Predictors of Conversion to Bipolar Disorder From Unipolar Depression

bipolar disorder
bipolar disorder
Investigators sought to determine identifiable predictors of conversion from unipolar depression to bipolar disorder.

Conversion from unipolar depression to bipolar disorder is more likely for individuals with a parental history of bipolar disorder, severe depression that required inpatient treatment, and comorbid psychotic symptoms, according to a recent study published in Acta Psychiatrica Scandinavica.1  

An earlier meta-analysis published in the same journal identified 3 strong predictors for conversion from unipolar depression to bipolar disorder: a family history of bipolar disorder, psychotic symptoms, and early onset of unipolar depression2. The current study sought to determine whether these previous results could be replicated using a large nationwide sample. To avoid potential confounding, study investigators included an extensive list of additional clinical covariates that could be linked to bipolar disorder.

Using data from the Danish Psychiatric Central Research Register (DPCRR) and the Danish Civil Registration System, investigators conducted a historical prospective cohort study of individuals with an initial unipolar depression diagnosis occurring after January 1, 1995, who had a subsequent bipolar disorder diagnosis occurring at least 8 weeks later. This 2-month gap ensured that analysis included diagnostic conversions from unipolar depression to bipolar disorder only, and not misclassified episodes of bipolar disorder.

Of the 91,587 cases of unipolar depression in the cohort sample, 3910 (8.4%) converted to bipolar disorder (95% CI, 7.9–8.8), 3092 died, and the rest were excluded due to end of follow-up or emigration. The strongest predictor of conversion was parental history of bipolar disorder (crude hazard ratio [HR] 2.99; 95% CI, 2.58–3.46; adjusted HR 2.60; 95% CI, 2.20–3.07), followed by depressive psychosis at diagnosis of unipolar depression (crude HR 2.31; 95% CI, 1.99–2.68; adjusted HR 1.73; 95% CI, 1.48–2.02), prior or concomitant nonaffective psychosis (crude HR 2.14; 95% CI, 1.88–2.43; adjusted HR 1.73; 95% CI, 1.51– 1.99), and inpatient treatment at diagnosis of unipolar depression (crude HR 1.91; 95% CI, 1.77–2.05; adjusted HR 1.76; 95% CI, 1.63–1.91).

Related Articles

Study limitations include a dataset restricted to diagnoses from psychiatric hospitals, which would naturally represent a population with more severe cases of unipolar depression than diagnoses from private practices would. These real-world diagnoses are also subject to bias; ie, patients with a family history of bipolar disorder have a greater chance of being clinically diagnosed with bipolar disorder. However, this approach has its strengths as well. Using DPCRR allowed investigators to follow a very large cohort of nationally representative individuals for 702,710 person-years, making the study one of the largest examining conversions from unipolar depression to bipolar disorder.

Investigators conclude: “It appears that these risk factors are now so well established that they can be used clinically to identify [unipolar depression] patients at elevated risk of developing [bipolar disorder].”


  1. Musliner KL, Østergaard SD. Patterns and predictors of conversion to bipolar disorder in 91 587 individuals diagnosed with unipolar depression [published online March 2, 2018].  Acta Psychiatr Scand. doi: 10.1111/acps.12869
  2. Ratheesh A, Davey C, Hetrick S et al. A systematic review and meta-analysis of prospective transition from major depression to bipolar disorder. Acta Psychiatr Scand. 2017;135:273–284.