Long term pharmacotherapy of depression

The high rates of persistent morbidity, recurrence, and death among patients treated for depression ^1–2 have highlighted the importance of long term psychological and pharmacological treatment. What evidence do we have that long term drug treatment is effective, and how should we choose between individual drugs?

Long term drug treatment comprises continuation and maintenance treatment. Continuation treatment helps to consolidate recovery from depressive episodes and prevent relapses (worsening of continuing or recently treated episodes). Maintenance or prophylactic treatment helps to prevent occurrence of new episodes and is usually recommended for patients who have had at least three depressive episodes in the preceding five years. The distinction between continuation and maintenance treatment is somewhat arbitrary and may not reflect underlying biological processes, but most researchers agree that four to six months’ remission should occur before a recurrence is diagnosed.

Differences in methods between the many trials of long term pharmacological treatment³,⁴ make interpretation difficult, but about 60% of patients who respond to an antidepressant and are then given a placebo remain in remission for up to two years. If instead of receiving a placebo they continue on the drug they have a 20-25% better chance of maintaining their improvement⁴—that is, twice as many relapses occur on placebo as on antidepressants (about 40% v 20%). Drugs also increase the time to onset of relapse or recurrence and reduce the severity of residual depressive symptoms in those who do not relapse. Nevertheless, studies have thrown little light on which patients benefit most from long term treatment, the comparative effects of different psychological and pharmacological treatments, and the optimum length of treatment. Moreover, few trials have continued for longer than two years.

The benefit of long term drug treatment has been clearly shown only in outpatients with major depression. We cannot assume that the same benefits will be achieved in the milder, heterogeneous cases of “depression” encountered in general practice. We should not prescribe long term for people with infrequent, short bouts of mild depression and those whose low mood reflects changing social circumstances unless there is convincing evidence in individual patients that they have benefited from such treatment. We should also hesitate to prescribe long term for patients whose depression is an episodic symptom of personality disorder, an effect of alcohol or drug abuse, or a phenomenon perpetuated by the desire to remain in the sick role.

Few long term comparative studies of the efficacy of different drugs have been carried out, and meta-analyses of the results of short term trials have failed to show important differences between different types of antidepressants.⁵ Factors such as tolerability, unwanted effects, toxicity in overdose, and cost must therefore determine which drug to use.

A common index of tolerability is the discontinuation rate in clinical trials. Despite claims that newer antidepressants are better tolerated than older tricyclics, such as amitriptyline and imipramine, only 1-5% fewer patients receiving selective serotonin reuptake inhibitors than receiving tricyclics drop out from trials—figures of doubtful practical importance.⁵ The difference between the drugs is largely attributed to fewer dropouts due to side effects, although it is often difficult to know why patients stop their treatment.⁶ Furthermore, discontinuation rates from trials may not accurately represent routine clinical practice or long term treatment (when adaptation to unwanted effects may occur). Meta-analyses of efficacy and discontinuation rates have suggested that significant differences do exist between individual drugs, although the methods of these meta-analyses have been criticised.⁵

Death is more likely to result from overdoses of older tricyclic drugs than newer compounds.⁷ However, only about 4% of all suicides are due to overdoses of single antidepressants, and it is not known what proportion of these overdoses are taken during treatment (when drug choice is relevant). Furthermore, a higher suicide rate among patients taking tricyclic drugs could be accounted for by doctors prescribing these drugs more often for patients prone to suicide,⁸ and the overall suicide rate (by any method) among patients treated with new and old antidepressants is similar.⁹

The average net ingredient cost of an NHS prescription for a selective serotonin reuptake inhibitor in 1995 was £27.21 compared with £0.77 for amitriptyline. If all patients were prescribed serotonin reuptake inhibitors the annual cost (at 1995 prices and consumption rates) would be £350m more than if they were all prescribed amitriptyline.⁶ The long term benefits purchased from this are slightly lower discontinuation rates,⁵ possibly with fewer relapses and recurrences, and fewer deaths from overdose.⁷ Conversely, the additional cost means there is less money available for other purchases—for example, four million psychiatric outpatient attendances or almost 22 million hours of community psychiatric nurse time.⁴

The results of cost effectiveness and cost benefit assessments depend on the model used. Recent overviews do not recommend expensive newer antidepressants as first line treatment,^4–6,¹⁰,¹¹ but these drugs should be prescribed both short and long term for patients who cannot tolerate older antidepressants and/or have a high risk of suicide by overdose. Because the newer antidepressants have less sedative and autonomic effects, they should also be given to patients with depressive disorders who are prone to accidents ¹² or have cardiovascular disease.⁴,⁶