Applying Antidepressant Study Results to Clinical Practice
Michael E. Thase, MD
Departments of Psychiatry, University of Pennsylvania School of Medicine, Philadelphia Veterans Affairs Medical Center, and the University of Pittsburgh Medical Center, Philadelphia and Pittsburgh
AV 1. Adequate Treatment Remains an Unmet Need in Depression (00:25)
The recent controversies about whether antidepressants really work and if they are being overprescribed1 should not obscure the fact that depression is one of the world’s great public health problems in terms of economic burden, social cost, and suicide.2 Indeed, although it seems that increasing numbers of people who are not diagnosed as having depression are being prescribed antidepressants, it is also still true that the average person with MDD receives either no treatment or an inadequate course of therapy ().3
Answers to questions pertaining to overprescription of antidepressants also hinge on knowing more about the conditions, symptoms, or complaints for which physicians are prescribing the medication. For example, some of these “nondepressed” people may have depressive symptoms that cause significant impairment without meeting the current DSM-IV-TR4 criteria. For some of these patients, the presence of other associated symptoms that are not included in the diagnostic criteria for a major depressive episode may have prompted their physicians to prescribe an antidepressant. Yet other “nondepressed” patients may be prescribed antidepressants as part of a management plan for conditions, such as chronic neuropathic pain, chronic low back pain, fibromyalgia, premenstrual dysphoric disorder, nicotine dependence, migraine, vasomotor symptoms of menopause, or insomnia, for which there is a long history of antidepressant use in clinical practice despite the fact that only a small number of antidepressants have formal FDA indications for these uses. Still other “nondepressed” people taking antidepressants may be receiving these medications for prevention of recurrence. In such cases, the very reason that the person is not depressed may be because he or she responded to pharmacotherapy some months or even years ago and is following a doctor’s recommendation.
Thus, although there is evidence that antidepressants are being prescribed to people who are currently not depressed, it is less clear whether those prescriptions are appropriate. Physicians are clearly on sounder empirical footing when prescribing an antidepressant for a patient who has a disorder with an approved indication for that medication, though they also should not be constrained to unnaturally limit their practice. Treatment decisions need to be made on a case-by-case basis, with the perceived benefits balanced against the costs and known risks associated with that medication and carefully monitored on an ongoing basis.
Applying Placebo-Controlled Study Results to Practice
If treatment for depression is warranted, clinicians may look to the literature for guidance in addition to considering their own experiences in prescribing antidepressants as well as each particular patient’s medication history. Knowing the limitations of randomized controlled trials (RCTs) also can help reconcile inconsistencies between clinical experience and the research literature. These limitations include the facts that RCTs use methods that do not reflect real-world practice and that there is a dearth of comparative and cost-effectiveness data available for most new antidepressants. Other topics that are relevant to practitioners include publication bias and the growing placebo effect.
New antidepressants. When a new antidepressant comes to market in the United States, relatively little is actually known about that medication. For example, at the time of the “launch” of a new antidepressant, there will be data from at least 2 positive RCTs (ie, a study is positive when the active drug is statistically significantly more effective than the placebo on the primary endpoint) and there will be safety data for about 1,000 patients. In addition, there will be assessments of safety across up to 1 year of treatment exposure for several hundred depressed patients. By convention, these data studies are sufficient to demonstrate efficacy and safety at a level that satisfies authorities at the FDA; the actual utility of the new compound in comparison to older, more widely used medications is largely unknown. Importantly, the controlled studies reviewed by the FDA have limited generalizability to everyday practice. Specifically, the samples enrolled in these studies usually comprise patients with less complex or fewer comorbid illnesses than are found in clinical practice, and of course, the study doctors have treated these individuals using relatively simple and standardized protocols. In essence, the evidence base is adequate to conclude only that the new drug is more effective than a placebo, to describe the common adverse events, and to reassure prescribers that the novel therapy is likely not to be associated with any marked safety concern. In 2011, it is also likely that there will be at least some information about the likelihood that the new drug will be liable to drug interactions mediated by the cytochrome P450 system.
However, much information relevant to both patients and clinicians is unknown at the time a new drug is launched, including important aspects of the long-term safety profile. For example, weight gain is often imperceptible in 6- or 8-week trials and sexual side effects tend to be underreported unless they are assessed prospectively by a direct, explicit inquiry based on an interview or questionnaire.
It is also possible that a novel antidepressant will ultimately be found to cause a relatively rare side effect, the discovery of which may need hundreds of thousands of prescriptions. For example, if a novel medication was found to cause liver failure in, say, 1 per 100,000 exposures, it is likely to take several years before that danger signal can be recognized.
At the time of launch, only general information is available to tell clinicians whether a novel medication will be useful in combination with standard medications, such as mood stabilizers or second generation antipsychotics, that are commonly used for treatment of patients with more complex psychiatric histories. Clinicians also typically do not have good information on how to cross-titrate when switching from an older, standard medication to a newer one.
For these reasons, most pharmacy benefit plans tend to place new medications in lower rungs of their formularies, which ensures that the vast majority of patients will receive at least several trials with older, less expensive (ie, generic) medications before a trial of the new medication can be considered. Such a policy ensures that newer medications are likely to be viewed as not cost-effective in comparison to generic alternatives. It will ultimately be necessary for the manufacturer of a new medication with particularly exemplary merit to conduct a large number of head-to-head studies with existing drugs to demonstrate that the novel medication deserves to be used sooner rather than later in a treatment hierarchy.
Methodology. Randomized controlled trials are designed not to inform public health but rather to obtain the requisite data necessary to obtain regulatory approval. Yet these very trials contribute the vast majority of data to meta-analyses of the efficacy of antidepressants. Thus, industry-sponsored RCTs provide an imperfect foundation for the practice of evidence-based medicine. As noted earlier, these studies use relatively extensive inclusion and exclusion criteria that greatly narrow the pool of patients who are available to participate. Careful review of the participants in the STAR*D trial, for example, indicated that very few of the study participants would have been eligible to enroll in an industry-sponsored RCT.5
Publication bias. Until recently, agencies and individuals that undertake systematic reviews and meta-analyses of RCTs have delimited their review to studies that have cleared the hurdle of peer review and are available in the published literature. However, it has long been known that the published literature may give an overly positive impression of particular interventions because of the so-called file-drawer effect. Also known as publication bias, the file-drawer effect results from the almost universal observation that studies that fail to show a significant effect are less likely to be published than studies that show the predicted positive effect for the novel intervention.6 In the case of RCTs of novel antidepressants, publication bias accounts for about a 33% inflation in the apparent efficacy of the active treatment.5
Placebo and antidepressant response rates. Remarkable temporal trends in the results of RCTs of antidepressants have occurred over the past 30 years.7,8 Specifically, the response to placebo has grown faster than the response to active antidepressants, with a progressively shrinking drug-versus-placebo difference or effect size as a result.7 As the chemical formulation of placebo has not changed across the decades, it is therefore certain that either the characteristics of the study participants have changed or the way studies are conducted has changed. In either case, the net result is that the failure rate of antidepressant RCTs has grown proportionally with the increase in the placebo response rate. In the second decade of the 21st century, it is axiomatic that at least one half of RCTs of depression will fail to show statistically significant differences favoring a drug with established efficacy compared to an inert placebo.9
AV 2. Outcomes for Usual Care Vs Telephone-Enhanced Care With or Without CBT (00:30)
One reason for higher placebo response rates may be longer duration of trials, allowing extra time not only for the medication to work or for the depression to spontaneously resolve but also for more contact between subjects and researchers.7 The level of monitoring required in trials means that patients receive considerable clinical support in research settings. Cumulative interaction between subjects and researchers (ie, the nonspecific aspect of treatment) appears to increase subjects’ response rates whether they are taking placebo or active medication.10 Although this observation may confound a research study, it can directly enhance the outcomes of patients seen in clinical practice because researchers see patients more frequently and monitor their conditions more closely. When more physician contact is not practicable, the involvement of other members of the healthcare team can make a tangible difference. For example, nurse-provided telephone collaborative care has been consistently found to improve outcomes compared with usual care, as has the addition of a formal psychotherapy such as CBT ().11 Even the availability of online support has been found to improve the outcomes of people receiving antidepressant therapy.12
The growth of the placebo response rate in RCTs over the past 3 decades also could indicate that the nature of study participants has changed. However, beyond the recognition that people may now have higher expectations for benefits from antidepressant therapy, it is not clear how study participants may actually differ. One possibility is that, since it has long been known that placebo response is in part moderated by symptom severity, one might strongly suspect that participants of contemporary studies are less severely depressed. However, inspection of symptom severity scores at study entry does not confirm this to be the case. An upward drift in scoring conventions may have occurred, such that depressive episodes that were scored in the mild-to-moderate range in the 1980s might now be scored in the moderate-to-severe range. Consistent with this hypothesis, there is evidence of rater scoring inflation in studies that require relatively high severity scores at study intake.13
Whatever the reasons for the growth in placebo response rates, it is clear that the relative contribution of the so-called specific effects of treatment to the overall outcome in placebo-controlled RCTs has declined to such a degree that it can now be said that antidepressants have small effects, perhaps as low as only a 2-point average advantage on a standard rating scale.14,15 From a more cynical vantage point, such a small effect can be declared to be trivial or below a threshold of clinical significance. However, the small average effect observed in RCT participants does not result from each depressed person improving a small amount, but rather from a modestly sized subgroup improving a large amount (ie, a 2- or 3-point average difference can result from a 20-point advantage among 10% to 15% of the patients).9 From this vantage point, the effects of antidepressants fall within the range that is considered clinically significant.
Even in an era of relatively small drug-versus-placebo differences, evidence shows that some patient groups are more likely to benefit from the specific effects of antidepressants than others. For example, in the meta-analysis of Kirsch et al,14 the specific effect of antidepressants grew progressively larger in the studies that enrolled patients with higher depression severity scores. Fournier et al10 observed the exact same finding in a meta-analysis of individual patient data drawn from 6 placebo-controlled studies of either imipramine or paroxetine. Some have argued that this finding means that antidepressants should be reserved for only those depressed people with very severe depression.1 This argument, however, ignores the fact that, outside of research settings, clinicians cannot use placebo to treat patients and the nonspecific aspects of care must be seamlessly woven within the overall treatment experience. That said, psychotherapy continues to be a viable alternative to antidepressants for many depressed patients.1
Applying Study Results When Treating Unresolved Symptoms
The STAR*D study,16 which was more inclusive and less restrictive in its enrollment than is typical for placebo-controlled trials, provides important data on the outcome of depressed patients treated in contemporary psychiatric and primary care practices. In the first level of STAR*D, all patients received an initial course of treatment with citalopram titrated up across up to 14 weeks of therapy. Only about 30% of patients remitted, while another 20% responded (the responders had too many residual symptoms to be considered in remission); the remainder received little or no benefit from the first course of pharmacotherapy. Among the latter group, up to 3 sequential treatment trials were available, with approximately one-half eventually responding if they did not drop out of the treatment process. There was evidence that the likelihood of benefit decreased and the risk of subsequent relapse increased with each subsequent trial17; thus, many patients have unresolved symptoms after initial treatment. Therefore, STAR*D17 demonstrated that, although the available antidepressants offer a clinically significant benefit to a substantial proportion of patients, more effective antidepressants that work by different mechanisms are needed to help people who do not respond to first- and second-line choices.
Several analyses of STAR*D examined factors associated with unresolved symptoms.16,18 Nonmodifiable factors that were associated with a lower likelihood of remission included demographic characteristics, such as lower income and unemployment, less education, minority status, single status, and male sex, as well as clinical characteristics, such as greater baseline severity, more comorbid illnesses, and a longer current episode. Other treatment factors that affect the likelihood of remission but can be modified by altering the approach to treatment included slow onset of response, short initial treatment, nonadherence, and poor tolerability.19,20 For example, addressing adverse events may require switching medications to find one with a better tolerability profile.
Ultimately, the outcomes of depressed patients can be improved by adopting a systematic approach to treatment using measurement-based care within the framework of contemporary guidelines and an algorithm provides better symptom reduction and greater improvement in mental health functioning than treatment as usual.
AV 3. Factors Mitigating Against and Contributing to Antidepressant Adherence (00:25)
Nonadherence to medication is common, and addressing it also can improve outcomes.12 One study21 showed that, by the fourth week of treatment, 28% of patients with newly prescribed antidepressants had stopped taking their medication and, after 16 weeks, nearly 50% of patients had discontinued their medications. Another recent large study22 found that 56% of patients discontinued antidepressant treatment during the first 4 months and that only 22% had good adherence to their prescribed medication. No medication, however good, will work if the patient fails to take it. Patient education, regular monitoring, and managing side effects can improve adherence ().23,24
Although RCTs have real limitations, applying the results of these studies can help clinicians ensure that patients with depression receive effective treatment. Antidepressants remain a cornerstone of treatment, despite considerable rates of nonresponse and nonadherence. As no ideal antidepressant exists, many patients will require sequential drug trials.
One lesson readily apparent in contemporary trials is that the so-called nonspecific aspects of treatment may outweigh the specific actions of the particular antidepressant. Clinicians, thus, can improve outcomes by keeping in mind that the way that they care for their patients may be more important than the actual medication that they pick. Care can be enhanced by working to strengthen the therapeutic alliance and by educating patients about why a particular treatment has been chosen, how it is expected to work, what common side effects might be expected, and what can be done if the treatment of first choice is either ineffective or poorly tolerated. Outcome can be gauged reliably by the use of relatively brief self-report symptom rating scales at each visit; side effects and adherence should likewise be assessed on an ongoing basis.
There is great room for improvement in the antidepressant arena, and although the pace of drug development is currently slowed, new medications are introduced almost every year. To understand how a newer medication compares with existing, first-line medications and to appreciate its relative strengths and limitations, clinicians have to look well beyond the initial studies and gain clinical experience with the newer medication.
For Clinical Use
- Apply antidepressant study results in clinical practice to help ensure that patients with depression receive the most effective treatment possible
- Monitor and address tolerability, lack of response, and adherence problems
- Provide nonspecific aspects of care that can enhance antidepressant treatment
citalopram (Celexa and others), imipramine (Tofranil and others), paroxetine (Paxil, Pexeva, and others)
CBT = cognitive-behavioral therapy; DSM-IV-TR = Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition, Text Revision; FDA = United States Food and Drug Administration; MDD = major depressive disorder; RCT = randomized controlled trial; SCL = Hopkins Symptom Checklist Depression Scale; STAR*D = Sequenced Treatment Alternatives to Relieve Depression
Take the online posttest.
- Nierenberg AA, Leon AC, Price LH, et al. The current crisis of confidence in antidepressants. J Clin Psychiatry. 2011;72(1):27–33. Abstract
- World Health Organization. The Global Burden of Disease: 2004 Update. Geneva, Switzerland: WHO Press; 2008. http://www.who.int/healthinfo/global_burden_disease/GBD_report_2004update_full.pdf. Accessed November 28, 2011.
- Kessler RC, Berglund P, Demler O, et al. The epidemiology of major depressive disorder: results from the National Comorbidity Survey Replication (NCS-R). JAMA. 2003;289(23):3095–3105. PubMed
- American Psychiatric Association. Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition, Text Revision. Washington, DC: American Psychiatric Association; 2000.
- Wisniewski SR, Rush AJ, Nierenberg AA, et al. Can phase III trial results of antidepressant medications be generalized to clinical practice? a STAR*D report. Am J Psychiatry. 2009;166(5):599–607. PubMed
- Turner EH, Matthews AM, Linardatos E, et al. Selective publication of antidepressant trials and its influence on apparent efficacy. N Engl J Med. 2008;358(3):252–260. PubMed
- Walsh BT, Seidman SN, Sysko R, et al. Placebo response in studies of major depression: variable, substantial, and growing. JAMA. 2002;287(14):1840–1847. PubMed
- Thase ME. The small specific effects of antidepressants in clinical trials: what do they mean to psychiatrists? Curr Psychiatry Rep. 2011;13(6):476–482. PubMed
- Thase ME, Larsen KG, Kennedy SH. Assessing the 'true' effect of active antidepressant therapy v placebo in major depressive disorder: use of a mixture model. Br J Psychiatry. In press.
- Posternak MA, Zimmerman M. Therapeutic effect of follow-up assessments on antidepressant and placebo response rates in antidepressant efficacy trials: meta-analysis. Br J Psychiatry. 2007;190:287–292. PubMed
- Simon GE, Ludman EJ, Tutty S, et al. Telephone psychotherapy and telephone care management for primary care patients starting antidepressant treatment: a randomized controlled trial. JAMA. 2004;292(8). PubMed
- Wade AG, Häring J. A review of the costs associated with depression and treatment noncompliance: the potential benefits of online support. Int Clin Psychopharmacol. 2010;25(5):288–296. PubMed
- Gelenberg AJ, Thase ME, Meyer RE, et al. The history and current state of antidepressant clinical trial design: a call to action for proof-of-concept studies. J Clin Psychiatry. 2008;69(10):1513–1528. Abstract
- Kirsch I, Deacon BJ, Huedo-Medina TB, et al. Initial severity and antidepressant benefits: a meta-analysis of data submitted to the Food and Drug Administration. PLoS Med. 2008;5(2):e45. PubMed
- Fournier JC, DeRubeis RJ, Hollon SD, et al. Antidepressant drug effects and depression severity: a patient-level meta-analysis. JAMA. 2010;303(1):47–53. PubMed
- Trivedi MH, Rush AJ, Wisniewski SR, et al. Evaluation of outcomes with citalopram for depression using measurement-based care in STAR*D: implications for clinical practice. Am J Psychiatry. 2006;163(1):28–40. PubMed
- Rush AJ, Trivedi JH, Wisniewski SR, et al. Acute and longer-term outcomes in depressed outpatients requiring one or several treatment steps: a STAR*D report. Am J Psychiatry. 2006;163(11):1905–1917. PubMed
- Rush AJ, Wisniewski SR, Warden D, et al. Selecting among second-step antidepressant medication monotherapies: predictive value of clinical, demographic, or first-step treatment features. Arch Gen Psychiatry. 2008;65(8):870–880. PubMed
- Berlanga C, Heinze G, Torres M, et al. Personality and clinical predictors of recurrence of depression. Psychiatr Serv. 1999;50(3):376–380. PubMed
- Kornstein SG, Schneider RK. Clinical features of treatment-resistant depression. J Clin Psychiatry. 2001;62(suppl 16):18–25. Full Text
- Lin EH, Von Korff M, Katon W, et al. The role of the primary care physician in patients' adherence to antidepressant therapy. Med Care. 1995;33(1):67–74. PubMed
- Serna MC, Cruz I, Real J, et al. Duration and adherence of antidepressant treatment (2003 to 2007) based on prescription database. Eur Psychiatry. 2010;25(4):206–213. PubMed
- Bull SA, Hu XH, Hunkeler EM, et al. Discontinuation of use and switching of antidepressants: influence of patient-physician communication. JAMA. 2002;288(11):1403–1409. PubMed
- Halaris A. A primary care focus on the diagnosis and treatment of major depressive disorder in adults. J Psychiatr Pract. 2011;17(5):340–350. PubMed