Are selective serotonin re-uptake inhibitors associated with an increased risk of self-harm by antidepressant...

To investigate likelihood of self-harm by overdose with antidepressant drugs of different types by examining hospital admission data and poisons inquiries and relating them to prescribing.

Retrospective analysis of prospectively collected data on overdose admissions, poisons inquiries and prescribing of antidepressants in Edinburgh and Scotland.

Poisons treatment unit of the Royal Infirmary of Edinburgh and its surrounding catchment for overdose cases and Scotland for poisons inquiries.

All patients admitted to the Royal Infirmary of Edinburgh between 1 January 2000 and 31 December 2002 with an overdose involving an antidepressant.

Overdose admissions (patients) in relation to prescribing in Edinburgh and poisons inquiries in relation to prescription rates in Scotland.

There were 1656 admissions involving 1343 patients. The likelihood of admission for an individual patient in relation to volume of prescribing (likelihood ratio: 95%CI) in the catchment was somewhat smaller for amitriptyline (0.83:0.74-0.92) and sertraline (0.79:0.63-0.99), and somewhat greater for mirtazapine (1.99:1.57-2.51), trazadone (1.30:1.09-1.54) and venlafaxine (0.97:1.81-1.16) [corrected] For poisons inquiries in Scotland, the excess for venlafaxine and mirtazapine was confirmed and likelihood of an inquiry lowest for selective serotonin re-uptake inhibitors (SSRIs).

There was no evidence of an excess likelihood of presentation with overdose with SSRIs, and the likelihood was reduced with sertraline. There was a small excess of both admissions and poisons inquiries for mirtazapine and venlafaxine. This is a concern in view of the increased toxicity of venlafaxine in overdose in comparison with SSRIs.