Dose-response relationship between number of comorbid anxiety disorders in adolescent bipolar/unipolar disorders, and psychosis, suicidality, substance abuse and familiality

J Affect Disord. 2006 Dec;96(3):249-58. doi: 10.1016/j.jad.2006.07.008. Epub 2006 Aug 10.


Objectives: To ascertain rates of panic, obsessive-compulsive (OCD) and social phobic disorders among adolescents with bipolar disorder (BP), unipolar major depressive disorder (MDD) and psychiatric comparison patients, to assess their relationships to suicidality, psychosis, comorbidity patterns and familiality.

Methods: The first author (SCD) interviewed 313 Latino adolescents using a structured interview based on the SCID. Family history was ascertained by live interview or interview by proxy. Patients were classified as BP, MDD, or non-affectively ill comparison controls (CC). Data regarding suicidality and psychosis were collected. Regression analysis was used to test associations and control for confounding effects. Positive likelihood ratios were used to measure the dose-response relationships between number of anxiety disorders and measures of severity of illness and familial loading for affective illness.

Results: Of the total sample, 36.7% were BP, 44.7% MDD and 18.5% CC. In BP vs. MDD the odds of panic disorder were 4.4, of OCD 5.1, and of social phobia 3.3. MDD, in turn, were more likely to have these disorders than CC. BP (but not MDD) with panic disorder and social phobia, were more likely to have suicidal ideation; among the anxiety disorders, only social phobia was associated with having greater odds of suicide attempts. Among BP and MDD, patients with all three anxiety disorders were more likely to be psychotic. Presence of any mood disorder among first-degree relatives substantially increased the odds of having panic disorder and social phobia. The presence of one comorbid anxiety disorder increased the odds of having another. Finally, there were dose-response relationships between number of anxiety disorders and measures of severity of illness and familial loading for affective illness.

Limitations: Single interviewer using the SCID; cross sectional exploratory study.

Conclusions: BP adolescents have a greater anxiety disorder burden than their MDD counterparts. The results are compatible with the hypothesis that heavy familial-genetic loading for affective illness in juveniles is associated with bipolarity, cumulative anxiety disorder comorbidity, suicidality and psychosis. These observations are in line with pioneering psychopathologic observation in the early 1900s by two French psychiatrists, Gilbert Ballet and Pierre Kahn, who saw common ground between what until then had been considered the distinct categories of the neuroses and cyclothymic (circular) psychoses. This perspective has much in common with current complex genetic models of anxious diatheses in bipolar disorder.

MeSH terms

  • Adolescent
  • Anxiety Disorders / diagnosis
  • Anxiety Disorders / epidemiology*
  • Anxiety Disorders / genetics*
  • Bipolar Disorder / diagnosis
  • Bipolar Disorder / epidemiology*
  • Bipolar Disorder / genetics*
  • Cross-Sectional Studies
  • Female
  • Humans
  • Interviews as Topic
  • Male
  • Obsessive-Compulsive Disorder / diagnosis
  • Obsessive-Compulsive Disorder / epidemiology
  • Prevalence
  • Psychotic Disorders / diagnosis
  • Psychotic Disorders / epidemiology*
  • Psychotic Disorders / genetics*
  • Public Sector
  • Severity of Illness Index
  • Substance-Related Disorders / diagnosis
  • Substance-Related Disorders / epidemiology*
  • Substance-Related Disorders / genetics*
  • Suicide, Attempted / psychology*
  • Suicide, Attempted / statistics & numerical data*