Like James Blackburn earlier, I wanted to take issue with one sentence in the Editors' Summary, "Depression is a serious medical illness caused by imbalances in the brain chemicals that regulate mood."

This proposition is not supported by references, nor by the paper being summarised. While one can find assertions that the neurotransmitter serotonin (also known as 5-hydroxytryptamine, 5-HT) is associated with a dimension of emotional stability vs "lability", such "monoamine hypotheses" actually rest on little replicable evidence and have been questioned from a historical context, for example by David Healy's The Antidepressant Era. I find the word "imbalance" misleading, suggesting comparison or balance with some other variable. The main supporting evidence would be that medications that increase availability of synaptic "brain chemicals" (most serotonin is actually in the gut) tend to reduce depression. However, that effect, although generally accepted, is suggestive only, and other hypothetical mechanisms exist, including that the medications affect the hypothalamus-pituitary-adrenal axis, have a knock-on effect of increasing GABA and thereby sedating and reducing anxiety, or most popularly now, increase hippocampal neurogenesis. Further, there are clearly counter-examples, (the antidepressant tianeptine inhibits 5-HT reuptake; reserpine passed an RCT for depression). In the context of the paper, while a strong effect of SSRIs would suggest a monoamine deficit, a weak or "clinically insignificant" effect such as that reiterated by Kirsch et al is active evidence against involvement of specific neurotransmitters. Or in other words, if the pain and debility of depression were "caused by" low synaptic serotonin/5-HT, all depressed people would have to do is take a regular pill to feel "normal", which clearly is not the case. Hence this paper is active evidence against the statement in the summary.

Other groups (including a minority of patients) would also argue with depression as a "medical illness", instead identifying it as a problem of the social environment, or occasionally as a rational response to the world (see eg Richard Bentall, David Smail and Mark Fisher).

The main conclusion to take away from Kirsch et al might be one of changing the idea of an antidepressant into one of a mild palliative for some people. Many people continue to take antidepressants regularly becuase they believe that, while they still have major problems with depression, they would be even worse off without them. What was new in Kirsch was trying to quantify the uptick of response in very high Hamilton scores. If Kirsch et al set out to assert that antidepressants have no effect on depression, however, the paper would have repeated criticism of the bias introduced generally by placebo washout, and the spontaneous unblinding that can occur during trials when patients become vary aware who is on the active drug because they can perceive its effect in ways other than mood. For this last point, see Moncrieff J, Wessely S, Hardy R. Active placebos versus antidepressants for depression. Cochrane Database of Systematic Reviews 2004, Issue 1. Art. No.: CD003012. DOI: 10.1002/14651858.CD003012.pub2.