The DSM-IV and ICD-10 categories of recurrent [major] depressive and bipolar II disorders: evidence that they lie on a dimensional spectrum

Abstract

Background: Presently it is a hotly debated issue whether unipolar and bipolar disorders are categorically distinct or lie on a spectrum. We used the ongoing Ravenna-San Diego Collaboration database to examine this question with respect to major depressive disorder (MDD) and bipolar II (BP-II).

Methods: The study population in FB's Italian private practice setting comprised consecutive 650 outpatients presenting with major depressive episode (MDE) and ascertained by a modified version of the Structured Clinical Interview for DSM-IV. Differential assignment of patients into MDD versus BP-II was made on the basis of discrete hypomanic episodes outside the timeframe of an MDE. In addition, hypomanic signs and symptoms during MDE (intra-MDE hypomania) were systematically assessed and graded by the Hypomania Interview Guide (HIG). The frequency distributions of the HIG total scores in each of the MDD, BP-II and the combined entire sample were plotted using the kernel density estimate. Finally, bipolar family history (BFH) was investigated by structured interview (the Family History Screen).

Results: There were 261 MDD and 389 BP-II. As in the previous smaller samples, categorically defined BP-II compared with MDD had significantly earlier age at onset, higher rates of familial bipolarity (mostly BP-II), history of MDE recurrences (>or=5), and atypical features. However, examining hypomania scores dimensionally, whether we examined the MDD, BP-II, or the combined sample, kernel density estimate distribution of these scores had a normal-like shape (i.e., no bimodality). Also, in the combined sample of MDE, we found a dose-response relationship between BFH loading and intra-MDE hypomania measured by HIG scores.

Limitations: Although the interviewer (FB) could not be blind to the diagnostic status of his private patients, the systematic rigorous interview process in a very large clinical population minimized any unintended biases.

Conclusions: Unlike previous studies that have examined the number of DSM-IV hypomanic signs and symptoms both outside and during MDE, the present analyses relied on the more precise hypomania scores as measured by the HIG. The finding of a dose-response relationship between BFH and HIG scores in the sample at large strongly suggests a continuity between BP-II and MDD. Our data indicate that even in those clinically depressed patients without past hypomanic episodes (so-called "unipolar" MDD), such scores are normally rather than bimodally distributed during MDE. Moreover, the absence of a 'zone of rarity' in the distribution of hypomanic scores in the combined total, MDD and BP-II MDE samples, indicates that MDD and BP-II exist on a dimensional spectrum. From a nosologic perspective, our data are contrary to what one would expect from a categorical unipolar-bipolar distinction. In practical terms, intra-MDE hypomania and BFH, especially in recurrent MDD, represent strong indicators of bipolarity.