The first modern antidepressants became available in the late 1950s, coinciding with introduction of randomized controlled trials in psychiatry. Tightening of licensing requirements has ensured good efficacy for new antidepressants. Overall efficacy and speed of response of most antidepressants are similar. In addition to metaanalyses of specific antidepressant classes, there have been metaanalyses confirming efficacy in specific patient groups or disorder subtypes, including dysthymia⁽¹⁾ and the elderly.⁽²⁾

There are moderate, but not very large, effect sizes. About 30 per cent more subjects respond well on antidepressants than on placebo, partly because good response is often seen in placebo groups, probably not due to the placebo but to spontaneous improvement and non-specific treatment effects such as those of seeing a helping figure, supportive psychotherapy, and hospitalization. Non-specific outcome tends to be worse among severely ill patients, but here also antidepressants may be less effective. Placebo-controlled trials are still mandated by regulatory authorities for new antidepressants, since comparisons between active drugs have high risk of type 2 error.

Table 4.5.8.1 lists the principal antidepressants and recommended doses, omitting minority drugs limited to a small number of countries. Some of the antidepressants listed are not available in all countries. Recommended doses depend partly on national authorities and readers should check the situation in their own country.

Tricyclic antidepressants. In addition to the tricyclics listed in Table 4.5.8.1, maprotiline and amoxapine are available in some countries. There have been many older efficacy reviews, with
further studies accumulating as tricyclics have been included in placebo-controlled trials of new antidepressants.

Earlier views that tricyclics were more effective in endogenous and psychotic depressives have not been confirmed and effects extend across a broad spectrum of depressives, extending more widely into anxiety disorders, panic disorder, and obsessivecompulsive disorder. Tricyclic use has lessened in recent years in favour of antidepressants with fewer side effects.

Serotonin and noradrenaline reuptake inhibitors (SNRIs). Selective SNRIs share the proposed therapeutic mechanisms of tricyclics but with fewer and different side effects, including the serotoninergic effect of nausea, and for venlafaxine at high dose, risk of blood pressure elevation and cardiac arrhythmia, with toxicity in overdose. There is good efficacy evidence for venlafaxine, with possible superiority over SSRIs,⁽³⁾ and emerging evidence for duloxetine.⁽⁴⁾ A further SNRI in some countries is milnacipran.

Selective serotonin reuptake inhibitors (SSRIs). Meta-analyses of SSRIs⁽⁵⁾ show efficacy comparable with tricyclics, but lower rates of side effects and discontinuation. There is conflicting evidence as to whether SSRIs are less effective than tricyclics and SNRIs in severe depression. SSRIs and clomipramine, the most serotoninergic tricyclic, are more effective than noradrenergic tricyclics in obsessivecompulsive disorder.

There have been vigorous debates regarding effects of antidepressants, particularly SSRIs, on suicidal behaviour, with many studies.⁽⁶⁾ Actual suicide does not appear to be increased and may be reduced. There is some evidence that suicidal feelings and attempts may be increased in the early weeks after starting, particularly for SSRIs, with development of tension or agitation, and in children and adolescents. On the other hand SSRIs and other newer antidepressants are considerably safer in overdose than tricyclics and SNRI. Warnings were issued by regulatory authorities in the United States and the United Kingdom when the evidence started to emerge. Most strongly the findings argue for careful clinical surveillance of patients prescribed antidepressants, particularly in the early phases of treatment when risk of suicidality has long been recognized as increased. This is in any case important in depressed patients.

Noradrenaline reuptake inhibitors. Only one newer selective NARI is available, reboxetine. It is clearly superior to placebo.⁽⁷⁾

Irreversible MAOIs. There are few MAOIs available, reflecting hypertensive and other interactions and limited use. The older MAOIs bind irreversibly to the enzyme, and new enzyme needs to be synthesized over 1–2 weeks to reverse the effect. Controlled trials⁽⁸⁾ show superiority to placebo in depression, and in anxiety disorders. Often high doses are necessary and the hydrazine MAOIs, phenelzine, and isocarboxazid, show better clinical response in slow acetylators. Tranylcypromine has additional stimulant effects, and has been viewed as more effective but with more risk of interactions.

From the late 1950s, there were suggestions of particular efficacy in atypical depression, variously regarded as non-endogenous depression, anxiety disorder with depression, or as a pattern of reversed vegetative symptoms with increased appetite, increased sleep, evening worsening, reactivity, and other features, a meaning currently predominant in the United States. Evidence is not strong, but comparative trials of phenelzine and tricyclics point to better effects than tricyclics with anxiety disorders and reversed vegetative symptoms.⁽⁸⁾ On the other hand MAOIs are mostly used as second-line drugs by psychiatrists, where other antidepressants have failed, irrespective of clinical picture.

Reversible competitive inhibitors. The reversible MAO-A selective drug moclobemide can dissociate from the enzyme and be displaced by substances with higher affinity, including tyramine. It shows superiority to placebo,⁽⁹⁾ but not at doses below 450 mg daily, and evidence is best for 600 mg. Many clinicians view moclobemide as less effective than older MAOIs.

Bupropion. Bupropion is relatively stimulant and is epileptogenic. It is licenced for depression in the United States, and in some other countries as an aid to smoking cessation.

Mianserin. Mianserin, an older drug which blocks alpha-2 autoreceptors, is sedative in side effects and carries a definite although low risk of agranulocytosis.

Mirtazepine. Mirtazepine blocks alpha-2, 5HT2, and 5HT3 receptors. It has been shown superior to placebo and is also a sedative.⁽¹⁰⁾

Trazodone. Trazodone, an older drug is relatively sedative carries a risk of priapism in males.

New classes of antidepressants, not yet licenced, are continually being sought.⁽¹¹⁾ Agomelatine, an agonist at melatonin MT1 and MT2 receptors and a 5HT2C antagonist has shown evidence of efficacy. Efforts have been made to develop drugs, which inhibit cortisol secretion. Development pathways for new antidepressants are long and failures due to weak treatment effects or major adverse effects are common.

The therapy models can broadly be divided into ‘process-orientated’ therapies such as psychodynamic or supportive–expressive therapies, and ‘outcome-orientated’ therapies, such as cognitive behaviour therapies (CBT) and interpersonal therapy (IPT), which primarily focus on symptom reduction. The latter EST are protocol or guideline driven, with a manual describing the therapy in detail. In large-scale randomized controlled treatment trials (RCTs) they have been found superior to pill or psychological placebocontrolled treatments or efficacious as antidepressants or other established treatments.⁽²⁸⁾ The EST are usually brief (12–20 sessions), delivered by a trained therapist, establishing a working alliance and using an individualized case formulation to focus on specific ‘here and now’ problems, encouraging between session ‘homework’ tasks to enable development of new coping skills.

The two best-evaluated therapies are CBT (including behaviour therapies) and IPT, with many meta-analyses of a large number of RCTs. There are fewer studies of dynamic therapies. Early outcome studies and resultant meta-analyses were hampered by use of ‘completer’ samples rather than ‘intent-to-treat’ analyses, but found specific psychological therapies, particularly CBT, more effective than other verbal therapies or waiting-list controls. A meta-analysis of 29 carefully selected controlled trials,⁽²⁹⁾ using intent to treat analyses, found efficacies of individual CBT (response rate, 50 per cent), behavioural therapy (55 per cent), and IPT (52 per cent) in the treatment of the acute episode were not significantly different to pharmacotherapy (58 per cent). Brief dynamic psychotherapies, mainly group therapies, were less effective (35 per cent) and only marginally superior to waiting-list controls (30 per cent). CBT was less effective in a group format, behavioural therapy was equally effective, and IPT was more beneficial when a significant other took part in therapy.

A recent meta-analysis⁽³⁰⁾ found CBT and IPT equally effective in reducing depressive symptoms in heterogeneous groups of patients with depressive symptoms and syndromes, and both were superior to counselling, other therapies or control treatments in primary care. When compared in meta-analyses to an active treatment, however, CBT and IPT show only equivalent efficacy or marginal superiority. Almost all RCTs of psychological therapies have been in outpatient depressives, not above moderate severity and comparisons with pharmacotherapy do not represent severe or inpatient disorder.

Earlier CBT studies suggested it was effective over a range of severity and endogenicity. However, in the three-centre NIMH study,⁽³¹⁾ which sought to overcome earlier methodological flaws, a secondary analysis found CBT less effective than pharmacotherapy in the more severe disorders within the outpatient range studied. Another study found that individuals with mild or moderately severe depression did equally well with either 8 or 16 sessions, but those with severe depression showed better response with 16 sessions. A large RCT in subjects with severe major depressive disorder⁽³²⁾ found no difference overall between antidepressants and CBT, but also significantly greater effects for antidepressants at the trial centre where the CBT therapists were less experienced.

IPT Has been found effective in moderate numbers of RCTs, in a variety of situations, including combination studies with antidepressants and some situations where medication as a first-line treatment for depression is not easily feasible (pregnancy, postpartum, adolescence, and in a developing country).⁽³³⁾ The NIMH study⁽³¹⁾ is the only one that has compared IPT with a pill-placebo control. Interpersonal therapy was nearest to the effectiveness of imipramine but a little weaker, and IPT was more beneficial than CBT in patients with more severe outpatient depression.