Specificity of Psychosis, Mania and Major Depression in A

Home , Cyclothymia, Dysthymia, Hypomania, Mania, Mood disorder, Psychiatry

ORIGINAL ARTICLE Speciﬁcity of psychosis, mania and major depression in a contemporary family study

CL Vandeleur1, KR Merikangas2, M-PF Strippoli1, E Castelao1 and M Preisig1

There has been increasing attention to the subgroups of mood disorders and their boundaries with other mental disorders, particularly psychoses. The goals of the present paper were (1) to assess the familial aggregation and co-aggregation patterns of the full spectrum of mood disorders (that is, bipolar, schizoaffective (SAF), major depression) based on contemporary diagnostic criteria; and (2) to evaluate the familial specificity of the major subgroups of mood disorders, including psychotic, manic and major depressive episodes (MDEs). The sample included 293 patients with a lifetime diagnosis of SAF disorder, bipolar disorder and major depressive disorder (MDD), 110 orthopedic controls, and 1734 adult first-degree relatives. The diagnostic assignment was based on all available information, including direct diagnostic interviews, family history reports and medical records. Our findings revealed specificity of the familial aggregation of psychosis (odds ratio (OR) ¼ 2.9, confidence interval (CI): 1.1–7.7), mania (OR ¼ 6.4, CI: 2.2–18.7) and MDEs (OR ¼ 2.0, CI: 1.5–2.7) but not hypomania (OR ¼ 1.3, CI: 0.5–3.6). There was no evidence for cross-transmission of mania and MDEs (OR ¼ .7, CI:.5–1.1), psychosis and mania (OR ¼ 1.0, CI:.4–2.7) or psychosis and MDEs (OR ¼ 1.0, CI:.7–1.4). The strong familial specificity of psychotic, manic and MDEs in this largest controlled contemporary family study challenges the growing assertion that the major types of mood disorders are manifestations of a common underlying diathesis.

Molecular Psychiatry (2014) 19, 209–213; doi:10.1038/mp.2013.132; published online 15 October 2013 Keywords: family aggregation patterns; mood disorders; psychosis; mania; hypomania; major depression

INTRODUCTION relationships of the full range of disorders across the mood Recent diagnostic systems in psychiatry include numerous disorder spectrum, including SAF disorder, psychotic and non- disorders and subtypes of schizophrenia (SZA), bipolar disorder psychotic BPD and major depression. Moreover, no previous study (BPD) and major depressive disorder (MDD) and their overlap to has examined the familial aggregation patterns of the core represent the spectrum of expression of mood and psychotic components of these conditions without applying the arbitrary disorders. However, the discrete boundaries between the broad distinctions specified in the diagnostic nomenclature. Therefore, categories of these disorders have been increasingly disputed the goals of this report are: (1) to assess the familial aggregation with the emergence of results from molecular genetic studies,1 and co-aggregation patterns of the full spectrum of mood particularly recent large genome-wide association studies that disorders based on contemporary diagnostic criteria; and (2) to demonstrate common polymorphisms underlying the broad evaluate the familial specificity of their major subgroups, including range of psychotic and mood disorders.2 Previous family studies psychotic, manic and major depressive episodes (MDEs). of the specificity of these conditions have been contradictory, with studies suggesting some specificity of these disorders (for 3–6 3,4 7 example, SZA, BPD and schizoaffective (SAF) disorder ), METHODS whereas others have demonstrated significant familial overlap Participants across these conditions.8–15 Likewise, overlap has been found There were a total of 403 probands, including 293 patients with mood between SZA and SAF.4,16 Therefore, SZA, BPD and MDD have 17 disorders and 110 controls selected from the local psychiatric and been postulated to belong to the same continuum. Most of this orthopedic departments. Probands with mood disorders were consecu- research was based on earlier diagnostic systems, including the tively recruited from the inpatient (61%) and outpatient (39%) facilities of Research Diagnostic Criteria (18), and the Diagnostic and Statistical the psychiatric departments of Lausanne and Geneva, Switzerland. These Manual of Mental Disorders (DSM)-III and III-R criteria, that psychiatric university departments cover the majority of inpatient specified different conventions for the diagnosis of psychotic treatments of the population of these respective regions. More severe symptoms that occurred in the context of mood disorders. Such patients and more patients who have been hospitalized are treated in the conventions as well as the exclusion of anxiety disorders that outpatient units of Lausanne and Geneva than by the private psychiatrists occur solely during mood disorders19 led to important differences in this area. Inclusion criteria for psychiatric probands were: (1) a lifetime diagnosis of SAF (SAF disorder, manic subtype (with at least one manic in the classification and familial aggregation of these disorders 20 episode); SAF disorder, depressive episode (with at least one MDE of over time. substantial duration of the total duration of psychosis)), BPD-I, BPD-II or There are few controlled direct interview family studies that MDD, (2) age between 18 and 65 years, (3) ability to speak sufficient used the most recent DSM-IV nomenclature to evaluate the French or English to complete the diagnostic interview, and (4) having a

1Department of Psychiatry, University Hospital of Lausanne, Prilly, Switzerland and 2Genetic Epidemiology Research Branch, Intramural Research Program, National Institute of Mental Health, Bethesda, MD, USA. Correspondence: Dr CL Vandeleur, Department of Psychiatry, University Hospital of Lausanne, Site de Cery, Prilly 1008, Switzerland. E-mail: [email protected] Received 9 April 2013; revised 28 July 2013; accepted 22 August 2013; published online 15 October 2013 Speciﬁcity of psychosis, mania and major depression CL Vandeleur et al 210 Table 1. Demographic characteristics and comorbidity in probands and relatives by proband disorder subgroup

Proband disorder

Schizoaffective Bipolar I Bipolar II MDD Controla Statistic P-value

Probands (N ¼ 403) 62 100 23 108 110 — — 2 Sex (% female) 38.7 60.0 39.1 49.1 45.5 X4 ¼ 8.9 NS Mean age (s.d.) 36.8 (10.8) 41.5 (11.8) 42.4 (10.7) 41.1 (11.5) 38.8 (12.9) F4 ¼ 2.3 NS Age range (years) 19.2–61.6 20.4–66.7 24.0–60.7 19.4–68.0 17.2–68.0 — — b Mean SES (s.d.) 3.0 (1.1) 3.5 (1.0) 3.0 (0.9) 2.9 (1.1) 3.3 (1.1) F4 ¼ 4.6 o0.01 c 2 Anxiety Dx (%) 45.2 29.0 47.8 55.6 9.1 X4 ¼ 59.1 o0.001 d 2 Alcohol Dx (%) 43.6 35.0 52.2 45.4 6.4 X4 ¼ 50.4 o0.001 e 2 Drug Dx (%) 40.3 21.0 43.5 28.7 7.3 X4 ¼ 32.9 o0.001 Relatives (N ¼ 1734) 251 432 98 460 493 — — 2 Sex (% female) 52.2 50.2 56.1 53.0 50.7 X4 ¼ 1.7 NS Mean age (s.d) 48.5 (17.4) 50.7 (18.4) 50.5 (19.2) 50.4 (18.1) 50.2 (18.3) F4 ¼ 0.7 NS Age range (years) 18.0–91.0 18.0–95.0 19.0–86.7 18.0–97.0 18.0–90.0 — — 2 Interviewed (%) 51.7 50.1 46.4 38.9 35.9 X4 ¼ 25.5 o0.001 Abbreviations: MDD, major depressive disorder; NS, not signiﬁcant; SES, socio-economic status. aIncludes probands with subthreshold bipolar disorder, cyclothymia, hyperthymic personality, subthreshold depression, dysthymia without MDD, depressive personality without MDD, anxiety disorders and alcohol or drug abuse/dependence when recruited in the orthopedic departments. bSocio-economic status of the family: a value of 3 represents an SES of III (middle class) on the Hollingshead scale. cAnxiety disorders, including generalized anxiety disorder, agoraphobia, panic disorder and/or social phobia. dAlcohol abuse/ dependence. eDrug abuse/dependence, including the following substances: marijuana, cocaine, narcotics, solvents, hallucinogens, stimulants and sedatives.

first-degree adult relative who agrees to participate in the study. Diagnoses information on non-interviewed relatives. The validity of the French of SAF, BPD-I, BPD-II and MDD were assigned according to the DSM-IV, with version of the FH-RDC has previously been established through the the exception of criterion A for SAF regarding the concomitant manifesta- assessment of agreement between diagnoses relying on direct interviews tion of a major mood disorder during an uninterrupted period of SZA. and family history reports for a series of diagnoses in adults26,27 and A sample of inpatients (10%) and outpatients (90%) were recruited from children.28 These analyses as well as similar analyses for mood diagnosis the orthopedic departments of hospitals in Lausanne and Geneva during have also allowed us to lower the diagnostic thresholds for disorders in the same time period to serve as a comparison or control group (n ¼ 110). order to minimize information bias due to the low sensitivity of the family Inclusion and exclusion criteria were the same as for the mood disorder history method. DSM-IV diagnoses were assigned according to a best- probands with the exception of the lack of a history of a mood or psychotic estimate procedure,29 which was based on review of information from disorder. Non-psychiatric medical controls were included in order to direct interviews, family history reports and medical records for subjects control for the inpatient setting of proband recruitment. Orthopedic who had been treated for a psychiatric disorder. The diagnoses of non- settings were selected for recruitment of controls because of their similar interviewed family members were based on all available family history age composition, and orthopedic conditions have not been shown to be reports based on the FH-RDC. differentially associated with mood disorders whereas disorders from most other specialties have high rates of comorbidity with mood disorders. There were a total of 1734 adult first-degree relatives (parents, siblings Data analysis and offspring 418 years) of the probands on whom diagnostic information Analyses were conducted for (1) specific DSM-IV major mood disorders, could be collected. This research project was approved by the local and (2) major mood or psychotic episodes independently of the disorder institutional review board. All participants gave written informed consent for during which they occurred. Univariate between-group analyses were their participation before the assessments. There were very few differences performed using chi-square tests or analysis of variance as appropriate. In between the patients who participated in the family study compared with order to simultaneously assess the effects of the probands’ mood disorders those who did not. Patients with mood disorders who had at least one or mood/psychotic episode types on the risk of disorders or episodes in participating adult relative (N ¼ 403) did not differ from those with the same family members, generalized linear mixed models (generalized estimating 30 mood disorders who had no participating family member (N ¼ 261) equation procedure ) were applied. These models account for the lack of regarding sex, age or the presence of comorbid alcohol or drug use independence of the observations (varying number of relatives across disorders. However, BPD-I patients with participating relatives were more families). All models were adjusted for the effects of sex and age in likely to have higher socio-economic status than the other BPD-I patients. relatives as well as for the effects of comorbid anxiety disorders (including Moreover, MDD patients with participating relatives were more likely to generalized anxiety disorder, agoraphobia, panic disorder and/or social suffer from comorbid anxiety disorders than the other MDD patients. phobia), alcohol abuse/dependence and drug abuse/dependence (marijuana, cocaine, narcotics, solvents, hallucinogens, stimulants and sedatives) in both probands and relatives. Statistical analyses were Procedures performed using the Statistical Analysis System, version 9.2 (SAS Institute, Inc., Cary, NC, USA). Participants were interviewed by masters-level psychologists or psychiatrists who completed intensive training over a 3-month period. Training included supervision of videotaped interviews by clinically experienced senior psychologists. Interviewers were blind to the disease status of the RESULTS family members in any given family. Diagnostic information on probands The demographic characteristics of the probands and the relatives and interviewed first-degree relatives was obtained using the Diagnostic as well as the prevalence of comorbid anxiety and SUD are 21 Interview for Genetic Studies (DIGS, ). There were high kappa coefficients provided in Table 1. The probands did not differ by either age for inter-rater reliability and slightly lower coefficients for test–retest (mean age of all probands ¼ 40.0 years, s.d. ¼ 11.9 years) or sex reliability for major Axis-I diagnoses, including major mood and SAF (48.6% females overall). However, a greater proportion of disorders22 as well as substance use disorders (SUD;23) on the French translation of the DIGS.24 The only disorder with poor test–retest reliability probands with mood disorders had a lifetime history of comorbid was BPD-II. anxiety and SUD compared with controls. Among the 100 BPD-I Family history information on probands and all first-degree relatives probands, 58% reported psychotic symptoms and 13 (13%) had was collected from all participants using the Family History-Research manic episodes without a history of a MDE (that is, unipolar Diagnostic Criteria (FH-RDC25), which allowed us to obtain diagnostic mania). Among the 62 SAF probands, 7 (11.3%) had only

Molecular Psychiatry (2014), 209 – 213 & 2014 Macmillan Publishers Limited Speciﬁcity of psychosis, mania and major depression CL Vandeleur et al 211

Table 2. Rates (%) and risk (OR, 95% CIa) of lifetime psychotic and mood disorders in relatives (N ¼ 1734) by proband disorder subgroup

Proband disorder

Relatives Schizoaffective Bipolar I Bipolar II MDD Controlsb

N 251 432 98 460 493

Relative % OR (95% CI) % OR (95% CI) % OR (95% CI) % OR (95% CI) % OR disorder

Schizoaff 2.8 6.4* (1.2; 33.5) 1.4 2.7 (0.6; 12.1) 0 — — 0.7 1.7 (0.3; 8.6) 0.8 Ref. Bipolar I 2.8 4.6* (1.2; 17.8) 6.9 12.2*** (4.2; 35.4) 2.0 2.4 (0.5; 13.0) 1.1 1.4 (0.4; 5.4) 0.4 Ref. Bipolar II 4.4 3.0* (1.0; 8.8) 2.3 2.0 (0.8; 5.0) 4.1 2.6 (0.8; 8.6) 2.4 1.6 (0.5; 4.5) 1.4 Ref. MDD 27.1 1.4 (0.9; 2.1) 22.9 1.2 (0.9; 1.8) 35.7 2.1* (1.2; 3.6) 38.9 2.4*** (1.7; 3.4) 19.5 Ref.

Abbreviations: CI, confidence interval; MDD, major depressive disorder; OR, odds ratio. *Po0.05; ***Po0.001. aModels adjusted for sex and age in relatives and anxiety disorders (generalized anxiety disorder, agoraphobia, panic disorder and/or social phobia), alcohol abuse/dependence and drug abuse/dependence in both probands and relatives. bThis category includes relatives of probands with subthreshold bipolar disorder, cyclothymia, hyperthymic personality, subthreshold depression, dysthymia without MDD, depressive personality without MDD, anxiety disorders, and alcohol or drug abuse/dependence when recruited to the orthopedic departments; Ref. ¼ reference group. experienced manic episode(s) but never a MDE. Table 1 also episodes in relatives. There was also a high rate of psychosis reveals that the relatives of the probands did not differ by either among the relatives of probands with mania, although this risk sex (51.7% female overall) or age (mean age of all relatives ¼ 50.1 was significantly diminished in the generalized linear mixed model years, s.d. 18.2 years), but the proportion of those with a direct after adjustment for comorbid mania in relatives. interview differed across groups. Indeed, the proportion of The non-hierarchical adjusted analyses also revealed a strong interviewed relatives was highest in the families of probands specific familial aggregation of manic episodes. In addition, there with SAF, followed by those of probands with BPD-I, BPD-II, MDD was a high rate of mania among the relatives of probands with and controls. psychosis, but this risk diminished to 1.0 when manic episodes in Table 2 provides the lifetime prevalence of DSM-IV mood probands were simultaneously entered into the generalized disorders in relatives according to the proband’s disorder status as linear mixed model. The risk of hypomania was not significantly well as the results from generalized linear mixed models that increased in the relatives of probands with any type of episode assessed the associations between disorders in probands and their compared with the relatives of controls according to the first-degree relatives, adjusting for demographic characteristics in generalized linear mixed models. Finally, the risk of MDE was relatives, anxiety and SUD in both probands and relatives and significantly elevated only among the relatives of probands with within-family correlations. The risk of SAF was largely increased in MDE after adjustment for demographic variables in relatives, the relatives of probands with the same disorder but not in the comorbid disorders in probands and relatives and for psychotic, relatives of probands with other mood disorders. Furthermore, the manic and hypomanic episodes in relatives. risk of BPD-I was largely increased in the relatives of probands Although the crude rates revealed a significant increase of MDE exhibiting the same disorder and significantly increased in the among the relatives of probands with all the other episode types relatives of probands with SAF. The risk of BPD-II was significantly as compared to those of probands with no mood episodes, this elevated only in relatives of probands with SAF but not in those of risk was almost entirely attributable to strong intra-individual probands with BPD-II or any other disorder. Moreover, the risk of associations between MDE and comorbid mania/hypomania in MDD was approximately doubled in the relatives of probands with these relatives. Additional analyses showed that, within the BPD-II and MDD. Additional analyses showed that the lack of an families of manic and psychotic probands, the relatives who association between MDD and BPD-I was independent of the exhibited mania or hypomania themselves revealed much higher presence of psychotic symptoms in BPD-I probands as the rates of rates of MDE than the other relatives of these probands. Similarly, MDD were almost identical (23.2% vs 22.5%) for relatives of BPD-I within the families of probands with hypomania, MDE and controls probands with and without psychotic features. the relatives affected with mania or hypomania themselves Table 3 presents the rates of specific DSM-IV mood and revealed very high rates of MDE as compared to relatives of these psychotic episodes in relatives by type of episode in probands as probands who had no comorbid mania/hypomania. well as the results of the generalized linear mixed models that assessed the associations between these episode types in probands and relatives. As psychotic, manic, hypomanic and DISCUSSION depressive episodes were analyzed in a non-hierarchical way, the To our knowledge, this is the first study to assess the familial number of relatives of probands with MDE included those of the aggregation patterns of the full spectrum of mood disorders, as probands with MDD and BPD-II as well as those of the probands well as their major components within the same sample, using with BPD-I and SAF who had also experienced MDE. Moreover, contemporary family study methods and classification systems. given that subjects with hypomania but no history of MDE were The most salient finding of this paper is the familial specificity of not excluded from the controls in the previous disorder-based psychotic, manic and MDEs, without evidence for shared familial analyses, the number of relatives of probands with hypomania risk across these manifestations of mood disorders. This suggests (Table 3) was larger than that of relatives of probands with BPD-II that there are distinct underlying etiological pathways for the disorder (Table 2). major subtypes of mood disorders that should be pursued in First, the risk of psychotic episodes was increased almost three future research on these disorders. times in the relatives of probands with psychosis, after adjustment The results of the present study confirm those of previous for demographic variables in relatives, comorbid disorders in family studies regarding the specificity of familial aggregation probands and relatives and for manic, hypomanic and depressive of psychosis,3–6 BPD,3,4,11 and major depression.11,12 However, our

& 2014 Macmillan Publishers Limited Molecular Psychiatry (2014), 209 – 213 Speciﬁcity of psychosis, mania and major depression CL Vandeleur et al 212

Table 3. Rates (%) and risk (OR, 95% CI)a of psychotic and mood episode types in relatives (N ¼ 1734) by proband episode type

Proband episode type

Psychosis Mania Hypomania MDE Other/nob

Relative N 505 628 113 1182 493

Relative episode type % OR (95% CI) % OR (95% CI) % OR (95% CI) % OR (95% CI) % OR (95% CI)

Psychosis 5.9 2.9* (1.1; 7.7) 5.1 0.9 (0.3; 2.6) 0.9 0.4 (0.1; 3.5) 3.6 1.0 (0.4; 2.4) 1.6 Ref. Mania 5.9 1.0 (0.4; 2.7) 6.2 6.4*** (2.2; 18.7) 1.8 1.6 (0.3; 7.3) 3.9 1.7 (0.5; 5.9) 0.4 Ref. Hypomania 5.7 1.2 (0.4; 3.3) 5.6 2.1 (0.8; 6.0) 4.4 1.3 (0.5; 3.6) 4.2 0.8 (0.4; 1.7) 2.4 Ref. MDE 35.3 1.0 (0.7; 1.4) 33.4 0.7 (0.5; 1.1) 39.8 1.0 (0.7; 1.6) 38.1 2.0*** (1.5; 2.7) 22.1 Ref.

Abbreviations: CI, conﬁdence interval; MDE, major depressive episode; OR, odds ratio. *Po0.05; ***Po0.001. aModels adjusted for sex and age in relatives and anxiety disorders (generalized anxiety disorder, agoraphobia, panic disorder and/or social phobia), alcohol abuse/dependence and drug abuse/dependence in both probands and relatives. bIncludes relatives of probands with subthreshold bipolar disorder, cyclothymia, hyperthymic personality, subthreshold depression, dysthymia without major depressive disorder (MDD), depressive personality without MDD, anxiety disorders and alcohol or drug abuse/ dependence when recruited to the orthopedic departments; Ref. ¼ reference group.

results also demonstrate the independence of the familial large efforts to directly interview all first-degree relatives, a transmission of mania and depression, providing a striking substantial proportion of them were not available for such an replication of the recent findings of Merikangas et al. (in press) assessment, and the diagnoses of these relatives needed to be that have critical implications for future treatment, biological and assigned according to family history reports. In addition, the molecular genetic studies of mood disorders. The substantial proportion of interviewed relatives varied across study subgroups. independence of mania and depression in the Maudsley twin However, we have minimized the risk of differential bias due to series31 suggests that the familial independence in the present differential participation of relatives by applying an extensively study is likely to tap differential genetic risk factors that contribute validated diagnostic best-estimate procedure that adjusted the to manic and MDEs. The lack of evidence for a common diathesis diagnostic threshold for specific disorders in function of the source underlying the ‘pure’ forms of these disorders9,17,32 supports the of available information.26,27 Second, the evaluation of subjects conclusion of a recent comprehensive review regarding the was cross-sectional, which may have diminished the accuracy of independence of BPD-I, SAF and SZA based on evidence for collected diagnostic data due to the risk of incomplete recall of differential treatment response and neuroimaging findings.33 The symptoms. Third, the study was based on treated probands who independence of psychosis from mania and major depression exhibited substantial comorbidity with anxiety and SUD, which further suggests that independent characterization of these limits the generalizability of our results to less severe forms of domains of cognitive and mood disturbances may be more mood disorders in the community. Fourth, our sample had fairly valid than the use of the categories of BPD and SZA specified in small numbers of family members of probands in the bipolar-II current diagnostic systems. diagnostic group, thereby limiting our ability to investigate the Although the present report focused on psychotic episodes that familial clustering of this disorder. Fifth, the small number of occurred within the bounds of SAF and other mood disorders, unipolar manic patients in this treatment sample reduced our inclusion of a small sample of families of patients with SZA yielded ability to study the independence of mania and depression.35 similar findings for psychosis (odds ratio (OR) ¼ 2.3, Po0.05; CI: This work highlights the importance of combining multiple 1.0–5.3), as well as for mania (OR=6.8, CI: 2.5; 18.7) and MDEs sources of evidence regarding etiological factors underlying (OR=2.0, CI: 1.5; 2.6), with no evidence for cross transmission of disorders in the development of diagnostic nomenclature and mania and MDEs (OR ¼ 0.7, CI: 0.5–1.0), psychosis and mania studies of their treatment, course and risk factors. The debate (OR ¼ 1.0, CI: 0.4–2.4) or psychosis and MDEs (OR ¼ 1.0, CI: 0.7–1.4). regarding common versus specific diatheses cannot be resolved Therefore, our findings of independence of categories may also by any one approach, such as genome-wide association studies generalize to families of probands with ‘pure’ psychotic disorders. that implicate numerous common polygenic markers across The lack of specific familial clustering together with the poor disorders,2 particularly in light of the low relative and test–retest reliability of the BPD-II subgroup defined by the DSM- attributable risk of such markers. The specificity of treatment IV22 (and unchanged in the DSM-534) cast doubt on the validity of and neuroimaging findings suggest the utility of retaining the BPD-II subtype. Rather, the evidence of shared underlying distinctions between these conditions.33 However, it is likely that diatheses between BPD-II and MDD suggests that hypomanic the multifactorial etiology of these conditions will be based on mood episodes that alternate with episodes of major depression both unique and common environmental and genetic risk factors. are likely to represent a particular clinical manifestation of MDD Our findings demonstrate that continued dissection of the core rather than a subgroup of BPD. features of these conditions will be essential to our future This is the first contemporary family study of a large clinical understanding of their etiology, course and treatment. sample of probands with a broad range of manifestations of the mood disorder spectrum and with comprehensive diagnostic evaluations on probands and a large proportion of first-degree CONFLICT OF INTEREST relatives. Whereas previous studies largely adhered to the The authors declare no conflict of interest. diagnostic conventions in place at the time of study enrollment, our analyses of the key components of these conditions demonstrate the importance of suspension of arbitrary diagnostic ACKNOWLEDGMENTS distinctions. There are also several limitations of this study that This research was supported by five grants from the Swiss National Foundation (SNF: should be considered in interpreting these findings. First, despite No.3200-040677, No.32003B-105969 and No.32003B-118326 to FF; No.3200-049746

Molecular Psychiatry (2014), 209 – 213 & 2014 Macmillan Publishers Limited Specificity of psychosis, mania and major depression CL Vandeleur et al 213 and No.3200-061974 to MP), and a grant from GlaxoSmithKline Clinical Genetics. The 16 Kendler KS, McGuire M, Gruenberg AM, O’Hare A, Spellman M, Walsh D. The funders had no involvement in any aspect of this study. We thank the participants Roscommon Family Study. I. Methods, diagnosis of probands, and risk of schi- and the collaborators who contributed to the coordination of the study and the zophrenia in relatives. Arch Gen Psychiatry 1993; 50: 527–540. collection of data. Special thanks to Professor Pierre-François Leyvraz, Dr Nicolas 17 Craddock N, Owen MJ. The beginning of the end for the Kraepelinian dichotomy. Favarger and Professor Daniel Egloff from the Orthopedic Department in Lausanne as Br J Psychiatry 2005; 186: 364–366. well as to Professor Pierre Hoffmeyer from the Orthopedic Department in Geneva for 18 Spitzer RL, Endicott J, Robins E. Research diagnostic criteria: rationale and relia- their help with the recruitment of the comparison subjects of this study. The National bility. Arch Gen Psychiatry 1978; 35: 773–782. Institute of Mental Health Intramural Program also contributed to this study. The 19 Weissman MM, Merikangas KR, Wickramaratne P, Kidd KK, Prusoff BA, Leckman JF views expressed herein are those of the author and do not represent those of the US et al. Understanding the clinical heterogeneity of major depression using family federal government. data. Arch Gen Psychiatry 1986; 43: 430–434. 20 Faraone SV, Blehar M, Pepple J, Moldin SO, Norton J, Nurnberger JI et al. Diag- nostic accuracy and confusability analyses: an application to the Diagnostic REFERENCES Interview for Genetic Studies. Psychol Med 1996; 26: 401–410. 1 Berrettini W. Bipolar disorder and schizophrenia: convergent molecular data. 21 Nurnberger Jr JI, Blehar MC, Kaufmann CA, York-Cooler C, Simpson SG, Harkavy- Neuromol Med 2004; 5: 109–117. Friedman J et al. Diagnostic interview for genetic studies. Rationale, unique 2 Cross-Disorder Group of the Psychiatric Genomics CSmoller JW, Craddock N, features, and training. NIMH Genetics Initiative. Arch Gen Psychiatry 1994; 51: Kendler K, Lee PH, Neale BM et al. Identification of risk loci with shared effects 849–859, discussion 863-844. on five major psychiatric disorders: a genome-wide analysis. Lancet 2013; 381: 22 Preisig M, Fenton BT, Matthey ML, Berney A, Ferrero F. Diagnostic interview for 1371–1379. genetic studies (DIGS): inter-rater and test-retest reliability of the French version. 3 Maier W, Lichtermann D, Minges J, Hallmayer J, Heun R, Benkert O et al. Continuity Eur Arch Psychiatry Clin Neurosci 1999; 249: 174–179. and discontinuity of affective disorders and schizophrenia. Results of a controlled 23 Berney A, Preisig M, Matthey ML, Ferrero F, Fenton BT. Diagnostic interview for family study. Arch Gen Psychiatry 1993; 50: 871–883. genetic studies (DIGS): inter-rater and test-retest reliability of alcohol and drug 4 Gershon ES, DeLisi LE, Hamovit J, Nurnberger Jr. JI, Maxwell ME, Schreiber J et al. A diagnoses. Drug Alcohol Depend 2002; 65: 149–158. controlled family study of chronic psychoses. Schizophrenia and schizoaffective 24 Leboyer M, Barbe B, Gorwood P, Teherani M, Allilaire JF, Preisig M et al. Interview disorder. Arch Gen Psychiatry 1988; 45: 328–336. Diagnostique Pour les Etudes Genetiques. INSERM: Paris, France, 1995. 5 Kendler KS, McGuire M, Gruenberg AM, O’Hare A, Spellman M, Walsh D. The 25 Andreasen NC, Endicott J, Spitzer RL, Winokur G. The family history method Roscommon Family Study. IV. Affective illness, anxiety disorders, and alcoholism using diagnostic criteria: reliability and validity. Arch Gen Psychiatry 1977; 34: in relatives. Arch Gen Psychiatry 1993; 50: 952–960. 1229–1235. 6 Maier W, Lichtermann D, Franke P, Heun R, Falkai P, Rietschel M. The dichotomy of 26 Rougemont-Buecking A, Rothen S, Jeanpretre N, Lustenberger Y, Vandeleur CL, schizophrenia and affective disorders in extended pedigrees. Schizophr Res 2002; Ferrero F et al. Inter-informant agreement on diagnoses and prevalence estimates 57: 259–266. of anxiety disorders: direct interview versus family history method. Psychiatry Res 7 Maier W, Lichtermann D, Minges J, Heun R, Hallmayer J, Benkert O. Schizoaffective 2008; 157: 211–223. disorder and affective disorders with mood-incongruent psychotic features: keep 27 Vandeleur CL, Rothen S, Jeanpretre N, Lustenberger Y, Gamma F, Ayer E et al. separate or combine? Evidence from a family study. Am J Psychiatry 1992; 149: Inter-informant agreement and prevalence estimates for substance use 1666–1673. disorders: direct interview versus family history method. Drug Alcohol Depend 8 Mortensen PB, Pedersen CB, Melbye M, Mors O, Ewald H. Individual and familial 2008; 92: 9–19. risk factors for bipolar affective disorders in Denmark. Arch Gen Psychiatry 2003; 28 Rothen S, Vandeleur CL, Lustenberger Y, Jeanpretre N, Ayer E, Gamma F et al. 60: 1209–1215. Parent-child agreement and prevalence estimates of diagnoses in childhood: 9 Lichtenstein P, Yip BH, Bjork C, Pawitan Y, Cannon TD, Sullivan PF et al. Common direct interview versus family history method. Int J Methods Psychiatr Res 2009; 18: genetic determinants of schizophrenia and bipolar disorder in Swedish families: a 96–109. population-based study. Lancet 2009; 373: 234–239. 29 Leckman JF, Sholomskas D, Thompson WD, Belanger A, Weissman MM. Best 10 Van Snellenberg JX, de Candia T. Meta-analytic evidence for familial coaggrega- estimate of lifetime psychiatric diagnosis: a methodological study. Arch Gen tion of schizophrenia and bipolar disorder. Arch Gen Psychiatry 2009; 66: 748–755. Psychiatry 1982; 39: 879–883. 11 Gershon ES, Hamovit J, Guroff JJ, Dibble E, Leckman JF, Sceery W et al. 30 Liang KY, Zeger SL. Longitudinal data analysis using generalized linear models. A family study of schizoaffective, bipolar I, bipolar II, unipolar, and normal control Biometrika 1986; 73: 13–22. probands. Arch Gen Psychiatry 1982; 39: 1157–1167. 31 McGuffin P, Rijsdijk F, Andrew M, Sham P, Katz R, Cardno A. The heritability of 12 Tsuang MT, Winokur G, Crowe RR. Morbidity risks of schizophrenia and affective bipolar affective disorder and the genetic relationship to unipolar depression. disorders among first degree relatives of patients with schizophrenia, mania, Arch Gen Psychiatry 2003; 60: 497–502. depression and surgical conditions. Br J Psychiatry 1980; 137: 497–504. 32 Craddock N, Owen MJ. The Kraepelinian dichotomy—going, going... but still not 13 Taylor MA, Berenbaum SA, Jampala VC, Cloninger CR. Are schizophrenia and gone. Br J Psychiatry 2010; 196: 92–95. affective disorder related? preliminary data from a family study. Am J Psychiatry 33 Cosgrove VE, Suppes T. Informing DSM-5: biological boundaries between bipolar I 1993; 150: 278–285. disorder, schizoaffective disorder, and schizophrenia. BMC Med 2013; 11:127. 14 Kendler KS, McGuire M, Gruenberg AM, Walsh D. Examining the validity of DSM- 34 Regier DA, Narrow WE, Clarke DE, Kraemer HC, Kuramoto SJ, Kuhl EA et al. DSM-5 III-R schizoaffective disorder and its putative subtypes in the Roscommon Family field trials in the United States and Canada, Part II: test-retest reliability of selected Study. Am J Psychiatry 1995; 152: 755–764. categorical diagnoses. Am J Psychiatry 2013; 170: 59–70. 15 Tsuang MT, Faraone SV, Fleming JA. Familial transmission of major affective 35 Solomon DA, Leon AC, Endicott J, Coryell WH, Mueller TI, Posternak MA et al. disorders. Is there evidence supporting the distinction between unipolar and Unipolar mania over the course of a 20-year follow-up study. Am J Psychiatry bipolar disorders? Br J Psychiatry 1985; 146: 268–271. 2003; 160: 2049–2051.

& 2014 Macmillan Publishers Limited Molecular Psychiatry (2014), 209 – 213