I just noticed that Psych Practice has a new article for our free online journal club, Antidepressants and suicide: risk-benefit conundrums.
As you can see, it’s embedded in a very interesting book review about the risk for suicide and suicidality thought to be associated with selective serotonin reuptake inhibitors (SSRIs), one example being fluoxetine (Prozac), Psych Practice: Book Review, Sort Of.
Dr. Dawson and Psych Practice have already made learned comments and I probably don’t have much new to add. However, I found reading Healey’s paper was rough going for me as well, possibly because I’m a geezer who didn’t pay as much attention as I should have in my medical school biostatistics course.
I also like Healey’s preference for open access journals, which fits with this being Open Access Week.
Therefore, I cheated and found an open access article on PLos ONE:
This paper also had mystifying statistical information, but the discussion seems to be easier to follow. The strength of the study was the use of a large population-based nationwide registry covering all suicides and all filled prescriptions of SSRI. A case-crossover design with matched-pair interval approach in which the cases acted as their own control seemed, to my untutored mind anyway, to eliminate a number of confounders.
The finding was that SSRIs probably do lead to an increased risk of suicide, probably from activation syndrome (which you probably could call akathisia), occurring generally in the first couple of weeks of therapy with an SSRI.
The authors acknowledge the inherent risk that severe depression plays in suicide. I think their elegant comparison study of the same design using tricyclics helped overcome this limitation. The problem I thought would be a deal breaker in a study which used prescription records was the difficulty in establishing whether patients actually took the medication. But the matched case-control design probably would have led to a misclassification bias in both case and control, leading to lower risk estimates without eliminating the finding of risk itself.
Another interesting point is the authors’ speculation about the actual cause of the overactivation, which could be reduced serotonin transporter availability in patients with impulsive aggression compared with healthy subjects. This could possibly be one sort of biomarker that could help clinicians decide on which antidepressant to prescribe to which patients–one day.
But then I’m over my head in this arena, anyway.