Patients prescribed mirtazapine or venlafaxine are more likely to attempt or complete suicide, a UK study has found.

Data were obtained from more than 600 contributing practices that use the EMIS medical records system. There were 238,963 participants who were aged between 20 and 64 years and had a first recorded diagnosis of depression between 1 January 2000 and 31 July 2011. The follow-up period ended on 1 August 2012.

Antidepressant exposure was measured from electronic prescribing data. Suicides were identified from linked death certificates; both open and suicide verdicts were classified as suicide. Attempted suicide or self-harm events recorded using the appropriate Read codes were identified from the medical records.

During follow-up (median 5.2 years), 87.7% of the cohort received at least one prescription for antidepressants. The median duration of treatment was 221 days (7.3 months). During the first five years of follow-up, there were 198 cases of suicide and 5,243 cases of attempted suicide or self harm.

• Compared with SSRIs, the suicide rate was significantly increased during periods of treatment with ‘other antidepressants’ (mainly mirtazapine and venlafaxine; adjusted hazard ratio (HR) 2.64 (95% CI: 1.74-3.99) but not with tricyclics

• The absolute annual risk of suicide was highest with mirtazapine (0.19%) and lowest with amitriptyline (0.02%)

• Similarly, when compared with SSRIs, the rate of attempted suicide or self harm was significantly increased with ‘other antidepressants’ adjusted HR 1.80 (95% CI: 1.61-2.00) but not with tricyclics

Compared with citalopram, the adjusted HRs for attempted suicide or self harm were significantly increased:

• Venlafaxine 1.85 (95% CI: 1.61-2.13)

• Trazodone 1.73 (95% CI: 1.26-2.37)

• Mirtazapine 1.70 (95% CI: 1.44-2.02)

The majority of open verdicts are thought to be suicides and it is standard practice to include them in studies of suicide. However, not all open verdicts are suicides and it has been suggested that those cases in which suicide is unlikely on the grounds of probability should be excluded.¹ This does not appear to have been attempted in the present study. Even with the inclusion of open verdicts, the total number of suicides was low and the results for completed suicide should be treated with caution. For example, the 95% confidence intervals for absolute suicide risk with mirtazapine and amitriptyline overlap.

The variation in suicide rates was not explained by the relative toxicity of antidepressants in overdose. A UK study found that tricyclics were by far the most toxic (case fatality rate ratio of 13.8 as compared with 2.5 for venlafaxine, 1.9 for mirtazapine and 0.5 for SSRIs).²

The authors acknowledge that the results may have been caused by channelling bias e.g. venlafaxine may have been more likely to be prescribed to patients at increased risk of suicide (as it is recommended as a third-line treatment for resistant depression) whereas tricyclics may have been less likely to be prescribed (because of their toxicity in overdose). They tried to reduce this bias by adjusting for attempted suicide at baseline, but were unable to adjust fully for depression severity.

NICE recommends generic SSRIs as first-line treatment for depression. For those who do not respond adequately, NICE recommends switching to either a different SSRI or a ‘better tolerated newer-generation antidepressant’ i.e. mirtazapine.

If patients still do not respond, third-line treatment with venlafaxine, a tricyclic or an MAOI can be considered.³

It is unclear how the results of this study should be interpreted. However, as the risk of suicidal behaviour associated with antidepressant use appears to be strongly age dependent, I would suggest that we avoid prescribing mirtazapine and venlafaxine to young adults (under 25 years).⁴ Lofepramine is thought to have a very low toxicity, comparable with that of the SSRIs,5 and may be preferred to mirtazapine as a second-line treatment.

Suicide rates were highest in the first 28 days after starting treatment: if we do prescribe mirtazapine or venlafaxine, we should monitor patients closely during this initial period.

Dr Phillip Bland