February 1995

Switching From 'Unipolar' to Bipolar IIAn 11-Year Prospective Study of Clinical and Temperamental Predictors in 559 Patients

Author Affiliations

From the National Institute of Mental Health (NIMH) Collaborative Program on the Psychobiology of Depression (Clinical Studies) (Drs Akiskal, Maser, Endicot, Coryell, Keller, Clayton, and Goodwin and Ms Warshaw) and DataMax, Inc, Rockville, Md (Dr Zeller). For other members of the NIMH Collaborative Program on the Psychobiology of Depression, see box at the end of the article.

Arch Gen Psychiatry. 1995;52(2):114-123. doi:10.1001/archpsyc.1995.03950140032004

Background:  Given the therapeutic and prognostic importance of the unipolar-bipolar dichotomy, predicting which patients will become bipolar subsequent to index diagnosis of major depressive disorder (MDD) is of paramount clinical significance. We sought to characterize the profile of patients with MDD who would convert to the more subtle bipolar subtype (known as BPII) on the basis of clinical and personality variables obtained during MDD episodes.

Methods:  A total of 559 patients, comprehensively evaluated with the Schedule of Affective Disorders and Schizophrenia and "unipolar" MDD at entry, were administered 17 self-report personality measures. Hypomanic and manic episodes were systematically recorded over a prospective observation period of up to 11 years. We compared 48 converters to BPII (8.6%) with 22 converters to bipolar I (BPI) (3.9%) and the remaining larger group of unipolar patients.

Results:  Except for greater acuteness, severity, and psychotic symptomatology, BPI converters were essentially similar to MDD nonconverters. By contrast, BPII converters were robustly distinguished from those with MDD who remained unipolar on the basis of self-report measures along the newly derived factors of Mood Lability, Energy-Activity, and Daydreaming. This profile was associated with early age at onset of MDD and pleomorphic psychopathology beyond the usual affective realm, high rates of substance abuse, as well as educational, marital, and occupational disruption and minor antisocial acts prior to discrete hypomanic episodes. Overall, BPII switchers had a more protracted and tempestuous course with shorter well intervals. "Habitual self" descriptions of temperamental instability during MDD episodes provided useful clinical information for predicting which depressed patients will switch to BPII, attaining a sensitivity of 91% for all three factors combined (23 items); Mood Lability alone (nine items) was the most specific predictor (86%), though of lower sensitivity (42%).

Conclusions:  The BPII subtype is best understood by such lability intruding into, and possibly its accentuation during, depressive episodes, thereby creating an intimate interweaving of trait and state. Clinicians must note that the foregoing temperamental profile appears more fundamental in defining the affective dysregulation of the BPII subtype than hypomanic episodes emphasized in DSM-IV.