follow-up Evaluation
(LIFE) or National Institute of Mental Health
(NIMH) life chart methodology used in other
prospective studies, both including and reporting
separately on BD types I and II (6, 7, 38–41). This
type of graphic life chart was also planned and
used in the Vantaa Depression Study (34).
Nevertheless, some methodological limitations
should be noted. First, we gathered no information
on lifetime aggression and impulsivity, which were
studies (9, 42). Aggression and impulsive behavior
are, however, commonly related to comorbid Clus-
ter B personality disorders (42) and substance or
alcohol abuse or dependence, neither of which
independently predicted suicide attempts in our
prospective study (10), or significantly modified risk
in the present study. Second, the life chart method-
ology is generally accepted as part of follow-up
studies on BD. Despite the special concern of
patients and treating personnel in recording mood
shifts during the mean time, we cannot fully exclude
(as cannot any other studies not based on daily
prospective mood ratings) (6, 7) the possibility that
or depressive mixed episodes, may be under-
reported. Third, circularity could in theory exist
between suicide attempts and diagnosing major
depressive phases in DSM-IV. However, as we
reported in our previous cross-sectional study, all
suicide attempters also reported suicidal ideation
(3), which alone is sufficient to fulfill criteria on A9
for a major depressive episode. Thus, this theoret-
ical circularity is unlikely to bias the relationship
between major depressive phases and suicide
0.2 1.0 5.0 25.0 100.0 10.0
Age (10 years)
Female versus male
Bipolar II versus I
Major depressive phase
versus other phase
Combined mixed phases
versus other phases
Fig. 2. Risk factors of suicide attempts. Hazard ratios with
BHS = Beck Hopelessness
Depression Rating Scale.
HAM-D = Hamilton
Incidence of suicide attempts in BD
attempts. Fourth, the study is naturalistic and
treatment was not controlled for. However, treat-
ment may influence risk for suicide attempts in two
principal ways: (i) by influencing (reducing) time
spent in illness phases, and possibly, (ii) by modi-
fying the risk in each or some of the illness phases
independent of the impact on mood. As our main
outcome measure is incidence of suicide attempts
per patient months, treatments that merely reduce
time spent ill [have effect (i)] should not significantly
bias incidence estimates. If they do not have other
risk-modifying effects [effect (ii)], they do reduce the
overall rate of suicide attempts as time at risk is
reduced, but should not markedly influence inci-
dence rates per month spent in each phase. Overall,
we believe that our incidence estimates represent
realistic and generalizable risk estimates under
usual treatment conditions in secondary-level psy-
chiatric care. Fifth, we could only investigate
average risk for time spent in the risk phases. This
relatively crude classification is, despite the marked
contrast observed between phases, unlikely to
reveal fully the true variations in risk that likely
depend on the severity of symptoms. Such expo-
sure)response gradients appear in Table 1 and
Fig. 1. The risk for suicide attempts likely covaries
with variations in levels of hopelessness, depression
and, perhaps, anxiety, none of which we could
measure on a daily basis. Future studies based on
daily prospective mood and symptom ratings may
provide a more accurate view of variations in risk
according to variations in clinical state.
Incidence of suicide attempts and risk factors of suicide
attempts during major depressive and combined mixed
In our study, the incidence of suicide attempts
during major depressive phases was 0.42 per
patient-year. In the Vantaa Depression Study (43),
a comparable study of unipolar patients, the inci-
No previous studies have examined variation in the
of BD in bipolar patients. We found, in agreement
with previous studies (9, 42, 44), that a history of
previous suicide attempts increased the risk of
futuresuicideattempts nearly4-fold. Wealso found
that combined mixed (mixed or depressive mixed)
or major depressive phases independently predicted
suicide attempts. It is clinically challenging to
diagnose these mixed states, which have been
previously associated with suicidal behavior in
retrospective and cross-sectional studies (11–20).
We found no evidence that static risk factors
significantly modify the effects of the high-risk
phases. This does not, however, exclude the possi-
bility that such factors exist, but remained unde-
power or because we were unable to measure them.
For example, we did not have a comparable life
chart regarding substance dependence or abuse.
why we could not find any association between
substance dependence or abuse and suicide at-
tempts. It is worth noting that omitting previous
suicide attempts from Cox?s model resulted in no
other potential risk factor emerging as a predictor.
Challenges for future studies include investigation,
in larger prospective cohort studies of bipolar
patients, of whether and how strongly various other
risk factors may modify risk during high-risk illness
phases. When assessing the impact of treatments on
suicidal behavior, future studies should take into
account changes in both the incidence of suicidal
behavior in various phases of illness, and overall
time spent in these high-risk phases.
Bipolar I and II disorders
The difference in rates of suicide attempts between
BD I and II is a controversial issue (45–47). Some
studies (48–51) have reported higher rates of
suicide attempts in BD II, whereas the Stanley
Foundation Bipolar Network study (52) and other
studies (53–55) found no difference. In our study,
BD II patients spent more time in risk phases than
did BD I patients, but the type of BD had no
significant modifying impact during the high-risk
phase. Some studies (7, 8), though not all (40, 41),
have reported that the longitudinal symptomatic
course of BD II is more dominated by the depres-
sive phases of illness. Overall, differences in risk for
suicide attempts between BD II and BD I patients
are likely to be a consequence largely of differences
in the distribution of time spent at risk, and thus
vary depending on the longitudinal course of
illness rather than on the type of disorder per se.
Spending more time depressed or in the mixed
illness phases would readily explain why BD II is
often associated with more frequent suicidal
behavior than BD I.
Time aspects of suicidal behavior
Research on suicidology has shown that time-
varying risk factors are important (56). In their
recent multi-center cohort study, Kapur et al. (57)
found that suicidal behavior tended to repeat
quickly: almost half of individuals who repeated
within one year did so within the first two months
after the index episode. Regarding bipolar patients,
Valtonen et al.
O¨sby et al. (58) found that standardized mortality
year after hospitalization, but declined over time.
Likewise, Høyer et al. (59) reported that the risk of
suicide was high immediately after admission
and immediately following discharge among hospi-
meta-analysis on psychological autopsy studies of
completed suicide, Arsenault-Lapierre et al. (60)
reported that on average 87% of subjects who
committed suicide had a current mental disorder.
Specifically, Isometsa ¨ et al. (61) reported that most
suicides of subjects with BD occur during a major
depressive episode (79%) or during a mixed state
(11%). We have previously reported that suicide
attempts in BD mostly occurred during depressive
a preceding decline in hopelessness and depressive
symptoms predicted a decline in suicidal ideation
among patients with major depressive disorder. All
these studies document that suicidal behavior is
strongly related to clinical state. Future suicide
studies should investigate not only static, but also
time-varying risk factors, and include time-at-risk
estimations when estimating overall suicide risk.
Among BD patients, the incidence of suicide
attempts varies remarkably depending on the type
of illness phase, with mixed and depressive phases
involving the highest risk per unit of time. Time
spent in high-risk illness phases is likely the major
determinant of overall risk for suicide attempts
among patients with BD. Future suicide studies
of time-varying risk factors for the overall risk of
suicidal behavior. Clinically, the fast and effective
management of mixed and depressive phases may
be crucial in reducing risk for attempting suicide.
The Jorvi Bipolar Study was funded by grants from the
Academy of Finland and the Helsinki University Central
Hospital (HUCH), Helsinki, Finland.
1. Jamison KR. Suicide and bipolar disorder. J Clin Psychi-
atry 2000; 61(Suppl 9): 47–51.
2. Slama F, Bellivier F, Henry C et al. Bipolar patients with
suicidal behavior: toward the identification of a clinical
sub-group. J Clin Psychiatry 2004; 65: 1035–1039.
3. Valtonen H, Suominen K, Mantere O, Leppa ¨ ma ¨ ki S,
Arvilommi P, Isometsa ¨ E. Suicidal ideation and attempts
in bipolar I and II disorders. J Clin Psychiatry 2005; 66:
4. Chen YW, Dilsaver SC. Lifetime rates of suicide attempts
among subjects with bipolar and unipolar disorders
relative to subjects with other Axis I disorders. Biol
Psychiatry 1996; 39: 896–899.
5. Hawton K, Sutton L, Haw C, Sinclair J, Harriss L. Suicide
and attempted suicide in bipolar disorder: a systematic
review of risk factors. J Clin Psychiatry 2005; 66: 693–704.
6. Judd LL, Akiskal HS, Schettler PJ et al. The long-term
natural history of the weekly symptomatic status of bipolar
I disorder. Arch Gen Psychiatry 2002; 59: 530–537.
7. Judd LL, Akiskal HS, Schettler PJ et al. A prospective
investigation of the natural history of the long-term weekly
symptomatic status of bipolar II disorder. Arch Gen
Psychiatry 2003; 60: 261–269.
8. Mantere O, Suominen K, Valtonen HM et al. Differences
in outcome of DSM-IV bipolar I and II disorders. Bipolar
Disord 2008; 10: 413–425.
9. Marangell LB, Bauer MS, Dennehy EB et al. Prospective
predictors of suicide and suicide attempts in 1,556 patients
with bipolar disorders followed for up to 2 years. Bipolar
Disord 2006; 8: 566–575.
10. Valtonen HM, Suominen K, Mantere O, Leppa ¨ ma ¨ ki S,
Arvilommi P, Isometsa ¨ ET. Prospective study of risk
factors for attempted suicide among patients with bipolar
disorder. Bipolar Disord 2006; 8: 576–585.
11. Goodwin FK, Jamison KR. Manic-Depressive Illness.
New York: Oxford University Press, 2007; 263.
12. Benazzi F. Bipolar disorder: focus on bipolar II disorder
and mixed depression. Lancet 2007; 369: 935–945.
13. Balazs J, Benazzi F, Rihmer R, Rihmer A, Akiskal KK,
Akiskal HS. The close link between suicide attempts and
mixed (bipolar) depression: implications for suicide pre-
vention. J Affect Disord 2006; 91: 133–138.
14. Akiskal HS, Benazzi F. Psychopathologic correlates of
suicidal ideation in major depressive outpatients: is it all
due to unrecognized (bipolar) depressive mixed states?
Psychopathology 2005; 38: 273–280.
15. Akiskal HS, Benazzi F. Delineating depressive mixed
states: their therapeutic significance. Clin Approaches
Bipolar Disord 2003; 2: 41–47.
16. Jameison GR.Suicide andmentaldisease:aclinical analysis
of 100 cases. Arch Neurol Psychiatry 1936; 36: 1–12.
17. Dilsaver SC, Chen YW, Swann AC, Shoaib AM, Krajew-
ski KJ. Suicidality in patients with pure and depressive
mania. Am J Psychiatry 1994; 151: 1312–1315.
18. Goldberg JF, Garno JL, Portera L, Leon AC, Kocsis JH,
Whiteside JE. Correlates of suicidal ideation in dysphoric
mania. J Affect Disord 1999; 56: 75–81.
19. Baldassano CF. Illness course, comorbidity, gender, and
suicidality in patients with bipolar disorder. J Clin Psychi-
atry 2006; 67 (Suppl 11): 8–11.
20. Valtonen HM, Suominen K, Mantere O, Leppa ¨ ma ¨ ki S,
Arvilommi P, Isometsa ¨ ET. Suicidal behavior during
different phases of bipolar disorder. J Affect Disord
2007; 97: 101–107.
21. Mantere O, Suominen K, Leppa ¨ ma ¨ ki S, Valtonen H,
Arvilommi P, Isometsa ¨ E. The clinical characteristics of
DSM-IV bipolar I and II disorders – baseline findings from
the Jorvi Bipolar Study (JoBS). Bipolar Disord 2004; 6:
22. Hirschfeld RMA, Williams JBW, Spitzer RL et al. Devel-
opment and validation of a screening instrument for
bipolar spectrum disorder: the Mood Disorder Question-
naire. Am J Psychiatry 2000; 157: 1873–1875.
23. First MB, Spitzer RL, Gibbon M, Williams JBW. Struc-
tured Clinical Interview for DSM-IV-TR Axis I Disorders,
Research Version, Patient Edition with Psychotic Screen
Incidence of suicide attempts in BD
(SCID-I ⁄P W ⁄PSY SCREEN) (Revision). New York: Download full-text
Biometrics Research, New York State Psychiatric Institute,
24. First MB, Gibbon M, Spitzer RL, Williams JBW. Struc-
tured Clinical Interview for DSM-IV Axis II Disorders
(SCID-II), Version 2. New York: New York Psychiatric
Institute, Biometrics Research, 1996.
25. Young RC, Biggs JT, Ziegler VE, Meyer DA. A rating
scale for mania: reliability, validity and sensitivity. Br J
Psychiatry 1978; 133: 429–435.
26. Hamilton M. A rating scale for depression. J Neurol
Neurosurg Psychiatry 1960; 23: 56–62.
27. Beck AT, Kovacs M, Weissman A. Assessment of suicidal
intention: the Scale for Suicide Ideation. J Consult Clin
Psychol 1979; 47: 343–352.
28. American Psychiatric Association: Diagnostic and statis-
tical manual of mental disorders, 4th Edition, Text
Revision, Washington, DC, American Psychiatric Associa-
tion, 2000: 817–818.
29. Beck AT, Ward CH, Mendelson M, Mock J, Erbaugh J.
An inventory for measuring depression. Arch Gen
Psychiatry 1961; 4: 561–571.
30. Beck AT, Epstein N, Brown G, Steer RA. An inventory
for measuring clinical anxiety: psychometric properties.
J Consult Clin Psychol 1988; 56: 893–897.
31. Beck AT, Weissman A, Lester D, Trexler L. The measure
of pessimism: the hopelessness scale. J Consult Clin
Psychol 1974; 42: 861–865.
32. BlumenthalJA,BurgMM,BraefootJ,Williams RB,Haney
T, Zimet G. Social support, type A behavior, and coronary
artery disease. Psychosom Med 1987; 49: 331–340.
33. Benazzi F, Akiskal HS. Delineating bipolar II mixed states
in the Ravenna)San Diego Collaborative Study: the
relative prevalence and diagnostic significance of hypo-
manic features during major depressive episodes. J Affect
Disord 2001; 67: 115–122.
34. Melartin TK, Rytsa ¨ la ¨ HJ, Leskela ¨ US, Lestela ¨ -Mielonen
PS, Sokero TP, Isometsa ¨ ET. Severity and comorbidity
predict episode duration and recurrence of DSM-IV major
depressive disorder. J Clin Psychiatry 2004; 65: 810–819.
35. American Psychiatric Association. Diagnostic and Statis-
tical Manual of Mental Disorders, 4th edn, text revision.
Washington, DC: American Psychiatric Association, 2000.
assessment and treatment of patients with suicidal behav-
iors. Am J Psychiatry 2003; 160 (November Suppl): 3.
37. Grambsch P, Therneau T. Proportional hazards tests and
diagnostics based on weighted residuals. Biometrika 1994;
38. Dittmann S, Biedermann NC, Grunze H et al. The Stanley
Foundation Bipolar Network: results of the naturalistic
follow-up study after 2.5 years of follow-up in the German
centers. Neuropsychobiology 2002; 46(Suppl 1): 2–9.
39. Nolen WA, Luckenbaugh DA, Altshuler LL et al. Corre-
lates of 1-year prospective outcome in bipolar disorder:
results from the Stanley Foundation Bipolar Network. Am
J Psychiatry 2004; 161: 1447–1454.
40. Post RM, Denicoff KD, Leverich GS et al. Morbidity in
258 bipolar outpatients followed for 1 year with daily
prospective ratings on the NIMH life chart method. J Clin
Psychiatry 2003; 64: 680–690.
41. Joffe RT, MacQueen GM, Marriott M, Young LT. A
prospective, longitudinal study of percentage of time spent
ill in patients with bipolar I or bipolar II disorders. Bipolar
Disord 2004; 6: 62–66.
42. Oquendo MA, Galfalvy H, Russo S et al. Prospective study
of clinical predictors of suicidal acts after a major depres-
sive episode in patients with major depressive disorder or
bipolar disorder. Am J Psychiatry 2004; 161: 1433–1441.
43. Sokero P, Melartin T, Rytsa ¨ la ¨ H, Leskela ¨ U, Lestela ¨ -
Mielonen P, Isometsa ¨ E. A prospective study of risk
factors for attempted suicide among psychiatric patients
with DSM-IV major depressive disorder. Br J Psychiatry
2005; 186: 314–318.
44. Galfalvy H, Oquendo MA, Carballo JJ et al. Clinical predic-
tors of suicidal acts after major depression in bipolar disorder:
a prospective study. Bipolar Disord 2006; 8: 586–595.
46. Vieta E, Benabarre A, Colom F et al. Suicidal behavior in
bipolar I and bipolar II disorder. J Nerv Ment Dis 1997;
47. Rihmer Z, Pestality P. Bipolar II disorder and suicidal
behavior. Psychiatr Clin North Am 1999; 22: 667–673.
48. Stallone F, Dunner DL, Ahearn J, Fieve RR. Statistical
predictions of suicide in depressives. Compr Psychiatry
1980; 21: 381–387.
49. Bulik CM, Carpenter LL, Kupfer DJ, Frank E. Features
associated with suicide attempts in recurrent major depres-
sion. J Affect Disord 1990; 18: 29–37.
50. Tondo L, Baldessarini RJ, Hennen J et al. Suicide attempts
use disorders. J Clin Psychiatry 1999; 60(Suppl 2): 63–69.
51. Balazs J, Lecrubier Y, Csiszer N, Kosztak J, Bitter L.
Prevalence and comorbidity of affective disorders in
persons making suicide attempts in Hungary: importance
of the first depressive episodes and of bipolar II diagnoses.
J Affect Disord 2003; 76: 113–119.
52. Leverich GS, Altshuler LL, Frye MA et al. Factors
associated with suicide attempts in 648 patients with
bipolar disorder in the Stanley Foundation Bipolar Net-
work. J Clin Psychiatry 2003; 64: 506–515.
53. Endicott J, Nee J, Andreasen N, Clayton P, Keller M,
Coryell W. Bipolar II. Combine or keep separate? J Affect
Disord 1985; 8: 17–28.
54. Coryell W, Andreasen NC, Endicott J, Keller M. The signif-
icance of past mania or hypomania in the course and outcome
of major depression. Am J Psychiatry 1987; 144: 309–315.
JL. Suicide risk in bipolar patients: the role of comorbid
substance use disorders. Bipolar Disord 2003; 5: 58–61.
56. Kraemer HC, Kazdin AE, Offord DR, Kessler RC, Jensen
PS, Kupfer DJ. Coming to terms with the terms of risk.
Arch Gen Psychiatry 1997; 54: 337–343.
57. Kapur N, Cooper J, King-Hele S et al. The repletion of
suicidal behavior: a multi-center cohort study. J Clin
Psychiatry 2006; 67: 1559–1609.
58. O¨sby U, Brandt L, Correia N, Ekbom A, Sparen P. Excess
mortality in bipolar and unipolar disorder in Sweden. Arch
Gen Psychiatry 2001; 58: 844–850.
59. Høyer EH, Mortensen PB, Olesen AV. Mortality and
causes of death in a total national sample of patients with
affective disorders admitted for the first time between 1973
and 1993. Br J Psychiatry 2000; 176: 76–82.
60. Arsenault-Lapierre G, Kim C, Turecki G. Psychiatric
diagnoses in 3275 suicides: a meta-analysis. BMC Psychi-
atry 2004; 4: 37.
61. Isometsa ¨ ET, Henriksson MM, Aro HM, Lo ¨ nnqvist JK.
Suicide in bipolar disorder in Finland. Am J Psychiatry
1994; 151: 1020–1024.
62. Sokero P, Eerola M, Rytsala H et al. Decline in suicidal
depression and hopelessness. J Affect Disord 2006; 95: 95–
Valtonen et al.