2008 Diagnostic Guidelines
from the
International Society for Bipolar Disorders


In recognition of continued controversy about several aspects of bipolar diagnosis, the International Society for Bipolar Disorders (ISBD) convened a group of 25 bipolar specialists from around the world.  They were charged with reviewing the available literature for seven different areas in which current diagnostic rules (the DSM and ICD) leave room for confusion and doubt. The idea was to present an interim solution, in the form of recommended revisions to current criteria, while we await the more thorough analyses of committees which are already underway formulating the next editions of these diagnostic systems.

Here is a summary of the seven reports.  Click on the individual report name at left for a few further details. The author name link leads to the article abstract.  An overview of the reports was prepared by the committee chairman, Dr. Nassir Ghaemi.

Report Authors Major Question Addressed Conclusion/Recommendation
Bipolar Depression Mitchell P, Goodwin G, Johnson G, Hirschfeld R In what ways does bipolar depression differ from unipolar "Major Depression"? This workgroup reviewed 50 different variables, looking to see what might help differentiate these two forms of depression.

Markers of bipolar disorder which are not manic symptoms are sometimes called "soft signs". The workgroup concluded that 10 such markers -- the same markers which have been featured on this website for five years -- are statistically helpful in recognizing bipolar depression.

Bipolar II Vieta E, Suppes T What are the boundaries of "hypomania?  What separates it from mania; and more importantly, from the absence of hypomania?  Is there a dividing line? Hypomania is "clearly dimensional", that is, it lacks clear boundaries separating it from its neighbors (mania and unipolar depression).

"Dichotomies are useful for education, communication, and simplification.  Unfortunately, simplicity is useful but untrue -- whereas complexity is true, but useless."

Bipolar Spectrum Model Phelps J, Katzow J, Angst J, 
Sadler J
Does the spectrum model help capture the reality of bipolar disorder as well or better than the DSM system of categories -- in which one either has unipolar or bipolar disorder, but nothing in between? The spectrum model is widely used, suggesting that it has been found useful.

The work group recommended adding one more additional example to the DSM- IV Bipolar NOS (Not Otherwise Specified) criteria. The effect of this addition is to formally recognize "sub-threshold" hypomania, in combination with episodes of severe depression, as "bipolar".

Children and Adolescents Youngstrom E, Birmaher B, Findling R Do children really get bipolar disorder?   Are their versions of bipolar disorder in children which do not meet current criteria yet still warrant consideration as "bipolar"? "There exists a group whose symptoms meet strict DSM-IV criteria for bipolar diagnoses..."

... "and there is a broader spectrum of cases that show symptoms of mania without meeting criteria for a DSM diagnosis."

Schizoaffective Disorder Malhi G, Green M, Fagiolini A, Peselow E, Kumari V Is schizoaffective disorder an independent entity, or a point on a continuum between schizophrenia and bipolar disorder? "Evidence favors the view of schizoaffective disorder as a point on a continuum" between schizophrenia and bipolar disorder. 

The workgroup even asked "is it not premature to suggest that the category be jettisoned from the DSM?"

Rapid Cycling Bauer M, Beaulieu S, Dunner D, Lafer B, Kupka R Is there a better point at which to differentiate "rapid cycling" from non-rapid cycling, versus the current criterion (more than four episodes in a year)?

"While episode cycling can be conceptualized as a dimensional phenomenon between the extremes of no cycling and continuous ultradian and cycling, there is insufficient new evidence to modify the existing definition in a manner that would be less arbitrary."

Mixed States (appears in mania paper) Cassidy F, Yatham L, Berk M, 
Grof P
Is the DSM mixed state to narrow?
"A less restrictive definition appears warranted.

"The DSM-IV may have fostered a false reliance on assumptions about diagnostic validity and clear diagnostic boundaries that are not sufficiently supported in the existing studies"


Bipolar Depression
If you are interested in this topic, you may want to dig up the article itself.  The master table of studies reviewed (Table 1), which covers two full pages, is an amazing feature in itself.

Summarizing their findings (from their Table 2), they list the three categories of markers which indicate a greater probability that a person with depression actually has bipolar depression, rather than unipolar:

  1. Symptom features: atypical depression symptoms (increased appetite, increased sleep, "leaden paralysis"); psychosis
  2. Course of illness: early onset of depression; more than five episodes of depression
  3. Family history: a relative with a bipolar diagnosis

These same features emerged in the paper on bipolar spectrum, through independent literature review by that team.  On that basis one might conclude that these particular "non-manic bipolar markers" have been established as worthy of incorporation in the diagnostic process.  Use of these markers in the diagnostic process is already routine at the mood disorders clinic associated with Harvard University, as an integral part of their Bipolarity Index.

Bipolar II
The authors note specifically that although many clinicians regard Bipolar II as a milder form of manic-depressive illness, "this is not only wrong but misleading".  They emphasize the evidence indicating that suicide rates are as high as in Bipolar I, and social disability resulting from Bipolar II is extremely high.also, they note that dysphoric or mixed states are the more common presentation of hypomania.Suppes

They review the data on duration of hypomania, concluding "while the 2-day requirement would likely be more sensitive [in detecting Bipolar II] than the 4-day, the ultimate distinction should likely be based more on quality of change in mood or behavior, rather than quantity."  Nevertheless, they also observe that "obviously a specific duration threshold needs to be set for practical purposes."

Bipolar Spectrum Model
This paper underwent at least 10 significant revisions, as points of view from the various authors, the editors, and the reviewers, were incorporated. Nevertheless, the resulting paper is not "watered down".  All the revisions made it more precise and representative of current work in the field. Here are a few more highlights :

1. Formal recognition of "Bipolar Spectrum Disorder", as originally characterized by Ghaemi and Goodwin in their important 2002 review, was not allowed in this paper.  There was not sufficient consensus that formalizing a version of bipolar disorder that lacks any features of hypomania at all is appropriate at this point.

2. Although the recommended revision in BP NOS criteria admittedly broadens the concept of bipolar disorder, we also attempted to make the diagnosis more rigorous by requiring that practitioners who invoke this form of bipolar disorder include in their assessment a detailed account of non-manic bipolar markers (roughly the same list as was developed by the Mitchell group, writing on bipolar depression).

3. The "bipolar spectrum" can be viewed in two ways.  These are shown in the figure below.  First, as shown in Part A, you can think of bipolar disorder as a continuum with discrete "nodes", like the visual spectrum -- which we call red, yellow, orange, green, and so forth.  Although the phenomenon being named is continuous, we can still name particular points along that spectrum. 

Alternatively, as in Part B  in the figure below, you can think of all of the diagnostic features of bipolar disorder as existing on a continuum.  Any given patient could have any admixture, e.g. Patient A might have three manic symptoms, a first-degree bipolar relative, but long episode durations; whereas Patient B might have one manic symptom, no relative with bipolar disorder, but early age of onset, frequent recurrences, and 4 "suggestive symptoms" .

Finally, here is the formal revision we recommended for the NOS criteria. Changes are shown in italics.

The Bipolar D isorder Not Otherwise Specified category includes disorders with bipolar features that do not meet criteria for any specific Bipolar D isorder.  Examples include

  1. Very rapid alternation (over days) between manic symptoms an d d epressive symptoms that do not meet minimal duration criteria for a Manic Episode or a Major D epressive Episode.
  2. Recurrent Hypomanic Episodes without intercurrent depressive symptoms
  3. A Manic or Mixed Episode superimposed on a D elusional D isorder, Residual Schizophrenia, or Psychotic D isorder Not Otherwise Specified
  4. Situations in which the clinician has concluded that a Bipolar D isorder is present but it unable to determine whether it is primary, due to a general medical condition, or substance induced
  5. Subthreshold Hypomanic Episodes in the context of multiple other signs of bipolarity.*

* Clinicians should specify precisely which such signs are present and include this list in their Assessment statement, as follows:

    1. Family history (bipolar diagnoses; multi-generational mental illness; alcohol and other substance use; suicides)
    2. D epressive symptom phenomenology (atypical, seasonal, psychomotor slowing, psychosis)
    3. Course of illness (early age of onset, post-partum onset, short duration of episodes, greater nu mb er of episodes)


Children and Adolescents
After a remarkably encyclopedic review, the authors were very cautious in offering any interpretation of their findings.  I think the summary above is just about all I can offer of their comments.

Schizoaffective Disorder
 The authors reviewed data on the classic diagnostic validators (phenomenology, epidemiology, diagnostic stability, and others); genetics, including family and molecular studies; data on the longitudinal course of illness, including response to treatment; and neurobiology, including neuroimaging and pathophysiologic findings. they conclude: "SAP is a prototypical boundary condition that highlights the disadvantages of a rigidly categorical classification system.  They advocate a "dimensional approach" (spectrum perspective), noting that such a view is more consistent with the genetic model and is supported by the brain imaging and other neurobiological findings. 

Nevertheless, they note difficulties with a dimensional approach, especially asking "how many dimensions suffice?" Typically the continuum is characterized as that between mood disorders and thought disorders (as effective-psychotic).  But there are other potential dimensions to be considered.  Motivation, energy, and even suicidal thinking might be separate "axes".  This would play havoc with attempts to define groups of patients for research studies.  That in turn would limit the applicability of research to clinical practice.  Their conclusion is very similar to that of the Bipolar II group noted above: "Dichotomies are useful for education, communication, and simplification.  Unfortunately, simplicity is useful but untrue -- whereas complexity is true, but useless."

Rapid Cycling
In examining the literature on this phenomenon, the authors offer numerous interesting details.  Here are a few which struck my eye:

Mixed States
A separate paper was prepared on mixed states, but it did not make the final edition of the ISBD Guidelines -- a pity, as Roger McIntyre, Franco Benazzi, and Mario Maj put a lot of work into it. I saw the paper. Looks like they got tangled up in the spectrum issue, to my eye, which makes this subject exceptionally difficult, if the "spectrum" concept is not the focus of the work. The committtee chair, Dr. Ghaemi, wrote an editorial addressing this issue: All Mixed Up -- on the absence of diagnostic guidelines for mixed states in the ISBD report.

For the table above, I extracted a position on mixed mania from the paper on manic subtypes -- as that is, in my opinion, the most controversial issue in this aspect of bipolar disorder.   Otherwise, although technically a masterful literature review worthy of the international reputation of the authors, the conclusions from the mania paper are not earthshaking nor paradigm-shifting. 

By contrast, Dr. Ghaemi's editorial on mixed states is typical, intriguing example of his philosophical approach to diagnosis (practitioners with such a bent will enjoy his Concepts of Psychiatry). He notes that diagnoses can be regarded as metaphors -- sometimes literally true, sometimes better regarded as metaphors per se.  He offers an interesting conclusion based on the failure of the mixed state paper to survive peer review, extrapolating  from experience getting papers published: "the idea of expanding mixed state definitions is either very bad, or excellent but ill timed."