Rationale for, Barriers to, and Appropriate Medication for the Long-Term Treatment of Depression

Rationale for, Barriers to, and Appropriate Medication for the Long-Term Treatment of Depression

David J. Nutt, DM, FRCP, FRCPsych, FMedSci

Antidepressants have proven efficacy in the treatment of acute depressive episodes and the prevention of relapse over the long-term. However, whether due to ignorance about the chronicity of depression, intolerable adverse effects, or an inappropriate fear of dependence, antidepressants are often discontinued after remission or recovery from an acute episode, which frequently leads to relapse or recurrence. This, in turn, increases the risk of subsequent poor treatment response and lifelong depressive chronicity. Clinicians should focus on preventing depressive relapse with long-term antidepressant pharmacotherapy, thereby improving patients’ overall outcomes, particularly with patients at high risk for relapse. When patients with depression have comorbid anxiety, benzodiazepines may be useful but should be used only as short-term augmentation during the beginning phase of antidepressant treatment; long-term treatment of comorbid anxiety is better managed by antidepressants that also treat anxiety disorders.

(J Clin Psychiatry 2010;71[suppl E1]:e02)

From the Neuropsychopharmacology Unit, Division of Experimental Medicine, Imperial College, London, United Kingdom.

This article is derived from the roundtable discussion "International Consensus Group on Depression," which was held on September 1, 2009, in Tokyo, Japan, and supported by an educational grant from GlaxoSmithKline.

Professor Nutt is a consultant for GlaxoSmithKline, Lundbeck, Pfizer, Servier, and sanofi-aventis; has received grant/research support from Organon and P1vital; and is a member of the speakers or advisory boards for Organon, GlaxoSmithKline, Lundbeck, Servier, and Bristol-Myers Squibb.

Corresponding author: Professor David J. Nutt, Imperial College London, Burlington Danes Bldg, Hammersmith Campus, 160 Du Cane Rd, London W12 0NN ([email protected]).

doi:10.4088/JCP.9058se1c.02gry

© Copyright 2010 Physicians Postgraduate Press, Inc.

Depression is a recurrent or chronic disorder with poor long-term outcomes, including ongoing morbidity and mortality from other psychiatric and medical conditions, impaired psychosocial function, and an increased tendency to relapse for patients who fail to reach remission. For patients who attain full symptomatic recovery, a high risk of depressive recurrence still exists, especially in the absence of long-term antidepressant therapy. In some populations, depression is treated with benzodiazepines, but this class of medications may actually exacerbate some of the problems associated with an illness already prone to multiple episodes and suboptimal patient outcomes. The need for and barriers to long-term treatment of depression, as well as evidence-based pharmacotherapy, are reviewed.

the Need for Long-Term Treatment With Antidepressants

Click figure to enlarge

Long-term treatment of depression with antidepressants benefits many patients by reducing the likelihood of relapse or recurrence and increasing the amount of time spent asymptomatic between episodes. Figure 1 illustrates the typical course of illness for patients with depression either receiving only acute-phase treatment or receiving longer-term treatment designed to prevent relapse.1 As shown, discontinuing antidepressant treatment in the acute phase before full symptomatic remission drastically increases the likelihood of relapse of the original episode, and discontinuing treatment during the continuation phase after symptomatic remission likewise increases the risk of relapse. Discontinuing treatment during the maintenance phase, which begins after full recovery, increases the risk of developing a new depressive episode (ie, recurrence).

Long-term antidepressant treatment could benefit not only individual patients but also society as a whole. Long-term treatment may reduce the significant public burden that is associated with depression by decreasing lost productivity, suicidality, morbidity, and the attendant costs of care.2,3

Barriers to Long-Term Antidepressant Treatment

Clinicians today have the means to treat many patients with depression to wellness with reasonable reliability. However, a focus on keeping patients well, which is essential for most patients and critical for those with a high risk of relapse, appears to be lacking in the clinical setting. Probably the most pervasive reason for this clinical lapse is a misunderstanding of the true nature of depression. Traditionally, pharmacotherapeutic strategies have proceeded from what might be called a "fracture model" of treatment for depression. That is, depression has often been treated like a broken bone, with antidepressants being analogous to plaster casts; when the injury appears to be healed, the treatment is removed. Unfortunately, this traditional mindset is at odds with the reality of the course of depression. Depression should be thought of as a chronic, relapsing illness similar to epilepsy, which, like depression, is a disorder of the brain. The goal in treating epilepsy is to stop patients’ epileptic fits and then to keep patients seizure-free for as long as possible. Over the course of long-term treatment for epilepsy, clinicians make reasoned judgments assessing the potential risks and benefits of reducing/discontinuing treatment versus maintaining current treatment levels. Treatment for depression should be reconceptualized and carried out in the same manner.

A fear of adverse events also contributes to the hesitance of some clinicians and patients in committing to a long-term antidepressant treatment strategy. Undoubtedly, a small proportion of the overall treatment population will have enduring side effects that limit compliance. Side effects are particularly problematic with tricyclic antidepressants (TCAs)—a class of medications associated with tachycardia, tremor, and sedation—which may be why they were often used at a subtherapeutic dosage.4-7 Impairments in cognitive function associated with TCAs have also been reported, although this may relate to sedation. Effective, newer antidepressants have better safety profiles than TCAs. One of the few significant long-term adverse events associated with newer antidepressants is weight gain.8,9 Newer antidepressants, such as selective serotonin reuptake inhibitors (SSRIs), also are much safer in overdose,10 which explains the disappearance of TCAs from use in many countries. A United Kingdom report of deaths from drug poisoning found that the death rate associated with TCA use from 1993 to 2004 was about 10 times greater than that with SSRIs.11 However, because of the recent concern about SSRIs and potential suicidal thoughts or actions in certain populations,12 some doctors have switched from prescribing SSRIs to TCAs, which may have had the effect of increasing completed suicide with these drugs.13

Fear of developing a chemical dependence to antidepressants is another potential barrier to long-term pharmacotherapeutic treatment for depression. In the United Kingdom, the widespread and mistaken belief that antidepressants are addictive is based, in part, on previous experience with benzodiazepines. The spate of lawsuits aimed at antidepressants is modeled after a similar campaign against benzodiazepines—a campaign that failed even though benzodiazepines, unlike antidepressants, may cause dependence in some patients. Press activity14 surrounding the lawsuits contributed to the spread of what is now a commonly held belief that antidepressants are addictive. Combating patients’ fears of chemical dependence, problems with tolerance, and withdrawal effects should improve antidepressant adherence and help clinicians successfully implement long-term treatment plans when appropriate.15

Evidence for Prevention of Relapse/Recurrence With Long-Term Antidepressant Treatment

European regulatory directives require placebo-controlled studies showing efficacy for a duration of at least 6 months before an antidepressant is approved for long-term use. As a result, a large amount of data has demonstrated that continuation and maintenance antidepressant pharmacotherapy reduces the risk of relapse and recurrence relative to placebo treatment (ie, discontinuation).

In 1992, Kupfer et al16 published the results of a 5-year landmark study of 128 subjects with recurrent depression who had recovered from an acute episode while being treated with the TCA imipramine and interpersonal psychotherapy and had then spent 3 years in recovery while continuing treatment with imipramine with or without psychotherapy. After 3 years, 20 subjects who remained in recovery were randomly assigned to take either imipramine (n = 11) or placebo (n = 9) for the next 2 years. During the 2-year continuation phase, two-thirds of the placebo group had a recurrence, but only 1 of the imipramine group completers (n = 10) had a recurrence, and that subject’s blood drug levels suggested noncompliance. This study demonstrated the prophylactic effect of long-term antidepressant treatment, which was, in this case, a particularly unpleasant and difficult-to-tolerate antidepressant given at a higher dose (mean = 236 mg/d) than most people in the United Kingdom or Japan would use. The majority of relapses in the placebo group occurred between about 10 and 30 weeks after placebo initiation, meaning that these relapses were not withdrawal reactions but a reemergence of an illness that had never gone away and had been prevented from expressing itself by the antidepressant.

Click figure to enlarge

One of the first long-term studies with newer antidepressants was a randomized, double-blind, placebo-controlled study17 of the SSRI paroxetine, which found that the antidepressant was effective in preventing relapse after recovery from an acute episode in subjects with major depression who had had at least 2 episodes in the previous 4 years. Subjects who responded to 8 weeks of treatment with paroxetine (N = 135) were randomly assigned to receive paroxetine or placebo for 1 year. Paroxetine was significantly more effective than placebo in preventing relapse (P < .01) and in extending time spent well prior to relapse (P < .001) (Figure 2).

Study designs for long-term efficacy of most modern antidepressants have typically followed the model of the previously described trials. Subjects are treated to remission or recovery and then randomly assigned to take either the same antidepressant they achieved wellness with or placebo. Results consistently demonstrate an advantage in relapse prevention for most modern antidepressants—certainly all the SSRIs—compared with placebo for at least 6 months.

In an influential and important meta-analysis,18 Geddes and colleagues analyzed long-term treatment data through August 2000 and found a significant benefit for antidepressants compared with placebo in relapse prevention (P < .00001). The meta-analysis reported that the antidepressants studied (including SSRIs, TCAs, monoamine oxidase inhibitors, and others) produced a relative risk reduction of about 20% over the course of 1 year, which is a number needed to treat (NNT) of about 5; ie, for every 5 patients treated with an antidepressant for at least 1 year, at least 1 of them will definitely have a better outcome than if he or she was not taking an antidepressant. The NNT for treatment of an acute episode is usually between 5 and 7,19 so an NNT of 5 for 12 months of continuation treatment demonstrates a profound pharmacologic and therapeutic effect. Among the studies in the meta-analysis that evaluated antidepressant relapse prevention for 3 years, the NNT was reduced to about 3, which is a more profound therapeutic benefit. Few treatments in medicine have a proven NNT of 3. A newer meta-analysis by Furukawa et al20 found that long-term antidepressant therapy is effective in relapse prevention after remittance from an acute phase, and a meta-analysis by Williams et al21 reported that the maintenance efficacy of antidepressants for at least 1 year was not affected by the patient’s initial depressive severity or age. Therefore, strong evidence supports the notion that clinicians should be using antidepressants to prevent relapse, which is a change in emphasis from the traditional view of antidepressant treatment.

Clinical Implications

The Geddes et al meta-analysis18 found an overall constant relative risk reduction of approximately 50% for antidepressant maintenance treatment regardless of the duration of previous treatment. Patients successfully maintained for 3, 6, 12, or 36 months will continue to derive a benefit from antidepressant therapy, but will return to baseline risk for relapse if treatment is discontinued. Of course, the benefit of continuation treatment depends on individual patients’ absolute risk of relapse, which is influenced by a number of factors, such as number of previous episodes and medical or psychiatric comorbidity. Individuals who are at high risk of relapse must be treated and appropriately maintained on an effective antidepressant. Furukawa and colleagues20 emphasized that applying maintenance treatment principles in an individualized way will best prevent relapse.

Discontinuing Antidepressant Pharmacotherapy

The epilepsy model may again be helpful when considering if or how to discontinue a patient’s antidepressant pharmacotherapy. In epilepsy, stopping treatment with an anticonvulsant is an active process dependent on several factors; the decision does not lie solely with the patient. Cessation of treatment with an antidepressant should be treated similarly, particularly for people who have had recurrent episodes of depression. The decision should be based on the patient’s current level of recovery, ongoing life stresses, and previous number of episodes and should come about, not as a choice made by either the patient or the doctor, but rather through doctor-patient consultation.

Click figure to enlarge

Paykel et al22 found that residual depressive symptoms predict relapse (Figure 3). In this particular study, two-thirds of patients with residual symptoms relapsed within 6 months of meeting remission criteria once treatment was stopped, whereas the relapse rate for those symptom-free was only 25%. Therefore, discontinuing antidepressant treatment in patients with residual depressive symptoms greatly increases their risk for relapse. A 4-year naturalistic study23 reported a 91% relapse rate during follow-up among patients with partial remission after acute treatment compared with a 51% rate among those with complete remission. Therefore, cessation of antidepressant pharmacotherapy should not be considered unless the current level of recovery is, at the very least, complete remission in a situation of relative life stability.

A history of multiple previous depressive episodes also contributes to a higher risk of relapse,23,24 which makes it all the more important that successful treatment of an acute episode in patients with this history be followed with appropriate continuation and maintenance therapy. Each new episode further increases the risk of another and promotes depressive chronicity.

Finally, patients and doctors must understand that current efficacy of an antidepressant does not guarantee future efficacy.25 In other words, if a patient recovers from a depressive episode while taking a particular antidepressant and, upon recovery, stops taking the medication, the same antidepressant may not work as well, or even at all, if the patient relapses. Therefore, discontinuation of antidepressants in patients with risk factors for relapse is probably unwise.

Use of Benzodiazepines in Depression

Benzodiazepines are recommended for symptoms of anxiety rather than depression and are not recommended for long-term use.19,26 Some have suggested that patients may develop tolerance to the effects of benzodiazepines, thus requiring higher doses, yet benzodiazepine tolerance probably occurs less often than has been claimed.27 The dependence-producing nature and abuse potential of benzodiazepines contribute to the diversion of the medication into the illegal drug market.28 In addict populations, people commonly abuse benzodiazepines if they cannot get heroin.

Long-term benzodiazepine therapy does not correlate to long-term treatment with antidepressants. Benzodiazepines work quickly; the effect builds up over a couple of weeks and persists for a period, but when the medication is discontinued, the patient may experience rebound or withdrawal symptoms, which, for some people, are so unpleasant that they have to resume taking the drug.28 When discontinuing an antidepressant, discontinuation symptoms may also occur but without the same immediate reemergence of the underlying disorder (or rebound) as is often experienced with benzodiazepines. Antidepressant discontinuation is associated with the risk of relapse but not the acute rebound of the underlying pharmacology that happens with the cessation of benzodiazepine therapy.29

Patients have claimed to have developed depression because they were taking benzodiazepines, and the claims were supported by some professionals in the field.30 But, this controversial topic is difficult to properly address because of vested interest and litigation. What can be said is that benzodiazepine withdrawal is unpleasant and stressful, and some of the symptoms of withdrawal can mimic depression. Theoretically, benzodiazepine therapy could inhibit recovery from depression because the blunting of memory and reactivity associated with benzodiazepines might impair the social learning skills that assist in overcoming depression.

Mechanism of Action of Benzodiazepines

Benzodiazepines target the brain’s calming system by binding to and enhancing γ-aminobutyric acid A (GABA-A) receptors, which are the main inhibitory system in the brain. Anxiety is associated with low GABA function. Benzodiazepines enhance GABA function, thereby reducing anxiety. However, long-term benzodiazepine use undermines their action by producing adaptive changes in GABA-A receptors such that, upon cessation of treatment, GABA function has been reduced, sometimes beyond what was present before benzodiazepine treatment began.29 This reduced GABA function leads to rebound anxiety and possibly even worse anxiety than the patient experienced prior to treatment initiation, which contributes to benzodiazepine dependence and is a fundamental difference between how benzodiazepines and antidepressants work. The process of adaptation and rebound is related to alterations in the expression of GABA-A receptor subunits.29 When subunits are altered, the function of the entire system is altered, which may help to explain some of the negative effects of long-term treatment with benzodiazepines.

Appropriate Short-Term Uses for Benzodiazepines

Initially, antidepressants may cause a worsening of anxiety in patients who have depression with comorbid anxiety. This phase of worsened anxiety is temporary but can be difficult for some patients. Benzodiazepines may be helpful in preventing antidepressant treatment dropout for patients experiencing this brief phase of delayed onset of antidepressant action.31 Antidepressants such as SSRIs have anxiolytic effects as well as antidepressant effects, and therefore benzodiazepines are not needed after the effects of antidepressant drugs begin. The United Kingdom guidelines for the treatment of depression19,26 recommend only short-term use of benzodiazepines, if they are to be used at all. Benzodiazepines may also be prescribed to lessen the occurrence of the emergent adverse events associated with the onset of antidepressant use in anxiety disorder, such as the so-called jitteriness syndrome.19,32 However, the benefits of benzodiazepine therapy must be weighed against the risks of long-term use.

Benzodiazepine Prescription Practices

The United States and the United Kingdom have had successful academically led campaigns to replace benzodiazepines with antidepressants, particularly SSRIs and other newer antidepressants. Unfortunately, a markedly higher rate of benzodiazepine use exists in Japan than in the West.33 Not only are benzodiazepines widely prescribed in Japan, but more than 70% of Japanese outpatients with mood disorders who are prescribed benzodiazepines continue taking them for more than 2 years.32 It is uncertain whether these findings are true for the rest of Asia, but in Japan there are still lessons to be communicated about the relative safety and benefits of both benzodiazepines and antidepressants.

SUmmary

Depression is often a life long illness and clinicians must treat it as such. Relapse of a depressive episode or occurrence of a new episode increases the risk of more subsequent episodes, and the occurrence of multiple episodes predicts poorer treatment response and outcomes. The traditional approach of treating only the acute episode ignores the true nature of depression. Overall patient outcomes and functioning can be improved if treatment is focused on the long-term prevention of relapse and recurrence. Antidepressants have demonstrated promising results in the prevention of relapse and should be included in appropriate long-term treatment strategies. Benzodiazepines, which have been used to treat depression, are more effective for anxiety than for depression and have potential abuse issues when used over the long-term. Physicians must overcome barriers to long-term antidepressant treatment, including fear of dependence or adverse events.

DISCUSSION

Dr Sakamoto: I believe that patients with a history of depression should receive long-term treatment. However, I wonder if long-term treatment is necessary for all patients with major depression. For instance, if a patient is experiencing his or her first episode and has only mild depression, could we discontinue pharmacotherapeutic treatment after full remission and institute some psychoeducation so that the patient knows to report back at the first sign of relapse?

Professor Nutt: With a single episode of mild depression, long-term treatment may not always be necessary, but all patients need to be made aware that the risk of a new episode is increased after just 1 episode,34 so education is important. If people start to get symptoms again after recovery, they should seek treatment quickly.

I saw a teacher who had retired due to ill health at 28 years of age. He had been teaching for 6 years and had experienced 9 depressive episodes. Each time the depression was treated, he got well, and treatment was discontinued. I asked why he did not continue taking the antidepressant after he remitted, and he said, "Well, my doctor stopped it because I got well." So, eventually he was so ill he had to retire. His career had been ruined within 6 years by repeated depressive episodes. Clearly, that is an extreme case. But there is no question that long-term treatment is absolutely required for some, if not many, patients.

Dr Higuchi: Recently, in Japan, television and newspaper reports about the predominant treatment methods used in England have said that you use drugs but you also use cognitive-behavioral therapy (CBT) for relapse prevention. Is it true that for first episodes of depression, especially with mild symptomatology, you use CBT only and use pharmacotherapy if CBT is ineffective?

Professor Nutt: No. We give patients a choice. They can use CBT or drugs depending on their preference. A strong push exists for patients to try nonpharmacologic treatment options,26 but the decision ultimately depends on the individual patient’s illness profile and treatment preferences. For milder depression, CBT is certainly an acceptable and sensible approach, if you have access to it, but even with the recent increase in qualified CBT administrators, only a very small proportion of patients can get access. I think most of us psychiatrists like the idea of trying both. When quality CBT is more readily available, I think many British psychiatrists would like to use a collaborative treatment approach using a combination of drugs and CBT. Problem-solving therapy or CBT may give patients skills in preventing negative thinking, which can lead to a recurrence of depression.

Dr Higuchi: When we use electroconvulsive therapy (ECT) in severe depression or refractory depression, maintaining the efficacy of ECT is difficult. Do you know if ECT efficacy can be maintained with drugs?

Professor Nutt: Few controlled data exist on the subject, but I tend to use an antidepressant in a class that is different from the ones that the patient has previously not responded to, usually an agent that has mixed activity covering at least 2 neurotransmitters, or possibly a drug combination. Some evidence suggests that noradrenergic-potentiating antidepressants may be better prophylactically after ECT than serotonergic agents.35 But, it is very much still trial and error, and occasionally we use maintenance ECT. As ECT becomes more acceptable, some patients are beginning to use it monthly or even twice monthly.

Dr Gelenberg: I have 2 responses to your earlier comments. The first is that the American Psychiatric Association’s new guideline—and I think it is the same as in the last guideline36—will say something similar about CBT, that the patient’s choice and the feasibility of access should determine treatment selection in the mild-to-moderate cases.

And, in terms of access to CBT-trained therapists, in the United States, it is difficult even in major metropolitan areas to find an adequately trained CBT therapist. Initial evidence indicates that computer-based systems work pretty well.37 Given the difficulty of access, digitally administered CBT may be beneficial, whether used as solo treatment or adjunctive to medication.

Professor Nutt: I am supportive of computer-based CBT, assuming the patient’s depression is mild enough for him or her to be able to make use of it.

Dr Davidson: The United Kingdom National Institutes of Health and Clinical Excellence (NICE) guidelines26 endorse computerized CBT for mild depression.

I wanted to make a comment about TCAs. The marketing data in the United States suggest that they are still being used a fair amount in primary care. I think that there are probably 2 reasons for this. One is the prevalence of pain with depression. The SSRIs are not very effective for pain, and the SNRIs could be better. And the second reason has to do with sleep problems. Low doses of TCAs enhance sleep. What do we advise primary care doctors or pain specialists whose patient with pain is depressed? Do we recommend that they continue with the TCAs or do they add an SSRI, and if so, which SSRI and which TCA?

Professor Nutt: Those are great unknowns. Hopefully, with the introduction of new guidelines, we will be able to work out some answers.

A final comment about mild-to-moderate depression is that the British Association for Psychopharmacology has dealt with this question in quite a clever way. They rely on 2 variables: severity and duration.19 Even if a patient has relatively less severe depression, if it has been chronic, drugs may offer benefit. It may be useful to think about mild depression not as a single construct relating to symptom severity but also in the context of duration.

Dr Terao: How do we teach primary care physicians to manage patients already dependent on benzodiazepines? For example, Lader and colleagues38 recommended switching short-acting benzodiazepines with long-acting benzodiazepines and then gradually reducing that benzodiazepine.

Professor Nutt: You and Lader and colleagues are absolutely right. However, the process of switching to longer-acting benzodiazepines sometimes is not easy because patients are very attached to their short-acting benzodiazepine and doubt another agent can be as effective.

Although trials are lacking, clinical experience has taught me to institute an antidepressant before stopping a benzodiazepine. I would stabilize the patient on an SSRI a few months beforehand to control any emergent depression or anxiety.

Drug Names: imipramine (Tofranil), paroxetine (Paxil, Pexeva, and others).

Disclosure of off-label usage: The author has determined that, to the best of his knowledge, no investigational information about pharmaceutical agents that is outside US Food and Drug Administration−approved labeling has been presented in this article.

References

1. Kupfer DJ. Long-term treatment of depression. J Clin Psychiatry. 1991;52(suppl 5):28-34. PubMed

2. Goldstein BJ, Goodnick PJ. Selective serotonin reuptake inhibitors in the treatment of affective disorders, pt. III: tolerability, safety and pharmacoeconomics. J Psychopharmacol. 1998;12(suppl B):S55-S87. PubMed

3. Sobocki P, Jönsson B, Angst J, et al. Cost of depression in Europe. J Ment Health Policy Econ. 2006;9(2):87-98. PubMed

4. Fujita A, Azuma H, Kitamura T, et al. Adequacy of continuation and maintenance treatments for major depression in Japan. J Psychopharmacol. 2008;22(2):153-156. doi:10.1177/0269881107079049 PubMed

5. Kusturica J, Zulić I, Loga-Zec S, et al. Frequency and characteristics of side effects associated with antidepressant drugs. Bosn J Basic Med Sci. 2002;2(1-2):5-11. PubMed

6. Degner D, Grohmann R, Bleich S, et al. New antidepressant drugs: what side effects and interactions are to be expected [in German]? MMW Fortschr Med. 2000;142(49-50):35-38, 40. PubMed

7. Motohashi N, Shioe K, Nakamura J, et al. Revised psychopharmacological algorithms for the treatment of mood disorders in Japan. Int J Psychiatry Clin Pract. 2008;12(1):11-18. doi:10.1080/13651500701330791

8. Gartlehner G, Thieda P, Hansen RA, et al. Comparative risk for harms of second-generation antidepressants: a systematic review and meta-analysis. Drug Saf. 2008;31(10):851-865. doi:10.2165/00002018-200831100-00004 PubMed

9. Hirschfeld RMA. Long-term side effects of SSRIs: sexual dysfunction and weight gain. J Clin Psychiatry. 2003;64(suppl 18):20-24. PubMed

10. Barbey JT, Roose SP. SSRI safety in overdose. J Clin Psychiatry. 1998;59(suppl 15):42-48. PubMed

11. Flanagan RJ. Fatal toxicity of drugs used in psychiatry. Hum Psychopharmacol. 2008;23(suppl 1):S43-S51. doi:10.1002/hup.916

12. Richmond TK, Rosen DS. The treatment of adolescent depression in the era of the black box warning. Curr Opin Pediatr. 2005;17(4):466-472. doi:10.1097/01.mop.0000166347.53102.e7 PubMed

13. Deaths related to drug poisoning in England and Wales, 2003–07. Health Stat Q. 2008;39:82-88. PubMed

14. Watts G. Is this journalism that makes a difference? BMJ. 2003;326(7398):1093. doi:10.1136/bmj.326.7398.1093

15. Lader M. Pharmacotherapy of mood disorders and treatment discontinuation. Drugs. 2007;67(12):1657-1663. doi:10.2165/00003495-200767120-00001 PubMed

16. Kupfer DJ, Frank E, Perel JM, et al. Five-year outcome for maintenance therapies in recurrent depression. Arch Gen Psychiatry. 1992;49(10):769-773. PubMed

17. Montgomery SA, Dunbar G. Paroxetine is better than placebo in relapse prevention and the prophylaxis of recurrent depression. Int Clin Psychopharmacol. 1993;8(3):189-196. doi:10.1097/00004850-199300830-00009 PubMed

18. Geddes JR, Carney SM, Davies C, et al. Relapse prevention with antidepressant drug treatment in depressive disorders: a systematic review. Lancet. 2003;361(9358):653-661. doi:10.1016/S0140-6736(03)12599-8 PubMed

19. Anderson IM, Ferrier IN, Baldwin RC, et al. Evidence-based guidelines for treating depressive disorders with antidepressants: a revision of the 2000 British Association for Psychopharmacology guidelines. J Psychopharmacol. 2008;22(4):343-396. doi:10.1177/0269881107088441 PubMed

20. Furukawa TA, Cipriani A, Barbui C, et al. Long-term treatment of depression with antidepressants: a systematic narrative review. Can J Psychiatry. 2007;52(9):545-552. PubMed

21. Williams N, Simpson AN, Simpson K, et al. Relapse rates with long-term antidepressant drug therapy: a meta-analysis. Hum Psychopharmacol. 2009;24(5):401-408. doi:10.1002/hup.1033 PubMed

22. Paykel ES, Ramana R, Cooper Z, et al. Residual symptoms after partial remission: an important outcome in depression. Psychol Med. 1995;25(6):1171-1180. doi:10.1017/S0033291700033146 PubMed

23. Pintor L, Torres X, Navarro V, et al. Is the type of remission after a major depressive episode an important risk factor to relapses in a 4-year follow up? J Affect Disord. 2004;82(2):291-296. doi:10.1016/j.jad.2003.11.008 PubMed

24. Bockting CL, Spinhoven P, Koeter MW, et al, for the Depression Evaluation Longitudinal Therapy Assessment (DELTA) Study Group. Prediction of recurrence in recurrent depression and the influence of consecutive episodes on vulnerability for depression: a 2-year prospective study. J Clin Psychiatry. 2006;67(5):747-755. doi:10.4088/JCP.v67n0508 PubMed

25. Kaymaz N, van Os J, Loonen AJ, et al. Evidence that patients with single versus recurrent depressive episodes are differentially sensitive to treatment discontinuation: a meta-analysis of placebo-controlled randomized trials. J Clin Psychiatry. 2008;69(9):1423-1436. doi:10.4088/JCP.v69n0910 PubMed

26. National Collaborating Centre for Mental Health. Depression (amended): Management of Depression in Primary and Secondary Care [NICE clinical guideline 23]. London, United Kingdom: National Institute for Health and Clinical Excellence; 2007.

27. Stevens JC, Pollack MH. Benzodiazepines in clinical practice: consideration of their long-term use and alternative agents. J Clin Psychiatry. 2005;66(suppl 2):21-27. PubMed

28. O’ brien CP. Benzodiazepine use, abuse, and dependence. J Clin Psychiatry. 2005;66(suppl 2):28-33. PubMed

29. Nutt DJ. Death and dependence: current controversies over the selective serotonin reuptake inhibitors. J Psychopharmacol. 2003;17(4):355-364. doi:10.1177/0269881103174019 PubMed

30. Patten SB. Still no evidence that benzodiazepines cause depression. Int J Psychiatry Clin Pract. 2008;12(1):85-88. doi:10.1080/13651500701419644

31. Furukawa TA, Streiner DL, Young LT. Is antidepressant-benzodiazepine combination therapy clinically more useful? a meta-analytic study. J Affect Disord. 2001;65(2):173-177. doi:10.1016/S0165-0327(00)00254-8 PubMed

32. Sinclair LI, Christmas DM, Hood SD, et al. Antidepressant-induced jitteriness/anxiety syndrome: systematic review. Br J Psychiatry. 2009;194(6):483-490. doi:10.1192/bjp.bp.107.048371 PubMed

33. Nakao M, Takeuchi T, Yano E. Prescription of benzodiazepines and antidepressants to outpatients attending a Japanese university hospital. Int J Clin Pharmacol Ther. 2007;45(1):30-35. PubMed

34. Furukawa TA, Yoshimura R, Harai H, et al. How many well vs. unwell days can you expect over 10 years, once you become depressed? Acta Psychiatr Scand. 2009;119(4):290-297. doi:10.1111/j.1600-0447.2008.01288.x PubMed

35. Sackeim HA, Haskett RF, Mulsant BH, et al. Continuation pharmacotherapy in the prevention of relapse following electroconvulsive therapy: a randomized controlled trial. JAMA. 2001;285(10):1299-1307. doi:10.1001/jama.285.10.1299 PubMed

36. American Psychiatric Association. Practice Guideline for the Treatment of Patients With Major Depressive Disorder, 2nd ed. Am J Psychiatry. 2000;157(suppl 4):1-45. http://www.psychiatryonline.com/pracGuide/pracGuideTopic_7.aspx. Accessed March 3, 2010.

37. van den Berg S, Shapiro DA, Bickerstaffe D, et al. Computerized cognitive-behaviour therapy for anxiety and depression: a practical solution to the shortage of trained therapists. J Psychiatr Ment Health Nurs. 2004;11(5):508-513. doi:10.1111/j.1365-2850.2004.00745.x PubMed

38. Lader M, Tylee A, Donoghue J. Withdrawing benzodiazepines in primary care. CNS Drugs. 2009;23(1):19-34. doi:10.2165/0023210-200923010-00002 PubMed