Antidepressants are no more effective than placebo as a treatment for patients with subthreshold depression, a meta-analysis has found.

Six double-blind randomised controlled trials met the inclusion criteria. There was a total of 468 participants (234 treated with antidepressants and 234 with placebo). Three studies compared paroxetine with placebo, and fluoxetine, amitriptyline and isocarboxacid were each compared with placebo in individual studies.

The duration of the trials ranged from 6 to 12 weeks and, overall, their quality was rated as low (high or unclear risk of bias).

Failure to respond to treatment was defined as a reduction of less than 50% on the Hamilton Rating Scale for Depression (HRSD), or a failure to score ‘much improved' on the Clinical Global Impression rating scale. Participants treated with antidepressants were no less likely to fail to respond than those treated with placebo, relative risk = 0.94 (95% CI: 0.81-1.08). The non-significant absolute risk reduction was 3.7%, which corresponds to a number needed to treat (NNT) of 27.

Three studies were excluded from the meta-analysis because the participants had a serious concomitant medical illness. This is a pity for two reasons: patients with subthreshold depression and chronic illness are an important sub-group, and inclusion would have boosted the low total number of participants.

For comparison, the NNT for patients with major depression is 5, with 50% of patients responding to antidepressants and 30% responding to placebo.1 A meta-analysis restricted to primary care studies reported an NNT of 7, with 58% responding to antidepressants and 44% responding to placebo.² The placebo response among patients with minor depression is even higher (49-66%).3

Minor (subthreshold) depression is defined by DSM IV as between two and four depressive symptoms, one of which must be depressed mood or loss of interest, which have been present for most of the time over a period of at least two weeks, and are associated with some impairment of functioning. It causes significant distress and in a minority of patients may progress to major depression.⁴

Both NICE4 and the British Association for Psychopharmacology³do not recommend the use of antidepressants for short-duration (

Of course, most patients with subthreshold depression will appear to respond to antidepressant treatment, but this is due to the placebo response.

It has been suggested that the placebo response is entirely due to spontaneous recovery, and that placebo treatment is no better than no treatment at all.¹ However, comparison of meta-analyses of waiting list controls and placebo control groups suggests that there is a small benefit with an NNT of 10.3 Even so, most of the benefit is likely to derive from the patient-practitioner interaction rather than the ritual of taking a pill,5 and there is no justification for prescribing antidepressants as a placebo.

It is important to apply strict diagnostic criteria to discriminate between major and minor depression before initiating antidepressant treatment. A severity rating score below the cut-off (PHQ-9

Dr Phillip Bland