The relationship between anxiety disorders and dimensional representations of DSM-IV personality disorders: A co-twin control study

• Published in: Journal of affective disorders 2016-01, Vol.190, p.349-356
• Main Authors: Welander-Vatn, A, Ystrom, E, Tambs, K, Neale, M.C, Kendler, K.S, Reichborn-Kjennerud, T, Knudsen, G.P
• Format: Article
• Language: English
• Subjects: Adult; Antisocial Personality Disorder - diagnosis; Anxiety disorders; Anxiety Disorders - diagnosis; Anxiety Disorders - epidemiology; Anxiety Disorders - genetics; Co-twin control design; Comorbidity; Cross-Sectional Studies; Diagnostic and Statistical Manual of Mental Disorders; Diseases in Twins - diagnosis; Diseases in Twins - epidemiology; Diseases in Twins - genetics; Female; Genetic epidemiology; Humans; Male; Multivariate Analysis; Norway; Paranoid Personality Disorder - diagnosis; Paranoid Personality Disorder - epidemiology; Paranoid Personality Disorder - genetics; Personality Disorders; Personality Disorders - diagnosis; Personality Disorders - epidemiology; Personality Disorders - genetics; Psychiatry; Regression Analysis; Risk Factors; Twin study; Twins, Dizygotic - psychology; Twins, Monozygotic - psychology
• ISSN: 0165-0327; 1573-2517
• DOI: 10.1016/j.jad.2015.09.038

Abstract Background There is substantial comorbidity between personality disorders (PDs) and anxiety disorders (ADs). Sharing of familial risk factors possibly explains the co-occurrence, but direct causal relationships between the disorders may also exist. Methods 2801 persons from 1391 twin pairs from the Norwegian Institute of Public Health Twin Panel were assessed for all DSM-IV PDs and ADs. Bivariate Poisson-regression analyses were performed to assess whether PDs predicted ADs at three different levels: All PDs combined, PDs combined within DSM-IV-clusters and each individual PD separately. Next, bivariate co-twin control analyses were executed within monozygotic (MZ) and dizygotic (DZ) twin pairs. A similar analytic strategy was employed in multivariate models including PDs as independent variables. Results PDs predicted ADs at all levels of analysis in bivariate regression models. Bivariate co-twin control analyses demonstrated an increased risk of ADs in all PDs combined, all PD-clusters and in schizotypal, paranoid, borderline, antisocial, avoidant and dependent PD. In the multivariate regression model, all PD-clusters and schizotypal, borderline, avoidant and obsessive-compulsive PD predicted ADs. Only borderline and avoidant PD predicted ADs in the multivariate co-twin control analysis. Limitations Over-adjustment may explain the results from the multivariate analyses. The cross-sectional study design hampers causal inference. Conclusions Comorbidity between ADs and PDs can be largely accounted for by shared familial risk factors. However, the results are also consistent with a direct causal relationship partly explaining the co-occurrence. Our results indicate specific environmental factors for comorbidity of ADs and borderline and avoidant PDs that are not shared with other PDs.