Onset polarity in bipolar disorder: A strong association between first depressive episode and suicide attempts

Author’s Accepted Manuscript Onset polarity in Bipolar Disorder: A strong association between first depressive episode and suicide attempts Laura Cremaschi, Bernardo Dell’Osso, Matteo Vismara, Cristina Dobrea, Massimiliano Buoli, Terence A. Ketter, A. Carlo Altamura www.elsevier.com/locate/jad

PII: DOI: Reference:

S0165-0327(16)31438-0 http://dx.doi.org/10.1016/j.jad.2016.11.043 JAD8658

To appear in: Journal of Affective Disorders Received date: 19 August 2016 Revised date: 22 October 2016 Accepted date: 28 November 2016 Cite this article as: Laura Cremaschi, Bernardo Dell’Osso, Matteo Vismara, Cristina Dobrea, Massimiliano Buoli, Terence A. Ketter and A. Carlo Altamura, Onset polarity in Bipolar Disorder: A strong association between first depressive episode and suicide attempts, Journal of Affective Disorders, http://dx.doi.org/10.1016/j.jad.2016.11.043 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting galley proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Onset polarity in Bipolar Disorder: a strong association between first depressive episode and suicide attempts Laura Cremaschi1, Bernardo Dell’Osso1,2*, Matteo Vismara1, Cristina Dobrea1, Massimiliano Buoli1, Terence A. Ketter2, A. Carlo Altamura1 1

Department of Psychiatry, University of Milan, Fondazione IRCCS Ca’ Granda, Ospedale Maggiore Policlinico,

Milan, Italy. 2

Department of Psychiatry and Behavioral Sciences, Bipolar Disorders Clinic, Stanford University, CA, USA.

*

Corresponding author. MD. Assistant Prof. of Psychiatry, Department of Psychiatry, University of Milan;Fondazione IRCCS Ca’ Granda,Ospedale Maggiore Policlinico, Milano. Via Francesco Sforza 35, 20122 Milano, Italy. Tel.: 02 55035994; fax: 02 55033140 [email protected]

Abstract

Background. The role of onset polarity (OP) in patients with bipolar disorder (BD) has been increasingly investigated over the last few years, for its clinical, prognostic, and therapeutic implications. The present study sought to assess whether OP was associated with specific correlates, in particular with a differential suicidal risk in BD patients. Methods. A sample of 362 recovered BD patients was dichotomized by OP: depressed (DO) or elevated onset (EO: hypomanic/manic/mixed). Socio-demographic and clinical variables were compared between the subgroups. Additionally, binary logistic regression was performed to assess features associated with OP. Results.

DO compared with EO patients had older current age and were more often female, but less often single and unemployed. Clinically, DO versus EO had a more than doubled rate of suicide attempts, as well as significantly higher rates of BD II diagnosis, lifetime stressful events, current psychotropics and antidepressants use, longer duration of the most recent episode (more often depressive), but lower rates of psychosis and involuntary commitments. Limitations. Retrospective design limiting the accurate assessment of total number of prior episodes of each polarity. Conclusions. Our results support the influence of OP on BD course and outcome. Moreover, in light of the relationship between DO and a higher rate of suicide attempts, further investigation may help clinicians in identifying patients at higher risk of suicide attempts.

Keywords: Bipolar Disorder (BD), onset polarity (OP), suicide attempt.

INTRODUCTION

Bipolar disorder (BD) is responsible for a chronic burden and a significantly increased suicide risk, compared with the general population (Simon et al., 2007). Recently, renewed interest has emerged on the role of predominant polarity over the course of BD and use of polarity index as a long-term course specifier, with meaningful therapeutic and prognostic implications (Carvalho et al., 2014). Moreover, in a recent review, depressive predominant polarity has been associated with depressive onset (DO), delayed BD diagnosis, diagnosis of BD II, and a higher rate of suicidal behaviors, whereas manic predominant polarity has been related to earlier onset, first manic/psychotic episode, and an increased rate of substance abuse (Carvalho et al, 2014). 1

However, to date, only limited evidence is available in relation to DO versus elevated (EO) onset polarity (OP) and their biological, socio-demographic, and clinical correlates (Azorin et al., 2011). For instance, OP appears to cluster in families (Kassem et al., 2006) and has been increasingly investigated for its potential role over BD course, outcome, prognosis, and impact on clinical and therapeutic decision-making (Daban et al., 2006). Furthermore, like for other clinical variables addressed in previous studies from our group (Dell’Osso et al., 2016a; 2016b), the assessment of OP may contribute to distinguish more homogeneous subgroups of bipolar patients (Daban et al., 2006), leading to more targeted interventions (Forty et al., 2009). Indeed, several studies over the last decades have supported the association between OP and selected clinical variables, such as long-term predominant polarity (Turvey et al., 1999; Perugi et al., 2000; Perlis et al., 2005; Kassem et al., 2006; Vieta et al., 2009), polarity at relapse (Calabrese et al., 2004), rapid cycling (Perugi et al., 2000), number of episodes and suicide attempts (Perlis et al., 2005), comorbidity with other psychiatric disorders (e.g., anxiety) (Azorin et al., 2011), lifetime psychotic symptoms (Daban et al., 2006), chronic severity of illness, and treatment response (Haag et al., 1987; Maj et al., 1989). Importantly, several studies reported that suicide risk was robustly associated with DO (Perris et al., 1966; Roy-Byrne et al., 1985; Perugi et al., 2000; Perlis et al., 2005). This has very substantial clinical relevance, since approximately 30% of individuals with BD attempt suicide in their lifespan (Chen et al., 1996), most often during depressive episodes (Goldberg et al., 2002) and dysphoric mania (Rihmer et al., 2002). Prior research on OP has been frequently limited by being restricted only to BDI patients (Forty et al., 2000; Perlis et al., 2005; Kassem et al., 2006; Azorin et al., 2011), focusing on the impact of first manic or psychotic episodes, mainly conducted with inpatients (who likely over-represent

2

the diagnosis of mania) (Sipos et al., 2001; Tohen et al., 2003), and not including onset with mixed features (Forty et al., 2000; Perlis et al., 2005; Daban et al., 2006; Kassem et al., 2006). Thus, the present study aimed to assess demographic and clinical associates of OP in a large sample of recovered Italian bipolar I and II patients. The inclusion of BDII patients in our study sample was intended to make it more representative of the real-world population of bipolar patients, providing new insight in the field. We hypothesized that OP might be significantly related with specific socio-demographic and clinical correlates, in particular that DO might be associated with an increased risk of suicidal behavior.

METHODS

Subjects The study sample consisted of 362 recovered BD patients, primarily referred to the University Department of Mental Health at the Fondazione IRCCS Ca’ Granda, Ospedale Maggiore Policlinico in Milan, Italy. We also included patients who sought evaluation and treatment at outpatient and community-based psychiatric services, in order to obtain a more representative BD phenomenology of the Northern Italian population. All subjects provided written informed consent to have their clinical charts and medical records used for research purpose.

Assessment Participants were assessed by psychiatrists or residents in psychiatry with specific training in mood disorders, with the Structured Clinical Interviews for Diagnostic and Statistical Manual of Mental Disorders, 4th Edition, Text Revision (DSM-IV-TR, APA 2000) (SCID I and II - First, 1997; 2002), in order to ascertain Axis I/II diagnoses of mood disorders and comorbid psychiatric conditions. To enhance diagnostic specificity, individuals with BD Not Otherwise 3

Specified were not included in the study sample, so that all participants had BD I or BD II. If patients had a comorbid psychiatric disorder, BD had to be the disorder primarily affecting their everyday functioning, quality of life, and help seeking. Individuals with mental retardation, neurological disorders, mental illnesses associated with an organic basis, or other disabling medical conditions were excluded. Data was gathered from patients and their relatives, and derived from available clinical records. Socio-demographic variables included: age, gender, education, current employment, and lifetime occupational functioning impairment, co-habitation, and marital status. Clinical variables included: diagnosis, age at onset (i.e., age of first mood episode of any polarity), duration of BD, duration of untreated BD illness (DUI), OP, most recent episode duration and polarity, lifetime number of psychiatric hospitalizations, involuntary commitments, suicide attempts, family history of psychiatric disorders, lifetime occurrence of psychosis, current subthreshold symptoms, history of stressful life events (e.g., catastrophes, serious accidents or illnesses, death of

a

loved

one,

significant

relationship

breakup,

financial

problems,

pregnancy),

psychiatric/medical comorbidity, and psychosocial rehabilitation, which includes communitybased interventions to improve social/working skills, reduce functional disability, and improve quality of life (Barbato, 2006). Moreover, the Global Assessment of Functioning (GAF) (Hall, 1995) was administered after the resolution of the most recent syndromal mood episode (to exclude potential mood episode-related bias), in order to evaluate patient’s current level of global functioning. Current pharmacological treatment was recorded, focusing in particular on the use of antidepressants, mood stabilizers, and antipsychotics, in mono- and poly-therapy. All patients were grouped according to OP, defined as DO or EO (i.e., hypomanic/manic/mixed first episode). Only subjects able to specifically provide the above OP information were included in our analyses.

4

Statistical analyses For continuous parameters, the DO and EO subgroups were compared using corrected Multivariate Analysis of Covariance (MANCOVA), with polarity of first episode as the independent variable and current age as covariate (to exclude its potential influence when analyzing other clinical parameters). The MANCOVA model proved to be valid (Wilks’ lambda test, p<0.001). For categorical parameters, the DO and EO subgroups were compared using Chisquare tests, with Bonferroni post-hoc analysis. A two-tailed significance threshold was set at p<0.05. Furthermore, to assess potential confounders, binary logistic regressions were performed to investigate significant relationships between OP (dependent variable) and other variables (included as covariates). Statistical analyses were performed using Statistical Package for the Social Sciences (SPSS), version 22.

RESULTS

Socio-demographic and clinical characteristics Table 1 and Table 2 show, respectively, socio-demographic and clinical data of the entire sample and the DO and EO subgroups. Overall 46/362 patients (13%) were excluded from our original sample due to inadequate information about OP. DO compared with EO was more common (binomial test, p<0.05). With respect to socio-demographic variables, patients with DO compared with EO had significantly older current age (F=10.8, p=0.001) and were significantly more often female (χ2=7.4, df=1, p=0.008), but less often single (χ2=8.4, df=3, p=0.04) and unemployed (χ2=5.9, df=2, p=0.05). With respect to clinical variables, BD I was the most common diagnosis in both DO and EO (62.6% and 86.6%). However, patients with DO vs EO had significantly higher rates of BD II 5

(χ2=22.3, df=1, p<0.001), lifetime suicide attempts (χ2=4.7, df=1, p=0.03), lifetime stressful events (χ2=5.5, df=1, p=0.02), longer duration of the most recent episode (F=5, p=0.03) - more often depressive (χ2=6.8, df=1, p=0.01) - as well as more current psychotropic drugs (F=6.4, p=0.01) and antidepressants use (χ2=13.6, df=1, p<0.001). In contrast, DO vs EO had lower rates of lifetime psychotic symptoms (χ2=30, df=1, p<0.001) and lifetime involuntary commitments (χ2=10.3, df=1, p=0.001) as well as less lifetime involuntary commitments (F=10.5, p=0.001). Patients with DO vs EO had no significant difference for any other demographic or clinical parameter in Table 1.

Logistic regression analysis/Correlates of OP The variables found to be significant after MANCOVAs were entered into binary logistic regression, in order to assess which features were best associated with the OP. These included: age, gender, employment, marital status, diagnosis, polarity and duration of the most recent episode, involuntary commitments, lifetime psychosis, lifetime suicide attempts, lifetime stressful events, current antidepressant treatment, and number of current psychotropic drugs. In addition, we included GAF among covariates, since according to the literature, the analogous Global Assessment Scale (GAS, Endicott et al., 1976) was found to be significantly associated with OP in BD patients (Forty et al., 2009). There was no evidence of lack of fit in our model (Hosmer & Lemeshow test: χ2=12.46, df=8, p=0.13), and binary logistic regression was significant overall (Omnibus test: χ2=52.57, df=18, p<0.001), predicting 71.4% of cases. DO was associated with a higher probability of suicide attempts [B=0.85, S.E.=0.43, Wald=3.88, df=1, OR=2.35, 95% CI: 1-5.48; p=0.049)] and current use of antidepressants [B=0.95, S.E.=0.47, Wald=4.06, df=1, OR=2.58, 95% CI: 1.03-6.47; p=0.04)]. 6

DISCUSSION

In the present study, we investigated potential differences in socio-demographic and clinical profiles of BD patients, according to their OP, and found DO vs EO significantly associated with older current age and female gender, but with inferior rates of unemployment and single marital status. From a clinical perspective, DO vs EO was associated with a more than twice as high rate of suicide attempts, as well as more BD II and lifetime stressful events, longer duration of the most recent episode (more often depressive), more current psychotropic drugs and antidepressants use. Conversely, individuals with DO showed a lower rate of psychosis and involuntary commitments, compared with EO patients. Our finding of a significantly higher rate of lifetime suicide attempts in DO vs EO is consistent with prior studies, reporting an association of DO with the lifetime history (Quitkin et al., 1986; Perugi et al., 2000; Perlis et al., 2005; Cha et al., 2009; Forty et al., 2009; Azorin et al., 2011; Ryu et al., 2010; Chaudhury et al., 2007; Tundo et al., 2015) and number of suicide attempts (Chaudhury et al., 2007; Neves et al., 2009), as well as with predominant depressive polarity (Colom et al., 2006; Rosa et al., 2008; Baldessarini et al., 2012; Carvalho et al., 2014; Popovic et al., 2014). Of note, our results from logistic regression yielded further support to this perspective, highlighting that DO was associated with more than doubled risk of suicide attempts. This aspect has relevant implications in clinical practice: in fact, in case of a depressive onset, clinicians should be aware of the enhanced risk of suicide attempts over the long-term, thus being more sensitive to the early identification and adequate treatment of high-risk patients. Of note, the increased risk of suicide attempts observed in the present study is lower than the eightfold risk increase previously reported (Chaudhury et al., 2007). Such discrepancy may be explained by the wider size of our sample and the different BD I-II ratio (with a lower percentage of BD II 7

patients) as well as by the inclusion of a larger number of patients with cluster B personality disorder in the study by Chaudhury and colleagues, potentially causing an increased suicide risk. Moreover, socio-cultural and geographic factors (i.e., American vs European bipolar population) may have in turn contributed to such a difference. Consistently with prior studies (Akiskal et al., 1985; Roy-Byrne et al., 1985; Perugi et al., 2000; Judd et al., 2003; Perlis et al., 2005; Daban et al., 2006; Kassem et al., 2006; Chaudhury et al., 2007; Forty et al., 2009; Neves et al., 2009; Tundo et al., 2015), though in contrast with other reports (Angst et al., 1978; Jagadheesan and Sinha et al., 2003; Cha et al., 2009; Azorin et al., 2011), we found DO was more common than EO, accounting for approximately two thirds of our sample. Even though BD course has been commonly characterized by episodes of the same polarity as the index one (Perris et al., 1966; Roy-Byrne et al., 1985; Turvey et al., 1999; Perugi et al., 2000; Perlis et al., 2005; Daban et al., 2006), the lack of adequate data on the number/polarity of prior episodes did not allow us to assess the predominant polarity in DO vs EO. As previously reported (Daban et al., 2006; Azorin et al., 2011), our bipolar patients with DO were older than those with EO, with females being overrepresented (Perlis et al., 2005; Forty et al., 2009; Neves et al., 2009; Azorin et al., 2011; Tundo et al., 2015). Furthermore, our DO vs EO patients were less often unemployed. However, earlier data on employment status in DO vs EO patients yielded different results, indicating either no difference (Daban et al., 2006; Cha et al., 2009) or an opposite trend, with a lower rate of unemployment/retirement in subjects with manic onset and a higher rate in subjects with mixed onset, compared with individuals with DO (Perugi et al., 2000). In relation to marital status, our DO vs EO patients were less often single. Despite most prior studies finding no difference in such regard (Perugi et al., 2000; Chaudhury et al., 2007; Cha et al., 2009; Neves et al., 2009), a more recent report found a lower prevalence of married 8

individuals among subjects with manic onset and observed that being married increased the odds of depressive vs manic/mixed onset (Azorin et al., 2011). In contrast with some earlier reports focusing on BD I patients (Perugi et al., 2000; Perlis et al., 2005; Cha et al., 2009; Forty et al., 2009; Azorin et al., 2011), both bipolar subtypes were included in our sample, although BD II subjects comprised only about one-quarter of the whole sample. Our BD II and BD I patients more frequently experienced DO and EO, respectively, consistently with prior studies involving both subtypes of bipolar subjects (Daban 2006; Chaudhury et al., 2007; Neves et al., 2009; Tundo et al., 2015). In our analysis, age at onset (i.e., age at first mood episode) did not vary according to OP, consistently with several previous reports (Perugi et al., 2005; Cha et al., 2009; Neves et al., 2009; Tundo et al., 2015). However, a few prior studies linked DO to earlier age at onset (Perlis et al., 2005; Kassem et al., 2006; Chaudhury et al., 2007; Forty et al., 2009), whereas other reports linked DO to later age at onset (Daban et al., 2006; Azorin et al., 2011). Moreover, we found no significant difference in duration of BD illness between patients with DO vs EO, in accordance with some prior studies (Neves et al., 2009; Cha et al., 2009), but not with others reporting a longer illness duration in DO vs EO (Perugi et al., 2005; Perlis et al., 2005; Daban et al., 2006; Forty et al., 2009; Azorin et al., 2011; Tundo et al., 2015). Focusing on the most recent episode, a correlation with OP emerged, with most recent episodes being more frequently depressive in DO vs EO. Additionally, the duration of the last episode was longer in our DO compared with EO patients. The presence of limited literature on these specific issues, with only one study reporting a mixed last episode occurring more frequently in patients with a mixed onset (Perugi et al., 2000), complicates the interpretation of our results. However, one previous study reported a strong correlation between OP and number of subsequent episodes of the same polarity (Rosa et al., 2007).

9

In line with prior studies (Perugi et al., 2000; Chaudhury et al., 2007), though in contrast with other reports showing higher hospitalization rates in patients with manic onset (Daban et al., 2006; Cha et al., 2009), we found that number of psychiatric hospitalizations did not differ between patients with DO vs EO. However, we found that the lifetime occurrence of involuntary commitments and their total number was higher in EO vs DO patients. This specific issue has not been addressed to date, in light of OP, and its further investigation may be of clinical relevance. Our finding of patients with DO being less likely to experience lifetime psychotic features is consistent with some studies (Perugi et al., 2000; Daban et al., 2006; Chaudhury et al., 2007; Forty et al., 2009; Neves et al., 2009; Tundo et al., 2015), but at variance with others reporting no difference in lifetime psychotic features, on the basis of the OP (Perlis et al., 2005; Kassem et al., 2006). With respect to pharmacological treatment, our DO vs EO patients more frequently showed current psychotropic drug use, being significantly (2.6 times) more likely to be currently taking an antidepressant, but not a mood stabilizer or an antipsychotic, in line with prior reports of increased rate of previous treatment (Perugi et al., 2000), particularly antidepressants (at least 1), during the last year in DO patients (Azorin et al., 2011). With regard to medical/psychiatric comorbidities, we found no significant difference between patients with DO vs EO. In literature, DO vs mixed onset has been associated, on one hand, with a greater prevalence of lifetime comorbid anxiety disorders - in particular, generalized anxiety disorder, post-traumatic stress disorder (Perlis et al., 2009), panic disorder (Azorin et al., 2011; Tundo et al., 2015), panic attacks (Kassem et al., 2006), and agoraphobia (Azorin et al., 2011) and alcohol abuse/dependence (Kassem et al., 2006). On the other hand, DO vs mixed onset has been linked to a less frequent history of alcohol (Chaudhury et al., 2007; Azorin et al., 2011) or substance use disorders (Azorin et al., 2011; Tundo et al., 2015). In contrast, some authors have

10

reported no difference in psychiatric comorbidities in patients with DO vs EO (Daban et al., 2006; Neves et al., 2009). Although in our study, patients with DO vs EO were comparable in terms of global functioning, a longitudinal assessment of more specific functional variables may be needed to adequately investigate the impact of OP on functional outcome. We may speculate that at least some of our findings of more challenging illness in patients with DO vs EO might be related to our observed higher rate of lifetime stressful events in DO patients, which is consistent with prior reports (Daban et al., 2006; Azorin et al., 2011). Furthermore, robust evidence supports the positive association between early life stress/abuse and suicide attempts in BD (Leverich et al., 2003; Carballo et al., 2008; Benedetti et al., 2011; Maniglio, 2013; Erten et al., 2014; Schaffer et al., 2015). On the other hand, we could hypothesize that the increased rate of involuntary commitments in the EO subgroup might have had a protective effect against suicidal behavior. Since previous suicide attempts are significantly associated with a greater risk of subsequent suicidal behavior and mortality (Tsai et al., 2002; Michaelis et al., 2003; Aizenberg et al., 2006; Valtonen et al., 2006; Guillaume et al., 2010), with no studies addressing the relationship between OP and suicide mortality published yet (Schaffer et al., 2015), such issue needs to be further explored. Some evidence suggests that DO vs EO patients could have some less severe symptoms, given the lower psychosis and hospitalization rates (Perugi et al., 2000; Daban et al., 2006). However, we argue that the significantly higher risk of suicide attempts in patients with DO could tip the balance towards, at least, an equally severe burden of illness, because of its relevant implications, both at a clinical and a pharmacological level. Finally, the last noteworthy finding from the logistic regression showed that DO was associated with a higher probability of current antidepressant use. This result may add further insight to the ongoing debate about the relationship between antidepressant use in BD and depressive 11

recurrence/rapid cycling, with some authors opining that the longitudinal depressive burden of some expressions of BD (BD II, in particular) may cause more antidepressant use (Amsterdam et al., 2015), while others expressing the belief that antidepressants cause more depressive recurrences and rapid cycling (El-Mallakh et al., 2015). In our view, our finding seems to support the former hypothesis. Several strengths and limitations of this study need to be considered. To the authors’ knowledge, our sample is the second largest among such studies including both BD I and BD II patients (Daban et al., 2006; Chaudhury et al., 2007; Neves et al., 2009; Tundo et al., 2015), though smaller than some other investigating OP but including exclusively BD I patients (Forty et al., 2000; Kassem et al., 2000; Perugi et al., 2000; Perlis et al., 2005; Cha et al., 2009; Azorin et al., 2011). Among limitations, the retrospective design of the study may have led to recall bias in collecting clinical data, although this has been partially overcome by performing further verification through relatives’ interviews and clinical records. Of note, although the adoption of DSM-IV criteria may implicate a loss of data regarding hypomanic onsets, frequently under- or misdiagnosed, the revision and confirmation of diagnoses according to DSM-5 (APA, 2013) allowed us to overcome, at least partially, such a limitation. Moreover, the retrospective nature of our study did not permit the assessment of medication effects over time, with a potential impact on the illness course of the two subgroups. In addition, as data about the total number of episodes and related polarity was lacking, we could not rule out the potential role of the OP as indicator of the likely predominant polarity and/or rapid cycling over the course of illness. Finally, the inclusion of mixed and hypo-/manic patients in the EO group may have adversely affected the statistical power of our analysis. Indeed, stratification by diagnostic subtype (BD I, BD II) may clarify how OP differentially moderates the clinical presentation.

12

Therefore, replication of our findings in future prospective studies is needed to provide more conclusive evidence regarding the clinical impact of OP in BD.

References 1. Aizenberg, D., Olmer, A., Barak, Y., 2006. Suicide attempts amongst elderly bipolar patients. J. Affect. Disord. 91, 91-94. 2. Akiskal, H.S., Downs, J., Jordan, P., Watson, S., Daugherty, D., Pruitt, D.B., 1985. Affective disorders in referred children and younger siblings of manic-depressives. Mode of onset and prospective course. Arch. Gen. Psychiatry 42, 996-1003. 3. American Psychiatric Association, 2000. Diagnostic and Statistical Manual of Mental Disorders (4 th ed., text rev.). American Psychiatric Association, Washington, DC. 4. American Psychiatric Association, 2013. Diagnostic and Statistical Manual of Mental Disorders (5th ed.). American Psychiatric Association, Washington, DC. 5. Amsterdam, J.D., Lorenzo-Luaces, L., Soeller, I., Li, S.Q., Mao, J.J., DeRubeis, R.J., 2015. Safety and effectiveness of continuation antidepressant versus mood stabilizer monotherapy for relapseprevention of bipolar II depression: a randomized, double-blind, parallel-group, prospective study. J Affect. Disord. 185, 31-37. 6. Angst, J., Felder, W., Frey, R., Stasssen, H.H., 1978. The course of affective disorders: I. Change of diagnosis of monopolar, unipolar, and bipolar illness. Arch. Psychiatr. Nervenkr. 226, 57-64. 7. Azorin, J.M., Kaladjian, A., Adida, M., Fakra, E., Hantouche, E., Lancrenon, S., 2011. Correlates of first-episode polarity in a French cohort of 1089 bipolar I disorder patients: role of temperaments and triggering events. J. Affect. Disord. 129, 39-46. 8. Baldessarini, R.J., Salvatore, P., Khalsa, H.M., Gebre-Medhin, P., Imaz, H., González-Pinto, A., Perez, J., Cruz, N., Maggini, C., Tohen, M., 2010. Morbidity in 303 first-episode bipolar I disorder patients. Bipolar Disord. 12, 264-270. 9. Barbato, A., 2006. Psychosocial rehabilitation and severe mental disorders: a public health approach. World Psychiatry 5, 162-163. 10. Benedetti, F., Radaelli, D., Poletti, S., Locatelli, C., Falini, A., Colombo, C., Smeraldi, E., 2011. Opposite effects of suicidality and lithium on gray matter volumes in bipolar depression. J. Affect. Disord. 135, 139-147.

13

11. Calabrese, J.R., Vieta, E., El-Mallakh, R., Findling, R.L., Youngstrom, E.A., Elhaj, O., Gajwani, P., Pies, R., 2004. Mood state at study entry as predictor of the polarity of relapse in bipolar disorder. Biol. Psychiatry 56, 957-963. 12. Carballo, J.J., Harkavy-Friedman, J., Burke, A.K., Sher, L., Baca-Garcia, E., Sullivan, G.M., Grunebaum, M.F., Parsey, R.V., Mann, J.J., Oquendo, M.A., 2008. Family history of suicidal behavior and early traumatic experiences: additive effect on suicidality and course of bipolar illness? J. Affect. Disord. 109, 57-63. 13. Carvalho, A.F., McIntyre, R.S., Dimelis, D., Gonda, X., Berk, M., Nunes-Neto, P.R., Cha, D.S., Hyphantis, T.N., Angst, J., Fountoulakis. K.N., 2014. Predominant polarity as a course specifier for bipolar disorder: a systematic review. J. Affect. Disord. 163, 56-64. 14. Cha, B., Kim, J.H., Ha, T.H., Chang, J.S., Ha, K., 2009. Polarity of the first episode and time to diagnosis of bipolar I disorder. Psychiatry Investig. 6, 96-101. 15. Chaudhury, S.R., Grunebaum, M.F., Galfalvy, H.C., Burke, A.K., Sher, L., Parsey, R.V., Everett, B., Mann, J.J., Oquendo, M.A., 2007. Does first episode polarity predict risk for suicide attempt in bipolar disorder? J. Affect. Disord. 104, 245-250. 16. Chen, Y.W., Dilsaver, S.C., 1996. Lifetime rates of suicide attempts among subjects with bipolar and unipolar disorders relative to subjects with other Axis I disorders. Biol. Psychiatry 39, 896-899. 17. Colom, F., Vieta, E., Daban, C., Pacchiarotti, I., Sánchez-Moreno, J., 2006. Clinical and therapeutic implications of predominant polarity in bipolar disorder. J. Affect. Disord. 93, 13-17. 18. Daban, C., Colom, F., Sanchez-Moreno, J., García-Amador, M., Vieta, E., 2006. Clinical correlates of first-episode polarity in bipolar disorder. Compr. Psychiatry 47, 433-437. 19. Dell'Osso, B., Dobrea, C., Cremaschi, L., Buoli, M., Miller, S., Ketter, T.A., Altamura, A.C., 2016. Italian Bipolar II vs I patients have better individual functioning, in spite of overall similar illness severity. CNS Spectr. 1-8 [Epub ahead of print]. a 20. Dell'Osso, B., Grancini, B., Vismara, M., De Cagna, F., Maggi, M., Molle, M., Cremaschi, L., Miller, S., Ketter, T.A., Altamura, A.C., 2016. Age at onset in patients with bipolar I and II disorder: a comparison of large sample studies. J. Affect. Disord. 201, 57-63. b 21. El-Mallakh, R.S., Vöhringer, P.A., Ostacher, M.M., Baldassano, C.F., Holtzman, N.S., Whitham, E.A., Thommi, S.B., Goodwin, F.K., Ghaemi, S.N., 2015. Antidepressants worsen rapid-cycling course in bipolar depression: A STEP-BD randomized clinical trial. J. Affect. Disord. 184, 318-321. 14

22. Endicott, J., Spitzer, R.L., Fleiss, J.L., Cohen, J., 1976. The global assessment scale. A procedure for measuring overall severity of psychiatric disturbance. Arch. Gen. Psychiatry 33, 766-771. 23. Erten, E., Funda Uney, A., Saatçioğlu, Ö., Özdemir, A., Fıstıkçı, N., Çakmak, D., 2014. Effects of childhood trauma and clinical features on determining quality of life in patients with bipolar I disorder. J. Affect. Disord. 162, 107-113. 24. First, M.B., Gibbon, M., Spitzer, R.L., Williams, J.B.W., Benjamin, L.S., 1997. Structured Clinical Interview for DSM-IV Axis II Personality Disorders (SCID-II). Washington, DC: American Psychiatric Press. 25. First, M.B., Spitzer, R.L., Gibbon, M., Williams, J.B.W., 2002. Structured Clinical Interview for DSM–IV–TR Axis I Disorders, Research Version, Patient Edition (SCID-I/P). New York: Biometrics Research, New York State Psychiatric Institute. 26. Forty, L., Jones, L., Jones, I., Smith, D.J., Caesar, S., Fraser, C., Gordon-Smith, K., Hyde, S., Craddock, N., 2009. Polarity at illness onset in bipolar I disorder and clinical course of illness. Bipolar Disord. 11, 82-88. 27. Goldberg, J.F., Ernst, C.L., 2002. Features associated with the delayed initiation of mood stabilizers at illness onset in bipolar disorder. J. Clin. Psychiatry 63, 985-991. 28. Guillaume, S., Jaussent, I., Jollant, F., Rihmer, Z., Malafosse, A., Courtet, P., 2010. Suicide attempt characteristics may orientate toward a bipolar disorder in attempters with recurrent depression. J. Affect. Disord. 122, 53-59. 29. Haag, H., Heidorn, A., Haag, M., Greil, W., 1987. Sequence of affective polarity and lithium response: preliminary report on Munich sample. Prog. Neuropsychopharmacol. Biol. Psychiatry 11, 205-208. 30. Hall, R.C., 1995. Global assessment of functioning: a modified scale. Psychosomatics 36, 267-275. 31. Jagadheesan, K., Sinha, V.K., 2003. Course and outcome of adolescent-onset bipolar affective disorder: a retrospective clinic based study. Hong Kong J. Psychiatry 13, 2-6. 32. Judd, L.L., Akiskal, H.S., Schettler, P.J., Coryell, W., Maser, J., Rice, J.A., Solomon, D.A., Keller, M.B., 2003. The comparative clinical phenotype and long term longitudinal episode course of bipolar I and II: a clinical spectrum or distinct disorders? J. Affect. Disord. 73, 19-32. 33. Kassem, L., Lopez, V., Hedeker, D., Steele, J., Zandi, P., 2006. Bipolar Disorder Consortium NIMH Genetics Initiative, McMahon FJ. Familiality of polarity at illness onset in bipolar affective disorder. Am. J. Psychiatry 163, 1754-1759. 15

34. Leverich, G.S., Altshuler, L.L., Frye, M.A., Suppes, T., Keck, P.E.Jr., McElroy, S.L., Denicoff, K.D., Obrocea, G., Nolen, W.A., Kupka, R., Walden, J., Grunze, H., Perez, S., Luckenbaugh, D.A., Post, R.M., 2003. Factors associated with suicide attempts in 648 patients with bipolar disorder in the Stanley Foundation Bipolar Network. J. Clin. Psychiatry 64, 506-515. 35. Maj, M., Pirozzi, R., Starace, F., 1989. Previous pattern of course of the illness as a predictor of response to lithium prophylaxis in bipolar patients. J. Affect. Disord. 17, 237-241. 36. Maniglio, R., 2013. The impact of child sexual abuse on the course of bipolar disorder: a systematic review. Bipolar Disord. 15, 341-358. 37. Michaelis, B.H., Goldberg, J.F., Singer, T.M., Garno, J.L., Ernst, C.L., Davis, G.P., 2003. Characteristics of first suicide attempts in single versus multiple suicide attempters with bipolar disorder. Compr. Psychiatry 44, 15-20. 38. Neves, F.S., Malloy-Diniz, L.F., Barbosa, I.G., Brasil, P.M., Corrêa, H., 2009. Bipolar disorder first episode and suicidal behavior: are there differences according to type of suicide attempt? Rev. Bras. Psiquiatr. 31, 114-118. 39. Perlis, R.H., Delbello, M.P., Miyahara, S., Wisniewski, S.R., Sachs, G.S., Nierenberg, A.A., 2005. Revisiting depressive-prone bipolar disorder: polarity of initial mood episode and disease course among bipolar I systematic treatment enhancement program for bipolar disorder participants. Biol. Psychiatry 58, 549-553. 40. Perris, C., d'Elia, G., 1966. A study of bipolar (manic-depressive) and unipolar recurrent depressive psychoses. X. Mortality, suicide and life-cycles. Acta Psychiatr. Scand. Suppl. 194, 172-189. 41. Perugi, G., Micheli, C., Akiskal, H.S., Madaro, D., Socci, C., Quilici, C., Musetti, L., 2000. Polarity of the first episode, clinical characteristics, and course of manic depressive illness: a systematic retrospective investigation of 320 bipolar I patients. Compr. Psychiatry 41, 13-18. 42. Popovic, D., Torrent, C., Goikolea, J.M., Cruz, N., Sánchez-Moreno, J., González-Pinto, A., Vieta, E., 2014. Clinical implications of predominant polarity and the polarity index in bipolar disorder: a naturalistic study. Acta Psychiatr. Scand. 129, 366-374. 43. Quitkin, F.M., Rabkin, J.G., Prien, R.F., 1986. Bipolar disorder: are there manic-prone and depressiveprone forms? J. Clin. Psychopharmacol. 6, 167-172. 44. Rihmer, Z., Kiss, K., 2002. Bipolar disorders and suicidal behaviour. Bipolar Disord. 4 Suppl 1, 21-25.

16

45. Rosa, A.R., Bonnín, C.M., Vázquez, G.H., Reinares, M., Solé, B., Tabarés-Seisdedos, R., BalanzáMartínez, V., González-Pinto, A., Sánchez-Moreno, J., Vieta, E., 2010. Functional impairment in bipolar II disorder: is it as disabling as bipolar I? J. Affect. Disord. 127, 71-76. 46. Roy-Byrne, P., Post, R.M., Uhde, T.W., Porcu, T., Davis, D., 1985. The longitudinal course of recurrent affective illness: life chart data from research patients at the NIMH. Acta Psychiatr. Scand. Suppl. 317, 1-34. 47. Ryu, V., Jon, D.I., Cho, H.S., Kim, S.J., Lee, E., Kim, E.J., Seok, J.H., 2010. Initial depressive episodes affect the risk of suicide attempts in Korean patients with bipolar disorder. Yonsei Med. J. 51, 641-647. 48. Schaffer, A., Isometsä, E.T., Azorin, J.M., Cassidy, F., Goldstein, T., Rihmer, Z., Sinyor, M., Tondo, L., Moreno, D.H., Turecki, G., Reis, C., Kessing, L.V., Ha, K., Weizman, A., Beautrais, A., Chou, Y.H., Diazgranados, N., Levitt, A.J., Zarate, C.A.Jr., Yatham, L., 2015. A review of factors associated with greater likelihood of suicide attempts and suicide deaths in bipolar disorder: Part II of a report of the International Society for Bipolar Disorders Task Force on Suicide in Bipolar Disorder. Aust. N. Z. J. Psychiatry 49, 1006-1020. 49. Simon, G.E., Bauer, M.S., Ludman, E.J., Operskalski, B.H., Unützer, J., 2007. Mood symptoms, functional impairment, and disability in people with bipolar disorder: specific effects of mania and depression. J. Clin. Psychiatry 68, 1237-1245. 50. Sipos, A., Harrison, G., Gunnell, D., Amin, S., Singh, S.P., 2001. Patterns and predictors of hospitalisation in first-episode psychosis. Prospective cohort study. Br. J. Psychiatry 178, 518-523. 51. Tohen, M., Zarate, C.A.Jr., Hennen, J., Khalsa, H.M., Strakowski, S.M., Gebre-Medhin, P., Salvatore, P., Baldessarini, R.J., 2003. The McLean-Harvard First-Episode Mania Study: prediction of recovery and first recurrence. Am. J. Psychiatry 160, 2099-2107. 52. Tsai, S.Y., Kuo, C.J., Chen, C.C., Lee, H.C., 2002. Risk factors for completed suicide in bipolar disorder. J. Clin. Psychiatry 63, 469-476. 53. Tundo, A., Musetti, L., Benedetti, A., Berti, B., Massimetti, G., Dell'Osso, L., 2015. Onset polarity and illness course in bipolar I and II disorders: the predictive role of broadly defined mixed states. Compr. Psychiatry 63, 15-21. 54. Turvey, C.L., Coryell, W.H., Arndt, S., Solomon, D.A., Leon, A.C., Endicott, J., Mueller, T., Keller, M., Akiskal, H., 1999. Polarity sequence, depression, and chronicity in bipolar I disorder. J. Nerv. Ment. Dis. 187, 181-187. 17

55. Valtonen, H.M., Suominen, K., Mantere, O., Leppämäki, S., Arvilommi, P., Isometsä, E.T., 2006. Prospective study of risk factors for attempted suicide among patients with bipolar disorder. Bipolar Disord. 8, 576-585. 56. Vieta, E., Berk, M., Wang, W., Colom, F., Tohen, M., Baldessarini, R.J., 2009. Predominant previous polarity as an outcome predictor in a controlled treatment trial for depression in bipolar I disorder patients. J. Affect. Disord. 119, 22-27.

Figure 1. Clinical variables with statistical significant difference between DO and EO bipolar

Centinaia

patients. 80%

**

70%

*

60%

*

50% 40%

** *

**

*

30%

DO

20%

EO

10% 0% BD II subtype

depressive most recent episode

involuntary commitments

psychosis

suicide attempts stressful life events

current antidepressant treatment

Legend: DO=depressive onset, EO=elevated (hypomanic/manic/mixed) onset. Statistics: *p<0.05; **p≤0.001

Table 1. Socio-demographic variables of the study sample and related sugroups according to the polarity of the first episode. All patients

Age (years, mean ± SD) Gender (%) Female Education (%) Secondary school High-school University Employment (%) Employed Unemployed Retired

(N=362, 100%)∆

Depressive onset (DO) (N=182, 57.6%)

Elevated onset (EO) (N=134, 42.4%)

p values

Effect size§

48.6±14.4

49.3±14.3 **

42.2±13.5

0.001

0.375

52.5

61.0 *

45.5

0.008 0.09

0.153 0.145

16.9 51.3 29.1

21.9 46.1 29.2

13 59.5 25.2

48.7 37.8 13.5

50.8 33.0 16.2

46.6 44.4 * 9

0.05

0.137

18

Lifetime occupational instability (%) Co-habitation (%) Family Family of origin Alone Other Marital status (%) Single Partner Divorced

37.1

32.3

38.6

45.7 24.7 22.9 6.7

48 24.6 24 3.4

41.4 25.8 21.9 10.9

43 42.1 12.3

37.8 44.4 13.9

53.4 * 35.9 9.2

0.27 0.07

0.065 0.154

0.04

0.165

Values for categorical and continuous variables are expressed in percentages and mean ± SD, respectively. Boldface indicates parameters with significant differences between the two subgroups. Missing data: 7.9% for lifetime occupational instability, 4.1% for co-habitation, 0-2.2% for other parameters. ∆ §

46/362 patients were excluded from comparative analyses due to inadequate information about OP. Effect size was expressed by means of Cramer’s phi for categorical variables and of d for continuous ones.

Table 2. Clinical variables of the study sample and related subgroups according to the polarity of the first episode.

All patients (N=362, 100%)∆ Diagnosis BD II Age at onset (years, mean ± SD) < 18 years ≥ 18 years Duration of illness (months, mean ± SD) Duration of Untreated Illness (months, mean ± SD) Family history of psychiatric disorder (%) Polarity of most recent episode Depressive most recent episode (%) Duration of most recent episode (days, mean ± SD) Psychiatric hospitalizations (lifetime nr., mean ± SD) Involuntary commitments (lifetime, %) Involuntary commitments (lifetime nr., mean ± SD) Psychosis (lifetime, %) Suicide attempts (lifetime, %) Suicide attempts (lifetime nr., mean ± SD) Subthreshold symptoms (lifetime, %) Stressful life events (lifetime, %) Psychiatric comorbidity (%) Any Any anxiety disorder Obsessive Compulsive Disorder Personality disorder Alcohol/Substance use disorder Eating disorder Psychiatric policomorbidity Medical comorbidity (lifetime, %) Psychosocial rehabilitation (lifetime, %) Global Assessment of Functioning (current, mean ±

Depressive onset (DO) (N=182, 57.6%)

Elevated p onset (EO) values (N=134, 42.4%)

Effect size§

25.7 28.7±11.4 12.5 87.5 241.8±155.6 58±101.2 65.6

37.4 ** 28.2±10.7 13.7 86.3 253±155.8 62.8±101.4 68.9

13.4 26.8±9.2 11.9 88.1 189.7±137.8 52.2±102.1 60.9

<0.001 0.27

0.266 0.127

0.44 0.98 0.14

0.088 0.004 0.083

44.3

51.1 *

36

40.6±54.9

51±68.2 *

31.1±36.4

0.01 0.03

0.150 0.255

3.1±5.1

3.3±6.7

3.3±4.7

0.49

0.079

28.6

22.1

39.3 *

0.6±1.4

0.4±0.9

1±2.2 *

0.001 0.001

0.187 0.37

57.5 26.1 0.4±0.8 47 58.3

45.3 31.9 * 0.4±0.8 50 64.5 *

75.9 ** 20.9 0.3±0.8 43.2 50.8

<0.001 0.03 0.44 0.25 0.02

0.307 0.122 0.088 0,067 0.137

47.4 28.8 1.4 5.1 11.6 3.1 14.7 48.3 12.7 65.3±14

49.7 32 1.7 5.6 9.5 2.8 13.4 50.3 10.8 67.3±13.8

45.1 24.2 0.8 5.3 12.8 4.5 16.7 44.4 16.4 68±14

0.41 0.13

0.127 0.085

0.42 0.38 0.15 0.08

0.045 0.166 0.082 0.033

19

SD) Current treatment (%) Mood stabilizers+antipsychotics Antidepressants Psychotropic drug (current, mean nr. ± SD)

59.8 35 2.3±1

64.8 44.7 ** 2.7±1 *

56.9 24.2 2.3±0.9

0.19 <0.001 0.01

0.080 0.210 0.288

Values for categorical and continuous variables are expressed in percentages and mean ± SD, respectively. Boldface indicates parameters with significant differences between the two subgroups. Legend: BD I=bipolar I disorder, BD II=bipolar II disorder. Missing data: 28% for duration of most recent episode, 0-7% for other parameters. ∆

46 patients were excluded from comparative analyses due to inadequate information about OP

§

Effect size was expressed by means of Cramer’s phi for categorical variables and of d for continuous ones.

Highlights

  

Onset polarity (OP) in BD has been increasingly investigated over the last years for its clinical, prognostic, and therapeutic implications. The present study aimed to assess OP correlates, in particular as regards suicidal risk, in a sample of 362 recovered bipolar patients. A strong association between depressive OP and suicide attempts emerged, thus supporting the influence of OP on BD course and outcome.

20