isotretinoin/Roaccutane causes depression, psychosis and suicide

After decades of case reports to regulators suggesting an increase in depression and suicide with the drug, a 2008 Canadian study (Azoulay et al) of more than 30,000 patients found that:

A quantified association with both depression and suicide had previously (2001) been detected in an analysis of case reports to the United States FDA , which Azoulay et al summarised:

It is well-established that depression greatly increases the risk of suicide. This passage in the 2009 NICE guidelines for depression is unchanged in the draft of the latest version, which will be published soon:

Therefore, although Azoulay et al did not specifically examine suicide, it is reasonable to infer that they confirmed the 2001 finding of an increase in suicide, as well as depression. In fact, they excluded patients with a previous history of depression, so their finding of a trebling of depression was probably an underestimate.

There is good evidence that isotretinoin and other retinoids cause depression and other neuropsychiatric disorders through effects on brain cells.

By how much might isotretinoin increase the suicide rate? Suppose severe acne itself doubles the suicide rate from 1/10,000 a year to 2/10,000 a year. Given that two out of three suicides, according to NICE, is caused by depression, and isotretinoin treatment trebled depression in Azoulay et al, it follows that the suicide rate was at least doubled.

But recent reporting of suicides on isotretinoin to the UK regulator, the MHRA, suggests that Wysowski et al’s 2001 estimate of a six-fold increase may better reflect the true rate. That would mean five out of six suicides are caused by isotretinoin. Patients need to be properly warned about this.

Depression and psychosis overlap. 3% of the patients with depression in Azoulay et al had ‘depressive psychosis’. There is evidence that isotretinoin causes ‘pure psychosis’ in some patients (link on neuropsychiatric disorders, above): they would have been excluded from the study, leading to an underestimate of the overall suicide risk caused by the drug.

A future blog piece will look at the 2010 study of ‘suicide attempts’ by Sundstrom et al, which used different methods, but also found an association with isotretinoin treatment. This paper has been cited by both the MHRA (in 2012 and 2014) and Roche, as showing ‘no causation’. We will argue that the 2010 paper is unreliable, but also uses faulty reasoning.

We will also argue that all patients should be fully informed of the evidence for an increased risk of depression, psychosis and suicide. In 2015 the UK Supreme Court ruled, in the Montgomery case, that doctors should inform patients of all ‘material risks, as well as any to which it would be reasonable for them to think the individual patient would attach significance.’ The GMC had set out very similar guidance in 2009.

This means that it is not for doctors, including the doctors in the MHRA, to decide that a small ‘absolute increase’ in risk – for example, an estimated extra risk of suicide between 2/10,000 (Azoulay et al) and 12/10,000 (Wysowski et al) with isotretinoin – would put a patient off the benefits of the drug, and not inform him or her.

The next blog piece will look at sexual dysfunction, and how it causes depression and other mental disorders in many patients.