Placebo Effects on Pharmacotherapy Outcomes in Major Depression

September 15, 2007
Aimee M. Hunter, PhD

Volume 24, Issue 10

If clinical trials data are any indication, the potential impact of placebo treatment on depression outcomes may be potent. Placebo response rates in clinical trials for depression average approximately 30%, with a top range beyond 50%-and the trend is upward.

If clinical trials data are any indication, the potential impact of placebo treatment on depression outcomes may be potent. Placebo response rates in clinical trials for depression average approximately 30%, with a top range beyond 50%-and the trend is upward.1 Granted, estimates of the magnitude of the placebo response are open to debate2,3 because placebo response rates do not account for statistical regression to the mean or spontaneous remission (an important consideration given the natural waxing and waning of symptoms in depression). In fact, there are no data to support the conclusion that the placebo response is not spontaneous remission. However, converging evidence suggests that placebo effects can exert a significant impact on both clinical symptoms and functional brain changes in depressive illness.

Measuring PLACEBO RESPONSE

The placebo response or placebo effect is increasingly being conceptualized as the effect of the environmental and psychosocial context surrounding treatment.4,5 Focus has shifted from physical features of the placebo (eg, the dummy sugar pill or sham treatment) toward the larger therapeutic setting,6 including the "meaning" that this setting holds for the individual.7 The beliefs and expectations of the patient and physician/clinician, as well as the nature of the patient-physician relationship, are of primary importance in the treatment context.

If one considers the placebo response to include classic conditioning effects that can occur in pharmacotherapy,8 then the patient's previous exposure to pharmacological agents also becomes an important factor. Following exposure to an active agent, the cues associated with drug administration (eg, a look-alike placebo pill or saline placebo injection) can come to elicit physiological responses.

Placebo research uses pill placebos or other sham interventions that do not have specific efficacy for the condition being treated, with the understanding that placebo administration serves as a proxy for the effects of the treatment setting. That is, a subject showing a response to placebo is thought to be responding to all the trappings of treatment and the meaning that it holds for him or her.

Studies designed to examine the so-called placebo response in pain conditions often do not use placebo at all. The open-hidden paradigm9 has offered a particularly productive approach to examining the effects of the treatment context. Numerous reports of clinical and experimental pain studies have demonstrated the greater efficacy of analgesics during open administration (in full view and with full knowledge of the patient or subject) compared with hidden administration, ie, usually delivered via a computer-driven infusion pump without the patient's awareness and without a physician present.

The difference in effectiveness between open and hidden administration of active drug is attributed to placebo mechanisms. The presence of a physician, and the patient's expectations of relief achieved with open administration, provide a context that enhances or augments the effect of medication. Although the open-hidden paradigm is relevant to acute pain conditions, this type of within-subjects approach is not applicable to placebo research in depression because treatments for depression require ongoing administration to achieve clinical results.

Placebo research in depression typically is conducted in conjunction with clinical efficacy studies that include a placebo control. Thus the primary study is usually designed to address the efficacy of one or more medications/ interventions, rather than to target placebo effects. Because of ethical concerns, studies in depression do not typically use a "no treatment" or "no treatment context" condition against which to compare the effects of placebo. Nonetheless, substantial placebo response rates in antidepressant trials of medication, psychotherapy, and somatic interventions for depression suggest that the treatment context itself does indeed influence the expression of depressive symptoms.1,10,11

One NIH-funded study currently in progress (A. F. Leuchter, personal communication, July 2007) is specifically designed to assess contributions of various aspects of the treatment setting to the placebo response in depression. This study assigns patients to 1 of 3 treatment conditions: interpersonal clinical interaction, in which patients interact with clinical research personnel on a fixed schedule; placebo treatment, in which patients have the same clinical interaction and placebo pills are added; and medication treatment, in which patients have the same clinical interaction and medication is added instead of placebo. By varying aspects of the treatment context, this study will provide information regarding the relative contributions of interpersonal interactions, expectations associated with taking a placebo pill, and medication effects, to the clinical outcomes in depression.

IMAGING THE PLACEBO RESPONSE

Neuroimaging studies provide evidence that the placebo effect in depression has demonstrable physiological underpinnings. Depressed individuals who respond during placebo treatment show brain functional changes that are not observed in nonresponders. Several reports have examined brain functional changes during placebo administration in double-blind trials of antidepressant medication. In a study of depressed males, fluorodeoxyglucose positron emission tomography (FDG-PET) scans were obtained at pretreatment baseline, and again after 1 and 6 weeks of treatment with fluoxetine or placebo.12 Both medication- and placebo-responders showed metabolic changes at week 6 (end of trial), including increases in activity in the prefrontal, parietal, and posterior cingulate regions, and decreases in the subgenual cingulate cortex. Additional changes were seen in the fluoxetine group in the subcortical and limbic regions, including increases in the pons and decreases in the striatum, hippocampus, and anterior insula.

Another functional imaging study used quantitative electroencephalography cordance to assess serial brain functional changes in patients with depression who were randomized to 8 weeks of treatment with antidepressant medication (fluoxetine or venlafaxine) or placebo.13 Electroencephalographic recordings were obtained at baseline, 48 hours after start of treatment, and at weekly intervals throughout randomized treatment. Analysis of predefined scalp regions revealed significant and unique baseline to end-of-trial increases in prefrontal cordance in placebo responders. In contrast, medication group responders showed early (week 2) decreases in prefrontal cordance, whereas nonresponder groups did not show significant changes from baseline prefrontal cordance at any point in the trial.

Although differences in imaging techniques and study designs do not allow for direct comparisons across these studies, both reports found significant frontal brain changes that distinguished between placebo responders and nonresponders. At minimum, this indicates that the placebo response in depression does not merely reflect a response bias wherein placebo responders simply report symptoms differently; placebo responders do in fact show underlying physiological changes that are not observed in nonresponders.

It appears that "pill placebo response" in depression has both shared and distinct neurophysiological features compared with medication response. However, as with all imaging studies in depression, it is difficult to isolate which functional brain changes reflect improvement in symptoms and which are related to treatment mechanisms. There is marked heterogeneity in findings regarding functional brain changes associated with remission of symptoms of depression even without considering placebo treatment conditions.

One framework that finds traction in the literature is the notion of "bottom-up" versus "top-down" neurophysiological effects of different treatment modalities. Pharmacotherapy interventions generally initiate a bottom-up cascade primarily targeting the brain stem, limbic, and subcortical regions with longer-term effects occurring in cortical areas. Conversely, psychotherapeutic interventions that directly address maladaptive cognitions and emotions are thought to work "top-down," first affecting changes in frontal cortical mechanisms.

FDG-PET imaging studies of cognitive-behavioral therapy14 or interpersonal psychotherapy15 provide an interesting basis of comparison with the pill placebo study findings. Results of cognitive-behavioral therapy and interpersonal psychotherapy studies show clinical response associated with substantive bilateral prefrontal decreases in metabolism.14,15 This is in contrast to the FDG-PET imaging of the placebo response in a placebo-controlled trial of fluoxetine that found prefrontal increases in both medication and placebo responders.12

With respect to FDG-PET studies to date, the functional neurophysiology of the pill placebo response appears to more closely resemble medication response than psychotherapy response. Further research is needed to delineate the neurophysiology of placebo mechanisms from those of specific interventions for major depressive disorder (MDD).

EXPECTATIONS AND PLACEBO RESPONSE

Most theoretical accounts of the placebo response (pure classic conditioning accounts notwithstanding) rely on the concept of patient expectations. The patient must have some expectation of getting better, or of experiencing relief of symptoms, in order to recruit the placebo response. This poses an interesting paradox or circularity for depression because one of the hallmark symptoms of depressive illness is a sense of hopelessness. To the extent that the treatment environment can help elicit positive expectations and a sense of hope, this in itself may make inroads in altering the course of depression, at least briefly.

Numerous studies have documented the importance of expectation-related factors for treatment outcomes in MDD. Negative expectations have been found to portend poorer outcomes. For example, higher "dysfunctional attitudes" before treatment were negatively associated with response to tricyclic antidepressants, fluoxetine, or placebo in patients with mixed depression.16,17 In a more recent study, baseline hopelessness (Beck Hopelessness Scale) was demonstrated to predict poorer response to fluoxetine in a clinical trial sample of 312 patients.18

The converse has been demonstrated for positive expectations. A study of 63 depressed inpatients receiving naturalistic treatment for depressive illness revealed that both the psychiatrist's optimism regarding the patient's treatment regimen and the degree to which patients believed that their treatment was right for them predicted severity of symptoms at discharge.19

In a study of reboxetine, higher baseline expectations of treatment outcomes were associated with better 8-week outcomes.20 As part of the NIMH Treatment of Depression Collaborative Research Program, a study of 162 patients with MDD revealed that pretreatment expectations of improvement were a significant predictor of patient response to pharmacotherapy.21

Imaging "expectations" in depression treatment

Is there a neurobiology of hopefulness or patient expectations in depression? It is interesting to consider that some brain functional changes that occur early in the course of treatment might reflect expectation-related processes that are associated with later clinical response. Because symptoms do not typically resolve in the first week of treatment, brain changes observed within this timeframe are less likely to be influenced by clinical improvement.

In an intriguing post hoc analysis of the FDG-PET study of pill placebo response in depression,4,12 investigators examined metabolic changes occurring in the first week of treatment for those patients who later were classified as medication or placebo responders or nonresponders. Both responder groups (but not nonresponders) were found to have metabolic increases in the ventral striatal and orbital frontal regions at week 1; moreover, this change pattern dissipated by week 6 when clinical response was established. Because of the brain regions involved and time course of this change pattern, the investigators speculated that such early metabolic changes might have reflected subject expectations of clinical benefit before the actualization of clinical response.4

Other studies using quantitative electroencephalography cordance have examined brain functional changes during an initial week of placebo treatment (a placebo lead-in phase) before randomized treatment with antidepressant medication or placebo, and similarly, researchers have speculated that these brain changes were related to patient expectations. Of importance, the placebo lead-in phase of the trial included all of the same contextual features as the medication treatment phase (eg, relationships with physicians and other health care staff, clinical assessments and monitoring, pill taking)-except for the active drug. Neurophysiological changes at this point could not have been influenced by medication and were not likely to be influenced by change in clinical status, because most patients did not show significant improvement during a 1-week placebo lead-in; those who did were discontinued from the study.

In a report of depressed patients randomized to 8 weeks of treatment with medication (fluoxetine or venlafaxine) or placebo, bilateral decreases in prefrontal cordance during placebo lead-in predicted 19% of the final clinical outcome in the medication-treated group.22 The authors suggested that the initial neurophysiological response to the treatment setting might have captured patient expectations that had a subsequent impact on medication outcome.

Another study examined brain functional changes during placebo lead-in in healthy never-depressed persons randomly assigned to 4 weeks of treatment with venlafaxine or placebo.23 Although clinical benefits were not expected in nondepressed persons, medication adverse effects were anticipated. In that study, decreases in left prefrontal cordance during placebo lead-in were associated with later adverse-effect burden in the medication group, suggesting that expectations during initial placebo treatment might have influenced the experience of medication adverse effects.

The preponderance of data suggests that expectations and subsequent outcomes can be predicted by the initial changes in brain physiology early in treatment. Further research is required to directly assess whether early brain functional changes reflect differences in patient expectations

Sustaining positive expectations

It is worth noting that the role of expectation in MDD has by and large been examined in relation to acute treatment outcomes. The potential for longer-term placebo effects in depression is less clear. Although placebo response rates in clinical trials can be quite high, maintenance studies examining longer-term outcomes of placebo treatment have found placebo to be substantially less effective than active medication as a prophylactic against relapse.24-27 Some of the loss of placebo effects over time may be due to decreased expectations. Positive expectations may be difficult to sustain because of the nature of depression, an illness that is characterized by hopelessness as well as disruptions in cognitive processing and frontal function.

A recent study in patients with Alzheimer disease showed that reduced cognitive status and reduced prefrontal connectivity were associated with decrements in the placebo component of response to analgesic, indicating that the loss of expectation-related mechanisms can decrease the effectiveness of pharmacological treatment. From a different but consistent perspective, a study of baseline predictors of the placebo outcomes in depression identified a frontocentral electroencephalographic marker, as well as higher neurocognitive processing speed, as significant positive predictors of placebo response.28

CONCLUSION

There is strong evidence that the treatment context, including patient and physician expectations, can have an important impact on clinical outcomes in depression, and that placebo effects in depression have demonstrable neurophysiological underpinnings. Placebo-related mechanisms may interact synergistically with specific interventions for MDD to achieve enhanced outcomes. Positive expectations and related placebo mechanisms are shown to enhance the effectiveness of medications and other specific interventions; symbiotically, medications and other specific interventions may support a neurophysiological substrate for fostering and maintaining positive expectations and related placebo mechanisms.

References:

References


1.

Walsh BT, Seidman SN, Sysko R, et al. Placebo response in studies of major depression: variable, substantial, and growing.

JAMA

. 2002;287:1840-1847.

2.

Hróbjartsson A, Gøtzsche PC. Is the placebo powerless? Update of a systematic review with 52 new randomized trials comparing placebo with no treatment.

J Intern Med

. 2004;256:91-100.

3.

Wampold BE, Minami T, Tierney SC, et al. The placebo is powerful: estimating placebo effects in medicine and psychotherapy from randomized clinical trials.

J Clin Psychol

. 2005;61:835-854.

4.

Benedetti F, Mayberg HS, Wager TD, et al. Neurobiological mechanisms of the placebo effect.

J Neurosci.

2005;25:10390-10402.

5.

Colloca L, Benedetti F. Placebos and painkillers: is mind as real as matter? Nature reviews.

Nat Rev Neurosci

. 2005;6:545-552.

6.

Price DD, Finniss DG, Benedetti F. A comprehensive review of the placebo effect: recent advances and current thought.

Annu Rev Psychol

. 2007 Jun 5; [Epub ahead of print].

7.

Moerman DE, Jonas WB. Deconstructing the placebo effect and finding the meaning response.

Ann Intern Med

. 2002;36:471-476.

8.

Ader R. The role of conditioning in pharmacotherapy. In: Harrington A, ed.

The Placebo Effect

. Cambridge, Mass: Harvard University Press; 1997:138-165.

9.

Levine JD, Gordon NC. Influence of the method of drug administration on analgesic response.

Nature

. 1984;312: 755-756.

10.

DeRubeis RJ, Hollon SD, Amsterdam JD, et al. Cognitive therapy vs medications in the treatment of moderate to severe depression.

Arch Gen Psychiatry

. 2005; 62:409-416.

11.

Koerselman F, Laman DM, van Duijn H, et al. A 3-month, follow-up, randomized, placebo-controlled study of repetitive transcranial magnetic stimulation in depression.

J Clin Psychiatry

. 2004;65:1323-1328.

12.

Mayberg HS, Silva JA, Brannan SK, et al. The functional neuroanatomy of the placebo effect.

Am J Psychiatry.

2002;159:728-737.

13.

Leuchter AF, Cook IA, Witte EA, et al. Changes in brain function of depressed subjects during treatment with placebo.

Am J Psychiatry.

2002;159:122-129.

14.

Goldapple K, Segal Z, Garson C, et al. Modulation of cortical-limbic pathways in major depression: treatment-specific effects of cognitive behavior therapy.

Arch Gen Psychiatry.

2004;61:34-41.

15.

Brody AL, Saxena S, Stoessel P, et al. Regional brain metabolic changes in patients with major depression treated with either paroxetine or interpersonal therapy: preliminary findings.

Arch Gen Psychiatry.

2001;58:631-640.

16.

Peselow ED, Robins C, Block P, et al. Dysfunctional attitudes in depressed patients before and after clinical treatment in normal control subjects.

Am J Psychiatry.

1990;147:439-444.

17.

Fava M, Bless E, Otto MW, et al. Dysfunctional attitudes in major depression: changes with pharmacotherapy.

J Nerv Ment Dis

. 1994;182:45-49.

18.

Papakostas GI, Petersen T, Homberger CH, et al. Hopelessness as a predictor of non-response to fluoxetine in major depressive disorder.

Ann Clin Psychiatry

. 2007;19:5-8.

19.

Priebe S, Gruyters T. The importance of the first three days: predictors of treatment outcome in depressed in-patients.

Br J Clin Psychol.

1995;34:229-236.

20.

Krell HV, Leuchter AF, Morgan M, et al. Subject expectations of treatment effectiveness and outcome of treatment with an experimental antidepressant.

J Clin Psychiatry.

2004;65:1174-1179.

21.

Sotsky SM, Glass DR, Shea MT, et al. Patient predictors of response to psychotherapy and pharmacotherapy: findings in the NIMH Treatment of Depression Collaborative Research Program.

Am J Psychiatry.

1991; 148:997-1008.

22.

Hunter AM, Leuchter AF, Morgan ML, et al. Changes in brain function (quantitative EEG cordance) during placebo lead-in and treatment outcomes in clinical trials for major depression.

Am J Psychiatry.

2006;163: 1426-1432.

23.

Hunter AM, Leuchter AF, Morgan ML, et al. Neurophysiologic correlates of side effects in normal subjects randomized to venlafaxine or placebo.

Neuropsycho-pharmacology

. 2005;30:792-799.

24.

Frank E, Kupfer DJ, Perel JM, et al. Three-year outcomes for maintenance therapies in recurrent depression.

Arch Gen Psychiatry

. 1990;47:1093-1099.

25.

Montgomery SA. Efficacy in long-term treatment of depression.

J Clin Psychiatry

. 1996;57:24-30.

26.

Stewart JW, Quitkin FM, McGrath PJ, et al. Use of pattern analysis to predict differential relapse of remitted patients with major depression during 1 year of treatment with fluoxetine or placebo.

Arch Gen Psychiatry.

1998;55: 334-343.

27.

McGrath PJ, Stewart JW, Petkova E, et al. Predictors of relapse during fluoxetine continuation or maintenance treatment of major depression.

J Clin Psychiatry

. 2000; 61:518-524.

28.

Leuchter AF, Morgan M, Cook IA, et al. Pretreatment neurophysiological and clinical characteristics of placebo responders in treatment trials for major depression.

Psychopharmacology (Berl).

2004;177:15-22.