The use of the term hypomania in this and other countries impressively follows the Humpty Dumpty principle. Indeed, in the UK many, perhaps a majority of, inpatients are diagnosed to be ‘hypomanic’, perhaps partly out of a sense of politeness. The term manic does, after all, have a pejorative flavour to some ears (pace Reference JamisonJamison, 1996). The original Greek sense of the word is that hypomania is hierarchically below or beneath mania. It fills a gap between the full syndrome and more everyday states of elation. The issue is simply where to draw the line with mania, on the one hand, and with normality, on the other. If we actually use them, operationalised diagnostic criteria should permit us to do this reliably.

DSM—IV AND ICD—10 DEFINITIONS OF ELATED MOOD DISORDER

Significantly, the definitions of affective disorder found in DSM—IV and ICD—10 differ for elated states and the differences are probably magnified by their even more different usage on the ground. In DSM—IV, mania describes all elevated mood states with functional impairment and is qualified by severity: mild, moderate and severe (with or without psychosis). The definition of hypomania introduced in DSM—IV is a sometimes short-lasting (4 days minimum) elevation of mood identified by the usual criteria for mania (see Appendix) but, crucially, without marked social or occupational dysfunction (and hence never when admitted to hospital) and without severe symptoms such as delusions and hallucinations (American Psychiatric Association, 1994). For many individuals, indeed, hypomania by this definition is a positively attractive state to be in if it is not followed by depression or mania itself. In fact, where DSM—IV adds that ‘hypomaniacan cause some social or occupational impairment’ it unnecessarily almost blurs an otherwise tidy distinction from mania.

In ICD—10, a manic episode is graded as hypomania, non-psychotic mania or psychotic mania (World Health Organization, 1993). The boundary between non-psychotic mania and hypomania is defined by ‘severe or complete’ disruption of work and social activity. However, ‘considerable interference with work or social activity is consistent with a diagnosis of hypomania’, although hypomania is stated to span ‘the range of disorders between cyclothymia and mania’ (World Health Organization, 1992). In ICD—10, cyclothymia is defined as a chronic and persistent pattern of mood instability. In defining the elation there is a requirement for at least three symptoms in addition to mood elevation (from a list overlapping markedly with DSM—IV hypomania). In DSM—IV, cyclothymia is based also on longitudinal course but requires an unspecified (thus, low) number of hypomanic symptoms.

In summary, DSM—IV defines hypomania to be a milder condition (literally, below mania) than does ICD—10 and cyclothymia in turn is also a milder diagnosis in DSM—IV. The upper boundary between hypomania and mania hinges on a definition of functional disturbance that is different between the two systems but depends upon the meaning of ill-defined qualifying words: marked, severe, complete, etc. Unfortunately, in UK psychiatry and possibly elsewhere hypomania is used as a term that receives little qualification at all. Fulford (Reference Fulford1995) has argued persuasively that we should be much more explicit in acknowledging the difficulty of defining social and occupational dysfunction. Such value judgement is critical to the diagnosis of illness in general and psychiatric disorder in particular. Hypomania and its related spectrum diagnoses illustrate this practical dilemma very clearly. However, in ICD—10 hypomania is an almost superfluous term that describes mild mania, whereas DSM—IV offers us something different.

IS IT USEFUL TO EMPLOY THE DSM—IV DEFINITION OF HYPOMANIA?

The value of DSM—IV hypomania as a diagnosis is enhanced because it appears to affect treatment in two important ways. First, bipolar II disorder usually presents with a depressive episode but the existence of a history of hypomania appears to make a real difference. This is particularly because antidepressants may induce hypomania and rapid mood cycling, which can be very troublesome. There is weak evidence that some antidepressants are particularly likely to do this (tricyclics > selective serotonin reuptake inhibitors > monoamine oxidase inhibitors) and if cycling occurs it will often be necessary either to add a mood stabiliser or to avoid antidepressants altogether. Thus, to ignore a history of hypomania in its DSM—IV version may be a significant clinical mistake.

The second issue is when to treat mood elevation. When a UK trainee tells me that a patient has hypomania, it is always necessary to clarify what he or she means because the current UK usage covers too wide a spread of manic illness. The term may be capturing an episode of mood elevation that is producing significant social and occupational dysfunction, or it may not. The great advantage of the DSM—IV split is that if you diagnose mania, you should be doing something about it.

THE BIPOLAR SPECTRUM

If we use the term hypomania as it is used in DSM—IV, we will detect more individuals who have had it. We depend heavily upon Angst's seminal community cohort study in Zurich for estimates of lifetime incidence (Reference AngstAngst, 1998). Angst described DSM—IV diagnoses of mania and hypomania in 5.5% of the population. However, relaxing the time and severity criteria for hypomania brings in a spectrum of additional minor states of elation. These additional categories were recurrent brief hypomania in 1.5%, sporadic brief hypomania in 1.3% and sub-diagnostic hypomanic symptoms in a further 11.3% of the population. Hypomania itself may occur either in isolation or more commonly in association with minor or major depressive symptoms. An illness course of hypomania and major depression (bipolar II disorder) showed a population rate in the Zurich cohort of 3%. The rate for bipolar I disorder (mania and usually major depression) was 0.5%.

The apparent size of this so-called bipolar spectrum invites disbelief. Can there be so large an iceberg of meaningful psychopathology below the rare syndrome that we all recognise as bipolar I disorder or manic—depressive psychosis? Similar figures for US populations support the conclusions of the Zurich study (Reference Carlson and KashaniCarlson & Kashani, 1988; Reference Lewinsohn, Klein and SeeleyLewinsohn et al, 1995). The reliability with which we can detect hypomania and particularly sub-syndromal manic symptoms may be poor and the published studies depend on relatively modest sample sizes. Something, however, is being detected that requires explanation and clarification. In the Zurich study the subjects with a ‘spectrum diagnosis’ showed higher rates of a family history of mood disorder and personal rates of substance misuse, deliberate self-harm and rates of treatment for depression. There was evidence for a gradient of risk, with manic symptoms alone rarely showing metamorphosis into something clearly syndromal. Hypomania itself, however, was associated with additional psychopathology in about 50% of cases. This is suggestive evidence that a hypomania diagnosis is not simply one of abundant good health: although it may be benign, it is commonly not. Nevertheless, DSM—IV's hypomania and, in particular, the hypomanic syndromes described by Angst pose new problems, especially at the boundary with normality; indeed, are such spectrum states illnesses at all?

Angst's study depends upon structured interview and operational use of DSM—IV. The potential for further inflation of the hypomania concept is obvious. Kraepelin (Reference Kraepelin1921) wrote of the spectrum of mood disorders: ‘we include here certain slight and slightest [sic] colourings of mood, some of them periodic, some of them continuously morbid, which on the one hand are to be regarded as the rudiment of more severe disorders, on the other pass over into the domain of personal predisposition’. His authority, no less, therefore sanctions efforts to pathologise even sub-syndromal manic symptoms when seen in a clinical context and to extend the spectrum concept to include drug-induced elated states, a variety of depressive states with activated components and even borderline personality disorder. Akiskal & Pinto (Reference Akiskal and Pinto1999) give a charismatic exposition of this approach and their ideas throw down a real challenge to those of us with more conservative instincts. There is a strong folk tradition within psychiatry everywhere to see manic patients as more likely to be temperamentally ‘hypomanic’ or hyperthymic, even if the retrospective identification of personality variables may be coloured by the illness itself or by the misuse of stimulants. The further argument is whether treatment paradigms worked out for bipolar I disorder may be appropriate for clinically significant spectrum conditions.

The bottom line is that accurate diagnosis, although always desirable, has become clinically important for elated states. To define where hypomanic features end and individual differences begin is a major contemporary challenge. However, to make the distinction between hypomania and mania exactly where it is made in DSM—IV seems to have important advantages. To describe most manic states loosely as hypomania has become, it seems to me, poor practice.

APPENDIX

DECLARATION OF INTEREST

None.