Morbidity in 303 first-episode bipolar I disorder patients

Transcription

1 Bipolar Disorders 2010: 12: ª 2010 The Authors Journal compilation ª 2010 John Wiley & Sons A/S BIPOLAR DISORDERS Original Article Morbidity in 303 first-episode bipolar I disorder patients Baldessarini RJ, Salvatore P, Khalsa H-MK, Gebre-Medhin P, Imaz H, González-Pinto A, Perez J, Cruz N, Maggini C, Tohen M. Morbidity in 303 first-episode bipolar I disorder patients. Bipolar Disord 2010: 12: ª 2010 The Authors. Journal compilation ª 2010 John Wiley & Sons A S. Objectives: To test the hypotheses that: (i) depressive-dysthymicdysphoric (D-type) morbidity is more prevalent than manic-hypomanicpsychotic (M-type) morbidity even from first episodes of bipolar I disorder (BPD-I) and despite treatment; (ii) initial presentations predict later morbidity; (iii) morbidity varies internationally; and (iv) early and later morbidity are similar. Methods: We followed SCID-based, DSM-IV BPD-I patients (n = 303) systematically and prospectively for two years to estimate the percent of weeks in specific morbid states from first lifetime major episodes. Results: Total morbidity accounted for 44% of the first two years, and D-type exceeded M-type illnesses by 2.1-fold (30% 14%) among morbidities ranking: mixed states (major + minor) dysthymia mania major depression > hypomania > psychosis. In 164 cases, morbidities at and years were very similar. Depressive or mixed initial episodes predicted a 3.6-fold excess of D-type morbidity, and initial M-type episodes predicted a 7.1-fold excess of M-type morbidity over two years. Morbidity in European (EU) sites was nearly half that in the U.S., and 22% greater overall among men than women. In five comparable studies, illness accounted for 54% of follow-up time, and the ratio of D M morbidity averaged 3.0. Conclusions: In accord with four midcourse studies, morbidity from BPD-I onset, despite treatment by community standards, averaged 44%, was 68% D-type morbidity, and was strongly predicted by first-episode polarity. Lower morbidity in EU than U.S. sites may reflect differences in healthcare or social systems. Ross J Baldessarini a,b, Paola Salvatore a,c, Hari-Mandir Kaur Khalsa b, Priscilla Gebre-Medhin b, Harkaitz Imaz d, Ana González- Pinto b,d, Jesus Perez b,e,núria Cruz b,f, Carlo Maggini a,c and Mauricio Tohen b,g a Department of Psychiatry, Harvard Medical School, Boston, b International Consortium for Bipolar Disorder Research, McLean Hospital, Belmont, MA, USA, c Department of Neuroscience, Section on Psychiatry, University of Parma, Parma, Italy, d Bipolar Disorder Program, Department of Psychiatry, Santiago Apóstol Hospital, Vitoria, Spain, e Department of Psychiatry, Cambridge University, Cambridge, UK, f Bipolar Disorder Program, Department of Psychiatry, Institute of Clinical Neuroscience, Hospital Clinic, IDIBAPS, CIBERSAM, University of Barcelona, Barcelona, Spain, g Department of Psychiatry, University of Texas Health Science Center at San Antonio, San Antonio, TX, USA doi: /j x Key words: bipolar disorder depression first episode follow-up mania mixed-states morbidity prediction Received 30 July 2009, revised and accepted for publication 29 December 2009 Corresponding author: Ross J. Baldessarini, M.D. Department of Psychiatry Mailman Research Center 306 McLean Hospital 115 Mill Street Belmont, MA 02478, USA Fax: rjb@mclean.org RJB has recently been a consultant or investigator-initiated research collaborator with AstraZeneca, Auritec, Biotrofix, Janssen, JDS-Noven, Eli Lilly & Co., Luitpold, NeuroHealing, Novartis, Pfizer, and SK-BioPharmaceutical Corporation; he is not a member of pharmaceutical speakers bureaus, nor does he or any family member hold equity positions in biomedical or pharmaceutical corporations. MT was an employee of Eli Lilly & Co. (to 2008), and has consulted to, or received speakers honoraria from AstraZeneca, Bristol Myers-Squibb, Eli Lilly & Co., GlaxoSmithKline, Johnson & Johnson, and Wyeth Corporations; his spouse is an Eli Lilly employee and stockholder. PS, H-MKK, PG-M, HI, AG-P, JP, NC, and CM have no relevant disclosures or potential conflicts of interest to report. 264

2 Morbidity in first-episode bipolar disorder Bipolar disorder (BPD) is typified by episodic major recurrences and variable subsyndromal morbidity, as well as considerable risk of cognitive impairment (1) and disability (2). Based mainly on midcourse analyses of patients ill for some time, there is growing evidence that depressive and dysthymic or dysphoric features continue to be a persistent type of long-term morbidity among bipolar I disorder (BPD-I) and bipolar II disorder patients (3 7), despite use of modern mood-stabilizing treatments and possible overuse of antidepressants (8, 9). Studies of long-term morbidity starting from illness onset, with prospective followup, are needed, especially in view of remaining uncertainty about whether BPD may or may not tend to worsen over time (10, 11). Depressive morbidity in BPD is particularly hard to treat effectively and has been implicated in poor clinical outcomes, disability, and very high suicidal risk in this disorder (12 17). It would be particularly useful clinically to be able to predict future illness from the onset of BPD to guide planning of treatment and rehabilitation. There are some indications that initial depression is followed by an excess of later depression among BPD patients (7, 18), but broader predictive associations of initial presentations with future morbidity types remain to be defined more clearly. Given these considerations, we now report on analyses of estimated proportions of time spent in specific long-term morbid states during the initial years from first lifetime major episodes of BPD-I in over 300 American and European (EU) patients followed systematically. Methods Subjects and treatments Patient subjects were evaluated at first lifetime hospitalizations for an episode of major affective or psychotic illness and followed prospectively and systematically at intervals ranging from one week to six months for a minimum of two years, with Structured Clinical Interview for DSM-IV Disorders (SCID)-based assessments at intake and at 24 months supporting final consensus diagnoses of DSM-IV BPD-I. Late diagnoses were employed to minimize effects of instability of some initial DSM-IV diagnoses over time (19). At follow-up assessments, interval weekly morbid status was reconstructed in life charts, based on both in-person and telephone semistructured interviews of patients, supplemented by interviews with relatives and treating physicians, and medical records (20), involving the years of follow-up following the first lifetime index illness episode. Patients were included in the same protocols, definitions, and methods at three collaborating sites over the past decade: McLean Hospital, Belmont, MA, USA, and the mood-disorder units of the University of Parma, Parma, Italy, and Vitoria Medical Center, Vitoria, Spain. Protocols were approved by ethics committees at each site and subjects provided written, informed consent for study participation and for anonymous and aggregate reporting of information obtained. Patients were followed and treated by their individual clinicians based on prevailing community standards. In this naturalistic follow-up study, treatments were not controlled by experimental protocols, nor was conformity to specified therapeutic standards required at any site. Diagnostic and assessment methods employed were detailed previously (20), including use of life-table methods of recording estimates and largely self- and family reports of interval morbidities (3 5). DSM-IV criteria were applied at all sites for major episodes, with consensus measures for subsyndromal morbidity not meeting full DSM-IV severity or duration criteria, and all sites defined functional recovery as a return to the best individual premorbid status of employment and living situation (20). Morbidity assessments Illnesses were categorized in major subtypes for simplicity. Depressive-dysthymic-dysphoric (D-type) morbidity included weeks in major depressive episodes, less severe dysthymia, or in major or minor mixed manic-depressive states of dysphoria and agitation. Manic-hypomanicpsychotic (M-type) morbidity included mania, hypomania, or psychosis. Mixed-dysphoric states were associated with D-type morbidity based on preliminary analyses indicating their far stronger association with depression than with mania, as reported below. In addition, a small proportion of cases presenting initially with unstable, polymorphic states that fluctuated among major depression, mixed states, and mania also were found in preliminary analyses to associate with later D-type morbidity, and so were included with initial D-type morbidity. Individual times in both full and subsyndromal, specific manic, hypomanic, mixed, depressive, or psychotic states were recorded during follow-up. Additional preliminary analyses indicated that findings were very similar whether or not initial illnesses or the initial six months of risk were deducted from two-year total morbidity estimates, and only totals are reported. 265

3 Baldessarini et al. To evaluate the potential impact of current mood states on reported morbidity during intervals between follow-up assessments, we rated current morbid states with standardized clinical scales (20), including a 36-item Expanded Brief Psychiatric Rating Scale (BPRS-E) (21) that includes symptoms typical of BPD patients, the 24-item Hamilton Depression Rating Scale (HDRS-24) (22) for depressive symptoms, and Young Mania Rating Scale (YMRS) (23) for manic features. We compared the percentage of time in D- or M-type and total morbid states for two years among: (i) all subjects; (ii) a random sample of subjects directly examined and rated for current morbidity (D-type if HDRS 10 and YMRS < 6; M-type if YMRS 6 and HDRS < 10); (iii) two-year morbidity in 50 patients followed with unusually close intervals, every 1 2 weeks; and (iv) in cases matched 2:1 with the latter patients for age and sex. In addition, we compared morbidity among 164 of the subjects followed for at least five years to compare initial morbidity following the index episode in two-year epochs, from years versus years. Data analyses Estimated times (percentage of weeks) in specific morbid states or in pooled D-type or M-type illnesses or in clinically stable or euthymic intervals were analyzed by ANOVA methods (F) with stated degrees of freedom (df). Statistical significance was set at two-sided p 0.01 to compensate for multiple comparisons. Data are reported as means ± SD unless indicated otherwise. We also compared long-term morbidity following either D- or M-type initial DSM-IV major episodes, as well as ratios of time in D M states, and pooled comparable data from previous studies in midcourse BPD patients. Results Patients and initial illnesses 266 A total of 303 patient subjects met final, two-year, consensus and double-scid-based (baseline and 24 months) DSM-IV diagnoses of BPD-I. Patients [216 U.S. patients from McLean Hospital, and 87 EU cases (50 from the University of Parma and 37 from Vitoria Medical Center)] included 164 men and 139 women, average age 30.6 ± 12.8 years. Of the total cohort, 164 patients were followed up to five years, and considered separately. Initial presenting illnesses included mania (49.5%), mixed manic-depressive states (22.4%), or major depression (16.2%), as well as Ôpolymorphic-unstableÕ states (shifting among depression, mixed states, and mania within two weeks; 10.6%) and psychosis without apparent initial major affective illness (1.32%). All diagnoses were DSM-IV BPD-I, verified by two-year SCID reassessments and by consensus from all available clinical information obtained over the initial 24 months. Initial mixed states were far more strongly associated with D-type than M-type morbidity over two years (57.4% versus 2.68%), as were initial polymorphic presentations (54.1% versus 5.80%). Moreover, mixed or polymorphic initial presentations were strongly related to twoyear percentages of weeks, specifically in DSM-IV major depressive versus manic-hypomanic morbidity: initial mixed states anticipated later depression 49.2 times more than mania, and initial polymorphic states 9.2 times more depression than mania (F =12.0 or 12.7, both p < ). Accordingly, both initial mixed and polymorphic illness types were categorized among D-type presentations. However, estimated time in each specific type of morbidity was considered separately and then secondarily gathered into D- or M-type morbid groups for simplicity. Distribution of morbidity The incidence of initial D- and M-type illnesses was similar (M D: 50.8% 49.2%). Estimated proportions (percent) of weeks over the initial two years of BPD illness in specific morbid states, as well as summaries of time spent in D-type or M-type states and total illness are shown for all subjects and those initially presenting in first episodes classified as either D- or M-type (Table 1). Overall, total morbidity averaged 44.4 ± 36.8% of weeks over two years. Proportions of time in specific morbid states ranked: mixed (12.1%) dysthymia (11.4%) > mania (9.89%) > major depression (6.70%) > hypomania (3.44%) > psychosis (0.90%). However, pooled D-type morbidity (depression, dysthymia, or mixed-dysphoric states) was 2.13 times more prevalent than M-type morbidity over two years [D M: 30.2% 14.2%; F (df = 301) = 40.1, p < ), and this D M preponderance was similar at U.S. (ratio = 2.1) and EU sites (ratio = 2.3). Similar proportions of time in D- or M-states were obtained over two years in the following circumstances: (i) in all 303 patients, (ii) at extensive follow-up assessments that included use of rating scales in a random subsample of 81 cases, and (iii) in 50 cases followed every 1 2 weeks and in 100 comparison subjects matched 2:1 for age and sex. For D-type morbidity, the proportions

4 Morbidity in first-episode bipolar disorder Table 1. Morbidity as percent of time during the first two years in 303 cases of DSM-IV bipolar I disorder followed from first episodes Morbidity All cases (n = 303) M-type (n = 154) D-type (n = 149) Relative risk D M Mixed states ± ± ± Dysthymia ± ± ± Mania 9.89 ± ± ± Major depression 6.70 ± ± ± Hypomania 3.44 ± ± ± Psychosis 0.90 ± ± ± All D-type ± 35.6 a ± ± b All M-type ± 25.8 a ± ± c Totals ± ± ± d M-type morbidity = mania + hypomania + psychosis; D-type morbidity = depression + dysthymia + mixed states + uncommon unstable initial episodes (mixing or switching among mania, mixed states, and depression within a single first episode). These combinations are supported by preliminary testing of highly significant associations of specific later morbidities with initial illness types. a Total time in D-type morbidity exceeded M-type by 2.13-fold (30.2% 14.2%; F=40; p < ). b D-type morbidity was 3.55 times more prevalent following D-type onsets [F (df = 301) = 61.0, p < ]. c M-type morbidity was 7.07 times more prevalent following M-type onsets (F =91.2; p < ). d Total 24-month morbidity was 1.35 times greater after D-type onsets (51.1% 30.2%; F=9.89, p = 0.002). were: (i) 30.2%, (ii) 35.8%, (iii) 32.3%; for M-type morbid states: (i) 14.2%; (ii) 6.2%; (iii) 4.2%; and for total morbidity: (i) 44.2%; (ii) 42.0%; (iii) 36.5%. In addition, total two-year morbidity in all cases versus a sample of 100 cases matched for age and sex to the 50 followed every 1 2 weeks also was similar (44.4% versus 50.1%). These findings support the similarity of distribution of morbid states based on direct and frequent assessments compared to estimates in the overall cohort evaluated with intermittent life-charting methods. Moreover, a comparison of 164 patients for morbidity at versus years yielded similar results as among all subjects in the first two years of illness. In this subsample followed for at least five years, the distribution of specific morbidity types agreed within an average of 7.0% [ratio = 1.07 (95% confidence interval: )] for early versus later years, and total morbidity averaged 43.9% versus 44.4%; the D M morbidity ratio averaged 1.96 early and 3.33 later (not shown), indicating that morbidity was similarly distributed over several years from onset. Comparisons with previous studies in midcourse patients The present findings of total and D- or M-type morbidity from initial major episodes of BPD-I accorded remarkably closely with comparable findings from four earlier studies (3 6) involving 746 such patients in the midcourse of BPD illness, plus our 303 first-episode cases, or a total of 1,049 in all five studies (Fig. 1). Overall, time in D-type morbidity (including dysthymic or dysphoric states) averaged 40.6 ± 5.83%, time in M-type morbidity (including hypomania and psychosis) averaged 13.6 ± 3.29%, and overall morbidity averaged 54.2 ± 2.59%, despite ongoing standard treatment (Fig. 1). The overall ratio of D M morbidity was 2.99, and D-type morbidity accounted for 74.9% of all illness during follow-up. Prediction of two-year morbidity from initial illness types Initial types of illnesses strongly predicted the majority illness types over two years, in that twoyear D- versus M-type morbidity among patients starting with D-type illnesses differed by 3.55-fold [D M: 47.6% 13.4%; F (df = 301) = 91.2, p < ], and differed by 7.07-fold in the opposite direction (M D: 24.6% 3.48%) among those starting with M-type illnesses [F (df = 301) = 91.2, p < ; Fig. 2]. We also computed time in D M states only during months following syndromal recovery from the initial episodes, and again found a 3.43-fold excess of D-type morbidity. These findings indicate that initial presentations were strongly predictive of future dominant morbidity types, and that D-type morbidity was twice as prevalent as M-type morbidity in BPD-I patients treated clinically by contemporary community standards (Table 1). We also considered long-term morbidities among patients presenting initially in: (i) mixed (n = 78), (ii) major depressive (n = 53), or (iii) manic (or psychotic) states (n = 172). In these subsamples, two years of D-type morbidity ranked: (i) 58.5% > (ii) 31.6% > (iii) 17.2%; and M-type morbidity ranked: (iii) 22.7% > (ii) 3.01% (i) 2.68%. The ratio of D M morbidity ranked: (i) 21.8 > (ii) 10.5 > (iii) Total morbidity ranked: (i) 61.2% > (iii) 40.0% (ii) 34.6%, indicating higher levels of long-term morbidity 267

5 Baldessarini et al. Judd et al (3) (n = 146) Post et al (4) (n = 258) Joffe et al (5) (n = 138) Paykel et al (6) (n = 204) Baldessarini et al a (n = 303) M-type D-type Total Percent time ill in initial two years ± SEM D-type 2-year morbidity M-type 2-year morbidity Overall, 5 studies (n = 1,049) (Total morbidity: 54%; D/M = 3.0) Percent time ill Fig. 1. Summary of findings from four earlier midcourse studies (3 6) plus the present first-episode study ( a ) of M-type (mania, hypomania, psychosis; white bars), D-type (depression, dysthymia, dysphoric mixed states; black bars), and Total morbidity (gray bars) in individual studies and an overall pool of all 1,049 bipolar I disorder patients, evaluated intermittently over a weighted average of 3.8 ± 0.9 years, following 19.8 ± 6.6 years of illness. Bars represent percentage of weeks in specified states. Overall, among 1,049 bipolar I disorder patients, mean D-type illness accounted for 74.9 ± 0.44% of total morbidity (which averaged 54.2 ± 2.59% of weeks of follow-up), and the D M morbidity ratio was 3.0, as indicated in the figure. 0 D-type (n = 149) M-type (n = 154) First episode type Fig. 2. Mean distribution of morbid states of D-type (depression, dysthymia, dysphoric mixed-states; black bars) or M-type (mania, hypomania, psychosis; white bars) during two years of repeated observations following first major episodes of the same types among 303 DSM-IV bipolar I disorder patients. Morbidity types (especially D-type) were strongly predicted by first lifetime episodes. (n = 164): 48.5 ± 37.7 versus 39.7 ± 35.3 (F = 4.33, p = 0.04). among patients presenting in mixed states. Twoyear morbidity among patients presenting in mixed states [subgroup (i)] was: mixed (37.8%) > depressive or dysthymic (20.7%), in contrast to initially depressed patients [subgroup (ii): mixed, 5.6% versus depressed or dysthymic, 26.0%], with very little mania or hypomania in either subgroup (1.70% and 2.98%, respectively). These findings further indicate a substantial degree of predictability of similar later illness types following initial mixed states, depression, or mania. Other factors associated with morbidity Total morbidity over two years among treated BPD-I patients was 1.90 times greater in the U.S. (n = 216) than in the EU centers (n = 87 cases): 51.1 ± 37.4% versus 26.9 ± 28.4% [F (df = 301) = 30.6, p < ]. In addition, total morbidity was only moderately (1.22 times) greater among men (n = 139) than women 268 Discussion There were several notable findings in this prospective study of DSM-IV BPD-I patients followed systematically over several years from first lifetime major episodes (Table 1). Of particular importance, we confirmed basic findings of four comparable earlier studies involving BPD-I patients in midcourse of illness (3 6), including similar observations of a large excess of depressive-type morbidity from the start of the illness (Fig. 1). These findings are consistent with emerging clinical and researchbased impressions that the D-type morbidity phases of BPD are not only strongly associated with comorbidity, disability, and premature death, as reviewed above, but evidently are poorly controlled by contemporary treatments that include routine and widespread use of mood-stabilizing agents and probably unrealistic overuse of antidepressants (8, 9, 17). Particularly prominent were mixeddysphoric, dysthymic, as well as depressive states, which together accounted for 41% of all weeks of

6 Morbidity in first-episode bipolar disorder illness, and accounted for 75% of average total morbidity in five comparable studies (Fig. 1). The particularly poor prognosis for patients presenting in mixed states might reflect the manic component of their initial and later illnesses. However, aside from additional mixed states (most prevalent), their morbidity during follow-up was far more depression or dysthymia than mania or hypomania (by more than 12:1). In addition, we found evidence that both initial depression-related and maniarelated presentations were strongly predictive of later, similar morbidity types (Fig. 2), and that initial mixed states, depression, and mania were followed by strong excesses of similar morbidity during follow-up. Morbidity also was very similar early and later among a subgroup of patients followed for at least five years. These observations together suggest that it may be feasible as well as clinically desirable to form prognoses early in the emergence of BPD illness, closer to onset, typically in adolescence or early adulthood, from the early age at onset in BPD-I patients (24, 25). Another observation that requires further study is objective evidence that mixed-dysphoric, manicdepressive states may be more predictive of similar and depression-related than of mania-related future illness. Additional current findings support the even less secure proposal that initial, unstable, polymorphous states in BPD may also predict an excess of future depression-related morbidity, particularly including mixed states. Currently, many studies include patient subjects presenting in mixed states as if they had a special form of mania, evidently by convention and perhaps in efforts to increase enrollment in clinical trials, rather than based on objective evidence. In general, the mixed states of BPD appear to be particularly ominous prognostically and remain inadequately recognized and poorly studied as a specific therapeutic target (26 31). A striking and unexplained incidental finding was that long-term morbidity at EU sites averaged about half that in the U.S. This preliminary observation encourages speculation that international cultural and healthcare-system differences may be involved. These might include possibly increased tolerance and social supports for the mentally ill, perhaps including greater availability of extended families in Europe. Other plausible factors are relatively reliable, systematic, and accessible treatment and follow-up programs at the EU sites. Such healthcare systems may yield superior results to what often appears to be a less reliable and dependable U.S. system of medical care, including in psychiatry. We suggest that this somewhat provocative observation urgently requires further study, especially in view of ongoing discussions of reformations of the American healthcare system. A notable limitation of the present study is that most patients were observed only at relatively extended intervals (months), and that some information represented in life charts may be unreliable. Moreover, most patients were followed systematically for only two years, although findings from the first and second two-year epochs for more than half of the patients were very similar to the overall results. We were particularly concerned with the clinical impression that BPD patients often emphasize depressive aspects of their long-term illnesses and find them particularly distressing, in addition to the objective evidence that these phases of contemporary BPD are least well managed with currently available treatments and particularly likely to result in poor outcomes (12 17). In contrast, manic phases of moderate severity, and especially hypomania, appear to be under-reported and often not recognized by patients as part of the spectrum of morbidity of BPD. We tried to support the validity of morbidity estimates provided by comparing results obtained by life-charting with those based on direct and frequent clinical observations and ratings. Nevertheless, there may yet be some risk of selective recall and reporting of depressive periods, and the possibility of reporting bias remains of concern for all follow-up studies of this type. In conclusion, this international, first-episode, follow-up study supports previous midcourse studies and indicates that, even from the start of BPD- I, depression-related morbidity was the dominant form in this early, and several other midcourse studies, despite availability of a growing range of mood-stabilizing and mood-altering treatments. The findings also indicate considerable predictability of later morbidity based on the type of initial episode. Acknowledgements This project was supported by NIH grants MH (RJB), MH and MH (MT), a grant from the Bruce J. Anderson Foundation, and the McLean Private Donors Research Fund for Bipolar Disorder Research (RJB), by a grant from NARSAD (PS), a Rio-Hortega fellowship from the Instituto de Salud Carlos III and CIBERSAM (NC), and by the Atlas Foundation (MT). References 1. Wingo AP, Harvey PD, Baldessarini RJ. Cognitive impairment in euthymic bipolar disorder patients: systematic 269

7 Baldessarini et al. review of functional implications, pathophysiology, and treatment. Bipolar Disord 2009; 11: Huxley N, Baldessarini RJ. Disability and its treatment in bipolar disorder patients. Bipolar Disord 2007; 9: Judd LL, Akiskal HS, Schettler PJ et al. Long-term natural history of the weekly symptomatic status of bipolar I disorder. Arch Gen Psychiatry 2002; 59: Post RM, Denicoff KD, Leverich GS et al. Morbidity in 258 bipolar outpatients followed for 1 year with daily prospective ratings on the NIMH life chart method. J Clin Psychiatry 2003; 64: Joffe RT, MacQueen GM, Marriott M, Young TL. Prospective, longitudinal study of percentage of time spent ill in patients with bipolar I or bipolar II disorders. Bipolar Disord 2004; 6: Paykel ES, Abbott R, Morriss R, Hayhurst H, Scott J. Sub-syndromal and syndromal symptoms in the longitudinal course of bipolar disorder. Br J Psychiatry 2006; 189: Forty L, Jones L, Jones I et al. Polarity at illness onset in bipolar I disorder and clinical course of illness. Bipolar Disord 2009; 11: Baldessarini RJ, Henk HJ, Sklar AR, Chang J, Leahy LF. Use of psychotropic medicines in bipolar disorder patients in the United States. Psychiatr Serv 2008; 59: Ghaemi SN, Wingo AP, Filkowski MA, Goodwin FK, Baldessarini RJ. Effectiveness of long-term antidepressant treatment in bipolar disorder: a meta-analysis. Acta Psychiatr Scand 2008; 118: Oepen G, Salvatore P, Baldessarini RJ. On the periodicity of manic-depressive insanity by Eliot Slater (1938): translation and commentary. J Affect Disord 2004; 78: Salvatore P, Tohen M, Khalsa HM, Baethge C, Tondo L, Baldessarini RJ. Longitudinal research on bipolar disorders. Epidemiol Psichiatr Soc 2007; 16: Tondo L, Lepri B, Baldessarini RJ. Risks of suicidal ideation, attempts and suicides among 2826 men and women with types I and II bipolar, and recurrent major depressive disorders. Acta Psychiatr Scand 2007; 116: Michalak EE, Murray G, Young AH, Lam RW. Burden of bipolar depression: impact of disorder and medications on quality of life. CNS Drugs 2008; 22: Bauer M, Glenn T, Grof P et al. Frequency of subsyndromal symptoms and employment status in patients with bipolar disorder. Soc Psychiatry Psychiatr Epidemiol 2009; 44: Kauer-SantÕAnna M, Bond DJ, Lam RW, Yatham LN. Functional outcomes in first-episode patients with bipolar disorder: prospective study from the Systematic Treatment Optimization Program for Early Mania project. Compr Psychiatry 2009; 50: Martino DJ, Marengo E, Igoa A et al. Neurocognitive and symptomatic predictors of functional outcome in bipolar disorders: a prospective 1 year follow-up study. J Affect Disord 2009; 116: Baldessarini RJ, Vieta E, Calabrese JR, Tohen M, Bowden C. Bipolar depression: overview and commentary. Harv Rev Psychiatry 2010; in press. 18. Colom F, Vieta E, Daban C, Pacchiarotti I, Sánchez- Moreno J. Clinical and therapeutic implications of predominant polarity in bipolar disorder. J Affect Disord 2006; 93: Salvatore P, Baldessarini RJ, Tohen M et al. McLean- Harvard International First-Episode Project: two-year stability of DSM-IV diagnoses in 500 first-episode psychotic disorder patients. J Clin Psychiatry 2009; 70: Tohen M, Zarate CA Jr, Hennen J et al. The McLean- Harvard First-Episode Mania Study: prediction of recovery and first recurrence. Am J Psychiatry 2003; 160: Ventura J, Lukoff D, Neuchterlein K. Manual for Expanded Brief Psychiatric Rating Scale. Schizophr Bull 1997; 7: Hamilton M. A rating scale for depression. J Neurol Neurosurg Psychiatry 1960; 23: Young RC, Biggs JT, Ziegler VE, Meyer DA. A rating scale for mania: reliability, validity and sensitivity. Br J Psychiatry 1978; 133: Baldessarini RJ, Bolzani L, Cruz N et al. Onset-age of bipolar disorders at six international sites. J Affect Disord 2010; 121: Tondo L, Baldessarini RJ, Lepri B. Age at onset of bipolar I and II, and recurrent unipolar major depression disorder patients. Acta Psychiatr Scand 2010; in press. 26. Salvatore P, Baldessarini RJ, Centorrino F et al. WeygandtÕs, The Manic-Depressive Mixed States: translation and commentary on its significance in the evolution of the concept of bipolar manic-depressive disorder. Harv Rev Psychiatry 2002; 10: Gonza lez-pinto A, Aldama A, Mosquera F, Gonza lez Go mez C. Epidemiology, diagnosis and management of mixed mania. CNS Drugs 2007; 21: Gonza lez-pinto A, Barbeito S, Alonso M et al. Poor longterm prognosis in mixed bipolar patients: 10-year outcomes in the Vitoria prospective naturalistic study in Spain. J Clin Psychiatry 2010; in press. 29. Bertschy G, Gervasoni N, Favre S et al. Frequency of dysphoria and mixed states. Psychopathology 2008; 41: Muzina DJ. Pharmacologic treatment of rapid cycling and mixed states in bipolar disorder: an argument for the use of lithium. Bipolar Disord 2009; 11 (Suppl. 2): Tohen M, Sutton VK, Calabrese JR, Sachs GS, Bowden CL. Maintenance of response following stabilization of mixed index episodes with olanzapine monotherapy in a randomized, double-blind, placebo-controlled study of bipolar 1 disorder. J Affect Disord 2009; 116: