Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2019 Jan 1.
Published in final edited form as: J Affect Disord. 2017 Aug 15;225:395–398. doi: 10.1016/j.jad.2017.08.023

Quantitative genetic analysis of anxiety trait in bipolar disorder

J Contreras 1, E Hare 3, G Chavarría 2, H Raventós 1,2
PMCID: PMC5626643  NIHMSID: NIHMS902372  PMID: 28850853

Abstract

Background

Bipolar disorder type I (BPI) affects approximately 1% of the world population. Although genetic influences on bipolar disorder are well established, identification of genes that predispose to the illness has been difficult. Most genetic studies are based on categorical diagnosis. One strategy to overcome this obstacle is the use of quantitative endophenotypes, as has been done for other medical disorders.

Methods

We studied 619 individuals, 568 participants from 61 extended families and 51 unrelated healthy controls. The sample was 55% female and had a mean age of 43.25 (SD 13.90; range 18–78).

Heritability and genetic correlation of the trait scale from the Anxiety State and Trait Inventory (STAI) was computed by using the general linear model (SOLAR package software).

Results

we observed that anxiety trait meets the following criteria for an endophenotype of bipolar disorder type I (BPI): 1) association with BPI (individuals with BPI showed the highest trait score (F=15.20 [5,24], p=0.009), 2) state-independence confirmed after conducting a test-retest in 321 subjects, 3) co-segregation within families 4) heritability of 0.70 (SE: 0.060), p=2.33×10−14 and 5) genetic correlation with BPI was 0.20, (SE=0.17, p=3.12×10−5).

Limitations

Confounding factors such as comorbid disorders and pharmacological treatment could affect the clinical relationship between BPI and anxiety trait. Further research is needed to evaluate if anxiety traits are specially related to BPI in comparison with other traits such as anger, attention or response inhibition deficit, pathological impulsivity or low self-directedness.

Conclusions

Anxiety trait is a heritable phenotype that follows a normal distribution when measured not only in subjects with BPI but also in unrelated healthy controls. It could be used as an endophenotype in BPI for the identification of genomic regions with susceptibility genes for this disorder.

Keywords: Bipolar disorder, Endophenotype, Genetics, Heritability, Anxiety, Central Valley of Costa Rica

Introduction

Estimates of the prevalence of bipolar I disorder have ranged from 0.8% to 1.6% of the general population (Berns and Nemeroff, 2003). Although the genetic participation is well established, the identification of genes has remained elusive. Imprecision of the phenotype might explain the failure of genetic research to identify genes that contribute to susceptibility of BPI. Psychiatric disorders lack objective clinical and biological markers, and there are substantial disagreements on specific criteria to define diagnostic categories (Freedman et al., 2013). The need for a new approach to psychiatric genetics has led to the increasing popularity of endophenotypes (internal phenotypes that lie intermediate between the gene and the disease itself) (Gottesman and Shields, 1973). It is assumed that genes involved in endophenotypic variation are likely to represent more elementary phenomena than those involved in complex psychiatric diagnostic entities. It is also used interchangeably with the term ‘intermediate trait,’ describing a heritable quantitative phenotype believed to be closer in the chain of causality to the genes underlying the disease (Bearden and Freimer, 2006).

Many patients with BPI show anxiety symptoms that can be very disabling. Anxiety symptoms are caused by an interaction of biopsychosocial factors, including genetic vulnerability, stress, and trauma, which produce clinically significant syndromes. High comorbidity rates for anxiety have been documented (Shim et al., 2016). However, subclinical levels of anxiety (defined as anxiety trait) have also been associated with BPI (Mantere et al., 2008). These individuals have poor outcomes with longer, more frequent, and more difficult to treat mood episodes, are less responsive to lithium therapy, have earlier onset of symptoms and have greater functional impairment (McElroy et al., 2001).

Many candidate endophenotypes for BPI (e.g. neurocognitive functions, behavioral traits, sleep abnormalities) have been proposed (Gottesman and Gould, 2003; Hasler et al., 2006). As reported by Glahn et al., 2014, the number of genetic research using endophenotype approach has rapidly increased during the last decade. To our knowledge, none of these studies has identified a gene with a significant major effect on any psychiatric disorder. To date, few studies have assessed anxiety as a quantitative phenotype for bipolar disorder? (Wozniak et al., 2002).

Our hypothesis is that subclinical anxiety as a heritable trait genetically correlates with BPI. To determine whether quantitative anxiety symptomatology is a candidate endophenotype for BPI, we tested heritability of the trait (measurement of the general level of anxiety experienced over the lifetime) from the STAI in a sample of extended pedigrees from the Central Valley of Costa Rica (CVCR) with BPI disorder, and the genetic correlation between anxiety and BPI. We followed the same methodology of our pilot study (Contreras et al., 2010).

Methods

Participants

Subjects were originally recruited for the study Anxiety traits in Bipolar I Disorder in the Costa Rican population (NIMH 1R01TW008290-01A1). This study was reviewed and approved by the Institutional Review Board of the University of Costa Rica following the international guidelines for genetic research with human samples. The research project was explained to each participant before obtaining written consent.

The total study sample was 619 individuals from the Central Valley of Costa Rica (CVCR). The sample was comprised of 61 extended families (568 subjects, average family size 31 members, range: 4–41) and 51 unrelated healthy controls. Each family had at least one member diagnosed with BPI. The sample was 55% female and had a mean age of 43.25 (SD 13.90; range 18–78).

Diagnostic assessment

The subjects were diagnosed based on the diagnostic criteria of DSM-IV through a best estimation process (Leckman et al., 1982), utilizing clinical information obtained from the Diagnostic Interview for Genetic Studies (Nurnberger et al., 1994), a Family Interview for Genetic Studies (Maxwell, 1992) and psychiatric records. Final diagnoses were determined through a consensus process where two independent psychiatrists reviewed all available information. The Lifetime Dimensions of Psychosis Scale (LDPS) was used to capture affective and psychotic symptoms over the lifetime in addition to the consensus diagnoses (Levinson et al., 2002). LDPS was utilized as a proxy for bipolarity using the item that provide a quantitative measurement of lifetime mania.

Assessment of the anxiety trait

We measured subclinical anxiety in each subject regardless of the categorical diagnosis using the STAI. The STAI is a self-rated instrument that contains two subscales to measure anxiety. Each subscale has 20 items (4 response choices per item, higher scores indicate higher anxiety) (Spielberger et al., 1983). One scale measures state anxiety (i.e. the extent to which respondents experience anxiety symptoms at the time of measurement) (Vigneau and Cormier, 2008). The second scale measures trait anxiety (i.e. the extent to which respondents generally experience of anxiety symptoms as a behavioral trait). This instrument has been validated in Spanish (Rodrigo and Lusiardo, 1988).

The instrument was applied within the same month period after the initial psychiatric evaluation in 390 subjects, while 229 individuals were assessed with the STAI at a different time (later than a month after psychiatric interview) due to availability of the participants. We were also able to reassess 321 subjects after at least one year of the first evaluation of anxiety. The goal of this second evaluation was to test independency of the trait at a different moment of life.

Statistical analysis

We used a general linear model (GLM) to test endophenotype criteria (e.g. heritability, genetic correlation). Heritability and genetic correlation was assessed with variance component methods implemented in the SOLAR package software. We created an index representing mania based on the quantitative symptoms score from the item M-1 (duration × severity) of the LDPS as described before. We included in each model the covariates age, sex, the square of age, and interactions between age and sex, to allow for different age effects in males and females and non-linear change with age. Bivariate analyses provide genetic and environmental correlations as a means of examining how BPI and anxiety varies together. Such analysis allowed us to ask whether anxiety score correlated with mania in subjects with BPI.

To test association of anxiety trait with BPI, we examined multiple dependent, independent and covariate variables (e.g. age, sex). We tested the anxiety scores of individuals with BPI and their relatives to determine whether their scores are different compared to controls. The differences between groups would suggest an underlying genetic correlation between anxiety and BPI. This hypothesis was tested through GLM methods, modeling anxiety as function of genetic proximity to an affected individual (BPI > relatives with other psychiatric illness - different than BPI- > healthy relatives > healthy unrelated controls). All calculations of clinical and demographic variables were adjusted for age and gender by using the Statistical Package for the Social Sciences (SPSS) Software v.20.

Results

Sample characteristics

The primary DSMIV psychiatric diagnosis of the 568 subjects from the extended pedigrees were: BPI 151 (26%), major depressive disorder 59 (10%), specific phobia 10 (1%), panic disorder 11 (1%) and no axis I disorder 100 (18%). Out of the 61 families, 90% had one or more relatives with BPI, 77% had more than two members with BPI and 26%, four or more affected members. Out of the 618 participants, 340 (55%) were female. The mean age at interview was 43.25 (SD 13.90; range 18–78). The average dimensional index of lifetime mania (M-1 severity × duration of the LDPS) was 2.61. The average anxiety trait score was 28.15 (SD = 7.1) out of a maximum of 60. Detailed clinical and demographic characteristics of the sample are described in Contreras et al., 2014.

Heritability analysis

Since we observed in our pilot study that only anxiety trait is independent of clinical status in subjects with BPI, we restricted all the following analyses to anxiety trait without considering anxiety state. The heritability is 0.70 (SE: 0.060), p=2.33×10−14 and the genetic correlation between anxiety and BPI is 0.20, (SE=0.17, p=3.12×10−05, p: correlation is different from zero = 0.02 and p: correlation is different from one = 1.10×10−06).

As seen in Figure 1, subjects with BPI showed significantly higher anxiety trait scores than their healthy relatives and healthy controls (F=15.20 [5,24], p=0.009), (BPI patients> healthy relatives>unrelated healthy controls). Anxiety trait correlated with lifetime mania (LDPS M-1 duration x severity) r=0.30 (p<0.0001) after controlling for sex and age.

Figure 1.

Figure 1

Open in a new tab

Trait score by group

Of the 618 individuals, 390 (90%) were evaluated at the same time (within a month period after psychiatric evaluation), 62 (10%) met criteria for a current depressive syndrome (five or more depressive symptoms within the last two weeks of psychiatric assessment) and 18 (3%) for a current manic syndrome (elated mood + three or more manic symptoms or irritability + four or more manic symptoms within the last week of psychiatric evaluation). We found no significant difference in the trait scores regarding current mood status (currently depressed or currently manic), (F = 0.432, df = 1, p = 0.211) and (F=8.31, df=1, p=0.822) respectively.

To assess independency of anxiety trait from affective status of the subject with BPI (depressive or manic episode), we conducted a second evaluation of anxiety. We could contact and reevaluate 321 subjects after at least one year of the first evaluation of anxiety. We did not find significant difference in the score among BPI mood status (F=2.23, df=1, p=0.112).

Discussion

Studies aimed to understand the genetic architecture of psychiatric disorders have been of limited success when using the categorical diagnostic classification. Many researchers are focusing their research on endophenotypes, as defined by Gottesman and Gould, 2003. Endophenotypes, such as the quantitative anxiety score, could point the underlying genetically influenced biological processes in BPI and facilitate the identification of susceptibility genes. With the current categorical diagnostic system, anxiety symptoms are often under-diagnosed.

As in our previous study (Contreras et al., 2010), subjects with BPI showed the highest scores, followed by their relatives with other psychiatric disorders, healthy relatives and the healthy unrelated controls. These results are also supported by the research from Vázquez et al., (2008) who found that healthy relatives of BPI exhibit higher scores on the anxious temperament subscale of the TEMPS-A than controls. In a study by Evans et al. (2005), healthy relatives of BPI and healthy controls also showed differences on the hyperthymic scale and on the first extracted factor, anxious reactive of the TEMPS-A and TCI-125. All the above evidence is consistent with temperament and personality anxiety related traits as being part of the pathophysiology of BPI. The trait of the STAI represents another approximation to anxiety-related behavior.

One of our goals was to replicate our pilot work conducted in the population of the CVCR. In the first study, we assessed 300 subjects from 30 extended families. We observed that anxiety trait met criteria for an endophenotype in BPI. In that study, we did not conduct test-retest of the STAI, so we could not conclude if the euthymic, manic or depressed state of the BPI patients influenced the score. In the current work, with a larger sample and a second evaluation in a subgroup of families, we could confirm independency of the anxiety score to the mood status after one year of the first assessment of anxiety.

We observed a stair-step distribution of the scores in the family members based on their genetic proximity to individuals with BPI. The trait showed normal distribution across all the study sample that included BPI and healthy subjects (healthy relatives and healthy unrelated matched controls). The genetic correlation was significant between trait scores and quantitative lifetime mania whereas no correlation was observed between anxiety trait scores and lifetime depression. BPI is defined as manic episodes with or without depressive episodes but depressive episodes can also be part of a unipolar depression. The observed correlation between anxiety trait with mania, but not with depression, supports our initial hypothesis of anxiety trait being an endophenotype for BPI. Since we did not conduct a categorical assessment for anxiety disorders, we are not able to compare trait score among anxiety diagnoses. This finding reflects what is frequently seen in clinical settings, where patients with BPI often complain of unspecified anxiety symptoms that do not meet full criteria for a full-blown anxiety disorder. Likewise, many healthy relatives of patients with BPI complain of anxiety symptoms, sometimes perceived as personality anxiety related traits.

Heritability was statistically significant in this specific set of data, which suggests that part of the variation of the trait is influenced by genetic factors. The genetic correlation between BPI and anxiety was significantly different from zero between these scores, suggesting that both traits share genetic factors (overlapping genetic influence). Since confounding factors such as substance abuse, medical illness and medication history were not assessed in our study, these conclusions should be taken as preliminary.

In summary, anxiety trait meets the criteria, for being an endophenotype of BPI as outlined by Gottesman and Gould (2003). The trait is associated with BPI, is state independent (Reichenberg and Harvey, 2007), is heritable and co-segregates with BPI within families. The current report is a replication of the initial findings with a larger sample of extended and multigenerational families with one or more BPI from the CVCR. We confirmed each of the criteria for an endophenotype before running the heritability analysis. We assessed subclinical anxiety in healthy relatives with high genetic risk for BPI and healthy unrelated controls. The diagnosis of each subject was obtained through the best-estimate procedure, considered the gold standard in psychiatric genetic research, and fully tested by our research group (Contreras et al., 2009). All participants were from the Central Valley of Costa Rica.

Quantitative anxiety measures as an endophenotype may facilitate the identification of genes which predispose individuals to develop BPI. Further research to identify the genetic structure of this endophenotype will aid researchers to understand the pathophysiology underlying bipolar spectrum disorders and improve diagnosis and treatment of BPI in a more reliable and biologically valid manner than our current classification allows. The next steps of our study will be the identification of quantitative trait loci (QTLs), that may harbor candidate genes for the risk for BPI. Some lines of research could be oriented to study anxiety-like behaviors in animal models as done by some researchers including Slattery and Neumann, (2009). It is likely that some innate fear responses play a role in the evolution of humans’ mood.

Highlights.

In summary, anxiety trait meets the criteria, for being an endophenotype of BPI as outlined by Gottesman and Gould (2003). The trait is associated with BPI, is state independent (Reichenberg and Harvey, 2007), is heritable and co-segregates with BPI within families. The current report is a replication of the initial findings with a larger sample of extended and multigenerational families with one or more BPI from the CVCR. We confirmed each of the criteria for an endophenotype before running the heritability analysis. We assessed subclinical anxiety in healthy relatives with high genetic risk for BPI and healthy unrelated controls. The diagnosis of each subject was obtained through the best-estimate procedure, considered the gold standard in psychiatric genetic research, and fully tested by our research group (Contreras et al., 2009). All participants were from the Central Valley of Costa Rica.

Acknowledgments

Role of the Funding Source

This work was funded by the NIMH Genetics Initiative to the Research Project: Anxiety trait in Bipolar Disorder type I in the Costa Rican population (NIMH 1R01TW008290-01A1, PI: Javier Contreras Rojas). This grant provided financial support for the recruitment, clinical assessment of the subjects, genetic and statistical analysis.

Footnotes

Conflicts of interest of all authors

none

Disclosure

Al authors of this manuscript have no conflict of interest including any financial, personal or other relationships with other people or organizations within three (3) years of beginning the work submitted that could inappropriately influence, or be perceived to influence this work.

Contributors

Dr. Javier Contreras (first author) conducted the recruitment and clinical assessment of participants of the study. Dr. Contreras wrote the first draft of the manuscript. Dr. Elizabeth Hare (second author) was in charge of the statistical analysis together with Dr. Contreras. Dr. Gabriela Chavarría (third author) conducted part of the genetic analysis and she wrote part of the discussion together with the other authors. Dr. Henriette Raventos (fourth author) is the senior researcher and she was a key person en every step of the research. Dr. Raventos supervised the overall study methodology, write and edit the manuscript with the other authors. All authors have approved the final manuscript.

Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

References

  1. Almasy L, Blangero J. Multipoint quantitative-trait linkage analysis in general pedigrees. Am J Hum Genet. 1998;62(5):1198–211. doi: 10.1086/301844. [DOI] [PMC free article] [PubMed] [Google Scholar]
  2. Berns GS, Nemeroff CB. The neurology of bipolar disorder. Am J Med Genet C (Semin Med Genet) 2003;123C:76–84. doi: 10.1002/ajmg.c.20016. [DOI] [PubMed] [Google Scholar]
  3. Contreras J, Hare E, Pacheco A, Escamilla M, Raventos H. Is subclinical anxiety an endophenotype for bipolar I patients? A study from a Costa Rican sample. J Affect Disord. 2010 May;122(3):267–72. doi: 10.1016/j.jad.2009.07.017. [DOI] [PMC free article] [PubMed] [Google Scholar]
  4. Contreras-Rojas Javier, Raventós-Vorst Henriette. Caracterización clínica de familias costarricenses con trastorno afectivo bipolar. Acta méd costarric [online] 2014;56(4):167–173. [Google Scholar]
  5. Evans L, Akiskal HS, Keck PE, Jr, McElroy SL, Sadovnick AD, Remick RA, Kelsoe JR. Familiality of temperament in bipolar disorder: support for a genetic spectrum. J Affect Disord. 2005;85(1–2):153–168. doi: 10.1016/j.jad.2003.10.015. [DOI] [PubMed] [Google Scholar]
  6. Freedman R, Lewis DA, Michels R, Pine DS, Schultz SK, Tamminga CA, Gabbard GO, Gau SS, Javitt DC, Oquendo MA, Shrout PE, Vieta E, Yager J. The initial field trials of DSM-5: new blooms and old thorns. Am J Psychiatry. 2013 Jan;170(1):1–5. doi: 10.1176/appi.ajp.2012.12091189. [DOI] [PubMed] [Google Scholar]
  7. Glahn DC, Knowles EE, McKay DR, Sprooten E, Raventós H, Blangero J, Gottesman II, Almasy L. Arguments for the sake of endophenotypes: examining common misconceptions about the use of endophenotypes in psychiatric genetics. Am J Med Genet B Neuropsychiatr Genet. 2014 Mar;165B(2):122–30. doi: 10.1002/ajmg.b.32221. [DOI] [PMC free article] [PubMed] [Google Scholar]
  8. Gottesman II, Gould TD. The endophenotype concept in psychiatry: Etymology and strategic intentions. Am J Psychiatry. 2003;160:636–645. doi: 10.1176/appi.ajp.160.4.636. [DOI] [PubMed] [Google Scholar]
  9. Gottesman II, Shields J. Genetic theorizing and schizophrenia. Br J Psychiatry. 1973;122:15–30. doi: 10.1192/bjp.122.1.15. [DOI] [PubMed] [Google Scholar]
  10. Hasler G, Drevets W, Gould T, Gottesman I, Manji H. Toward constructing an endophenotype strategy for bipolar disorders. Biol Psychiatry. 2006;60(2):93–105. doi: 10.1016/j.biopsych.2005.11.006. [DOI] [PubMed] [Google Scholar]
  11. Leckman J, Sholomskas D, Thompson W, Belanger A, Weissman M. Best estimate of lifetime psychiatric diagnosis: a methodological study. Arch Gen Psychiatry. 1982;39:879–883. doi: 10.1001/archpsyc.1982.04290080001001. [DOI] [PubMed] [Google Scholar]
  12. Levinson DF, Mowry BJ, Escamilla MA, Faraone SV. The Lifetime Dimensions of Psychosis Scale (LDPS): Description and interrater reliability. Schizophr Bull. 2002;28:683–695. doi: 10.1093/oxfordjournals.schbul.a006972. [DOI] [PubMed] [Google Scholar]
  13. Mantere O, Suominen K, Valtonen HM, Arvilommi P, Isometsä E. Only half of bipolar I and II patients report prodromal symptoms. J Affect Disord. 2008;111(2–3):366–371. doi: 10.1016/j.jad.2008.03.011. [DOI] [PubMed] [Google Scholar]
  14. Maxwell M. Intramural Research Program. National Institute of Mental Health; Wasington, DC: 1992. Manual for the FIGS. [Google Scholar]
  15. McElroy SL, Altshuler LL, Suppes T, Keck PE, Jr, Frye MA, Denicoff KD, Nolen WA, Kupka RW, Leverich GS, Rochussen JR, Rush AJ, Post RM. Axis I psychiatric comorbidity and its relationship to historical illness variables in 288 patients with bipolar disorder. Am J Psychiatry. 2001;158:420–426. doi: 10.1176/appi.ajp.158.3.420. [DOI] [PubMed] [Google Scholar]
  16. Munoz R, Escamilla M. Diagnosis of schizophrenia in Latino populations: a comparison of direct interview and consensus based multi-source methods. J Nerv Ment Dis. 2009;197(7):530–535. doi: 10.1097/NMD.0b013e3181aac935. [DOI] [PubMed] [Google Scholar]
  17. Nurnberger JI, Jr, Blehar MC, Kaufmann CA, York-Cooler C, Simpson SG, Harkavy-Friedman J, Severe JB, Malaspina D, Reich T. Diagnostic interview for genetic studies. Rationale, unique features, and training. NIMH Genetics Initiative. Arch Gen Psychiatry. 1994;51:849–859. doi: 10.1001/archpsyc.1994.03950110009002. [DOI] [PubMed] [Google Scholar]
  18. Reichenberg A, Harvey PD. Neuropsychological impairments in schizophrenia: Integration of performance-based and brain imaging findings. Psychol Bull. 2007;133:833–858. doi: 10.1037/0033-2909.133.5.833. [DOI] [PubMed] [Google Scholar]
  19. Rodrigo G, Lusiardo M. Note on the reliability and concurrent validity of the Spanish version of the State–Trait Anxiety Inventory. Percept Mot Skills. 1988;67(3):926. doi: 10.2466/pms.1988.67.3.926. [DOI] [PubMed] [Google Scholar]
  20. Shim IH, Bae DS, Bahk WM. Anxiety or agitation in mood disorder with mixed features: A review with a focus on validity as a dimensional criterion. Ann Clin Psychiatry. 2016 Aug;28(3):213–20. [PubMed] [Google Scholar]
  21. Slattery DA, Neumann ID. Chronic icv oxytocin attenuates the pathological high anxiety state in female high anxiety-related behaviour rats. Neuropharmacology. 2009 doi: 10.1016/j.neuropharm.2009.06.038. [DOI] [PubMed] [Google Scholar]
  22. Spielberger CD, Gorsuch RL, Lushene R, Vagg PR, Jacobs GA. Manual for the State–Trait Anxiety Inventory (Form Y) Consulting Psychologists Press; Palo Alto, CA: 1983. [Google Scholar]
  23. Vázquez GH, Kahn C, Schiavo CE, Goldchluk A, Herbst L, Piccione M, Saidman N, Ruggeri H, Silva A, Leal J, Bonetto GG, Zaratiegui R, Padilla E, Vilapriño JJ, Calvó M, Guerrero G, Strejilevich SA, Cetkovich-Bakmas MG, Akiskal KK, Akiskal HS. Bipolar disorders and affective temperaments: a national family study testing the “endophenotype” and “subaffective” theses using the TEMPS-A Buenos Aires. J Affect Disord. 2008;108(1–2):25–32. doi: 10.1016/j.jad.2007.09.011. [DOI] [PubMed] [Google Scholar]
  24. Vigneau F, Cormier S. The factor structure of the State–Trait Anxiety Inventory: an alternative view. J Pers Assess. 2008;90(3):280–285. doi: 10.1080/00223890701885027. [DOI] [PubMed] [Google Scholar]
  25. Wozniak J, Biederman J, Monuteaux MC, Richards J, Faraone SV. Parsing the comorbidity between bipolar disorder and anxiety disorders: a familial risk analysis. J Child Adolesc Psychopharmacol. 2002;12(2):101–111. doi: 10.1089/104454602760219144. [DOI] [PubMed] [Google Scholar]

RESOURCES