A significant minority of depressed patients have unrecognised bipolar disorder, a primary care study from Wales has found.

Patients were eligible to participate if they had a recorded diagnosis of depression within the past five years. Participants were asked to complete two screening questionnaires for bipolar disorder: the Hypomania Checklist (HCL-32) and the Bipolar Spectrum Diagnostic Scale (BSDS). Half of those with a positive screen and all of those with a negative screen were then invited to attend for a diagnostic and clinical assessment with a research psychologist.

Of those who were invited to participate, fewer than one in five returned their questionnaires. Of those who were then invited to attend for an assessment, less than half were interviewed.

Of the 154 patients who did attend for interview, 29 fulfilled the DSM IV criteria for bipolar disorder (two with type I, 15 with type II and 12 with subthreshold disorder). The highest estimate of the prevalence of undiagnosed bipolar disorder (assuming the same prevalence among those who dropped out, either at the screening or interview stages) was 21.6%. The lowest estimate (assuming that none of those who dropped out at either stage had bipolar disorder) was 3.3%.

The positive predictive values (PPV) for the HCL-32 and BSDS were 50% and 30% respectively. However, these values were obtained with a prevalence of 21.6%. For the middle and lowest estimates of prevalence (9.6 and 3.3%), I calculate the PPV of a BSDS score ≥12 to be 17% and 6% respectively.

The BSDS has been recommended as the screening tool of choice in primary care1 and is freely available on line. In practice, however, its value will be severely limited by the low PPV, although participants with a false-positive screen nevertheless had more severe and persistent depression, and more functional impairment than those with a negative screen.

The lifetime prevalence estimates for DSM IV bipolar disorder type I (depression alternating with mania), bipolar disorder type II (depression alternating with hypomania) and subthreshold bipolar disorder are 1%, 1.1% and 2.4% respectively.2 Bipolar disorder is often misdiagnosed as unipolar depression: the present study suggests that between 1 in 30 and 1 in 5 depressed patients have unrecognised bipolar disorder, mostly type II or subthreshold bipolar disorder.

There are a number of reasons why bipolar disorder is often not recognised. Patients may perceive hypomania as normal wellbeing, distinction from personality disorder (particularly borderline personality disorder) can be difficult, and comorbid substance misuse may be misdiagnosed as the primary disorder.³

At the same time, early recognition is important. There are concerns that antidepressant monotherapy may be harmful in bipolar disorder, with a risk of mood destabilisation, rapid cycling and treatment resistance.¹

Therefore, we should raise our awareness of bipolar disorder, particularly among those with early-onset, severe or treatment-resistant depression. However, I would argue against the routine use of screening questionnaires as this is likely to replace the problem of underdiagnosis with that of overdiagnosis.⁴

Dr Phillip Bland