Kraepelin established manic–depressive disorder on the psychiatric map as a remitting but recurrent disorder. As the concept of affective disorder became broader, its potentially chronic nature was also recognised in long-term follow-up studies. However, the advent of antidepressants in the late 1950s gave rise to more therapeutic optimism, and a tendency to regard unremitting and chronic features as reflecting characterological rather than affective disorders. In the late 1970s, an era of reappraisal started, and it is now clear from long-term follow-up studies that recurrence is common and chronicity occurs, although probably in a smaller proportion than previously. The wheel has come full circle recently in Britain, as the Department of Health and the Medical Research Council have established relapse and chronicity in depression as major health service research priorities.

The realisation that major depression is often both chronic and recurrent has slowly begun to change the way that depression is diagnosed and treated. In particular, the need for continuation and maintenance treatment is an issue that now deserves increased attention, especially with the availability of new classes of antidepressant treatments, which have excellent efficacy and more favourable side-effect profiles. Although the serious consequences of depressive disorders clearly indicate the need for effective and prompt intervention on the part of clinicians, the results of several studies indicate that patients with depression consistently receive no or low levels of antidepressant therapy. It is hoped that, through continued education of health care providers and patients about the consequences of depression, the issue of undertreatment of this serious illness will be resolved.

Over the past decade, major epidemiological studies have been conducted to determine the prevalence of depressive syndromes, primarily major depression or dysthymia. The highest prevalences occur in younger cohorts (18–29 years); considerably lower prevalences are found in older individuals (45 years and above), with the lowest in those aged 65 and older. Several studies have confirmed an increase in the cumulative lifetime estimates of major depression in successively younger birth cohorts during this century. At the same time, questions have been raised about the low prevalence of depression in the elderly, including the role of confounding factors (e.g. differential morbidity and response-biased memory). Standardised diagnostic assessment procedures may be insufficiently adapted for use in the elderly. It has also been recognised that a substantial number of elderly individuals suffer from clinically relevant symptoms of depression but do not meet the criteria for major depression. Future research will be required to elucidate fully the apparently changing rates of depression.

The separation of persistent depression into meaningful and useful subcategories, including major depression, dysthymia, recurrent brief depression, and depressive personality disorder, is the subject of much debate. Depressions can be grouped on the basis of their type and severity of symptoms, aetiology, clinical course, or their association with other psychiatric illnesses. Several investigators have conducted epidemiologic and family studies to evaluate the prevalence of depressive disorders, their diagnostic stability over time, and the amount of overlap among the disorders. Although progress has been made toward a better understanding of the different disorders, insufficient evidence exists to support the hypothesis that these disorders are separate and distinct from one another. However, preliminary data suggest that depressive personality disorder is separate from the other disorders. Additionally, several questions have been raised, particularly the extent to which differentiation between the depressive disorders, specifically major depression and dysthymia, has an impact on treatment decisions.

Long-term treatment of depression encompasses two separate phases: relapse and recurrence prevention. Relapse prevention aims to consolidate the response to acute treatment. Some tricyclic antidepressants (TCAs) have been shown to be effective, possibly in lower than standard acute treatment doses. The selective serotonin reuptake inhibitors (SSRIs) have been shown to be effective at the same minimum effective doses used to treat acute depression, or in a lower dose as with citalopram. Recurrence prevention aims to reduce the risk of onset of a new episode of depression in patients with recurrent depression. Imipramine has been thoroughly studied in unipolar depressed patients in full therapeutic doses for up to five years and is clearly effective. Other TCAs have not been adequately tested and may not all be equally effective. The SSRIs fluoxetine, paroxetine and sertraline have also been shown to be effective in reducing the risk of new episodes of depression.

Although most studies of chronic depression show significant improvements with antidepressants versus placebo, the average Hamilton Depression Rating Scale results attained in the active-treatment group range between 10 and 14, suggesting that many patients only partially responded and failed to reach premorbid levels of symptom remission. Studies on the fate of these patients suggest that they are much more vulnerable to relapse, work impairment and suicide. Thus, partial response may be one form of treatment resistance, falling between total failure of response in a minority of patients, and a tendency to relapse or recur despite adequate maintenance treatment. Further study is needed to address the problem of improving the quality of response and attempting to reduce the detrimental effects of depressive illness in terms of relapse and recurrence.

A review of epidemiologic and clinical data on depression suggests that psychotherapy is both an important alternative as well as a supplement to medication for the maintenance treatment of depression. Psychotherapy is an alternative for patients during periods when medication may not be suitable or feasible (e.g. pregnancy, nursing, before or during major surgery, or in the elderly). Psychotherapy also has a role in maintenance treatment in dealing with the social and interpersonal consequences or triggers of recurrent depression. Although the number of continuation or maintenance treatment trials that include psychotherapy is quite limited, the efficacy of maintenance treatment in the delay of recurrence and enhancement of social functioning is best established for interpersonal psychotherapy. There are also some results concerning cognitive and behavioural therapies for maintenance treatment of depression.