Molecular (2014) 19, 214–219 & 2014 Macmillan Publishers Limited All rights reserved 1359-4184/14 www.nature.com/mp

ORIGINAL ARTICLE Independence of familial transmission of and : results of the NIMH family study of disorders

KR Merikangas1, L Cui1, L Heaton1, E Nakamura1, C Roca1, J Ding2, H Qin3, W Guo3, Y Yao-Shugart3, C Zarate4 and J Angst5

The goal of this study is to investigate the familial transmission of the spectrum of in a nonclinical sample of probands with a broad range of manifestations of disorders. The sample included a total of 447 probands recruited from a clinically enriched community screening and their 2082 adult living and deceased first-degree relatives. A best estimate diagnostic procedure that was based on either direct semistructured interview or structured family history information from multiple informants regarding non-interviewed relatives was employed. Results revealed that there was specificity of familial aggregation of bipolar I (BP I; odds ratio (OR) ¼ 8.40; 3.27–20.97; h2 ¼ 0.83) and major depressive disorder (OR ¼ 2.26; 1.58–3.22; h2 ¼ 0.20), but not BP II. The familial aggregation of BP I was primarily attributable to the familial specificity of manic episodes after adjusting for both proband and relative comorbid and substance use disorders. There was no significant cross-aggregation between subtypes suggesting that the familial transmission of manic and major depressive episodes is independent despite the high magnitude of comorbidity between these mood states. These findings confirm those of earlier studies of the familial aggregation of bipolar disorder and major depression in the first nonclinical sample, and the largest family study of bipolar disorder in the USA using contemporary nonhierarchical diagnostic criteria for mood and anxiety disorders. The results suggest that these major components of bipolar disorder may represent distinct underlying pathways rather than increasingly severe manifestations of a common underlying diathesis. Therefore, dissection of the broad bipolar phenotype in genetic studies could actually generate new findings that could index novel biologic pathways underlying bipolar disorder.

Molecular Psychiatry (2014) 19, 214–219; doi:10.1038/mp.2013.116; published online 15 October 2013 Keywords: anxiety; bipolar; depression; family study; genetics; mania

INTRODUCTION Comprehensive reviews of aggregate data from numerous 13 The lack of rapid success of the application of advances in sources suggest that mania and depression may be etiologically molecular genetics in identifying genes underlying bipolar distinct despite their high co-occurrence. Our finding that major disorder has been disappointing in light of the compelling depressive episodes (MDE) were no more strongly associated with evidence for the high heritability of this condition from controlled manic episodes in mid- than with generalized anxiety family studies1–4 and twin studies.5–7 The slow progress and disorder, posttraumatic disorder or eating disorders in a nationally representative US sample of adolescents14 provided preliminary evidence for some common genetic variants across compelling rationale for distinguishing between mania and the diagnostic boundaries has led to renewed concern that the depression in our family study analyses. major impediment to progress may lie in the lack of validity of the 8 The goal of the present study is to examine the familial links current diagnostic classification systems. However, most of the between manic and depressive episodes in the NIMH Family Study evidence regarding the role of genes in bipolar disorder was 1,2,5,9 of Affective Spectrum Disorders, a community-based family study collected more than 25 years ago in clinical samples that may that evaluates the core features, diagnostic thresholds, subtypes not reflect the full spectrum of mood disorder in the general and boundaries of mood disorders with other mental and physical 10 population that is increasingly being included in genetic studies. disorders and their association with a comprehensive set of Application of the hierarchical diagnostic systems in these studies biologic measures as potential endophenotypes for mood precluded simultaneous classification of and bipolar disorders.15 The major aims of this paper are: disorder, mood disorders and anxiety disorders8,11 or behavior disorders, particularly deficit hyperactivity disorder.12 1. To investigate the familial transmission of mood disorder sub- Moreover, only a few of these studies investigated the specificity types (for example, bipolar I (BP I), BP II and MDD) and their of the familial aggregation of subtypes of bipolar disorder and association with anxiety disorders and substance use disorders. their relationship to major depressive disorder (MDD), with 2. To examine the specificity of manic, hypomanic and inconclusive results.1,3 depressive episodes in probands and relatives.

1Genetic Epidemiology Research Branch, Intramural Research Program, National Institute of , Bethesda, MD, USA; 2Computational Biology Core, Laboratory of Neurogenetics, National Institute of Aging, Bethesda, MD, USA; 3Unit on Statistical Genomics, Intramural Research Program, National Institute of Mental Health, Bethesda, MD, USA; 4Experimental Therapeutics and Pathophysiology Research Branch, National Institute of Mental Health, Bethesda, MD, USA and 5Department of Psychiatry, Zurich University Hospital, Zurich, Switzerland. Correspondence: Dr KR Merikangas, Genetic Epidemiology Research Branch, National Institute of Mental Health, Building 35, Room 1A201, 35 Convent Drive, MSC #3720, Bethesda, MD 20892, USA. E-mail: [email protected] Received 8 April 2013; revised 12 May 2013; accepted 24 May 2013; published online 15 October 2013 Familial transmission of mania and depression KR Merikangas et al 215 MATERIALS AND METHODS Table 1. N and demographic characteristics of probands (n ¼ 447) and Sample first-degree relatives by proband mood disorder diagnostic groups Probands. A total of 447 probands was recruited from a community (n ¼ 2082) screening of the greater Washington, DC metropolitan area (n ¼ 199), the National Institutes of Health (NIH) Clinical Center general volunteer referral Bipolar I Bipolar II MDD Controls core and local health newsletters and announcements (n ¼ 219) and from screens or participants in the NIMH Mood and Program Probands (n ¼ 29). N 62 66 162 157 The community sample was recruited through screening a sample of Sex (% male) 33.9 31.8 22.2 43.3 11 000 households within 50 miles of Washington, DC with a letter Age (years) followed by reminders by mail. Twenty percent of households that did not Mean 43.6 47.3 48.7 52.9 respond to any of the mailings were contacted by study recruiters to s.d. 11.9 12.7 13.2 17.7 Range 22–69 19–83 23–80 18–90 encourage their participation. Inclusion criteria were an ability to speak English, and availability and consent to contact at least two living first- Source (%) degree relatives. Clinical 12.9 13.6 5.6 1.9 Other 74.2 42.5 50.0 40.8 Community 12.9 43.9 44.4 57.3 Relatives. Seventy-three percent of the probands had at least one first- Comorbid disorders (%) degree adult relative with a diagnostic interview (n ¼ 559). The interviewed Anxiety disorder 79.0 75.8 61.7 11.5 relatives represent 71% of the first-degree relatives who were alive and Substance use 48.4 43.9 25.9 16.6 could be located, 73% of whom enrolled in the study. Family history disorder information was systematically collected from probands and interviewed Interviewed 64.5 74.2 74.1 76.4 relativesa relatives regarding a total of 1523 living and deceased adult first-degree relatives, yielding a total of 2082 first-degree relatives for these analyses. Relatives Multiple family history reports were available for 36% of the relatives. N 246 311 726 799 There was an average of one adult and offspring per proband. All Sex (% male) 45.1 50.8 45.5 48.1 analyses were controlled for interview status, and stratified models by Age (years) interview status were also run to test the stability of the results for Mean 51.8 53.3 54.2 55.4 interviewed and non-interviewed relatives. s.d. 18.2 18.2 19.0 19.0 Range 18–92 18–96 18–99 18–99

Procedures Abbreviation: MDD, major depressive disorder. aProportion with at least one relative interviewed. Standard family study methodology was employed including direct interviews of probands and relatives by experienced clinicians, systematic enumeration of first-, second- and third-degree relatives including children and blind assessment of relatives.16 The study was approved by the mood disorders. A random intercept was included to test for the variability Combined Neuroscience IRB at the NIH. All subjects provided written among family members in each family. Adjusted models included the key informed consent. proband and relative mood episode subtype, sociodemographic factors including sex of the proband and sex and age of the relative, comorbid Diagnostic assessments. The NIMH Family Study Diagnostic Interview for anxiety and substance use disorders in probands and relatives, and Affective Spectrum Disorders (DIAS) was based on the adaptation of the methodologic variables including source of proband recruitment (clinic diagnostic interview used in our earlier family studies of anxiety disorders versus community/other) and interview status of relative in models that and substance use disorders at the Yale University School of included diagnoses of non-interviewed relatives obtained by the family Genetic Epidemiology Research Unit based on the SADS/DIGS.17,18 The history method because of the well-documented false negative rate. DIAS ascertains diagnostic criteria for current and lifetime DSM-IV-TR Analyses were completed using SAS 9.2 (SAS Institute, Cary, NC, USA). The heritability of mania, and major depression episodes disorders, but does not adhere to strict diagnostic criteria for skip-outs 23 based on frequency or duration at the probe level in order to capture based on the threshold liability model of Falconer was estimated using SOLAR software, San Antonio, TX, USA (Southwest Foundation for Bio- subthreshold phenomenology across the key domains of 24 for multiple diagnostic systems.19,20 Inter-rater reliability was excellent medical Research). Polygenic heritability estimates defined as the (intraclass coefficients exceeded 0.87 or above for all major diagnostic proportion of the total phenotypic variance explained by additive categories) and the DIAS detected all cases of mood disorders derived genetic (or familial) effects were calculated, adjusted for age and gender from structured clinical interview for DSM-IV interviews on the inpatient using the variance components approach implemented in the SOLAR package.25 Information on the covariance among relatives was used to unit of the NIH Clinical Center. The NIMH Family Study Family History 24 Interview (DIAS-FHX) was used to assess a family history of psychiatric estimate the polygenic (or additive genetic) component of variance. disorders. The DIAS-FHX was based on modifications of the family history interview from our previous family study research.17,18,21 The systematic inter-rater reliability of all diagnostic categories in the present study was RESULTS very high with an intraclass correlation of 0.87 and above across all Table 1 presents the demographic and clinical characteristics of diagnostic categories. probands and relatives by proband mood disorder subtypes. Best estimate diagnoses for this study were based on all available Proband mood disorders are presented in mutually exclusive information by a team of experienced clinicians (psychologists and a psychiatrist) using a best estimate procedure.22 The current analyses assess subgroups of the lifetime mood disorder subtypes. There were 62 BP I and II disorder subtypes and MDD as defined by DSM-IV, and then probands with BP I, 66 with BP II, 162 with MDD and 157 with non- distinguish mania and hypomania episodes and major depressive episodes mood disorders (anxiety, eating and substance use disorders) or independent of the bipolar disorder subtype. Anxiety disorders include no disorder (controls). The average age of the probands was 49 generalized anxiety disorder, disorder and social . The years, with more females than males across the diagnostic groups. sensitivity estimates of direct interview versus best estimate diagnoses About 14% of those with BP I or BP II were identified in clinical that also included family history information were: 62% for BP I, 75% for BP settings compared with 5.6% of those with MDD and 1.9% with II and 86% for MDD. This indicates the importance of family history other disorders or no disorders. More than half of the probands information for the diagnosis of BP I. had graduated from college. Consistent with comorbidity patterns in population-based samples, about 70% of the bipolar and 62% Statistical methods of the MDD probands met criteria for comorbid anxiety disorders, Mixed effects logistic regression, which includes both fixed and random and nearly half of the bipolar probands met criteria for substance effects, was used to evaluate the association between proband and relative use disorders. Direct interviews were conducted with at least one

& 2014 Macmillan Publishers Limited Molecular Psychiatry (2014), 214 – 219 Familial transmission of mania and depression KR Merikangas et al 216 Table 2. Lifetime prevalence rates and adjusted odds ratios (ORs)a of mood disorders in first-degree adult relatives by mood disorders in probands (n ¼ 2082)

Proband mood disorder diagnostic group

Controls (n ¼ 799) Bipolar I (n ¼ 246) Bipolar II (n ¼ 311) MDD (n ¼ 726)

Relatives (n ¼ 2082) % % OR (95% CI) % OR (95% CI) % OR (95% CI)

Bipolar I 1.6 20.7 8.40 (3.37, 20.97) 6.1 1.61 (0.62, 4.14) 3.3 1.07 (0.46, 2.46) Bipolar II 1.9 4.9 1.79 (0.60, 5.35) 4.8 2.55 (0.93, 7.00) 3.7 1.84 (0.80, 4.22) MDD 15.3 26.0 1.34 (0.81, 2.23) 31.2 2.24 (1.43, 3.52) 30.3 2.26 (1.58, 3.22) Abbreviations: CI, confidence interval; MDD, major depressive disorder. aControls for sex, source of ascertainment, and anxiety and substance use disorders in probands, and sex, age, interview status, and anxiety and substance use disorders in relatives.

Figure 1. (a) Transmission of mania, hypomania and major depression in probands and relatives (n ¼ 2082). (b) Cross-transmission of mania, hypomania and major depression in probands and relatives (n ¼ 2082).

additional relative in about 70% of the probands. The bottom half associated with all mood disorders subtypes in relatives. However, of the table shows the number of relatives of probands by mood proband anxiety (OR ¼ 1.46, 0.74–2.87 for BP I; OR ¼ 0.88, 0.43– disorder diagnostic groups. The average age of the relatives was 1.80 for BP II; OR ¼ 0.99, 0.72–1.36 for MDD) and substance use 54.2 years, with an approximately equal sex distribution across the disorders (OR ¼ 1.10, 0.64–1.89 for BP I; OR ¼ 0.77, 0.39–1.52 for BP proband groups. II; OR ¼ 0.87, 0.65–1.18 for MDD) were not associated with Table 2 shows the lifetime prevalence rates of mood disorder elevated rates of any of the mood disorder subtypes in relatives. subgroups by mutually exclusive subgroups of proband mood All of these models also controlled for proband sex and source disorders and the adjusted odds ratios (ORs) derived from that were not significantly associated with any of the outcomes, as nonlinear mixed models of the associations between proband well as relative sex, age and interview status. Interview status was and relative mood disorder subtypes. The prevalence rates of strongly associated with all of the outcomes in these analyses. We mood disorders were elevated among relatives of probands with also controlled for proband major depressive episodes in the mood disorders compared with those of controls. The results of mania aggregation analyses and manic episodes in the depressive adjusted models revealed specificity of familial aggregation of the episode aggregation analyses and the significance of the subtypes of mood disorders, particularly for BP I (OR ¼ 8.40, 3.37– aggregation of mania and of depressive episodes remained 20.97) and MDD (OR ¼ 2.26, 1.58–3.22). Although there was a significant. We also ran the analyses controlling for proband trend for increased rates of BP II among relatives of BP II probands, education and did not find significant effects in any of the the association did not reach statistical significance (OR ¼ 2.55, models. 0.93–7.00). There was no significant degree of cross-aggregation Figure 1a shows the results of the analyses of the specificity of of BP I with either BP II or MDD in relatives, or between MDD and familial aggregation of mania, hypomania and major depressive BP I in relatives. However, there was a twofold increased rate of episodes after controlling for demographic characteristics, comor- MDD among the relatives of probands with BP II (OR ¼ 2.24, 1.43– bid disorders and interview status of relatives. There is a strong 3.52). Not shown, anxiety disorders (OR ¼ 5.11, 3.10–8.42 for BP I; association between proband and relative mania (OR ¼ 8.27, 3.82– OR ¼ 2.93, 1.63–5.25 for BP II; OR ¼ 3.71, 2.85–4.85 for MDD) and 17.91) and a less potent but still significant familial association substance use disorders (OR ¼ 2.93, 1.79–4.78 for BP I; OR ¼ 2.15, between major depression in probands and relatives (OR ¼ 2.45, 1.20–3.96 for BP II; OR ¼ 1.58, 1.19–2.10 for MDD) were strongly 1.66–3.62). The heritability coefficients for the mood disorder

Molecular Psychiatry (2014), 214 – 219 & 2014 Macmillan Publishers Limited Familial transmission of mania and depression KR Merikangas et al 217 subgroups that indicate the extent to which the variance is This work also illustrates the importance of cross-fertilization of attributable to familiality are also shown in Figure 1a. research across clinical and community samples with a develop- The heritability of each of the mood disorder subgroups was: mental perspective to provide a comprehensive portrait of mania ¼ 0.83 (s.e. ¼ 0.08, P ¼ 1.69E À 26), MDD ¼ 0.54 (s.e. ¼ 0.05, manifestations of these conditions. The critical evidence that led P ¼ 1.38E À 42) and hypomania ¼ 0.20 (s.e. ¼ 0.09, P ¼ 0.012), and us to distinguish between mania and depression emanated from a not shown, any bipolar ¼ 0.61 (s.e. ¼ 0.06, P ¼ 9.39E À 25); and BP large population study that investigated patterns of mania and II ¼ 0.20 (s.e. ¼ 0.10, P ¼ 0.001269). All of these coefficients depression close to their onset in the developmental course.14 demonstrate familiality, but the heritability of mania was greatest With the strong association between manic and depressive with the very small P-value and narrow confidence intervals. episodes, it is not surprising that bipolar disorder appears to be Figure 1b presents the cross-aggregation of mania, hypomania a unitary diagnostic entity in the clinical samples that have been and major depressive episodes in probands and relatives. There the basis for the establishment of diagnostic criteria in psychiatry. were no significant associations between pairs of proband and However, the general belief that unipolar mania is extremely rare, relative mood disorder subtypes, particularly MDE among relatives could actually be, in part, an artifact of clinical sampling.32 of probands with mania (OR ¼ 1.00) or between manic episodes This study has incorporated several features that advance our among relatives and MDE in probands (OR ¼ 0.67). understanding of the familial nature of bipolar disorder. First, the recruitment of probands from predominantly nontreatment settings minimizes the growing bias that characterizes specialty DISCUSSION treatment settings from which the bulk of family and genetic These findings confirm earlier studies of the familial aggregation of studies have recruited probands with bipolar disorder. The bipolar disorder and major depression1,2,9 in the first nonclinical community sampling also provided a broad range of severity of sample, and the largest family study of bipolar disorder in the USA mood and anxiety disorders as well as a representative sample of using contemporary nonhierarchical diagnostic criteria for mood unaffected people to serve as controls. Second, the inclusion of and anxiety disorders. The most important finding is that mania and people with a range of conditions enabled us to examine patterns major depression are largely transmitted independently in families, of comorbidity that had only been available in a few prior studies suggesting that these major components of bipolar disorder may that had directly recruited probands with the full range of mood represent distinct underlying pathways rather than increasingly and anxiety subtypes. Third, the diagnostic interview employed in severe manifestations of a common underlying diathesis. This work this study captures the full range of phenomenology of mood and has major implications for future genetic studies as well as relevance anxiety disorders without strict adherence to a particular for classification, etiology and treatment of mood disorders. diagnostic classification system to tap their core dimensional With regard to the mood disorder subtypes, the eightfold underpinnings and core components. association between manic episodes in probands and relatives Limitations include the lack of generalizability to more severely that was highly specific and heritable suggests that the familial affected persons with bipolar disorder with overlapping schizoaf- aggregation of bipolar disorder can be primarily attributed to fective and/or severe who were only detected rarely in mania. In fact, early descriptions of mania as a ‘functional distur- this community-based sample; the cross-sectional interviews that bance of brain with increased excitability’ manifest by increased generally underestimate prevalence rates of mental disorders; and speed of all processes, did not even consider a mood change as a despite extensive efforts to obtain direct interviews with all core feature.26 By contrast, the lack of familial clustering of enrolled relatives, there was still a substantial proportion about hypomania, the core feature of the BP II subtype, suggests that BP whom we had to rely on reports by family members. Another I and II disorders do not represent more manifestations of a potential limitation is the lack of reliability of hypomania that has common underlying liability with respectively more severe been a controversial subgroup particularly because of the manifestations that had been demonstrated in early studies of requirement of a lack of impairment, which is a rather counter- bipolar disorder.1,2 In fact, the BP II subtype appears to be more intuitive inclusion criterion for a psychiatric diagnostic category. closely aligned with major depression than with mania. These Our reliance on multiple family histories to report hypomania, effects were not moderated by either proband anxiety disorders changes in activity as well as the traditional elevated or irritable or substance use disorders, indicating their familial independence mood changes as an index of hypomania, and a conservative from mood disorders. Therefore, the requirement of major depre- approach to the diagnosis in the initial analyses was intended to ssion as a criterion for BP II but not for BP I in contemporary reduce false-positive diagnosis of this condition. Finally, the diagnostic nomenclature requires further investigation. independence of mania and depression may not apply to These findings highlight the importance of the distinction subgroups, such as the ‘atypical’ subtype of depression that may between mania and depression as separate processes in actually be a manifestation of the same underlying dimension as accordance with the conclusion of a recent comprehensive review that of mania. Such subgroup analyses and follow-up of the of evidence that bipolar disorder comprises two distinct yet highly sample are currently ongoing. comorbid syndromes.13 Evidence for this distinction is provided by With familial aggregation as the primary source of evidence different patterns of course, symptomatology, neurobiology and regarding the role of genetic factors in disease etiology, the results psychosocial risk factors. For example, manic and depressive provide compelling evidence that mania and major depression episodes were associated with differential levels of disability and should also be evaluated as distinct phenotypes in future genetic impairment in long-term follow-up studies of patients with mood studies. Although a family study cannot provide presumptive disorders, such as the 15- and 20-year follow-ups of the NIMH evidence that the independence of etiology of mania and Collaborative Study of Depression.27–29 Likewise, there is depression can be attributed to genetic factors, the heritability compelling evidence for differential treatment strategies for BP I estimate of 0.83 for mania in the present study is remarkably and II disorders.30 The high comorbidity between mania and major similar to the 0.85 heritability of mania estimated in the Maudsley depression could result from depression as a consequence of the Twin Registry study by McGuffin et al.6 If genes underlying mania excitatory processes associated with mania,31 differences in the are partially distinct from those for depression, current knowledge characteristics of depression associated with the underlying regarding the genetic pathways for bipolar disorder may reflect a diathesis of mania and bipolarity from those that give rise to missed opportunity to characterize the molecular basis of mania. major depression without mania or a third factor that may influence For example, the definition of the phenotype in many of the the expression of both mania and depression such as anxiety or samples included in the large collaborative genetic studies attentional disorders.13 of bipolar disorder include the full spectrum of bipolar disorders

& 2014 Macmillan Publishers Limited Molecular Psychiatry (2014), 214 – 219 Familial transmission of mania and depression KR Merikangas et al 218 33–35 (BP I, BP II and bipolar NOS) as affected cases. The indepen- 3 Heun R, Maier W. The distinction of bipolar II disorder from bipolar I and recurrent dence of mania and depression, as well as the lack of shared unipolar depression: results of a controlled family study. Acta Psychiatr Scand familial risk for the bipolar I and II subtypes, suggests that use of 1993; 87: 279–284. this broad phenotypic definition may mask potential genetic 4 Kendler KS, McGuire M, Gruenberg AM, O’Hare A, Spellman M, Walsh D. The associations with the most heritable subgroup of mania that was Roscommon Family Study. IV. Affective illness, anxiety disorders, and identified in this study. In fact, the heritability of any bipolar in relatives. Arch Gen Psychiatry 1993; 50: 952–960. 5 Bertelsen A, Harvald B, Hauge MA. Danish twin study of manic-depressive disorder of 0.61 in our study was substantially lower than that of disorders. Br J Psychiatry 1977; 130: 330–351. mania (0.83). Therefore, dissection of the broader bipolar 6 McGuffin P, Rijsdijk F, Andrew M, Sham P, Katz R, Cardno A. The heritability of phenotype in these studies could actually generate new findings bipolar affective disorder and the genetic relationship to unipolar depression. that could index novel biologic pathways underlying this Arch Gen Psychiatry 2003; 60: 497–502. condition. 7 Kendler KS, Pedersen NL, Neale MC, Mathe AA. A pilot Swedish twin study of This study also illustrates the value of the family study approach affective illness including hospital- and population-ascertained subsamples: (and its extension to twins) in the era of molecular genetics. results of model fitting. Behav Genet 1995; 25: 217–232. Despite the long history of information provided by family studies 8 Goldberg DP, Andrews G, Hobbs MJ. Where should bipolar disorder appear in the regarding the genetic architecture of Mendelian as well as meta-structure? Psychol Med 2009; 39: 2071–2081. 9 Maier W, Lichtermann D, Minges J, Hallmayer J, Heun R, Benkert O et al. Continuity the heterogeneity of complex diseases such as breast and and discontinuity of affective disorders and schizophrenia. Results of a controlled , the family study approach has largely been abandoned in family study. Arch Gen Psychiatry 1993; 50: 871–883. psychiatry in favor of very large case–control studies of diagnosed 10 Low NC, Cui L, Merikangas KR. Community versus clinic sampling: effect on the patients from clinical samples or registries. Yet, family studies still familial aggregation of anxiety disorders. Biol Psychiatry 2008; 63: 884–890. have an essential role in identifying cross-generational transmission 11 Goes FS, McCusker MG, Bienvenu OJ, Mackinnon DF, Mondimore FM, Schweizer B of phenotypes and genotypes. With their assumption of homo- et al. Co-morbid anxiety disorders in bipolar disorder and major depression: geneity of etiology within families, family and twin studies can be familial aggregation and clinical characteristics of co-morbid , social employed to identify the sources of disease heterogeneity because phobia, specific phobia and obsessive-compulsive disorder. Psychol Med 2012; 42: of the assumption of homogeneity of etiology within families, to 1449–1459. investigate mechanisms for comorbidity and boundaries of dis- 12 Wozniak J, Faraone SV, Mick E, Monuteaux M, Coville A, Biederman J. A controlled family study of children with DSM-IV bipolar-I disorder and psychiatric orders by distinguishing causal associations from common etiologic co-morbidity. Psychol Med 2010; 40: 1079–1088. factors underlying comorbid disorders and to discriminate the core 13 Cuellar AK, Johnson SL, Winters R. Distinctions between bipolar and unipolar components of the phenotype, phenotypic covariates (for example, depression. Clin Psychol Rev 2005; 25: 307–339. early age of onset) and/or traits or markers (that is, endopheno- 14 Merikangas KR, Cui L, Kattan G, Carlson GA, Youngstrom EA, J. Mania with types)36 that may represent intermediate forms of expression of and without depression in a community sample of adolescents. Arch Gen underlying genes and the broader disease phenotype. The latter Psychiatry 2012; 69: 943–951. approach has been widely adopted in the schizophrenia field as 15 Bearden CE, Freimer NB. Endophenotypes for psychiatric disorders: ready for demonstrated by studies such as the Collaborative Study on the primetime? Trends Genet 2006; 22: 306–313. Genetics of Schizophrenia37 that have identified neurocognitive 16 Weissman MM, Merikangas KR, John K, Wickramaratne P, Prusoff BA, Kidd KK. and neurophysiologic domains that may have genetic under- Family-genetic studies of psychiatric disorders. Developing technologies. Arch Gen Psychiatry 1986; 43: 1104–1116. pinnings. However, until recently, there are few comparable efforts 38 17 Merikangas KR, Stevens DE, Fenton B, Stolar M, O’Malley S, Woods SW et al. underway for affective disorders. Family-based studies will be of Co-morbidity and familial aggregation of alcoholism and anxiety disorders. particular relevance in distinguishing de novo from heritable Psychol Med 1998; 28: 773–788. structural variants as well as in interpreting results of the growing 18 Merikangas KR, Stolar M, Stevens DE, Goulet J, Preisig MA, Fenton B et al. data from next generation sequencing efforts. Familial transmission of substance use disorders. Arch Gen Psychiatry 1998; 55: 973–979. 19 Angst J, Gamma A, Neuenschwander M, Ajdacic-Gross V, Eich D, Rossler W et al. CONFLICT OF Prevalence of mental disorders in the Zurich Cohort Study: a twenty year prospective study. Epidemiol Psichiatr Soc 2005; 14: 68–76. Dr Angst has served in advisory or speaking capacities for AstraZeneca, Eli Lilly, 20 Angst J, Dobler-Mikola A, Binder J. The Zurich study—a prospective epidemio- Janssen Cilag, and Sanofi-Aventis. The remaining authors declare no conflict of logical study of depressive, neurotic and psychosomatic syndromes. I. Problem, interest. methodology. Eur Arch Psychiatry Neurol Sci 1984; 234: 13–20. 21 Andreasen NC, Rice J, Endicott J, Reich T, Coryell W. The family history approach to diagnosis. How useful is it? Arch Gen Psychiatry 1986; 43: 421–429. ACKNOWLEDGMENTS 22 Leckman JF, Sholomskas D, Thompson WD, Belanger A, Weissman MM. Best This work was supported by the Intramural Research Program of the National Institute estimate of lifetime psychiatric diagnosis: a methodological study. Arch Gen of Mental Health (Z01 MH002804). The authors wish to thank the following individuals Psychiatry 1982; 39: 879–883. who contributed to the development, implementation and analysis of the study: Shelli 23 Falconer DS. The inheritance liability to certain diseases, estimated from the Avenevoli, PhD, Christina Beccue, PsyD, Francesca Belouad, MA, Marcy Burstein, PhD, incidence among relatives. Ann Genetics 1965; 29: 51–76. Susan Goo, MSN, Jian-ping He, MSc, Suzan Khoromi, MD, Femke Lamers, PhD, 24 Almasy L, Blangero J. Multipoint quantitative-trait linkage analysis in general Tarannum Lateef, MD, MPH, Nancy CP Low, MD, Andrea Marques, MD, PhD, Vetisha pedigrees. Am J Hum Genet 1998; 62: 1198–1211. McClair, PhD, Suzanne McGarrity, PhD, Judith Mulvihill, LLD, Kelly Murphy, MSW, Anja 25 de Andrade M, Amos CI, Thiel TJ. Methods to estimate genetic components of Schmitz, PhD, Barbara Usher, PhD, Shreeram Srirangam, MD, April Szilagyi, MSW, variance for quantitative traits in family studies. Genet Epidemiol 1999; 17: 64–76. Shannon Winters, MSW and Jihui Zhang, MD, PhD. The views and opinions expressed 26 Mendel E. Die Manie: Eine Monographie. Urban & Schwarzenberg: Wien, Austria, in this article are those of the authors and should not be construed to represent the 1881. views of any of the sponsoring organizations, agencies, or US Government. 27 Coryell W, Turvey C, Endicott J, Leon AC, Mueller T, Solomon D et al. Bipolar I affective disorder: predictors of outcome after 15 years. J Disord 1998; 50: 109–116. 28 Judd LL, Akiskal HS, Schettler PJ, Endicott J, Leon AC, Solomon DA et al. REFERENCES Psychosocial disability in the course of bipolar I and II disorders: a prospective, 1 Gershon ES, Hamovit J, Guroff JJ, Dibble E, Leckman JF, Sceery W et al. A family comparative, longitudinal study. Arch Gen Psychiatry 2005; 62: 1322–1330. study of schizoaffective, bipolar I, bipolar II, unipolar, and normal control 29 Judd LL, Schettler PJ, Akiskal HS, Maser J, Coryell W, Solomon D et al. Long-term probands. Arch Gen Psychiatry 1982; 39: 1157–1167. symptomatic status of bipolar I vs. bipolar II disorders. Int J Neuropsychopharmacol 2 Tsuang MT, Winokur G, Crowe RR. Morbidity risks of schizophrenia and affective 2003; 6: 127–137. disorders among first degree relatives of patients with schizophrenia, mania, 30 Swartz HA, Thase ME. Pharmacotherapy for the treatment of acute bipolar II depression and surgical conditions. Br J Psychiatry 1980; 137: 497–504. depression: current evidence. J Clin Psychiatry 2011; 72: 356–366.

Molecular Psychiatry (2014), 214 – 219 & 2014 Macmillan Publishers Limited Familial transmission of mania and depression KR Merikangas et al 219 31 Koukopoulos A, Ghaemi SN. The primacy of mania: a reconsideration of mood 35 Chen DT, Jiang X, Akula N, Shugart YY, Wendland JR, Steele CJ et al. Genome-wide disorders. Eur Psychiatry 2009; 24: 125–134. association study meta-analysis of European and Asian-ancestry samples identi- 32 Solomon DA, Leon AC, Endicott J, Coryell WH, Mueller TI, Posternak MA et al. fies three novel loci associated with bipolar disorder. Mol Psychiatry 2011; 18: Unipolar mania over the course of a 20-year follow-up study. Am J Psychiatry 195–205. 2003; 160: 2049–2051. 36 Gottesman II, Gould TD. The endophenotype concept in psychiatry: etymology 33 Sklar P, Ripke S, Scott LJ, Andreassen OA, Cichon S, Craddock N et al. Large-scale and strategic intentions. Am J Psychiatry 2003; 160: 636–645. genome-wide association analysis of bipolar disorder identifies a new suscep- 37 Braff DL, Greenwood TA, Swerdlow NR, Light GA, Schork NJ. Advances in endo- tibility locus near ODZ4. Nat Genet 2011; 43: 977–983. phenotyping schizophrenia. World Psychiatry 2008; 7: 11–18. 34 Ferreira MA, O’Donovan MC, Meng YA, Jones IR, Ruderfer DM, Jones L et al. 38 Glahn DC, Curran JE, Winkler AM, Carless MA, Kent Jr. JW, Charlesworth JC et al. Collaborative genome-wide association analysis supports a role for ANK3 and High dimensional endophenotype ranking in the search for major depression risk CACNA1C in bipolar disorder. Nat Genet 2008; 40: 1056–1058. genes. Biol Psychiatry 2012; 71: 6–14.

& 2014 Macmillan Publishers Limited Molecular Psychiatry (2014), 214 – 219