As many as one in ten patients aged 16-40 prescribed antidepressants for depression or anxiety have unrecognised bipolar disorder, a UK study has found.

A total of 21 general practices in West Yorkshire participated in the study. Patients aged 16-40 who had been prescribed antidepressant drugs were identified by a search of the primary care database. Of the 2,341 eligible patients who were invited to participate 236 agreed to do so.

Participants were asked to complete the Mood Disorder Questionnaire (MDQ). They attended a semi-structured diagnostic interview based on DSM IV criteria and those with possible bipolar disorder were subsequently discussed at a consensus meeting.

A total of 17 participants were found to have unrecognised bipolar II disorder; prevalence 7.3% (95% CI: 4.4-11.1%). Adjustment for differences between the sample and a national database increased the prevalence to 10%.

Of the 17 patients with unrecognised bipolar disorder, 11 had been diagnosed with depression, one with anxiety disorder and five with mixed anxiety and depression.

Unrecognised bipolar disorder was associated with younger age and more severe lifetime depression.

The positive predictive value of the MDQ was low. The MDQ takes about 5 minutes to complete. Using the recommended cut-off score of ≥ 7, the present study found a PPV of only 14% which is too low for this to be used as a screening tool. NICE is unequivocal in recommending that we should not use questionnaires to identify bipolar disorder.¹ I agree but suggest that, as with the PHQ-9, the MDQ may have some value as a rule-out test² i.e. if you suspect bipolar disorder but the MDQ score is less than 7, you are probably wrong (negative predictive value = 97%).

Both this and a previous study of unrecognised bipolar disorder in UK general practice³ are limited by low participation rates (10.1% and 18.5% respectively). The true prevalence is likely to be somewhere between 1% (assuming all of those with unrecognised bipolar disorder volunteered to participate) and 10% (assuming the same prevalence rate among non-participants). The previous study estimated a prevalence of between 3.3% and 21.6%.³

Antidepressants are thought to be less effective in bipolar depression and may destabilise mood. For patients with moderate to severe bipolar depression NICE recommends either a combination of fluoxetine and olanzapine, or quetiapine on its own.1 The BAP guideline also cautions against antidepressant monotherapy. In particular, tricyclics and venlafaxine are associated with a greater risk of causing a switch to mania and should not be used.⁴

It is important to try to identify depressed patients with unrecognised bipolar disorder both because of the risks associated with antidepressant monotherapy and because these patients may have a higher risk of suicide. NICE recommends that we should ask depressed patients whether they have had previous periods of overactivity or disinhibited behaviour.1 We should consider the possibility of bipolar disorder in depressed patients with any of the following: early onset; frequent relapses; treatment resistance; family history of bipolar disorder.

Dr Phillip Bland