Review

Risk factors for suicide in bipolar disorder: A systematic review

Lucas da Silva Costa a,n, Átila Pereira Alencar a, Pedro Januário Nascimento Neto a,Maria do Socorro Vieira dos Santos a, Cláudio Gleidiston Lima da Silva b,Sally de França Lacerda Pinheiro b, Regiane Teixeira Silveira b, Bianca Alves Vieira Bianco b,Roberto Flávio Fontenelle Pinheiro Júnior c, Marcos Antonio Pereira de Lima c,Alberto Olavo Advincula Reis d, Modesto Leite Rolim Neto e

a Laboratório de Escrita Científica, Faculdade de Medicina, Universidade Federal do Cariri, UFCA, Barbalha, Ceará, Brazilb Programa de Pós-Graduação em Ciências da Saúde, Faculdade de Medicina do ABC, Santo André, São Paulo, Brazilc Faculdade de Medicina, Universidade Federal do Cariri, UFCA, Barbalha, Ceará, Brazild Programa de Pós-Graduação em Saúde Pública, Faculdade de Saúde Pública, Universidade de São Paulo, USP, São Paulo, São Paulo, Brazile Líder de Grupo de Pesquisa em Suicidologia, Universidade Federal do Ceará, UFC/Conselho Nacional de Desenvolvimento Científico e Tecnológico, CNPq,Faculdade de Medicina, Universidade Federal do Cariri, UFCA, Barbalha, Ceará, Brazil

a r t i c l e i n f o

Article history:Received 30 July 2014Received in revised form18 August 2014Accepted 2 September 2014Available online 16 September 2014

Keywords:BipolarSuicideRisk factors

a b s t r a c t

Background: Bipolar disorder confers the highest risk of suicide among major psychological disorders.The risk factors associated with bipolar disorder and suicide exist and are relevant to clinicians andresearchers.Objective: The aim of the present study was to conduct a systematic review of articles regarding thesuicide risk factors in bipolar disorder.Methods: A systematic review of articles on suicide risk factors in bipolar disorder, published fromJanuary 1, 2010 to April 05, 2014, on SCOPUS and PUBMED databases was carried out. Search terms were“Suicide” (medical subject headings [MeSH]), “Risk factors” (MeSH), and “Bipolar” (keyword). Of the 220retrieved studies, 42 met the eligibility criteria.Results: Bipolar disorder is associated with an increased rate death by suicide which contributes tooverall mortality rates. Studies covered a wide range of aspects regarding suicide risk factors in bipolardisorder, such as risk factors associated to Sociodemographic conditions, Biological characteristics, DrugsRelationships, Psychological Factors, Genetic Compound, Religious and Spirituals conditions. Recentscientific literature regarding the suicide risk factors in bipolar disorder converge to, directly or indirectly,highlight the negative impacts of risk factors to the affected population quality of life.Conclusion: This review demonstrated that Bipolar disorders commonly leads to other psychiatricdisorders and co-morbidities involving risk of suicide. Thus the risk factors are relevant to have a betterdiagnosis and prognosis of BD cases involving risk of suicide.

& 2014 Elsevier B.V. All rights reserved.

Contents

1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2382. Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2383. Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2384. Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 239

4.1. Risk factors associated with sociodemographic components . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 246

Contents lists available at ScienceDirect

journal homepage: www.elsevier.com/locate/jad

Journal of Affective Disorders

http://dx.doi.org/10.1016/j.jad.2014.09.0030165-0327/& 2014 Elsevier B.V. All rights reserved.

n Correspondence to: Laboratório de Escrita Científica, Faculdade de Medicina, Universidade Federal do Cariri, UFCA, Rua Divino Salvador, 284, 63180-000, Barbalha, Ceará,Brazil. Tel.: þ55 88 3312 5000; fax: þ55 88 3312 5001.

E-mail address: lucas_10_00@hotmail.com (L.d.S. Costa).

Journal of Affective Disorders 170 (2015) 237–254

4.2. Risk factors associated with genetic components . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2474.3. Risk factors associated with medicines and drugs in general that interfere with bipolar disorder . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2484.4. Risk factors associated with biological components . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2484.5. Risk factors associated with psychological causes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2494.6. Risk factors associated with components of religious and spiritual components. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 250

5. Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 251Role of funding source . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 251Conflict of interest. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 251Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 251References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 251

1. Introduction

Bipolar disorder confers the highest risk of suicide amongmajor psychological disorders (Goldstein et al., 2012; Goodwinand Jamison, 2007). Suicide attempts and completed suicide aresignificantly more common in patients with bipolar disorder whencompared with the general population (Eroglu et al., 2013; Simonet al., 2007; Weissman et al., 1999).

Bipolar spectrum disorders, especially recurrent depressive epi-sodes, is the major risk of repeated suicide attempt and co-morbidityof another psychiatric disorders increase highly the risk of suicidereattempt (Kheirabadi et al., 2012). In particular, among mentaldisorders, bipolar disorder is one of the leading causes of suicidalbehaviors and this is a major issue in the management of the disease.About 50% of patients with bipolar disorder will experience at leastone suicide attempt (Jamison, 2000) and 11–19% will commit suicide(Goodwin and Jamison, 2007; Abreu et al., 2009; Angst et al., 2005;Harris and Barraclough, 1997; Parmentier et al., 2012).

This study is based on the following research question: what isthe main suicide risk factors associated with bipolar disorder? Thisissue has gained great impact in recent years with the establish-ment of new risk factors for suicide and bipolar disorder. Thus, thissystematic review aims to present the main risk factors andcompares them, since the applicant was disagreement amongauthors. Therefore, it is suggested, that further studies are neededin order to establish a stronger relationship between bipolardisorder and its risk factors that culminate in suicide.

2. Methods

We performed a qualitative systematic review of articles aboutsuicide risk factors in bipolar affective disorder in previouslychosen electronic databases.

A search of the literature was conducted via PubMed andSCOPUS online databases in April 2014 and was limited to articlespublished from January 1, 2010 to April 6, 2014. The reason forlimiting the search to 2010–2014 was that, during this period,there was an expansion of research into new types of comorbid-ities that influence the risk of suicide in Bipolar disorder, such ashopelessness, altitude and religiosity. Therefore, the Bipolar Affec-tive Disorder and its association with suicide had greater relevancein the scientific community.

Initially, the search terms browsed in SCOPUS database were

1. “bipolar” (keyword);2. “suicide” (Medical Subject Headings [MeSH] term); and3. “risk factors” (MeSH term).

The following searches were performed: 1 AND 2 AND 3.In addition to MeSH terms, we opted to add the keyword “bipolar”to the search strategy, because, despite not being included in theMeSH thesaurus, it is frequently used to describe studies that deal

with the theme object of the present review. The search strategyand the retrieved articles were reviewed on two separate occa-sions to ensure adequate sampling. A similar search strategy wasperformed in the PubMed database, using the aforementionedterms and their correspondent terms.

The article analysis followed previously determined eligibilitycriteria. We adopted the following inclusion criteria: Goldsteinet al. (2012) references written in English; Goodwin and Jamison(2007) studies pertaining suicide risk factors in bipolar affectivedisorder; Eroglu et al., (2013) original articles with online acces-sible full text available in database SCOPUS, PubMed or CAPES(Higher Education Co-ordination Agency) Journal Portal(Periodicos.capes.gov.br, 2014), a virtual library linked to Brazil'sMinistry of Education and subjected to content subscription;(Simon et al., 2007) articles that included in the title at least onecombination of terms described in the search strategy; (Weissmanet al., 1999) case reports, cohort studies, controlled clinical trialsand case-control studies; Kheirabadi et al. (2012) articles thatappear in more than one database will be included only once,giving priority to the SCOPUS database. Exclusion criteria were:Goldstein et al. (2012) studies that did not include the proposedtopic; Goodwin and Jamison (2007) non-original studies, includingeditorials, reviews, prefaces, brief communications and letters tothe editor.

Then, each paper in the sample was read in entirety, and dataelements were then extracted and entered into a matrix thatincluded authors, journal, description of the study sample, andmain findings. Some of the studies dealt not only with the riskfactors associated with bipolar disorder, but also to the risk factorsin other psychiatric disorders, such as schizophrenia and mooddisorder; because the focus of this study was the risk factorsassociated with suicide in bipolar disorder, studies related topsychiatric disorders in general were not recorded or analyzedfor this study.

To provide a better analysis, the next phase involved comparingthe studies and grouping. For heuristic reasons, the results regardingthe studied subject into six categories: Risk factors associated withsociodemographic components; Risk factors associated with geneticcomponents; Risk factors associated with Medicines and Drugs ingeneral that interfere with bipolar disorder; Risk factors associatedwith Biological components; Risk factors associated with Psycholo-gical causes; and Risk factors associated with components of Reli-gious and Spiritual components.

3. Results

Initially, the aforementioned search strategies resulted in 220references. After browsing the title and abstract of the retrievedcitations for eligibility based on study inclusion criteria, 178 articleswere excluded and 42 articles were further retrieved and included

L.d.S. Costa et al. / Journal of Affective Disorders 170 (2015) 237–254238

in the final sample Fig. 1). Articles from SCOPUS and PubMeddatabase matched the inclusion criteria of the present study.

Table 1 provides an overview of all studies included in the finalsample and of all data elements used during the data analysisprocess. Study designs included one case report (Kerner et al.,2013), seven transversal studies (Goldstein et al., 2012; Kheirabadiet al., 2012; Undurraga et al., 2012; De Abreu et al., 2012; Algortaet al., 2011; Evans et al., 2012; Gomes et al., 2010), nineteen cohortstudies (Parmentier et al., 2012; Huber et al., 2014; Ruengorn et al.,2012; Etain et al., 2013; Cassidy, 2011; Baldessarini et al., 2012;Bellivier et al., 2011; Sears et al., 2013; Jiménez et al., 2013; Leonet al., 2012; Finseth et al., 2012; Oquendo et al., 2010; Kennesonet al., 2013; Gilbert et al., 2011; Shabani et al., 2013; Pompili et al.,2012; Acosta et al., 2012; Song et al., 2012; Suttajit et al., 2013), andfifteen case-control studies (Eroglu et al., 2013; Antypa et al., 2013;Ryu et al., 2010; Manchia et al., 2013; Neves et al., 2010; Magnoet al., 2011; Yoon et al., 2011; Arias et al., 2013; Clements et al.,2013; Pawlak et al., 2013; Kamali et al., 2012; de Moraes et al.,2013; Azorin et al., 2013; Dervic et al., 2011; Pawlak et al., 2013).The 42 studies were distributed into the previously determined sixcategories as follows:

Risk factors associated with sociodemographic components(seven studies) (Huber et al., 2014; De Abreu et al., 2012;Ruengorn et al., 2012; Algorta et al., 2011; Cassidy, 2011; Antypaet al., 2013; Ryu et al., 2010); Risk factors associated with geneticcomponents (six studies) (Manchia et al., 2013; Sears et al., 2013;Neves et al., 2010; Magno et al., 2011; Jiménez et al., 2013; Kerneret al., 2013); Risk factors associated with Medicines and Drugs ingeneral that interfere with bipolar disorder (eight studies)

(Bellivier et al., 2011; Yoon et al., 2011; Leon et al., 2012; Ariaset al., 2013; Clements et al., 2013; Finseth et al., 2012; Oquendoet al., 2010; Kenneson et al., 2013); Risk factors associated withBiological components (three studies) (Kamali et al., 2012; Evanset al., 2012; Gomes et al., 2010); Risk factors associated withPsychological causes (seven studies) (Parmentier et al., 2012;Shabani et al., 2013; Pompili et al., 2012; Acosta et al., 2012;Song et al., 2012; Suttajit et al., 2013; Stewart et al., 2009); and Riskfactors associated with components of Religious and Spiritual(three studies) components (Azorin et al., 2013; Dervic et al.,2011; Pawlak et al., 2013). Among the 42 studies, 8 discussed about“suicide risk factors in Bipolar Affective disorder” — more broadly(Goldstein et al., 2012; Eroglu et al., 2013; Kheirabadi et al., 2012;Undurraga et al., 2012; Etain et al., 2013; Baldessarini et al., 2012;Pawlak et al., 2013; Gilbert et al., 2011), being refered in more thanone category. The categorization of studies aims to a betterorganizational quality systematic review and it is not compulsorythat each article must be referenced only in their respectivecategory.

4. Discussion

Bipolar disorder (BD) is a major public health concern worldwide,and is associated with significant morbidity and mortality (Kupfer,2005). In addition to an increased rate of death by suicide, communityand clinical studies indicate that bipolar patients usually present abroadrange of comorbid general medical conditions, which contribute

Fig. 1. Flow chart showing study selections for the review. Abbreviations MeSH, Medical Subject Headings.

L.d.S. Costa et al. / Journal of Affective Disorders 170 (2015) 237–254 239

Table 1Suicide risk factors in bipolar affective disorder: studies and main findings.

Authors Journal Sample Main findings

Goldstein et al.(2012)

Archives of GeneralPsychiatry

A total of 413 youths (mean [SD] age, 12.6 [3.3] years) who receiveda diagnosis of bipolar I disorder (n¼244), bipolar II disorder(n¼28), or bipolar disorder not otherwise specified (n¼141).

Of the 413 youths with bipolar disorder, 76 (18%)made at least 1 suicide attempt within 5 years ofstudy intake; of these, 31 (8% of the entire sampleand 41% of attempters) made multiple attempts. Girlshad higher rates of attempts than did boys, but rateswere similar for bipolar subtypes. The most potentpast and intake predictors of prospectively examinedsuicide attempts included severity of depressiveepisode at study intake and family history ofdepression. Follow-up data were aggregated over8-week intervals; greater number of weeks spentwith threshold depression, substance use disorder,and mixed mood symptoms and greater number ofweeks spent receiving outpatient psychosocialservices in the preceding 8-week period predictedgreater likelihood of a suicide attempt.

Eroglu et al.(2013)

Dusunen Adam One hundred twenty two consecutive patients, from BipolarDisorder Unit of Çukurova University, Faculty of Medicine,Department of Psychiatry, are included in this study.

The prevalence of suicide attempt was 19.7% in theoutpatient group. Lifetime history of suicidalbehavior was significantly associated with followingcharacteristics: being a woman, depression as a firstepisode and indicators of severity of bipolar disorderincluding duration of illness, duration of untreatedillness (latency), number of hospitalization, numberof total mood episodes, number of depressiveepisodes, number of mixed episodes, positive familialpsychiatric disorder history.

Kheirabadiet al. (2012)

Iranian Journal ofEpidemiology

Participants consisted of 703 individuals (424 of themwere female)with mean age of 25.979.7.

Bipolar spectrum disorders, unipolar depression andadjustment disorders were the more frequentspsychiatric disorders respectively. Age, family historyof suicide, kind of diagnosed psychiatric disorder andmethod of attempted suicide were meaningfullyrelated to mean of attempt suicide frequency.

Parmentieret al. (2012)

European Psychiatry In a sample of 652 euthymic bipolar patients, we assessed clinicalfeatures with the Diagnostic Interview for Genetics Studies (DIGS)and dimensional characteristics with questionnaires measuringimpulsivity/hostility and affective lability/intensity.

Of the 652 subjects, 42.9% had experienced at leastone suicide attempt. Lifetime history of suicidalbehavior was associated with being a woman, ahistory of head injury, tobacco misuse and indicatorsof severity of bipolar disorder including early age atonset, high number of depressive episodes, positivehistory of rapid cycling, alcohol misuse and socialphobia. Indirect hostility and irritability weredimensional characteristics associated with suicidalbehavior in bipolar patients, whereas impulsivity andaffective lability/intensity were not associated withsuicidal behavior.

Huber et al.(2014)

Medical Hypotheses Data were available for 16 states for the years 2005–2008,representing a total of 35,725 completed suicides in 922 U.S.counties.

Altitude was a significant, independent predictor ofthe altitude at which suicides occurred (F¼8.28,p¼0.004 and Wald chi-square.¼21.67, po 0.0001).Least squares means of altitude, independent ofother variables, indicated that individuals with BDcommitted suicide at the greatest mean altitude.Moreover, the mean altitude at which suicidesoccurred in BD was significantly higher than indecedents whose mental health diagnosis was majordepressive disorder (MDD), schizophrenia, or anxietydisorder.

Undurragaet al. (2012)

Journal of ClinicalPsychiatry

Accordingly, we compared selected demographic and clinicalfactors for long-term association with nonlethal suicidal acts orideation in 290 DSM-IV bipolar I (n¼204) and II (n¼86) disorderpatients followed for a mean of 9.3 years at the University ofBarcelona, using preliminary bivariate comparisons followed bymultivariate logistic regression modeling.

Rates of suicidal ideation (41.5%) and acts (19.7%)were similarly prevalent with bipolar I and IIdisorders and somewhat more common amongwomen. Factors significantly and independentlyassociated with suicidal acts were determined bymultivariate modeling and ranked in order of theirstrength of association: suicidal ideation, moremixed episodes, Axis II comorbidity, female sex,more antidepressant trials, rapid cycling,predominant lifetime depression, having beenhospitalized, older onset, and longer delay ofdiagnosis.

De Abreu et al.(2012)

Comprehensive Psychiatry One hundred eight patients with Diagnostic and Statistical Manualof Mental Disorders, Fourth Edition BD type I (44 with previous

Patients with BD and previous suicide attempts hadsignificantly lower scores in all the 4 domains of the

L.d.S. Costa et al. / Journal of Affective Disorders 170 (2015) 237–254240

Table 1 (continued )

Authors Journal Sample Main findings

suicide attempts, 64 without previous suicide attempts) werestudied.

World Health Organization's Quality of LifeInstrument-Short Version scale than did patientswith BD but no previous suicide attempts (physicaldomain P¼0.001; psychological domain Po0.0001;social domain P¼0.001, and environmental domainP ¼0.039). In the euthymic subgroup (n¼70),patients with previous suicide attempts hadsignificantly lower scores only in the psychologicaland social domains (P¼0.020 and P¼0.004).Limitations: This was a cross-sectional study, and nocausal associations can be assumed.

Ruengorn et al.(2012)

Psychology Research andBehavior Management

Medical files of 489 patients diagnosed with BD at SuanprungPsychiatric Hospital between October 2006 and May 2009 werereviewed.

Six statistically significant indicators associated withsuicide attempts were included in the risk-scoringscheme: depression, psychotic symptom(s), numberof previous suicide attempts, stressful life event(s),medication adherence, and BD treatment years. Atotal risk score (possible range �1.5–11.5) explainedan 88.6% probability of suicide attempts based on thereceiver operating characteristic (ROC) analysis.Likelihood ratios of suicide attempts with low riskscores (below 2.5), moderate risk scores (2.5-8.0),and high risk scores (above 8.0) were 0.11 (95% CI0.04-0.32), 1.72 (95% CI 1.41–2.10), and 19.0 (95% CI6.17-58.16), respectively.

Algorta et al.(2011)

Bipolar Disorders Participants were 138 youths aged 5–18years presenting tooutpatient clinics with DSM-IV diagnoses of bipolar I disorder(n¼27), bipolar II disorder (n¼18), cyclothymic disorder (n¼48),and bipolar disorder not otherwise specified (n¼45).

Twenty PBD patients had lifetime suicide attempts,63 had past or current suicide ideation, and 55 werefree of suicide ideation and attempts. Attempterswere older than nonattempters. Suicide ideation andattempts were linked to higher depressivesymptoms, and rates were even higher in youthsmeeting criteria for the mixed specifier proposed forDSM-5. Both suicide ideation and attempts wereassociated with lower youth QoL and poorer familyfunctioning. Parent effects (with suicidality treatedas outcome) and child effects (where suicide was thepredictor of poor family functioning) showed equallystrong evidence in regression models, even afteradjusting for demographics.

Etain et al.(2013)

Journal of ClinicalPsychiatry

587 patients with DSM-IV-defined bipolar disorder were recruitedfrom France and Norway between 1996-2008 and 2007-2012,respectively.

Multivariate analyses investigating trauma variablestogether showed that both emotional and sexualabuse were independent predictors of lower age atonset (P¼0.002 for each) and history of suicideattempts (OR¼1.60 [95% CI, 1.07 to 2.39], P¼0.023;OR¼1.80 [95% CI, 1.14–2.86], P¼0.012, respectively),while sexual abuse was the strongest predictor ofrapid cycling (OR¼2.04 [95% CI, 1.21–3.42],P¼0.007). Females reported overall higher childhoodtrauma frequency and greater associations to clinicalexpressions than males (P valueso0.05).

Cassidy (2011) Suicide and Life-Threatening Behavior

The study cohort included 87 males and 70 females. Ninety-sixwere White and sixty-one were Black.

Gender, nicotine use, medical comorbidity, andhistory of alcohol and other drug abuse were not,although a trend was noted for a history ofbenzodiazepine abuse.

Blair-Westet al. (1999)

Journal of PsychiatricResearch

Data was collected from the Systematic Treatment EnhancementProgram for Bipolar Disorder (STEP-BD) study. 3083 bipolarpatients were included in this report, among these 140 (4.6%) had asuicide event (8 died by suicide and 132 attempted suicide).

The strongest predictor of a suicide event was ahistory of suicide attempt (hazard ratio¼2.60, p-valueo0.001) in line with prior literature. Additionalpredictors were: younger age, a high total score onthe personality disorder questionnaire and a highpercentage of days spent depressed in the year priorto study entry.

Gould et al.(1996)

Yonsei Med J. A total of 579 medical records were retrospectively reviewed. The prevalence of suicide attempt was 13.1% in ourpatient group. The presence of a depressive firstepisode was significantly different betweenattempters and nonattempters. Logistic regressionanalysis revealed that depressive first episodes andbipolar II disorder were significantly associated withsuicide attempts in those patients.

Arató et al.(1988)

Acta PsychiatricaScandinavica

We tested factors for association with predominantly (Z2:1)depressive vs. mania-like episodes with 928 DSM-IV type-I BPDsubjects from five international sites.

Factors preliminarily associated with predominant-depression included: electroconvulsive treatment,longer latency-to-BPD diagnosis, first episodedepressive or mixed, more suicide attempts, moreAxis-II comorbidity, ever having mixed-states, ever

L.d.S. Costa et al. / Journal of Affective Disorders 170 (2015) 237–254 241

Table 1 (continued )

Authors Journal Sample Main findings

married, and female sex. Predominant-mania wasassociated with: initial manic or psychotic episodes,more drug abuse, more education, and more familypsychiatric history. Of the 47.3% of subjects withoutpolarity-predominance, risks for all factorsconsidered were intermediate. Expanding thedefinition of polarity-predominance to Z51% addedlittle, but shifting mixed-states to 'predominant-depression' increased risk of suicidal acts from 2.4-to 4.5-fold excess over predominant-mania-hypomania, and suicidal risk was associatedcontinuously with increasing proportions ofdepressive or mixed episodes.

Akiskal et al.(1995)

Bipolar Disorders European Mania in Bipolar Longitudinal Evaluation of Medication(EMBLEM) is a two-year, prospective, observational study thatenrolled 3,684 adult patients with bipolar disorder and initiated orchanged oral treatment for an acute manic/mixed episode.

Of the 2,219 patients who provided data on theirlifetime history of suicide attempts, 663 (29.9%) hada history of suicidal behavior (at least one attempt).Baseline factors associated with a history of suicidalbehavior included female gender, a history of alcoholabuse, a history of substance abuse, young age at firsttreatment for a mood episode, longer diseaseduration, greater depressive symptom severity(HAMD-5 total score), current benzodiazepine use,higher overall symptom severity (CGI-BP: mania andoverall score), and poor compliance.

McIntyre et al.(2008)

Bipolar disorders We studied 737 families of probands with MAD with 4919 first-degree relatives (818 affected, 3948 unaffected, and 153 subjectswith no information available).

The estimated lifetime prevalence of suicidalbehavior (attempted and completed suicides) in 737probands was 38.473.0%. Lithium treatmentdecreased suicide risk in probands (p¼0.007). Infirst-degree relatives, a family history of suicidalbehavior contributed significantly to the joint risk ofMAD and suicidal behavior (p¼0.0006).

Vaccari et al.,(1978)

Journal of AffectiveDisorders

We used family-based association testing in a cohort of 130multiplex bipolar pedigrees, comprising 795 individuals, to look forassociations between suicidal behavior and 32 single nucleotidepolymorphisms (SNPs) from across the genes brain-derivedneurotrophic factor (BDNF), cholecystokinin (CCK) and thecholecystokinin beta-receptor (CCKBR).

We found associations (pr0.05) between suicideattempt and 12 SNPs of CCKBR and five SNPs ofBDNF. After correction for multiple testing, sevenSNPs of CCKBR remained significantly associated. Noassociation was found between CCK and suicidalbehavior.

WHO (2011) Journal of AffectiveDisorders

We evaluated 198 bipolar patients and 103 health controls, using astructured interview according to DSM-IV criteria.

We found that 26.77% and 16.67% had a lifetimehistory of non violent suicide attempt and violentsuicide attempt, respectively. The clinical factorsassociated with violent and non violent suicideattempt had several differences. Violent suicideattempters had an earlier illness onset and had ahigher number of psychiatric comorbidities(borderline personality disorder, panic disorder andalcoholism). The frequency of S allele carriers washigher only in those patients who had made a violentsuicide attempt in their lifetime (x2¼16.969;p¼0.0001). In a logistic regression model includingthese factors, S allele carrier (5-HTTLPR) was the onlyfactor associated with violent suicide attempt.

Baldessariniet al. (2006)

Journal of AffectiveDisorders

TaqMan genotyping was used to detect FOXO3A SNPs in 273 BDpatients and 264 control subjects.

Three SNPs (rs1536057, rs2802292 and rs1935952)were associated with BD, but none was positivelylinked with suicidal behavior.

Dwivedi et al.(2003)

EuropeanNeuropsychopharmacology

Polymorphisms at the IMPA1 (rs915, rs1058401 and rs2268432)and IMPA2 (rs66938, rs1020294, rs1250171 and rs630110), INPP1(rs3791809, rs4853694 and 909270), GSK3α (rs3745233) and GSK3β(rs334558, rs1732170 and rs11921360) genes were genotyped.

Single SNP analyses showed that suicide attemptershad higher frequencies of AA genotype of thers669838-IMPA2 and GG genotype of the rs4853694-INPP1gene compared to non-attempters. Results alsorevealed that T-allele carriers of the rs1732170-GSK3β gene and A-allele carriers of the rs11921360-GSK3β gene had a higher risk for attempting suicide.Haplotype analysis showed that attempters hadlower frequencies of A:A haplotype (rs4853694:rs909270) at the INPP1 gene. Higher frequencies ofthe C:A haplotype and lower frequencies of the A:Chaplotype at the GSK-3β gene (rs1732170:rs11921360) were also found to be associated to SB inBP. Therefore, our results suggest that geneticvariability at IMPA2, INPP1 and GSK3β genes isassociated with the emergence of SB in BP.

L.d.S. Costa et al. / Journal of Affective Disorders 170 (2015) 237–254242

Table 1 (continued )

Authors Journal Sample Main findings

Polter et al.(2009)

Front Psychiatry Here, we describe a family with four siblings, three affected femalesand one unaffected male

Our results support a new model for psychiatricdisorders, in which multiple rare, damagingmutations in genes functionally related to a commonsignaling pathway contribute to the manifestation ofbipolar disorder.

Kerner et al.(2013)

Compr Psychiatry. This study is based on the US Multiple Cause of Death public-usedata files for 1999 to 2006. Secondary data analysis was conductedcomparing decedents with unipolar/bipolar disorders anddecedents with all other causes of death, based on the deathrecords of 19,052,468 decedents in the Multiple Cause of Deathdata files who died at 15 years and older

Prevalence of comorbid SUDs was higher amongunipolar and bipolar disorder deaths than thatamong all other deaths. Among unipolar and bipolardisorder deaths, comorbid SUDs were associatedwith elevated risks for suicide and other unnaturaldeath in both men and women (prevalence ratiosranging 1.49–9.46, Po0.05). They also wereassociated with reductions in mean ages at death(ranging 11.7–33.8 years, Po0.05). In general, theseeffects were much stronger for drug use disordersthan for alcohol use disorders. Both SUDs hadstronger effects on suicide among women, whereastheir effects on other unnatural deaths were strongeramong men.

Nilsson et al.,(2002)

American Journal ofPsychiatry

Analyses included 199 participants with bipolar disorder for whom1077 time intervals were classified as either exposed to anantiepileptic (carbamazepine, lamotrigine, or valproate) or notexposed to an antiepileptic, an antidepressant, or lithium during30 years of follow-up.

Participants who had more severe manic symptomswere more likely to receive antiepileptic drugs.Mixed-effects grouped-time survival modelsrevealed no elevation in risk of suicide attempt orsuicide during periods when participants werereceiving antiepileptics relative to periods when theywere not (hazard ratio¼0.93, 95% CI¼0.45–1.92),controlling for demographic and clinical variablesthrough propensity score matching.

Ratcliffe et al.,(2008)

Mental Health andSubstance Use: DualDiagnosis

The sample consisted of 837 outpatients from Madrid, Spain. Wecompared 528 subjects with a lifetime diagnosis of alcohol abuse ordependence and 182 with other substance use disorders (SUDs) notinvolving alcohol.

It was considered that 76.1% of the alcohol addictshad a current dual diagnosis, the most prevalentbeing mood and anxiety disorders. Fifty-two percenthad a personality disorder and most of them (81.6%)had other SUDs. There was a greater prevalence ofdual pathology in the alcohol addict subgroup thanin the subgroup without problems of alcohol abuseor dependence. Alcohol addicts were associated withdiagnoses of several types of personality disorderand bipolar disorder and presented a greater suiciderisk than the subgroup of other SUDs.

Tsai et al.(2002)

Psychological Medicine During the study period 1489 individuals with BD died by suicide,an average of 116 cases/year.

Compared to other primary diagnosis suicides, thosewith BD were more likely to be female, more than5 years post-diagnosis, current/recent in-patients, tohave more than five in-patient admissions, and tohave depressive symptoms. In BD suicides the mostcommon co-morbid diagnoses were personalitydisorder and alcohol dependence. Approximately40% were not prescribed mood stabilizers at the timeof death. More than 60% of BD suicides were incontact with services the week prior to suicide butwere assessed as low risk.

Clements et al.(2013)

General HospitalPsychiatry

The aim of the study was to look for suicide risk factors amongsociodemographic and clinical factors, family history and stressfullife events in patients with diagnosis of unipolar and bipolaraffective disorder (597 patients, 563 controls).

In the bipolar and unipolar affective disorderssample, we observed an association between suicidalattempts and the following: family history ofpsychiatric disorders, affective disorders andpsychoactive substance abuse/dependence;inappropriate guilt in depression; chronic insomniaand early onset of unipolar disorder. The risk ofsuicide attempt differs in separate age brackets (it isgreater in patients under 45 years old). No differencein family history of suicide and suicide attempts;marital status; offspring; living with family;psychotic symptoms and irritability; and coexistenceof personality disorder, anxiety disorder or substanceabuse/dependence with affective disorder wasobserved in the groups of patients with and withoutsuicide attempt in lifetime history.

Pawlak et al.(2013)

Bipolar Disorders A total of 206 consecutive patients (mean age 42715years; 54.9%women) with DSM-IV diagnosed BD-I (n¼140) and BD-II (n¼66)acutely admitted to a single psychiatric hospital department fromNovember 2002 through June 2009 were included.

Ninety-three patients (45.1%) had a history of one ormore serious suicide attempts. These constituted 60(42.9%) of the BD-I patients and 33 (50%) of the BD-IIpatients (no significant difference). Lifetime suicideattempt was associated with a higher number ofhospitalizations due to depression (po0.0001),

L.d.S. Costa et al. / Journal of Affective Disorders 170 (2015) 237–254 243

Table 1 (continued )

Authors Journal Sample Main findings

antidepressant (AD)-induced hypomania/mania(p¼0.033), AD- and/or alcohol-induced affectiveepisodes (p¼0.009), alcohol and/or substance use(p¼0.002), and a family history of alcohol abuse and/or affective disorder (p¼0.01). Suicide attempt wasnegatively associated with a higher Positive andNegative Syndrome Scale for Schizophrenia (PANSS)Positive Subscale score (p¼0.022) and morehospitalizations due to mania (p¼0.006).

Sublette et al.(2009)

Journal of ClinicalPsychiatry

1,643 individuals with a DSM-IV lifetime diagnosis of bipolardisorder were identified from 43,093 general-populationrespondents who were interviewed in the 2001-2002 NationalEpidemiologic Survey on Alcohol and Related Conditions.

More than half of the respondents (54%) who metcriteria for bipolar disorder also reported alcohol usedisorder. Bipolar individuals with comorbid alcoholuse disorder were at greater risk for suicide attemptthan those individuals without alcohol use disorder(adjusted odds ratio¼2.25; 95% CI, 1.61–3.14) andwere more likely to have comorbid nicotinedependence and drug use disorders.

Lopez et al.,(2001)

Comprehensive Psychiatry Using data collected from the National Comorbidity SurveyReplication study, we identified 158 individuals with childhood-onset (o13 years) or adolescent-onset (13-18 years) primarybipolar disorder (I, II or subthreshold).

Compared to adolescent-onset, people withchildhood-onset bipolar disorder had increasedlikelihoods of attention deficit hyperactivity disorder(ADHD) (adjusted odds ratio¼2.81) and suicideattempt (aOR¼3.61). Males were more likely thanfemales to develop SUD, and did so at a faster rate.Hazard ratios of risk factors for SUD were: lifetimeoppositional defiant disorder (2.048), any lifetimeanxiety disorder (3.077), adolescent-onset bipolardisorder (1.653), and suicide attempt (15.424). SUDwas not predicted by bipolar disorder type, familyhistory of bipolar disorder, hospitalization for a moodepisode, ADHD or conduct disorder.

Oquendo et al.(2010)

Journal of ClinicalPsychiatry

Participants included 67 adult inpatients and outpatients aged18-60 years meeting DSM-IV criteria for bipolar disorder (bipolar Iand II disorders, bipolar disorder not otherwise specified).

We found that nonattempters reported significantlyhigher trait impulsivity scores on the BarrattImpulsiveness Scale compared to attempters (t57¼2.2, P¼0.03) and that, among attempters, lowertrait impulsivity score was associated with higherscores of lethality of prior attempts (r25¼�0.53,P¼0.01). Analyses revealed no other groupdifferences on demographic, clinical, orneurocognitive variables when comparingattempters versus nonattempters. Regression modelsfailed to identify any significant predictors of pastsuicide attempt.

Pfennig et al.,(2005)

Journal of AffectiveDisorders

Salivary cortisol was collected for three consecutive days in 29controls, 80 bipolar individuals without a history of suicide and56 bipolar individuals with a past history of suicide. Clinical factorsthat affect salivary cortisol were also examined.

A past history of suicide was associated with a 7.4%higher bedtime salivary cortisol level in bipolarindividuals. There was no statistical differencebetween non-suicidal bipolar individuals andcontrols in bedtime salivary cortisol and awakeningsalivary cortisol was not different between the threegroups.

Yerevanianet al., (2004)

PLoS ONE We studied 27 bipolar subjects using the NEO-PI We found positive associations between personalityfactors and ratios of n-3 PUFA, suggesting thatconversion of short chain to long chain n-3s and theactivity of enzymes in this pathway may associatewith measures of personality. Thus, ratios ofdocosahexaenoic acid (DHA) to alpha linolenic acid(ALA) and the activity of fatty acid desaturase 2(FADS2) involved in the conversion of ALA to DHAwere positively associated with openness factorscores. Ratios of eicosapentaenoic acid (EPA) to ALAand ratios of EPA to DHA were positively associatedwith agreeableness factor scores. Finally, serumconcentrations of the n-6, arachidonic acid (AA),were significantly lower in subjects with a history ofsuicide attempt compared to non-attempters.

Evans et al.(2012)

Acta Neuropsychiatrica Two hundred fifty-five DSM-IV out-patients with bipolar disorderwere consecutively recruited from the Bipolar Disorder Program atHospital das Clínicas de Porto Alegre and the University Hospital atthe Universidade Federal de Santa Maria, Brazil.

Over 30% of the sample was obese and over 50% hada history of suicide attempt. In the multivariatemodel, obese patients were nearly twice (OR¼1.97,95% CI: 1.06–3.69, p¼0.03) as likely to have a historyof suicide attempt(s).

Azorin et al.(2009)

Iranian Journal ofPsychiatry and BehavioralSciences

One hundred patients were followed for 2–42 months (mean:20.6712.5 months).

Only one patient attempted suicide during thefollow-up period. 33% of the patients had history ofprevious suicide attempts. Female gender, divorce,

L.d.S. Costa et al. / Journal of Affective Disorders 170 (2015) 237–254244

Table 1 (continued )

Authors Journal Sample Main findings

and early age at onset of the disease wereindependently correlated with suicide attempt.

Undurragaet al. (2011)

Comprehensive Psychiatry Participants were 216 consecutive inpatients (97 men and 119women) with a Diagnostic and Statistical Manual of MentalDisorders, Fourth Edition, Text Revision (DSM-IV-TR), BD who wereadmitted to the Sant'Andrea Hospital's psychiatric ward in Rome(Italy).

Patients with BD-II had higher scores on the BHS(9.7875.37 vs. 6.8774.69; t143.59¼�3.94; Po0.001) than patients with BD-I. Hopelessness wasassociated with the individual pattern oftemperament traits (i.e., the relative balance ofhyperthymic vs. cyclothymic-irritable-anxious-dysthmic). Furthermore, patients with higherhopelessness (compared with those with lowerlevels of hopelessness) reported more frequentlymoderate to severe depression (87.1% vs. 38.9%;Po0.001) and higher MINI suicidal risk.

Shabani et al.(2013)

Comprehensive Psychiatry A sample of 102 outpatients with a diagnosis of bipolar disorderaccording to International Classification of Diseases, 10th Revisioncriteria during nonsyndromal stage were evaluated.

As compared with the nonsuicidal group, female sex,combined psychopharmacologic treatment, andhopelessness were independently associated withsuicide attempt. Hopelessness and insight intohaving a mental disorder were independentlyassociated with history of suicidal ideation.

Akiskal (2007) Journal of Nervous andMental Disease

Among 212 patients with bipolar disorder, 44 (21.2%) patients hadhistories of suicide attempts.

The variables that differentiated those who did fromthose who did not attempt suicide included age atfirst contact, lifetime history of antidepressant use,major depressive episode, mixed episode, auditoryhallucinations, rapid cycling, the number of previousmood episodes, age of first depressive episode, andage of first psychotic symptoms.

Weinstock andMiller (2008)

Neuropsychiatric Diseaseand Treatment

The data of 383 bipolar I disorder patients were included in theanalyses.

The demographic/clinical variables significantlyassociated with the MINI suicide risk scores includedage, number of overall previous episodes, the YoungMania Rating Scale score, the Montgomery AsbergDepression Rating Scale scores, and the ClinicalGlobal Impression Severity of Illness Scale for BipolarDisorder mania score, depression score, and overallscore. The variables affecting the differences ofsuicide risk scores between or among groups weretype of first mood episode, a history of rapid cycling,anxiety disorders, and alcohol use disorders.

Pompili et al.(2012)

Psicologia: Reflexao eCritica

The Iowa Gambling Task and the Conner's Continuous PerformanceTest evaluated impulsivity in 95 euthymic bipolar patients -42suicide attempters and 115 normal control participants.

A factorial analysis evaluated the adequacy of theinstruments. Furthermore, a multiple regressionanalysis was done in order to develop a model topredict suicide attempts. Our results point to aspecific type of impulsivity related to makingdecisions, lack of planning and borderlinepersonality disorder comorbidity. This type ofimpulsivity is a risk factor for suicide attempts inpatients with bipolar disorder.

AmericanPsychiatricAssociation(2003)

Journal of AffectiveDisorders

As part of the EPIDEP National Multisite French Study of 493consecutive DSM-IV major depressive patients evaluated in at leasttwo semi-structured interviews 1 month apart, 234 (55.2%) couldbe classified as with high religious involvement (HRl), and 190(44.8%) as with low religious involvement (LRl), on the basis oftheir ratings on the Duke Religious lndex (DRl).

Compared to LRl, HRl patients did not differ withrespect to their religious affiliation but had a laterage at onset of their affective illness with morehospitalizations, suicide attempts, associatedhypomanic features, switches under antidepressanttreatment, prescription of tricyclics, comorbidobsessive compulsive disorder, and family history ofaffective disorder in first-degree relatives.

World HealthOrganization(WHO)(2003)

Journal of ClinicalPsychiatry

A retrospective case control study of 149 depressed bipolar patients(DSM-III-R criteria) in a tertiary care university research clinic wasconducted. Patients who reported religious affiliation werecompared with 51 patients without religious affiliation in terms ofsociodemographic and clinical characteristics and history ofsuicidal behavior.

Religiously affiliated patients had more children andmore family-oriented social networks thannonaffiliated patients. As for clinical variables,religiously affiliated patients had fewer past suicideattempts, had fewer suicides in first-degree relatives,and were older at the time of first suicide attemptthan unaffiliated patients. Furthermore, patientswith religious affiliation had comparatively higherscores on the moral or religious objections to suicidesubscale of the RFLI, lower lifetime aggression, andless comorbid alcohol and substance abuse andchildhood abuse experience. After controlling forconfounders, higher aggression scores (P¼0.001)and lower score on the moral or religious objections

L.d.S. Costa et al. / Journal of Affective Disorders 170 (2015) 237–254 245

to overall mortality rates (Angst et al., 2002; Roshanaei-Moghaddamand Katon, 2009).

The risk of suicide for individuals with BD is approximately 60times greater than that of the general population (Simon et al.,2007). Fifteen to twenty percent of individuals with BD completesuicide and up to 40% report at least one suicide attempt duringtheir lifetime (Simon et al., 2007).

The ratio of suicide attempts to completed suicides for thegeneral population is 35:1, but for individuals with BD, the sameratio is 3:1 (Simon et al., 2007). In fact, it is estimated that BD mayaccount for one-quarter of all completed suicides (AmericanPsychiatric Association, 2013; Huber et al., 2014).

However, unlike other authors included in this review,Undurraga et al., (2012) concluded that suicidal risk-factors foundto be independent of bipolar disorder. This fact, as well as thedifferent conclusions reached by the authors, which will bedemonstrated below, exposes the need for further research.

4.1. Risk factors associated with sociodemographic components

Quality of life seems to be associated with suicidal behaviors(i.e., suicidal ideation, suicide attempts, and complete suicide) inthe general population and in psychiatric patients (De Abreu et al.,2012).

A recent review showed that Quality of Life (QoL) is markedlyimpaired in patients with BD, even when they are clinicallyeuthymic (De Abreu et al., 2012; Michalak et al., 2005). Also,stressful life event(s) was another preponderant factor predictingsuicide attempts, and has played an important role in predictingsuicide attempts among BD patients in many studies, particularlyduring depressive phases (Azorin et al., 2009; Ruengorn et al.,2012).

De Abreu et al. (2012) hypothesized that patients with BD andprevious suicide attempts would have worse QoL than patientswith BD but no previous suicide attempts . It is possible that lowQoL may reflect the existence of poor coping skills and inadequatesocial support, which in turn may increase the risk for suicideattempts (De Abreu et al., 2012). Further prospective studies areneeded to clarify the causal and temporal relationships betweenlow QoL and suicide attempts (De Abreu et al., 2012).

Pediatric Bipolar Disorder (PBD), for example, is associatedwith substantially lower average QoL than found with many othermajor medical illnesses, and worse than other mental illnesses inyouth except for major depression (Freeman et al., 2009; Algortaet al., 2011). Suicidality and lower youth QoL both were signifi-cantly associated with worse family functioning (Algorta et al.,2011). Poor family functioning, poor youth QoL, and mixedfeatures will each make unique contributions to suicidality as anoutcome variable (Algorta et al., 2011). Also, Etain et al. (2013)

demonstrated consistent associations between childhood traumaand more severe clinical characteristics in bipolar disorder .

Moreover, family history of completed suicide had the highestodds ratio of significant findings. In previous reports in bipolarcohorts (Galfavy et al., 2006; Valtonen et al., 2006), family historyof suicide was no different between bipolar patients with andwithout histories of attempts, although one study reported familyhistory was predictive of earlier attempts (Galfavy et al., 2006;Cassidy, 2011). Also, the higher frequency of bipolar disorderfamily history in agitated depression suggests that a bipolarvulnerability may be required to obtain such clustering of hypo-manic symptoms (Akiskal et al., 2005). Overall, findings support anassociation between family functioning and suicidality withinfamilies where youths have bipolar disorder (Miklowitz andChang, 2008). Results suggest that it is plausible that the youth'sillness may play an active role in disrupting family processes.Bipolar disorder may involve a potent combination of mooddysregulation and interpersonal processes where threats of harm— against oneself or another — may occur both impulsively and/orinstrumentally (Algorta et al., 2011). In fact, bipolar disorderimparts the greatest risk for completed suicide among youth(Goldstein et al., 2012).

With regard to age groups at risk for suicidal behavior, priorliterature suggests that young-aged patients are at higher risk ofsuicide compared to older patients, in line with the finding of thisreport. Studies with depressed patients have shown that youngpatients report a higher number of suicide attempts (Blair-Westet al., 1999; Antypa et al., 2013; Azorin et al., 2010). Although noassociation was found between age and suicide attempts (Azorinet al., 2009). Studies of adolescent suicide completers documentthe substantial contribution of parental depression to offspringsuicide risk (Gould et al., 1996), even after accounting for thechild's depressive severity (Brent et al., 1993). It is possible thatfamilial depression contributes to offspring suicide risk via multi-ple avenues, including decreased familial support and increasedconflict (Goldstein et al., 2012; Brent et al., 1994).

Ryu et al. (2010) investigated the descriptive characteristics ofsuicide attempts and the risk factors for suicide attempts in Koreanbipolar patients by assessing sociodemographic factors, clinicalfactors, and the methods of suicide attempts using retrospectivereviews of medical records . Ryu et al. (2010) reviewed medicalrecords of all 601 patients who were admitted to the psychiatricwards in one mental hospital and three general hospitals (Ryuet al., 2010). The 579 subjects who were included in the finalanalysis was comprised of 262 (45.3%) men and 317 (54.7%)women (Ryu et al., 2010).

Ryu et al. (2010) found two significant risk factors associated.First, they found that patients with depressive first episodesappear to be higher in suicide attempters. Bipolar patients with

Table 1 (continued )

Authors Journal Sample Main findings

to suicide subscale of the RFLI (Po0.001) weresignificantly associated with suicidal behavior indepressed bipolar patients. Moral or religiousobjections to suicide mediated the effects of religiousaffiliation on suicidal behavior in this sample.

Suttajit et al.(2013)

General HospitalPsychiatry

The aim of the study was looking for suicide risk factors amongpersonality dimensions and value system in patients with diagnosisof unipolar and bipolar affective disorder (n¼189 patients, n¼101controls).

The main limitations of the study are number ofparticipants, lack of data about stressful life eventsand treatment with lithium. Novelty seeking andharm avoidance dimensions constituted suicideattempt risk factors in the group of patients withaffective disorders. Protective role of cooperativenesswas discovered. Patients with and without suicideattempt in lifetime history varied in self-esteemposition in Value Survey.

L.d.S. Costa et al. / Journal of Affective Disorders 170 (2015) 237–254246

a depressive episode at their initial admission or first episode tendto have a depressed mood at the next episode (Daban et al., 2006;Perugi et al., 2000). Moreover, 60% of suicide attempters withdepressive episodes commit suicide at the first mood episode(Balázs et al., 2003). Prolonged exposure to depressive episodesmight increase the risk of suicide attempts in bipolar patients andpoor prognostic factor in suicide related behavior (Valtonen et al.,2006; Ryu et al., 2010). Second it was observed that bipolar IIpatients have a higher risk for suicide attempts. Bipolar II patientsare known to have greater risk of suicide than bipolar I patients(Balázs et al., 2003; Arató et al., 1988). Bipolar II patients show apredominantly depressive mood, mood lability, and mixed nature(Akiskal et al., 1995; Benazzi 2007). Bipolar II patients are likely tohave depressive or mixed episodes at hospital admission ratherthan bipolar I patients (Ryu et al., 2010). Also, there is a strongercontinuous relationship of suicidal risk with the proportion of totalrecurrences that were depressive-or-mixed vs. depressive(Baldessarini et al., 2012). Baldessarini et al. (2012) has shown astrong association of predominant depression, especially withmixed-episodes included, with suicidal behavior.

Furthermore, suicide attempts were observed in both males andfemales at similar rates. Moreover, females who attempted suicidewere as likely to have attempted suicide by a more violent method asmales (Cassidy, 2011). Although, it was observed that some authorscategorized the female gender as a risk factor (Bellivier et al., 2011).Parmentier et al. (2012) observed that reported rates of suicideattempts among women with bipolar disorders are about twice ashigh as among men with bipolar disorders, suggesting greaterlethality of suicide attempts in men (Parmentier et al., 2012;Suicidology AAo, 2000; Tondo et al., 2006).

Ruengorn et al. (2012) proposed a risk-scoring scheme forsuicide attempts in Thai patients with BD. Ruengorn et al. (2012)conducted a study at Suanprung Psychiatric Hospital, a total of 489patients' medical files were reviewed and included in the finalanalysis. Results revealed that suicide attempters were younger,single, did not have children, and had little or very little socialsupport. They reported experiencing more stressful life events,reported being depressed, had suffered from BD at an early age,had a family history of suicide, had previously attempted suicide,had previous suicidal ideation, alcohol use, and were prescribedantipsychotics, antidepressants, anxiolytics, and mood stabilizers(Ruengorn et al., 2012). Using multivariate logistic regression, theauthor found six indicators of suicide attempts: depressive epi-sodes, previous suicide attempt(s), stressful life event(s), inter-mittent or poor medication adherence, and shorter duration of BDtreatment. Psychotic symptom(s) provided an inverse associationwith suicide risk (Ruengorn et al., 2012).

Moreover, the altitude was seen as a significant risk factor(Huber et al., 2014). Several biological theories may explain analtitude-suicide association. Dopamine and serotonin are neuro-transmitters associated with pleasure, reward, and mood.Decreased levels of serotonin and increased levels of dopamineand norepinephrine associated with hypoxia at higher altitudesmay lead to increased irritability, depression, and suicide (Huberet al., 2014; Trouvin et al., 1986; Jou et al., 2009).

Converging lines of evidence also indicate that mitochondrialdysfunction plays a role in the pathophysiology of BD and mayinfluence the severity of episodes (Scaglia, 2010; Quiroz et al.,2008; Kato, 2006). Studies of patients with mitochondrial diseaseshow that both adults (Fattal et al., 2006, 2007) and children(Morava et al., 2010; Koene et al., 2009) have elevated rates ofdepressive symptoms. Metabolic stress due to hypoxia may haveimportant considerations for individuals with BD. Hypoxia due toreduced oxygen partial pressure at higher altitudes may furtherdecrease mitochondrial function in individuals with BD (McIntyreet al., 2008; Rezin et al., 2009). For these individuals, metabolic

changes associated with hypoxia may lead to depression, instabil-ity of mood, and increased risk of suicide (Huber et al., 2014;Vaccari et al., 1978).

4.2. Risk factors associated with genetic components

Suicide attempt was defined as an intentional self-inflictedinjury with self-destructive intent (Manchia et al., 2013). Globally,approximately one million individuals commit suicide each year(WHO, 2011; Sears et al., 2013). Suicidal behavior is a matter ofmajor concern in the management of BD patients for manyreasons. First, their suicide rates are about 60 times higher thanthat observed in general population. Second, about one third tohalf of these patients will make at least one suicide attempt duringtheir disease. Finally, their suicide acts have a higher lethality assuggested by a much lower ratio of attempted suicide (approxi-mately 3:1) than in the general population (approximately 30:1)(Baldessarini et al., 2006; Neves et al., 2010).

Although many data have suggested that BD confers a higherrisk of suicide than other psychiatric illnesses (Sajatovic, 2005),few studies have yet been conducted to investigate the contribu-tion of the genetic component (Magno et al., 2011).

Genetic variation plays an important role in BD and suicidesusceptibility. However, little is known about the genetic influenceon the risk of suicide, particularly in BD patients (Magno et al.,2011). The liability to suicidal behavior is influenced by geneticfactors (particularly family history of suicidal behavior and MajorAffective Disorders) (Manchia et al., 2013).

In addition, genetic determinants such as polymorphismswithin the tryptophan hydroxylase 1 (TPH1; Gene ID 7166 in11p15.3–p14) and the tryptophan hydroxylase 2 (TPH2; Gene ID121278 in 12q21.1) genes were found to be associated with suicideattempts of high lethality and with completed suicides, respec-tively (Manchia et al., 2013; Galfalvy et al., 2009; Lopez et al.,2007). These findings are of interest considering the associationbetween altered serotonin system function in the brain andsuicide (Manchia et al., 2013). Neves et al. (2010) showed thatserotonin polymorphism (5-HTTLPR; Gene ID 6532 in 17q11.2) isstrongly associated with violent suicidal behavior in BD patients.Their results could be an important step to create a genetic tool forlong-term suicide prediction (Neves et al., 2010). Biological mar-kers, such as 5-HTTLPR (Gene ID 6532 in 17q11.2), could help foridentification of potential suicide attempters (Neves et al., 2010).

Several lines of evidence indicate that brain-derived neuro-trophic factor (BDNF; Gene ID 627 in 11p13) is a good candidategene for involvement in suicidal behavior. Post-mortem studieshave shown that the expression of BDNF (Gene ID 627 in 11p13) issignificantly reduced in individuals that have committed suicide,regardless of psychiatric diagnosis (Sears et al., 2013; Dwivediet al., 2003; Karege et al., 2005). Moreover, an association betweenBDNF (Gene ID 627 in 11p13) gene and violent Suicide Attempt (SA)has been also detected in a sample of this patients (Neves et al.,2011; Jiménez et al., 2013). In addition, brain-derived neurotrophicfactor (BDNF; Gene ID 627 in 11p13) and lithium, well knowntherapeutic drug in mood disorder (Fountoulakis et al., 2008),reduces FoxO3a (Gene ID 2309 in 6q21) transcriptional activity(Magno et al., 2011; Mao et al., 2007; Zhu et al., 2004). FoxO3a(Gene ID 2309 in 6q21) influences distinct behavioral processeslinked to anxiety and depression. Recently, a study using aknockout (KO) mice model suggested that FoxO3a (Gene ID 2309in 6q21) may be a transcriptional target for anxiety and mooddisorder treatment (Magno et al., 2011; Polter et al., 2009). Thesedata suggest that FOXO3A (Gene ID 2309 in 6q21) is a novelsusceptibility locus for BD, but not for suicidal behavior in BDpatients. These results may contribute to a better understanding ofthe BD genetics (Magno et al., 2011).

L.d.S. Costa et al. / Journal of Affective Disorders 170 (2015) 237–254 247

Some evidence links phosphosinositol pathway to suicidalbehavior (Jiménez et al., 2013). Jiménez et al. (2013) suggest thatgenetic variability at rs669838-IMPA2 (Gene ID 3613 in 18p11.2),rs4853694-INPP1 (Gene ID 3628 in 2q32), rs1732170- GSK3b(Gene ID 2932 in 3q13.3) and rs11921360-GSK3b (Gene ID 2932in 3q13.3) genes is associated with a higher risk of attemptingsuicide in bipolar patients (Jiménez et al., 2013). It is known that attherapeutic concentrations, lithium immediately inhibits severalenzymes, such as both isoenzymes (1 and 2) of inositolmonopho-sphatase (IMPA), inositolpolyphosphate-1 phosphatase (INPP1),phosphoglucomutase and glycogen synthasekinase-3b (GSK3b)(Jiménez et al., 2013; Quiroz et al., 2004; Serretti et al., 2009).The phosphoinositol pathway is associated with cellular activitiessuch as metabolism, secretion, phototransduction, cell growth anddifferentiation (Jiménez et al., 2013; Serretti et al., 2009).

The question remains how genetic risk factors contribute to themanifestation of bipolar disorder. If we could answer this question,early intervention and effective treatment could become a reality(Kerner et al., 2013). Heritable factors have important effect onsusceptibility to suicidal behavior, which is supported by severalstudies showing that genetic polymorphisms play a role in suiciderisk (Magno et al., 2011; Galfalvy et al., 2009; Magno et al., 2010;Roy and Segal, 2001). Strong heritability of bipolar disorder hasbeen supported by many studies, but the identification of causalvariants has been challenging (Kerner et al., 2013).

4.3. Risk factors associated with medicines and drugs in general thatinterfere with bipolar disorder

Prevalence of comorbid substance use disorders was higheramong unipolar and bipolar disorder deaths than that among allother deaths. Among unipolar and bipolar disorder deaths, comor-bid substance use disorders were associated with elevated risks forsuicide and other unnatural death in both males and females(Yoon et al., 2011). Antiepileptic drugs are approved for thetreatment of epilepsy, bipolar disorder, and neuropathic pain. Eachof these conditions is associated with an elevated risk of suicide(Simon et al., 2007; Christensen et al., 2007; Nilsson et al., 2002;Ratcliffe et al., 2008; Tsai et al., 2002; Leon et al., 2012). Alcoholaddicts were associated with diagnoses of several types of person-ality disorder and bipolar disorder and presented a greater suiciderisk than the subgroup of other substance use disorders (SUDs)(Arias et al., 2013). Personality disorder and alcohol dependencewere the most common secondary diagnoses in the BD group(Clements et al., 2013). Twenty-five percent of persons consumealcohol prior to suicidal attempt (Raja and Azzoni, 2004). Leverichet al. (2003) also point to family history of abuse of medicinaldrugs as a suicide risk factor (Leverich et al., 2003; Pawlak et al.,2013).

The risk of suicidal behavior or ideation was significantlyelevated in patients who received an antiepileptic compared withthose who received placebo when no adjustments were made fortrial differences (Leon et al., 2012). This warning was based on anU.S. Food and Drug Administration (FDA) examination of data from199 randomized clinical trials of 11 antiepileptic medications(carbamazepine, divalproex, felbamate, gabapentin, lamotrigine,levetiracetam, oxcarbazepine, pregabalin, tiagabine, topiramate,and zonisamide) (Leon et al., 2012). The role of antidepressants(AD) in suicide risk is important, and this has received muchattention in recent years (McElroy et al., 2006). There is anassociation between the use of AD and a risk of acute manicswitch in BD (Ghaemi et al., 2003), and McElroy et al. (2006)concluded that AD may induce suicidal intention by manic con-version in a subset of depressive presentations (Finseth et al.,2012). In general, these effects were much stronger for Drug UseDisorders (DUD) than for Alcohol Use Disorders (AUD). Both

substance use disorders had stronger effects on suicide amongfemales, whereas their effects on other unnatural deaths werestronger among males (Yoon et al., 2011).

According to the National Epidemiologic Survey on Alcohol andRelated Conditions (NESARC) estimates, among individuals with12-month unipolar depressive disorder, 14.1% had alcohol usedisorders (AUD) and 4.6% had drug use disorders (DUD). Theserepresented significantly elevated risks for the comorbid sub-stance use disorders (Yoon et al., 2011; Hasin et al., 2005). In asample consisted of 837 outpatients from Madrid, Spain. Ariaset al., (2013) compared 528 subjects with a lifetime diagnosis ofalcohol abuse or dependence and 182 with other substance usedisorders (SUDs) not involving alcohol. The Mini InternationalNeuropsychiatric Interview (MINI) was used to evaluate Axis Idisorders and the Personality Disorder Questionnaire to evaluatepersonality disorders. It was considered that 76.1% of the alcoholaddicts had a current dual diagnosis, the most prevalent beingmood and anxiety disorders. Fifty-two percent had a personalitydisorder and most of them (81.6%) had other SUDs Compared topeople with bipolar disorder alone, those who have bipolardisorder with comorbid SUD have an increased prevalence ofsuicide attempts (Sublette et al., 2009; Dalton et al., 2003; Lopezet al., 2001; Oquendo et al., 2010; Potash et al., 2000; Kennesonet al., 2013).

Gilbert et al. (2011) identified the extremely difficult to predictsuicidal behavior, even when comprehensive clinical informationis available. However, empirical evidence has shown that peoplewith mood disorders and/or substance use disorders experienceexcess mortality (Yoon et al., 2011; Amaddeo et al., 1995; Blacket al., 1985; Bruce et al., 1994; Cuijpers and Smit, 2002; Harris andBarraclough, 1998; Hiroeh et al., 2001; Mykletun et al., 2007;Wulsin et al., 1999). Actually, only a few studies have examined theassociation between mood disorders and other causes of unnaturaldeath (Black et al., 1985; Hiroeh et al., 2001; Mykletun et al., 2007;Gau and Cheng, 2004; Joukamaa et al., 2001; Ösby et al., 2001),despite the fact that individuals with mood disorders, especiallythose with bipolar disorder, are more likely to engage in fatalaccidents due to impaired attention and concentration (Stahl,2000) or to be victims of homicide due to affective psychoses(Yoon et al., 2011; Hiroeh et al., 2001). These findings suggest thatabuse of alcohol or drugs could be considered as an importantcharacteristic to identify subgroups at risk for suicidal behavior(Akiskal et al., 1995; Maremmani et al., 2007). Leverich et al.(2003) have found a correlation between suicidal behaviors andthe family history of suicide attempts or committed suicides, aswell as the family history of abuse of medicinal drugs (Leverichet al., 2003; Pawlak et al., 2013). Interventions to reduce suiciderisk in bipolar disorder need to address the common and high riskcomorbidity with alcohol use disorders (Oquendo et al., 2010).

4.4. Risk factors associated with biological components

Altered functioning of the Hypothalamic-pituitary-adrenal(HPA) axis has been reported in suicidal behavior and in BipolarDisorder (BD) (Daban et al., 2005; Mann, 2003). However, manystudies of HPA axis function in bipolar disorder have not examinedthe potential effects of Suicidal Behavior (SB) (Cassidy et al., 1998;Cervantes et al., 2001; Cookson et al., 1985; Godwin, 1984;Linkowski et al., 1994; Rybakowski and Twardowska, 1999;Schmider et al., 1995) and studies of the association betweenHPA axis activity and suicidal behavior in varied diagnostic groupshave had mixed results (Black et al., 2002; Coryell and Schlesser,2001; Dahl et al., 1991; Duval et al., 2001; Jokinen and Nordström,2008; Jokinen and Nordström, 2009; Jokinen et al., 2009; Lindqvistet al., 2008; Pfennig et al., 2005; Pitchot et al., 2008; Tripodianakiset al., 2000; Yerevanian et al., 2004; Kamali et al., 2012).

L.d.S. Costa et al. / Journal of Affective Disorders 170 (2015) 237–254248

The HPA axis has been examined using a number of methods.Basal cortisol secretion has been measured with 24 h urinarycortisol secretion and serum or salivary cortisol levels. The feed-back and suppression mechanisms of the HPA axis have beeninvestigated with the dexamethasone suppression test (DST)(Kamali et al., 2012) or the dexamethasone/corticotropin-releasinghormone (DEX/CRH) challenge test (Carroll et al., 1981; Heuseret al., 1994).

Kamali et al. (2012) examined HPA axis activity as a traitmarker for bipolar disorder and suicide by measuring salivarcortisol in a bipolar cohort with a history of suicide and comparesit with non-suicidal bipolar individuals and unaffected controls.Kamali et al., (2012) hypothesis was that those with bipolardisorder would have elevated basal salivary cortisol compared tounaffected controls, and that the suicidal bipolar individuals(defined by a lifetime history of attempted suicide) would havehigher levels of salivary cortisol compared to those with no historyof suicidal behavior and unaffected controls.

A total of 185 individuals participated in the study andprovided at least one salivary cortisol sample. The majority (152individuals) were enrolled in the Prechter Longitudinal Study. Onesubject subsequently retracted their consent, the salivary samplesfrom one subject were missing at time of analysis, the salivavolume from two individuals was insufficient for analysis andthree individuals did not complete enough of the diagnosticinterview to reach a diagnosis. Of the remaining 178, 118 (66.3%)had a diagnosis of bipolar I, 14 (7.9%) had bipolar II with recurrentdepression, 7 (3.9%) had schizoaffective disorder—bipolar type, 8(4.5%) had other affective diagnosis (depressive disorder NOS,MDD, Bipolar II with single depressive episode), 2 (1.1%) had onlynon-affective diagnoses (alcohol abuse and dependence) and 29(16.3%) were unaffected controls. For the purpose of this study,those with bipolar I, bipolar II with recurrent depression andschizoaffective disorder bipolar type were grouped together as thebipolar group (N¼139) and were categorized based on reportedsuicide history obtained during the Diagnostic Interview forGenetic Studies (DIGS) (Kamali et al., 2012).

Kamali et al., (2012) found elevated bedtime salivary cortisol inbipolar individuals with a history of suicide attempts compared tononsuicidal bipolar individuals. Secondary analysis of the intensityof suicidal behavior and level of bedtime cortisol indicated apositive correlation, with the highest cortisol levels reported inindividuals that had made a past serious suicide attempt. Thedifference in bedtime salivary cortisol between suicidal and non-suicidal bipolar individuals remained significant even after con-trolling for age and sex, body mass index (BMI), smoking status,childhood sexual abuse, medications, mood state at time ofsampling and several clinical factors related to course and severityof illness (substance use disorders, chronicity of illness, rapidcycling, mixed states, years of illness, age of onset, anxiety andpsychosis). This is a strong indicator that their finding is related tothe presence of a past history of suicidality and not related toseverity of illness, mood state, or demographic confounders. Thepresence of this finding during different mood states and also inthe euthymic state indicates that hyperactivity of the HPA axis is abiological marker related to suicidality in bipolar disorder andwarrants more detailed investigation (Kamali et al., 2012).

The difference between bipolar participants with and withoutsuicidal behavior was only 0.05 μg/dl. Currently, the test has lowsensitivity and specificity in detecting individuals with suicidalhistory in practical clinical applications. However, the observationof a sustained correlation between increasing suicidality andcortisol levels while controlling for confounding clinical andbiological factors clearly indicates the relevance of HPA axisabnormalities in this potentially lethal clinical condition (Kamaliet al., 2012).

Several potential risk factors have been linked to suicidalbehavior. Two of these include personality factors and Polyunsa-turated fatty acids (PUFA) serum levels. It is unknown whetherPUFA serum levels are associated with personality factors and ifthese may interact to affect suicidal behavior (Evans et al., 2012).Supplementation with the long-chain n-3 (n-3) fatty acids, doc-osahexaenoic acid (DHA) and eicosapentaenoic acid (EPA), eitheras stand alone or adjunctive therapies have shown efficacy in thetreatment of bipolar disorder (Evans et al., 2012). Epidemiologicalstudies have pointed to an association between n-3 and n-6dietary intake and lifetime prevalence of bipolar disorder. Popula-tions that consume greater long-chain n-3 s and less long chainn-6 s have a lower incidence of bipolar disease (Evans et al., 2012;Hibbeln et al., 2006).

N-3 intake inversely associates with violent behavior andsuicidality, Evan et al. hypothesize that serum levels of the longchain n-3 s, DHA and EPA, may positively associate with person-ality factors that may be protective against suicide behavior and/ornegatively associate with personality factors that, themselves,associate with increased risk of suicide behavior (Evans et al.,2012). Several studies suggest that BD patients' previous suicideattempt(s) may indicate that they are more than 50% more likelyto go on to complete suicide (Ruengorn et al., 2012; Tsai et al.,2002; Isometsa et al., 1994).

Identifying metabolic or dietary factors that influence factorsassociated with psychiatric illness may provide a path to improv-ing therapeutic tools. Evans et al. found associations between lipidprofiles and suicidal history in bipolar subjects (Evans et al., 2012).These data further support a link between essential fatty acidmetabolism and mood disorders. While the current pilot study isan observational, cross-sectional study, it raises important ques-tions regarding potential causative roles for lipid profiles inregulating personality phenotypes that may impact the treatmentof bipolar disorder. Nevertheless, the fact that personality factors,promoted as trait markers in bipolar disorder are not entirelystable (Barnett and Huang, 2010) and longer-term longitudinalstudies are necessary to examine the relationship between per-sonality traits and fatty acid profiles. Evans et al., (2012)

Of significant interest is the co-occurrence of metabolic dis-turbances in bipolar disorder, particularly obesity (Gomes et al.,2010). Gomes, et al. find adds to the notion that obesity is acorrelate of severity in patients with bipolar disorder. Obesepatients usually have more markers of illness severity, such asmore previous affective episodes (Fagiolini et al., 2002) and suicideattempts (Fagiolini et al., 2004; Fagiolini et al., 2005; Wang et al.,2006). Recent data have stressed common features in the under-lying pathophysiology of obesity and bipolar disorder. Leptin, a keyhormone in regulation of adiposity has been shown to be posi-tively associated with risk for depression in a prospective study(Pasco et al., 2007). Disturbances in metabolic pathways such asinsulin-mediated glucose homeostasis, overactivation of thehypothalamic–pituitary–adrenal axis, dysregulated immune andinflammatory processes and adipocytokines profiles are present inboth conditions (Gomes et al., 2010; McIntyre et al., 2007).

4.5. Risk factors associated with psychological causes

BD is a frequent and chronic psychiatric disorder associated withan increase in all-cause mortality (Ösby et al., 2001; McIntyre andKonarski, 2004; McIntyre et al., 2008). In particular, among mentaldisorders, BD is one of the leading causes of suicidal behaviors andthis is a major issue in the management of the disease (Parmentieret al., 2012). Mixed-states and well as depressions are stronglyassociated with suicidal behavior in patients with BD (Algortaet al., 2011; Baldessarini et al., 2012; Azorin et al., 2009; Pompiliet al., 2009; Baldessarini et al., 2010; Undurraga et al., 2011).

L.d.S. Costa et al. / Journal of Affective Disorders 170 (2015) 237–254 249

Bipolar disorder is strongly associated with suicidal ideations,attempts and commissions (Shabani et al., 2013). Cyclothymictemperament may influence suicide risk on multiple levels, fromdetermining emotional reactivity in stressful situations at the levelof the personality, through determining illness and illness coursecharacteristics, to influencing within-episode dynamics (Rihmeret al., 2013).

There are also no previous studies investigating the role ofpossible mediating factors, such as hopelessness, in the associationbetween affective temperaments and suicidal behavior. Because itis well known that patients with BD-II are at a higher risk forattempting and completing suicide (Pompili et al., 2009; Rihmerand Pestality, 1999) and that hopelessness has been found to be agood predictor of suicidal behavior (Beck et al., 1990; Akiskal,2007), understanding the relationship between these factors, andthe possibly differential association of these factors, in patientswith BD-I and BD-II disorders would give us better insight in to thenature of the emergence of suicidal behavior (Pompili et al., 2012).

The finding of hopelessness as the most important variablewhen compared with depression is consistent with the nature ofthese psychopathological features (Acosta et al., 2012). A recentmeta-analysis revealed that previous suicide attempts and hope-lessness were the main risk factors for suicide, and that earlyonset, depressive symptoms, and family history of suicide werethe main risk factors for nonfatal suicide related behavior (Ryuet al., 2010; Hawton et al., 2005). We cannot dismiss the possibilitythat hopelessness may also, at least in part, represent a conse-quence of a more severe course of illness, especially those withlifetime depressive burden, and predispose to suicidality from thatperspective as well (Acosta et al., 2012). Hopelessness about thefuture in suicidal individuals is a multi-faceted construct but lackof positive future thinking is more important than presence ofnegative future thinking (Fountoulakis et al., 2012).

Patients with bipolar disorder have recurrent fluctuating moodepisodes with functional impairment, (Weinstock and Miller, 2008)which might induce chronic distress and increase suicide relatedbehaviors (Ryu et al., 2010; MacKinnon et al., 2003). Because suicideand suicidal behaviors are the result of a combination of individual riskfactors, precipitating stressors, and current disease features, theprediction of a suicide attempt for a given patient on the basis of riskfactors statistically associated with suicide or suicide attempts inpopulations of patients with bipolar disorder is difficult (Song et al.,2012). For Pompili et al. not only the absolute elevations of eachtemperament may be associated with psychopathological symptomsbut also that the individual pattern of temperaments may beassociated with a higher suicidal risk (Pompili et al., 2012). However,what we so far know about the risk factors associated with suicidalthinking and behavior in bipolar disorder has overwhelmingly beenderived by studying individuals who are in the “acute” phase of theirdisorder (Acosta et al., 2012).

Many studies have investigated clinical characteristics asso-ciated with suicidal behavior. Gender has been associated withsuicidal behavior in BD: men have a 4-fold greater risk for suicidethan women (Suicidology AAo, 2000; American PsychiatricAssociation, 2003; World Health Organization WHO, 2003). Rela-tive to the risk in the general population, BD is associated with anincreased risk of suicidal behavior in women and a higher lethalityin men (Parmentier et al., 2012).

In particular, careful evaluation and effective management ofbipolar depression among patients with mood disorder duringmajor depressive episodes is necessary to prevent suicide attemptsin bipolar disorder (Ryu et al., 2010). While suicidal ideation and ahistory of attempted suicide are among the most important risksfor suicide (Osman et al., 2001; Kuo et al., 2001), only a few studieshave taken into account both suicidal ideas and attempts inassessing the risk factors (Suttajit et al., 2013). The characteristics

associated with suicidal behavior in patients with BD may stimu-late the development of specific therapeutic strategies; these mayinclude emotional and hostility regulation and problem-solvingtherapies or specific treatment of comorbid social phobia oraddiction (Parmentier et al., 2012; Gray and Otto, 2001; Stewartet al., 2009). Suicide prevention strategies are currently based onscreening for the numerous risk factors (de Moraes et al., 2013).

4.6. Risk factors associated with components of religious andspiritual components

Religiosity and Spirituality are important aspects to identifygroups at risk of suicide in BD. However, there is a lack of studieson their impact on bipolar disorder and little is known about them(Azorin et al., 2013).

Dervic et al. (2011) related higher score on the moral orreligious objections to suicide subscale of the Reason for LivingInventory (RFLI) with fewer suicidal acts in depressed bipolarpatients. The strength of this association was comparable to that ofaggression scores and suicidal behavior, and had an independenteffect. A possible protective role of moral or religious objections tosuicide deserves consideration in the assessment and treatment ofsuicidality in bipolar disorder. In this study (Dervic et al., 2011),patients who reported religious affiliation were compared with 51patients without religious affiliation in terms of sociodemographicand clinical characteristics and history of suicidal behavior. Theresults were patients with religious affiliation had comparativelyhigher scores on the moral or religious objections to suicidesubscale of the RFLI, lower lifetime aggression, and less comorbidalcohol and substance abuse and childhood abuse experience(Dervic et al., 2011).

In another hand, Azorin et al. (2013) identified another point ofview. In their sample, Compared to Low Religious Involvement(LRI), High Religious Involvement (HRI) patients did not differ withrespect to their religious affiliation but had a later age at onset oftheir affective illness with more hospitalizations, suicide attempts,associated hypomanic features, switches under antidepressanttreatment, prescription of tricyclics, comorbid obsessive compul-sive disorder, and family history of affective disorder in first-degree relatives. The following independent variables were asso-ciated with religious involvement: age, depressive temperament,mixed polarity of first episode, and chronic depression. The studyconcluded that in depressive patients belonging to the bipolarspectrum, high religious involvement associated with mixedfeatures may increase the risk of suicidal behavior, despite theexistence of religious affiliation.

The current study (Azorin et al., 2013) may help understandsome potential negative effects of religious involvement in depres-sive patients belonging to the bipolar spectrum. First of all, theirfindings may be in line with the hypothesis of Cruz et al. (2010)that higher levels of distress as such caused by mixed episodesand/or chronic depression, would prompt patients to seek relieffrom religion, and therefore increase the frequency of theirreligious behaviors. However, it could be that, once depressed,HRI patients become the victims of their religious commitmentand that, in this case, religion exerts harmful effects on health.

Actually, for an individual with depressive temperament char-acterized by a rigid duty - orientation of his behavior, whichdistinguishes itself by an overidentification with what is norma-tively expected or by a meticulous fulfillment of social norms(Tellenbach, 1974), experiencing hypomanic social desinhibitionmay hardly be assimilated in his usual way of life. This “egodys-tonic” experience could therefore appear to consciousness underthe form of obsessive thoughts such as the “fear of committing asin” or an “excessive guilt” and give rise to compulsive religiousbehaviors, such as those found in our HRI patients. In those cases,

L.d.S. Costa et al. / Journal of Affective Disorders 170 (2015) 237–254250

it is likely that religious involvement may aggravate their guiltfeelings or the sinful character of their hypomanic experiences,enhancing thereby the suicidal tendencies. This may contribute toexplain why in the case of some mixed depressive patients,religiosity is unlikely to be associated with less suicidal behavior(Azorin et al., 2013).

The last study (Pawlak et al., 2013) confirms what Dervic et al.analyzed, affirming that subjective sense of religious involvementmay play a protective role in some communities (Sisask et al.,2010). Patients, who have not declared their commitment to anyreligion, undertook suicidal attempts significantly more often andhad more relatives, who had died of suicide, in comparison withthose involved in religion (Dervic et al., 2004).

5. Conclusion

The results of the studies in the literature show that the riskfactors associated with bipolar disorder and suicide exist and arerelevant to clinicians and researchers, whereas knowledge of suchinfluence better diagnosis and prognosis of BD cases involvingsuicide risk. Notwithstanding the differences in some points of thestudies, research becomes important to maintain the high qualityof knowledge of the disorder and its peculiarities, seekingimproved quality of life for people suffering from bipolar disorder.

Role of funding sourceWe have no foundation source.

Conflict of interestMr. Costa, Mr. Alencar, Mr. Nascimento and Drs. Maria do Socorro, Cláudio,

Sally, Regiane, Bianca, Roberto, Marcos Antonio, Alberto and Modesto have noconflicts of interest or financial ties to report.

AcknowledgmentsThe authors of this review would like to thank the support of the Suicidology

Research Group, Federal University of Ceará (UFC/Brazil)/Conselho Nacional deDesenvolvimento Científico e Tecnológico (CNPq/Brazil) and the Laboratório deEscrita Científica (LABESCI/Brazil)—Medical School of Federal University of Cariri(UFCA/Brazil).

References

Abreu, L.N., Lafer, B., Baca-Garcia, E., et al., 2009. Suicidal ideation and suicideattempts in bipolar disorder type I: an update for the clinician. Rev. Bras.Psiquiatr. 31 (3), 271–280.

Acosta, F.J., Vega, D., Navarro, S., et al., 2012. Hopelessness and suicidal risk inbipolar disorder. A study in clinically nonsyndromal patients. Compr. Psychiatry53 (8), 1103–1109.

Akiskal, H., 2007. Targeting suicide preventing to modifiable risk factors: hasbipolar II been overlooked? Acta Psychiatr. Scand. 116 (6), 395–402.

Akiskal, H.S., Maser, J.D., Zeller, P.J., et al., 1995. Switching from unipolar‘ to bipolarII. An 11-year prospective study of clinical and temperamental predictors in 559patients. Arch. Gen. Psychiatry 52, 114–123.

Akiskal, H.S., Benazzi, F., Perugi, G., et al., 2005. Agitated “unipolar” depression re-conceptualized as a depressive mixed state: Implications for theantidepressant-suicide controversy. J. Affect. Disord. 85 (3), 245–258.

Algorta, G.P., Youngstrom, E.A., Frazier, T.W., et al., 2011. Suicidality in pediatricbipolar disorder: predictor or outcome of family processes and mixed moodpresentation? Bipolar Disord. 13 (1), 76–86.

Amaddeo, F., Bisoffi, G., Bonizzato, P., et al., 1995. Mortality among patients withpsychiatric illness. A ten-year case register study in an area with a community-based system of care. Br. J. Psychiatry 166, 783–788.

American Psychiatric Association, 2003. Practice guideline for the assessment andtreatment of patients with suicidal behaviors. Am. J. Psychiatry 160 (Suppl. 11),S1–S60.

American Psychiatric Association, 2013. Diagnostic and statistical manual of mentalDisorders, 5th ed. American Psychiatric Association, Arlington, VA (DSM-5).

Angst, F., Stassen, H.H., Clayton, P.J., et al., 2002. Mortality of patients with mooddisorders: follow-up over 34–38 years. J. Affect. Disord. 68, 167–181.

Angst, J., Angst, F., Gerber-Werder, R., et al., 2005. Suicide in 406 mood-disorderpatients with and without long-term medication: a 40 to 44 years' follow- up.Arch. Suicide Res. 9 (3), 279–300.

Antypa, N., Antonioli, M., Serretti, A., 2013. Clinical, psychological and environ-mental predictors of prospective suicide events in patients with BipolarDisorder. J. Psychiatr. Res. 47 (11), 1800–1808.

Arató, M., Demeter, E., Rihmer, Z., et al., 1988. Retrospective psychiatric assessmentof 200 suicides in Budapest. Acta Psychiatr. Scand. 77, 454–456.

Arias, F., Szerman, N., Vega, P., et al., 2013. Alcohol abuse or dependence and otherpsychiatric disorders. Madrid study on the prevalence of dual pathology. Ment.Health Subst. Use: Dual Diagn. 6 (4), 339–350.

Azorin, J.-M., Kaladijan, A., Fakra, E., et al., 2013. Religious involvement in majordepression: protective or risky behavior? The relevance of bipolar spectrum. J.Affect Disord. 150 (3), 753–759.

Azorin, J.M., Kaladjian, A., Adida, M., et al., 2009. Risk factors associated withlifetime suicide attempts in bipolar I patients: findings from a French nationalCohort. Compr. Psychiatry 50 (2), 115–120.

Azorin, J.M., Aubrun, E., Bertsch, J., et al., 2009. Mixed states vs. pure mania in theFrench sample of the EMBLEM study: results at baseline and 24 months–European mania in bipolar longitudinal evaluation of medication. BMC Psy-chiatry 9, 33–40.

Azorin, J.M., Kaladjian, A., Adida, M., et al., 2009. Risk factors associated withlifetime suicide attempts in bipolar I patients: findings from a French NationalCohort. Compr. Psychiatry 50 (2), 115–120.

Azorin, J.M., Kaladjian, A., Besnier, N., et al., 2010. Suicidal behavior in a FrenchCohort of major depressive patients: characteristics of attempters and non-attempters. J. Affect Disord. 123 (1–3), 87–94.

Balázs, J., Lecrubier, Y., Csiszér, N., et al., 2003. Prevalence and comorbidity of affectivedisorders in persons making suicide attempts in Hungary: importance of the firstdepressive episodes and of bipolar II diagnoses. J. Affect Disord. 76, 113–119.

Baldessarini, R.J., Pompili, M., Tondo, L., 2006. Suicide in bipolar disorder: risks andmanagement. CNS Spectr. 11, 465–471.

Baldessarini, R.J., Salvatore, P., Khalsa, H.-M.K., 2010. Dissimilar morbidity followinginitial mania versus mixed-states in type-I bipolar disorder. J. Affect Disord.126, 299–302.

Baldessarini, R.J., Undurraga, J., Vázquez, G.H., et al., 2012. Predominant recurrencepolarity among 928 adult international bipolar I disorder patients. ActaPsychiatr. Scand. 125 (4), 293–302.

Barnett, J.H., Huang, J., Perlis, R.H., 2010. Personality and bipolar disorder: dissectingstate and trait associations between mood and personality. Psychol. Med., 1–12.

Beck, A.T., Brown, G., Berchick, R.J., et al., 1990. Relationship between hopelessnessand ultimate suicide: a replication with psychiatric outpatients. Am. J. Psychia-try 147, 190–195.

Bellivier, F., Yon, L., Luquiens, A., et al., 2011. Suicidal attempts in bipolar disorder:results from an observational study (EMBLEM). Bipolar Disord. 13 (4), 377–386.

Benazzi, F., 2007. Bipolar disorder–focus on bipolar II disorder and mixed depres-sion. Lancet 369, 935–945.

Black, D., Monahan, P., Winokur, G., 2002. The relationship between DST results andsuicidal behavior. Ann. Clin. Psychiatry 14, 83–88.

Black, D.W., Warrack, G., Winokur, G., 1985. Excess mortality among psychiatricpatients. The Iowa record-linkage study. J. Am. Med. Assoc. 253, 58–61.

Blair-West, G.W., Cantor, C.H., Mellsop, G.W., et al., 1999. Lifetime suicide risk inmajor depression: sex and age determinants. J. Affect. Disord. 55, 171e8.

Brent, D.A., Perper, J.A., Moritz, G., et al., 1993. Stressful life events, psychopathol-ogy, and adolescent suicide: a case control study. Suicide Life Threat. Behav. 23(3), 179–187.

Brent, D.A., Perper, J.A., Moritz, G., et al., 1994. Familial risk factors for adolescentsuicide: a case-control study. Acta Psychiatr. Scand. 89 (1), 52–58.

Bruce, M.L., Leaf, P.J., Rozal, G.P., et al., 1994. Psychiatric status and 9-year mortalitydata in the New haven epidemiologic catchment area study. Am. J. Psychiatry151, 716–721.

Carroll, B.J., Feinberg, M., Greden, J.F., et al., 1981. A specific laboratory test for thediagnosis of melancholia. Standardization, validation, and clinical utility. Arch.Gen. Psychiatry 38, 15–22.

Cassidy, F., 2011. Risk factors of attempted suicide in bipolar disorder. Suicide LifeThreat. Behav. 41 (1), 6–11.

Cassidy, F., Ritchie, J.C., Carroll, B.J., 1998. Plasma dexamethasone concentration andcortisol response during manic episodes. Biol. Psychiatry 43, 747–754.

Cervantes, P., Gelber, S., Kin, F.N., et al., 2001. Circadian secretion of cortisol inbipolar disorder. J. Psychiatry Neurosc. 26, 411–416.

Christensen, J., Vestergaard, M., Mortensen, P.B., et al., 2007. Epilepsy and risk ofsuicide: a population-based case-control study. Lancet Neurol. 6, 693–698.

Clements, C., Morriss, R., Jones, S., et al., 2013. Suicide in bipolar disorder in anational English sample, 1996–2009: frequency, trends and characteristics.Psychol. Med. 43 (12), 2593–2602.

Cookson, J.C., Silverstone, T., Williams, S., et al., 1985. Plasma cortisol levels inmania: associated clinical ratings and changes during treatment with haloper-idol. Br. J. Psychiatry 146, 498–502.

Coryell, W., Schlesser, M., 2001. The dexamethasone suppression test and suicideprediction. Am. J. Psychiatry 2001 (158), 748.

Cruz, M., Pincus, H.A., Welsh, D.E., et al., 2010. The relationship between religiousinvolvement and clinical status of patients with bipolar disorder. BipolarDisord. 12, 68–76.

Cuijpers, P., Smit, F., 2002. Excess mortality in depression: a meta-analysis ofcommunity studies. J. Affect. Disord. 72, 227–236.

de Moraes, P.H.P., Neves, F.S., Vasconcelos, A.G., et al., 2013. Relationship betweenneuropsychological and clinical aspects and suicide attempts in euthymicbipolar patients. Psicologia: Reflexao e Critica 26 (1), 160–167.

L.d.S. Costa et al. / Journal of Affective Disorders 170 (2015) 237–254 251

Daban, C., Vieta, E., Mackin, P., et al., 2005. Hypothalamic-pituitaryadrenal axis andbipolar disorder. Psychiatric Clin N. Am. 28, 469–480.

Daban, C., Colom, F., Sanchez-Moreno, J., et al., 2006. Clinical correlates of first-episode polarity in bipolar disorder. Compr. Psychiatry 47, 433–437.

Dahl, R.E., Ryan, N.D., Puig-Antich, J., et al., 1991. 24-hour cortisol measures inadolescents with major depression: a controlled study. Biol. Psychiatry 1991(30), 25–36.

Dalton, J.E., Cate-Carter, T.D., Mundo, E., et al., 2003. Suicide risk in bipolar patients:the role of co-morbid substance use disorders. Bipolar Disord. 5, 58–61.

De Abreu, L.N., Nery, F.G., Harkavy-Friedman, J.M., et al., 2012. Suicide attempts areassociated with worse quality of life in patients with bipolar disorder type I.Compr. Psychiatry 53 (2), 125–129.

Dervic, K., et al., 2004. Religious affiliation an suicide attempt. Am. J. Psychiatry 161(12), 2303–2308.

Dervic, K., Carballo, J.J., Baca-Garcia, E., et al., 2011. Moral or religious objections tosuicide may protect against suicidal behavior in bipolar disorder. J. Clin.Psychiatry 72 (10), 1390–1396.

Duval, F., Mokrani, M.C., Correa, H., et al., 2001. Lack of effect of HPA axishyperactivity on hormonal responses to d-fenfluramine in major depressedpatients: implications for pathogenesis of suicidal behavior. Psychoneuroendo-crinology 26, 521–537.

Dwivedi, Y., Rizavi, H.S., Conley, R.R., et al., 2003. Altered gene expression of brain-derived neurotrophic factor and receptor tyrosine kinase B in postmortem brainof suicide subjects. Arch. Gen. Psychiatry 60 (8), 804–815.

Eroglu, M.Z., Karakus, G., Tamam, L., 2013. Bipolar disorder and suicide. DusunenAdam. 26 (2), 139–147.

Etain, B., Aas, M., Andreassen, O.A., et al., 2013. Childhood trauma is associated withsevere clinical characteristics of bipolar disorders. J. Clin. Psychiatry 74 (10),991–998.

Evans, S.J., Prossin, A.R., Harrington, G.J., et al., 2012. Fats and factors: lipid profilesassociate with personality factors and suicidal history in bipolar subjects. PLoSOne 7 (1), e29297.

Fagiolini, A., Frank, E., Houck, P.R., et al., 2002. Prevalence of obesity and weightchange during treatment in patients with bipolar I disorder. J. Clin. Psychiatry63, 528–533.

Fagiolini, A., Kupfer, D.J., Rucci, P., et al., 2004. Suicide attempts and ideation inpatients with bipolar I disorder. J. Clin. Psychiatry 65, 509–514.

Fagiolini, A., Frank, E., Scott, J.A., et al., 2005. Metabolic syndrome in bipolardisorder: findings from the Bipolar Disorder Center for Pennsylvanians. BipolarDisord. 7, 424–430.

Fattal, O., Budur, K., Vaughan, A.J., et al., 2006. Review of the literature on majormental disorders in adult patients with mitochondrial diseases. Psychosomatics47, 1–7.

Fattal, O., Link, J., Quinn, K., et al., 2007. Psychiatric comorbidity in 36 adults withmitochondrial cytopathies. CNS Spectr. 12, 429–438.

Finseth, P.I., Morken, G., Andreassen, O.A., et al., 2012. Risk factors related to lifetimesuicide attempts in acutely admitted bipolar disorder inpatients. Bipolar Disord.14 (7), 727–734.

Fountoulakis, K.N., Grunze, H., Panagiotidis, P., et al., 2008. Treatment of bipolardepression: an update. J. Affect. Disord. 109, 21–34.

Fountoulakis, K.N., Pantoula, E., Siamouli, M., et al., 2012. Development of the RiskAssessment Suicidality Scale (RASS): a population-based study. J. Affect. Disord.138 (3), 449–457.

Freeman, A.J., Youngstrom, E.A., Michalak, E., et al., 2009. Quality of life in pediatricbipolar disorder. Pediatrics 123, e446–e452.

Galfalvy, H., Huang, Y.Y., Oquendo, M.A., et al., 2009. Increased risk of suicideattempt in mood disorders and TPH1 genotype. J. Affect. Disord. 115, 331–338.

Galfavy, H., Oquendo, M.A., Carballo, J.J., et al., 2006. Clinical predictors of suicidalacts after major depression in bipolar disorder: a prospective study. BipolarDisord. 8 (5 Pt 2), 586–595.

Gau, S.S., Cheng, A.T., 2004. Mental illness and accidental death. Case-controlpsychological autopsy study. Br. J. Psychiatry 185, 422–428.

Ghaemi, S.N., Hsu, D.J., Soldani, F., et al., 2003. Antidepressants in bipolar disorder:the case for caution. Bipolar Disord. 5, 421–433.

Gilbert, A.M., Garno, J.L., Braga, R.J., et al., 2011. Clinical and cognitive correlates ofsuicide attempts in bipolar disorder: Is suicide predictable? J. Clin. Psychiatry72 (8), 1027–1033.

Godwin, C.D., 1984. The dexamethasone suppression test in acute mania. J. Affect.Disord. 7, 281–286.

Goldstein, T.R., Ha, W., Axelson, D.A., et al., 2012. Predictors of prospectivelyexamined suicide attempts among youth with bipolar disorder. Arch. Gen.Psychiatry 69 (11), 1113–1122.

Gomes, F.A., Kauer-Sant‘Anna, M., Magalhães, P.V., et al., 2010. Obesity is associatedwith previous suicide attempts in bipolar disorder. Acta Neuropsychiatr. 22 (2),63–67.

Goodwin, F.K., Jamison, K.R., 2007. Manic-Depressive Illness: Bipolar Disorders andRecurrent Depression. Oxford University Press, New York, NY.

Gould, M.S., Fisher, P., Parides, M., et al., 1996. Psychosocial risk factors of child andadolescent completed suicide. Arch. Gen. Psychiatry 53 (12), 1155–1162.

Gray, S.M., Otto, M.W., 2001. Psychosocial approaches to suicide prevention:applications to patients with bipolar disorder. J. Clin. Psychiatry 62 (Suppl.25), S56–S64.

Harris, E.C., Barraclough, B., 1997. Suicide as an outcome for mental disorders. Ametaanalysis. Br. J. Psychiatry 170, 205–228.

Harris, E.C., Barraclough, B., 1998. Excess mortality of mental disorder. Br. J.Psychiatry 173, 11–53.

Hasin, D.S., Goodwin, R.D., Stinson, F.S., et al., 2005. Epidemiology of majordepressive disorder: results from the National Epidemiologic Survey onAlcoholism and Related Conditions. Arch. Gen. Psychiatry 62, 1097–1106.

Hawton, K., Sutton, L., Haw, C., et al., 2005. Suicide and attempted suicide in bipolardisorder: a systematic review of risk factors. J. Clin. Psychiatry 66, 693–704.

Heuser, I., Yassouridis, A., Holsboer, F., 1994. The combined dexamethasone/CRHtest: a refined laboratory test for psychiatric disorders. J. Psychiatr. Res. 28(341–356), 138.

Hibbeln, J.R., Nieminen, L.R., Blasbalg, T.L., et al., 2006. Healthy intakes of n-3 and n-6 fatty acids: estimations considering worldwide diversity. Am. J. Clin. Nutr. 83,1483S–1493S.

Hiroeh, U., Appleby, L., Mortensen, P.B., et al., 2001. Death by homicide, suicide, andother unnatural causes in people with mental illness: a population- basedstudy. Lancet 358, 2110–2112.

Huber, R.S., Coon, H., Kim, N., et al., 2014. Altitude is a risk factor for completedsuicide in bipolar disorder. Med. Hypotheses 82 (3), 377–381.

Isometsa, E.T., Henriksson, M.M., Aro, H.M., et al., 1994. Suicide in bipolar disorderin Finland. Am J Psychiatry. 151 (7), 1020–1024.

Jamison, K.R., 2000. Suicide and bipolar disorder. J. Clin. Psychiatry 61 (Supp l9),S47–S51.

Jiménez, E., Arias, B., Mitjans, M., et al., 2013. Genetic variability at IMPA2, INPP1and GSK3b increases the risk of suicidal behavior in bipolar patients. Eur.Neuropsychopharmacol. 23 (11), 1452–1462.

Jokinen, J., Nordström, P., 2008. HPA axis hyperactivity as suicide predictor inelderly mood disorder inpatients. Psychoneuroendocrinology 33, 1387.

Jokinen, J., Nordström, P., 2009. HPA axis hyperactivity and attempted suicide inyoung adult mood disorder inpatients. J. Affective Disord. 2009 (116), 117.

Jokinen, J., Nordström, A., Nordström, P., 2009. CSF 5-HIAA and DST nonsuppres-sion–orthogonal biologic risk factors for suicide in male mood disorderinpatients. Psychiat. Res. 165, 96.

Jou, S.H., Chiu, N.Y., Liu, C.S., 2009. Mitochondrial dysfunction and psychiatricdisorders. Chang Gung. Med. J. 32, 370–379.

Joukamaa, M., Heliövaara, M., Knekt, P., et al., 2001. Mental disorders and cause-specific mortality. Br. J. Psychiatry 179, 498–502.

Kamali, M., Saunders, E.F.H., Prossin, A.R., et al., 2012. Associations between suicideattempts and elevated bedtime salivary cortisol levels in bipolar disorder. J.Affect. Disord. 136 (3), 350–358.

Karege, F., Vaudan, G., Schwald, M., et al., 2005. Neurotrophin levels in postmortembrains of suicide victims and the effects of antemortem diagnosis andpsychotropic drugs. Mol. Brain Res. 136 (1–2), 29–37.

Kato, T., 2006. The role of mitochondrial dysfunction in bipolar disorder. Drug NewsPerspect. 19, 597–602.

Kenneson, A., Funderburk, J.S., Maisto, S.A., 2013. Risk factors for secondarysubstance use disorders in people with childhood and adolescent-onset bipolardisorder: opportunities for prevention. Compr. Psychiatry 54 (5), 439–446.

Kerner, B., Rao, A.R., Christensen, B., et al., 2013. Rare genomic variants link bipolardisorder with anxiety disorders to CREB-regulated intracellular signalingpathways. Front Psychiat. 4, 154.

Kheirabadi, G.R., Hashemi, S.J., Akbaripour, S., et al., 2012. Risk factors of suicidereattempt in patients admitted to khorshid hospital, Isfahan, Iran, 2009. Iran. J.Epidemiology. 8 (3), 39–46.

Koene, S., Kozicz, T.L., Rodenburg, R.J., et al., 2009. Major depression in adolescentchildren consecutively diagnosed with mitochondrial disorder. J. Affect. Disord.114, 327–332.

Kuo, W.H., Gallo, J.J., Tien, A.Y., 2001. Incidence of suicide ideation and attempts inadults: the 13-year follow-up of a community sample in Baltimore, Maryland.Psychol. Med. 31 (7), 1181–1191.

Kupfer, D.J., 2005. The increasing medical burden in bipolar disorder. J. Am. Med.Assoc. 293, 2528–2530.

Leon, A.C., Solomon, D.A., Li, C., et al., 2012. Antiepileptic drugs for bipolar disorderand the risk of suicidal behavior: a 30-year observational study. Am. J.Psychiatry 169 (3), 285–291.

Leverich, G.S., et al., 2003. Factors associated with suicide attempts in 648 patientswith bipolar disorder in the Stanley Foundation Bipolar Network. J. Clin.Psychiatry 64 (5), 506–515.

Lindqvist, D., Isaksson, A., Trskman-Bendz, L., et al., 2008. Salivary cortisol andsuicidal behavior—a follow-up study. Psychoneuroendocrinology 33, 1061.

Linkowski, P., Kerkhofs, M., Van Onderbergen, A., et al., 1994. The 24-hour profilesof cortisol, prolactin, and growth hormone secretion in mania. Arch. Gen.Psychiatry 51, 616–624.

Lopez, D.L., Brezo, J., Rouleau, G., et al., 2007. Effect of tryptophan hydroxylase-2gene variants on suicide risk in major depression. Biol. Psychiatry 62, 72–80.

Lopez, P., Mosquera, F., deLeon, J., et al., 2001. Suicide attempts in bipolar patients. J.Clin. Psychiatry 62, 963–966.

MacKinnon, D.F., Zandi, P.P., Gershon, E., et al., 2003. Rapid switching of mood infamilies with multiple cases of bipolar disorder. Arch. Gen. Psychiatry 60,921–928.

Magno, L.A., Miranda, D.M., Neves, F.S., et al., 2010. Association between AKT1 butnot AKTIP genetic variants and increased risk for suicidal behavior in bipolarpatients. Genes Brain Behav. 9, 411–418.

Magno, L.A.V., Santana, C.V.N., Rezende, V.B., et al., 2011. Genetic variations inFOXO3A are associated with Bipolar Disorder without confering vulnerabilityfor suicidal behavior. J. Affect. Disord. 133 (3), 633–637.

Manchia, M., Hajek, T., O‘Donovan, C., et al., 2013. Genetic risk of suicidal behaviorin bipolar spectrum disorder: analysis of 737 pedigrees. Bipolar Disord. 15 (5),496–506.

L.d.S. Costa et al. / Journal of Affective Disorders 170 (2015) 237–254252

Mann, J.J., 2003. Neurobiology of suicidal behavior. Nat. Rev. Neurosci. 4, 819.Mao, Z., Liu, L., Zhang, R., et al., 2007. Lithium reduces FoxO3a transcriptional

activity by decreasing its intracellular content. Biol. Psychiatry 62, 1423–1430.Maremmani, I., Pani, P.P., Canoniero, S., et al., 2007. Is the bipolar spectrum the

psychopathological substrate of suicidality in heroin addicts? Psychopathology40 (5), 269–277.

McElroy, S.L., Kotwal, R., Kaneria, R., et al., 2006. Antidepressants and suicidalbehavior in bipolar disorder. Bipolar Disord. 8, 596–617.

McIntyre, R.S., Konarski, J.Z., 2004. Bipolar disorder: a national health concern. CNSSpectr. 9 (11 Suppl 12), S6–S15.

McIntyre, R.S., Soczynska, J.K., Konarski, J.Z., et al., 2007. Should depressivesyndromes be reclassified as "metabolic syndrome type II"? Ann. Clin. Psy-chiatry 19, 257–264.

McIntyre, R.S., Soczynska, J.K., Mancini, D., et al., 2008. The relationship betweenchildhood abuse and suicidality in adult bipolar disorder. Violence Vict. 23 (3),361–372.

McIntyre, R.S., Muzina, D.J., Kemp, D.E., et al., 2008. Bipolar disorder and suicide:researchsynthesis and clinical translation. Curr. Psychiatry Rep. 10, 66–72.

Michalak, E.E., Yatham, L.N., Lam, R.W., 2005. Quality of life in bipolar disorder: areview of the literature. Health Qual. Life Outcomes 15, 3–72.

Miklowitz, D.J., Chang, K.D., 2008. Prevention of bipolar disorder in at-risk children:theoretical assumptions and empirical foundations. Dev. Psychopathol. 20,881–897.

Morava, E., Gardeitchik, T., Kozicz, T., et al., 2010. Depressive behavior in childrendiagnosed with a mitochondrial disorder. Mitochondrion 10, 528–533.

Mykletun, A., Bjerkeset, O., Dewey, M., et al., 2007. Anxiety, depression, and cause-specific mortality: the HUNT study. Psychosom. Med. 69, 323–331.

Neves, F.S., Malloy-Diniz, L.F., Ma, Romano-Silva, et al., 2010. Is the serotonintransporter polymorphism (5-HTTLPR) a potential marker for suicidal behaviorin bipolar disorder patients? J. Affect. Disord. 125 (1-3), 98–102.

Neves, F.S., Malloy-Diniz, L., Romano-Silva, M.A., et al., 2011. The role of BDNFgenetic polymorphisms in bipolar disorder with psychiatric comorbidities.J. Affect. Disord. 131, 307–311.

Nilsson, L., Ahlbom, A., Farahmand, B.Y., et al., 2002. Risk factors in suicide inepilepsy: a case-control study. Epilepsia 43, 644–651.

Oquendo, M.A., Currier, D., Liu, S., et al., 2010. Increased risk for suicidal behavior incomorbid bipolar disorder and alcohol use disorders. J. Clin. Psychiatry 71 (7),902–909.

Ösby, U., Brandt, L., Correia, N., et al., 2001. Excess mortality in bipolar and unipolardisorder in Sweden. Arch. Gen. Psychiatry 58, 844–850.

Osman, A., Bagge, C.L., Gutierrez, P.M., et al., 2001. The Suicidal BehaviorsQuestionnaire-Revised (SBQ-R): validation with clinical and nonclinical sam-ples. Assessment 8 (4), 443–454.

Parmentier, C., Etain, B., Yon, L., et al., 2012. Clinical and dimensional characteristicsof euthymic bipolar patients with or without suicidal behavior. Eur. Psychiatry27 (8), 570–576.

Pasco, J., Jacka, F., Williams, L.J., et al., 2007. Leptin in depressed women: cross-sectional and longitudinal data from an epidemiologic study. J. Affect. Disord.107, 211–225.

Pawlak, J., Dmitrzak-Weglarz, M., Skibińska, M., et al., 2013. Suicide attempts andpsychological risk factors in patients with bipolar and unipolar affectivedisorder. Gen. Hosp. Psychiatry 35 (3), 309–313.

Pawlak, J., Dmitrzak-Weglarz, M., Skibińska, M., et al., 2013. Suicide attempts andclinical risk factors in patients with bipolar and unipolar affective disorders.Gen. Hosp. Psychiatry 35 (4), 427–432.

Periodicos.capes.gov.br [homepage on the Internet]. 2014. Brasília: Higher Educa-tion Co-ordination Agency of Brazil‘s Ministry of Education; 2000. Availablefrom: ⟨http://www.periodicos.capes.gov.br/⟩. (accessed 6.6.2014.

Perugi, G., Micheli, C., Akiskal, H.S., et al., 2000. Polarity of the first episode, clinicalcharacteristics, and course of manic depressive illness: a systematic retro-spective investigation of 320 bipolar I patients. Compr Psychiatry. 41, 13–18.

Pfennig, A., Kunzel, H., Kern, N., et al., 2005. Hypothalamus-pituitary-adrenalsystemregulation and suicidal behavior in depression. Biol. Psychiatry 57,336–342.

Pitchot, W., Scantamburlo, G., Pinto, E., et al., 2008. Vasopressin–neurophysin andDST in major depression: relationship with suicidal behavior. J. Psychiatr. Res.42, 684–688.

Polter, A., Yang, S., Zmijewska, A.A., et al., 2009. Forkhead box, class O transcriptionfactors in brain: regulation and behavioral manifestation. Biol. Psychiatry 65,150–159.

Pompili, M., Rihmer, Z., Innamorati, M., et al., 2009. Assessment and treatment ofsuicide risk in bipolar disorders. Expert Rev. Neurother. 9, 109–136.

Pompili, M., Rihmer, Z., Akiskal, H., et al., 2012. Temperaments mediate suicide riskand psychopathology among patients with bipolar disorders. Compr. Psychiatry53 (3), 280–285.

Pompili, M., Rihmer, Z., Akiskal, H., et al., 2012. Temperaments mediate suicide riskand psychopathology among patients with bipolar disorders. Compr. Psychiatry53 (3), 280–285.

Potash, J.B., Kane, H.S., Chiu, Y.-F., et al., 2000. Attempted suicide and alcoholism inbipolar disorder: clinical and familial relationships. Am. J. Psychiatry 157,2048–2050.

Quiroz, J.A., Gould, T.D., Manji, H.K., et al., 2004. Molecular effects of lithium. Mol.Interv. 4, 259–272.

Quiroz, J.A., Gray, N.A., Kato, T., et al., 2008. Mitochondrially mediated plasticity inthe pathophysiology and treatment of bipolar disorder. Neuropsychopharma-cology 33, 2551–2565.

Raja, M., Azzoni, A., 2004. Suicide attempts: differences between unipolar andbipolar patients and among groups with different lethality risk. J. Affect. Disord.82 (3), 437–442.

Ratcliffe, G.E., Enns, M.W., Belik, S.L., et al., 2008. Chronic pain conditions andsuicidal ideation and suicide attempts: an epidemiologic perspective. Clin. J.Pain 24, 204–210.

Rezin, G.T., Amboni, G., Zugno, A.I., et al., 2009. Mitochondrial dysfunction andpsychiatric disorders. Neurochem. Res. 34, 1021–1029.

Rihmer, Z., Pestality, P., 1999. Bipolar II disorder and suicidal behavior. Psychiatr.Clin. N. Am. 22 (667–73), ix–x.

Rihmer, Z., Gonda, X., Torzsa, P., et al., 2013. Affective temperament, history ofsuicide attempt and family history of suicide in general practice patients.J. Affect. Disord. 149 (1–3), 350–354.

Roshanaei-Moghaddam, B., Katon, W., 2009. Premature mortality from generalmedical illnesses among persons with bipolar disorder: a review. Psychiatr.Serv. 60, 147–156.

Roy, A., Segal, N.L., 2001. Suicidal behavior in twins: a replication. J. Affect. Disord.66, 71–74.

Ruengorn, C., Sanichwankul, K., Niwatananun, W., et al., 2012. A risk-scoringscheme for suicide attempts among patients with bipolar disorder in a Thaipatient cohort. Psychol. Res. Behav. Manage. 5, 37–45.

Rybakowski, J.K., Twardowska, K., 1999. The dexamethasone/corticotropinreleasinghormone test in depression in bipolar and unipolar affective illness. J. Psychiatr.Res. 33, 363–370.

Ryu, V., Jon, D.I., Cho, H.S., et al., 2010. Initial depressive episodes affect the risk ofsuicide attempts in Korean patients with bipolar disorder. Yonsei. Med. J. 51 (5),641–647.

Sajatovic, M., 2005. Bipolar disorder: disease burden. Am. J. Manag. Care 11,S80–S84.

Scaglia, F., 2010. The role of mitochondrial dysfunction in psychiatric disease. Dev.Disabil. Res. Rev. 16, 136–143.

Schmider, J., Lammers, C.H., Gotthardt, U., et al., 1995. Combined dexamethasone/corticotropin-releasing hormone test in acute and remitted manic patients, inacute depression, and in normal controls: I. Biol. Psychiatry 38, 797–802.

Sears, C., Wilson, J., Fitches, A., et al., 2013. Investigating the role of BDNF and CCKsystem genes in suicidality in a familial bipolar cohort. J. Affect. Disord. 151 (2),611–617.

Serretti, A., Drago, A., De, R.D., et al., 2009. Lithium pharmacodynamics andpharmacogenetics: focus on inositol mono phosphatase (IMPase), inositolpoliphosphatase (IPPase) and glycogen sinthase kinase 3 beta (GSK-3 beta).Curr. Med. Chem. 16, 1917–1948.

Shabani, A., Teimurinejad, S., Koka, S., et al., 2013. Suicide risk factors in iranianpatients with bipolar disorder: A 21- month follow-up from BDPF study. Iran. J.Psychiatry Behav. Sci. 7 (1), 16–23.

Simon, G.E., Bauer, M.S., Ludman, E.J., et al., 2007. Mood symptoms, functionalimpairment and disability in people with bipolar disorder: specific effects ofmania and depression. J. Clin. Psychiatry 68, 1237–1245.

Simon, G.E., Hunkeler, E., Fireman, B., et al., 2007. Risk of suicide attempt andsuicide death in patients treated for bipolar disorder. Bipolar Disord. 9,526–530.

Sisask, M., et al., 2010. Is religiosity a protective factor against attempted suicide: acrosscultural case–control study. Arch. Suicide Res. 14 (1), 44–55.

Song, J.Y., Hy, Y.u., Kim, S.H., et al., 2012. Assessment of risk factors related to suicideattempts in patients with bipolar disorder. J. Nerv. Ment. Dis. 200 (11), 978–984.

Stahl, S.M., 2000. Essential psychopharmacology of depression and bipolar dis-order. Cambridge University Press, NewYork, NY.

Stewart, C.D., Quinn, A., Plever, S., et al., 2009. Comparing cognitive behaviortherapy, problem solving therapy, and treatment as usual in a high riskpopulation. Suicide Life Threat Behav. 39 (5), 538–547.

Sublette, M., Carballo, J.J., Moreno, C., et al., 2009. Substance use disorders andsuicide attempts in bipolar subtypes. J. Psychiatry Res. 43 (3), 230–238.

Suicidology AAo. 2000. American Association of Suicidology [AAS]. Offical 1998statistics.

Suttajit, S., Paholpak, S., Choovanicvong, S., et al., 2013. Correlates of current suiciderisk among Thai patients with bipolar I disorder: findings from the Thai BipolarDisorder Registry. Neuropsychiatry Dis. Treat. 9, 1751–1757.

Tellenbach, 1974. Melancholie; Problemgeschichte; Endogenit ̈ at, Typologie, Patho-genese, Klinik. Springer, Berlin.

Tondo, L., Albert, M.J., Baldessarini, R.J., 2006. Suicide rates in relation to health careaccess in the United States: an ecological study. J. Clin. Psychiatry 67 (4),517–523.

Tripodianakis, J., Markianos, M., Sarantidis, D., et al., 2000. Neurochemical variablesin subjects with adjustment disorder after suicide attempts. Eur. Psychiatry 15,190–195.

Trouvin, J.H., Prioux-Guyonneau, M., Cohen, Y., et al., 1986. Rat brain monoaminemetabolism and hypobaric hypoxia: a new approach. Gen. Pharmacol. 17,69–73.

Tsai, S.Y., Kuo, C.J., Chen, C.C., et al., 2002. Risk factors for completed suicide inbipolar disorder. J. Clin. Psychiatry 63, 469–476.

Undurraga, J., Baldessarini, R.J., Valentı´, M., et al., 2011. Dissimilar suicidal riskfactors in bipolar I and II disorders. J. Clin. Psychiatry.

Undurraga, J., Baldessarini, R.J., Valenti, M., et al., 2012. Suicidal risk factors inbipolar I and II disorder patients. J. Clin. Psychiatry 73 (6), 778–782.

Vaccari, A., Brotman, S., Cimino, J., et al., 1978. Adaptive changes induced by highaltitude in the development of brain monoamine enzymes. Neurochem. Res. 3,295–311.

L.d.S. Costa et al. / Journal of Affective Disorders 170 (2015) 237–254 253

Valtonen, H.M., Suominen, K., Mantere, O., et al., 2006. Prospective study of riskfactors for attempted suicide among patients with bipolar disorder. BipolarDisord. 8 (5 Pt 2), 576–585.

WHO, 2011. Causes of Death 2008: Data Sources and Methods Department ofHealth Statistics and Informatics. World Health Organization, Geneva.

Wang, P.W., Sachs, G.S., Zarate, C.A., et al., 2006. Overweight and obesity in bipolardisorders. J. Psychiatr. Res. 40, 762–764.

Weinstock, L.M., Miller, I.W., 2008. Functional impairment as a predictor of short-term symptom course in bipolar I disorder. Bipolar Disord. 10, 437–442.

Weissman, M.M., Bland, R.C., Canino, G., et al., 1999. Prevalence of suicide ideationand suicide attempts in nine countries. Psychol. Med. 29, 9–17.

World Health Organization (WHO). International suicide rates. 2003.

Wulsin, L.R., Vaillant, G.E., Wells, V.E., 1999. A systematic review of the mortality ofdepression. Psychosom. Med. 61, 6–17.

Yerevanian, B., Feusner, J., Koek, R., et al., 2004. The dexamethasone suppressiontest as a predictor of suicidal behavior in unipolar depression. J. Affect. Disord.83, 103.

Yoon, Y.H., Chen, C.M., Moss, H.B., 2011. Effect of comorbid alcohol and drug usedisorders on premature death among unipolar and bipolar disorder decedentsin the United States, 1999–2006. Compr. Psychiatry 52 (5), 453–464.

Zhu, W., Bijur, G.N., Styles, N.A., et al., 2004. Regulation of FOXO3a by brainderivedneurotrophic factor in differentiated human SH SY5Y neuroblastoma cells.Brain Res. Mol. Brain Res. 126, 45–56.

L.d.S. Costa et al. / Journal of Affective Disorders 170 (2015) 237–254254