“Depression” is not a diagnosis. It is a syndrome, a collection of signs and symptoms, like fever. It can happen in different diseases, and it can happen without any disease – just like fever.
The biggest mistake made today is to treat “depression” with “antidepressants”, as if we should always treat fever with “antifever” drugs (antipyretics) like Tylenol, without caring about any diseases that might be causing fever and treating those diseases (such as penicillin for pneumonia).
This doesn’t mean that antidepressants don’t “work”, in the same sense that we don’t say that Tylenol doesn’t “work” for fever. Antidepressants provide some symptomatic relief of depression, in some kinds of depression, but sometimes they provide little or no benefit – and sometimes they can even worsen depression if the underlying disease (like bipolar illness) is left inadequately treated.
So the first step in the treatment of depression is to realize that we aren’t treating “depression” itself, but whatever is causing the depression. There are a few types of depression with different likely causes, and different benefits with antidepressants versus psychotherapies or other treatments. They are:
1. Melancholic depression – this is a severe depression where patients do not eat or talk or move much. This condition responds best to electroconvulsive treatment (ECT) and some antidepressant medications (older agents like tricyclic antidepressants, TCAs, or monoamine oxidase inhibitors, MAOIs, are more effective than newer agents, like serotonin reuptake inhibitors, SRIs).
2. Mixed depression – this is a severe depression mixed with manic symptoms of marked agitation and rage, with mood swings of anger and anxiety mixed with the sadness of depression, The mood is “labile”, moving around a lot, unlike melancholia, where the mood is always down and sad. This kind of depression likely worsens with antidepressants, and is better treated with dopamine blockers (antipsychotics/neuroleptics) or with anticonvulsants (like valproate, Depakote, or carbamazepine, Tegretol) or lithium.
3. Neurotic depression – this is a moderate depression in which anxiety symptoms are always present (unlike the above two types which are episodic, come and go) to a mild degree and can worsen briefly for a few weeks or so to a moderate degree under stress. It responds to anything temporarily (antidepressants or psychotherapies) and nothing long-term (symptoms are never completely gone since they are always present to a mild degree).
4. Pure depression – none of the above. This is “really” depression, or to put it another way, “just” depression. It is like melancholia, but less severe, not mixed, and not neurotic. It may be that this kind of depression best responds to “antidepressants.”
In “major depressive disorder”, antidepressant medications work best in pure depression, less in mixed depression, and although they help neurotic depression, the benefit in the latter case is the same as placebo, meaning that anything would help, whether or not a specific pill is given. These conclusions are based on our interpretations of controversial meta-analyses of the topic, which are explained here. Melancholic depression responds well to some antidepressants (but not all), and electroconvulsive treatment (ECT).
These conclusions all apply to short-term treatment of a current depressive episode (they usually last up to a few months in most people before they become less severe or even completely go away). Once the depressive episode is over, most clinicians continue antidepressants indefinitely. Evidence for future benefit in prevention of new episodes with antidepressants is not as obvious or scientifically strong as the evidence for short-term benefit in specific subtypes. Long-term treatment is a more complex decision. Lithium and other agents may be better proven to prevent depressive episodes (irrespective of whether or not the depression happens in “major depressive disorder” or bipolar disorder).
Bipolar depression can also be of any of the four subtypes above, but it is associated with past manic or hypomanic episodes. Antidepressants are, based on the best scientific studies, basically ineffective in bipolar depression even short-term, and certainly long-term. They have little benefit, and, in about one-quarter of patients, they worsen the course of bipolar illness, causing more depressive episodes over time. Their use in bipolar depression is further discussed here.