Editorial Credits/ Crédits éditorials

Subscription Rates /Prix d'abonnements

Advertising Rates / Tarifs publicitaires (PDF)

Editorial
Thanks to Our Reviewers in 2005
Joel Paris, MD
(PDF)

In Debate
Do Many Patients With Depression Suffer From Bipolar Disorder?
Jules Angst, MD
(PDF)

Does Almost Everybody Suffer From a Bipolar Disorder?
Scott B Patten, MD, FRCPC, PhD
(PDF)

Original Research Community Survey of Bipolar Disorder in Canada: Lifetime Prevalence and Illness Characteristics
Ayal Schaffer, John Cairney, Amy Cheung, Scott Veldhuizen, Anthony Levitt,
(PDF)

Correlates of Methylphenidate Use in Canadian Children: A Cross-Sectional Study
Alice Charach, Hongmei Cao, Russell Schachar, Teresa To
(PDF)

Polysomnographic and Symptomatological Analyses of Major Depressive Disorder Patients Treated With Mirtazapine
Jianhua Shen, Sharon A Chung, Leonid Kayumov, Henry Moller, Naheed Hossain, Xuehua Wang, Prativa Deb, Frank Sun, Xin Huang, Marta Novak, Darryl Appleton, Colin M Shapiro
(PDF)

Suicidality, Depression, and Mental Health Service Use in Canada
Anne E Rhodes, Jennifer Bethell, Susan J Bondy
(PDF)

Status of First-Episode Psychosis Patients Presenting for Routine Care in a Defined Catchment Area
Jennifer Payne, Ashok Malla, Ross Norman, Deborah Windell, Nicole Brown
(PDF)

Test Wisconsin chez les patients souffrant de schizophrénie, et leurs frères et soeurs
Youssef El Hamaoui, Meryem Elyazaji, Sakina Yaalaoui, Linda Rachidi, Mohamed Saoud, Thierry d’Amato, Driss Moussaoui, Jean Dalery, Omar Battas
(PDF)

Book Reviews
(PDF)

Clinical Handbook of Psychotropic Drugs. 14th revised edition
Review by
Nicholas Delva

Marijuana and Madness
Review by
Nady el-Guebaly

Manual of Psychiatric Care for the Medically Ill
Review by
François M Mai

Guidebook on Helping Persons with Mental Retardation Mourn
Review by
Bruce D McCreary

Ethical Issues in Forensic Mental Health Research
Review by
Paul Fedoroff

Letters to the Editor
(PDF)

Re: Listening to the Past: History, Psychiatry, and Anxiety

Reply: Listening to the Past: History, Psychiatry, and Anxiety

In Debate

Does Almost Everybody Suffer From a Bipolar Disorder?

Scott B Patten, MD, FRCPC, PhD¹

(Can J Psychiatry 2006:51:6–8)

Click here for author affiliations. 

Click here for information on funding and support. 

The concept of a broad and inclusive bipolar spectrum disorders appear to be gaining momentum in the psychiatric literature, although cautionary comments have previously been made (for example, by Baldessarini; 1). The possibility that a large proportion of people diagnosed with depression actually have BD is an important clinical consideration and a natural corollary of this trend. I will not focus specifically on the unipolar-bipolar distinction, however, since the general trend toward broadening the diagnostic boundaries of BD raises similar issues for many other conditions (such as impulse control disorders and Axis II pathology). Rather, I will direct my comments toward the broad, ongoing debate about the bipolar spectrum concept. My intention is not to argue that existing diagnostic conventions are perfect or that a bipolar spectrum does not exist but to draw attention to certain issues that have not received adequate emphasis in the literature. These issues include setting appropriate diagnostic thresholds, the inevitable tension in psychiatry between empirically oriented and theory-driven concepts, and debate over the relative merits of categorical and dimensional measurement.

The arguments put forward in favour of a broad bipolar spectrum are protean. In 2000, Akiskal and colleagues provided a detailed review (2). Some of the more notable arguments are that hypomanic episodes may last less than the 4-day threshold specified in the DSM-IV, that certain subsets of subjects without a DSM-IV BD have a higher-than-expected proportion of relatives with BDs, that mixed states occur frequently and are underrecognized clinically, that patients with manic or hypomanic episodes induced by antidepressants often later manifest bipolar states, and that some subjects who do not meet DSM-IV criteria for a BD respond to mood stabilizers.

There is no formally accepted definition of what is meant by the term bipolar spectrum. The published works of various authors have adopted quite different emphases. To some authors, the idea of a spectrum implies a spectrum of symptom severity, resulting in the argument that a larger or milder set of bipolar symptoms should be accepted as fulfilling diagnostic criteria for hypomanic or mixed episodes. For some, the emphasis has been more deeply conceptual, typically targeting what is seen as an artificial distinction in the DSM-IV between unipolar and bipolar disorders. In this latter instance, the spectrum concept can be understood not simply as the broadening of a diagnostic category but as a more basic reappraisal, perhaps one that favours dimensional measurement.

The DSM-IV includes a kind of bipolar spectrum: BD I and BD II, as well as cyclothymic disorder. These are classified together with other mood disorders. Mixed states are treated much like manic states in the 6 DSM-IV coding categories for BD I. Further, the categories are fluid: a manic episode can move an afflicted individual from one category (such as major depressive disorder, recurrent) to another (such as BD I, most recent episode manic). However, the movement occurs as a result of signs and symptoms identified according to explicit diagnostic criteria. The classification operates within a fruitful empirical tradition established by the DSM-III, which attempts to be neutral with respect to etiologic theory. Some authors supporting the concept of a broad bipolar spectrum have emphasized that some patients diagnosed originally as suffering from unipolar depression later develop manic, hypomanic, or mixed episodes. Often, this observation is offered as if it provided self-evident confirmation of problems with the existing nosology—but it is rarely acknowledged that this is how the criteria are designed to work. The DSM-IV makes “no assumption that all individuals described as having the same disorder are alike in all important ways” (3, Introduction, p xxxi). Returning to a diagnostic approach that depends on theoretical assumptions about what bipolarity theoretically means (including assumptions about underlying pathophysiology) can jeopardize the advantages that an empirical approach offers. Similarly, modification of the diagnostic criteria for BD that results in the inclusion of people without actual evidence of the illness but whose clinical features suggest they may be at higher risk of developing it later (for example, by virtue of having a family history of BD or some other clinical feature that may be predictive of later development of a BD) would amount to backing away from the DSM-IV’s empirical stance and moving back toward the use of diagnostic definitions that are more subjective and more dependent on theoretical and etiologic beliefs.

Another set of issues revolves around the idea of a spectrum itself. The argument has been made that the bipolar spectrum spans a dimension ranging from acute psychopathology all the way to temperamental differences (2). These arguments challenge the utility of categorization in diagnostic practice, usually without offering much in the way of boundaries for the new entities being proposed. The most developed expression of this thrust is bipolarity scaling, several examples of which have appeared in the literature in recent years. The developers of these instruments have, however, typically presented the new scales as a means of augmenting, rather than replacing, categorical diagnoses.

Dimensional strategies have been recommended as an alternative to categorical diagnoses in almost every area of psychiatry. However, dimensional approaches have drawbacks. Dimensions tend to interface poorly with diagnostic practice and are more difficult to integrate, for example, with clinical practice guidelines. In the area of depression, dimensional symptom ratings are considered poor guides to clinical action, since a high score on a symptom rating scale may indicate a completely normal occurrence, such as bereavement, or a serious disorder. The same ambiguities are raised by dimensional evaluation of bipolarity, whether this uses explicit scaling or, more implicitly, the “soft bipolar spectrum” concept. A spectrum that does not differentiate clearly between, for example, differences in temperament and personality and overt psychopathology may not be useful in clinical practice.

Almost all measurable human characteristics distribute along spectrums. This is not de facto a critical flaw for diagnostic categorization. Blood pressure is an obvious example. In juxtaposition to the physical measurement of blood pressure, essential hypertension is a diagnostic construction that helps to bridge the gap between what can be evaluated in clinical practice and what should be done therapeutically. Nevertheless, like any diagnosis, the concept of essential hypertension is a construction that is vulnerable to rhetorical attack. It is even easier to argue that the diagnosis of essential hypertension should be replaced with a spectrum concept than it is to make the same case for bipolarity. Additional questions need to be answered, though. In the case of bipolarity, one question is of paramount importance: Can current treatment strategies for BD be generalized to the broader spectrum?

It has been asserted that epidemiologic studies, because of their reliance on “traditional” concepts of BD, have produced unrealistically low prevalence estimates for BD (2,4). The reality is that most studies attempting to apply DSM criteria by structured interview have found a very high prevalence of BD. Although instruments such as the DIS and the CIDI use existing diagnostic criteria, reappraisal studies have consistently found that clinicians conducting validation interviews believe that most of the bipolar syndromes identified are not clinically relevant (5,6). This same phenomenon probably accounts for the high prevalence of BD recently reported in Canada from the Canadian Community Health Survey: Mental Health and Well-Being, which used an interview based on the CIDI. The prevalence of BD I among people aged 25 to 64 years was 2.6% (7). These epidemiologic data suggest that manic and hypomanic symptoms blend to an appreciable extent with (what many psychiatrists would regard as being) normal experience, such that great caution must be used to constrain diagnostic definitions to what truly constitutes a disorder.

At this stage in the scientific development of psychiatry, diagnostic definitions and clinical guidelines should certainly remain somewhat fluid. However, changes should be driven by evidence, and the level of evidence must be appropriate to the question being addressed. In epidemiology, the evidence should derive from data showing enhanced performance (such as improved predictive validity) of reconstructed diagnostic categories or proposed dimensional alternatives. In the realm of therapeutics, randomized controlled clinical trials (including pragmatic trials) are the required standard of evidence. Some enthusiastic proponents of a broadened bipolar spectrum concept have seemed quick to assume that their arguments for redefinition, relabelling, and dimensional conceptualization have straightforward, self-evident therapeutic implications. This undercurrent is evident in frequent claims of vast underrecognition and undertreatment of BDs (2). These claims seem to imply that a large proportion of the population, including currently untreated people or people currently being treated with antidepressants, should be managed instead with mood stabilizers. However, basing treatment recommendations on theoretical, historical, and epidemiologic evidence exceeds the boundaries of what is now regarded as evidence-based practice.

When should diagnostic definitions be modified? Recently, it has been argued that the concept of clinical utility may provide an answer (8). Another consideration is scientific utility. If one conceptualizes an observable phenotype as being an indicator of some underlying etiologic disturbance, then definitions that are not specific for that disturbance will slow the rate of scientific progress by introducing bias toward the null, a dilution of observed effect known in epidemiology as nondifferential misclassification bias (9). Broadening diagnostic categories can be scientifically risky.

Proponents of the spectrum concept have raised interesting hypotheses that warrant continued evaluation in appropriately designed studies. Longitudinal studies, as pioneered and championed by Professor Angst, provide invaluable data that are capable of informing debate about many of the issues outlined above. Epidemiologic data, however, should not inform therapeutic decisions. When therapeutic decisions are involved, randomized controlled trials provide the required standard of evidence. Appropriate constraint in the bipolar spectrum literature will help to ensure continued progress.

Funding and Support

Dr Patten is supported by an Alberta Heritage Foundation for Medical Research Health Scholar Award and is a Fellow with the Institute of Health Economics.