Research Article
Effect of placebo groups on blood pressure in hypertension: a meta-analysis of beta-blocker trials

https://doi.org/10.1016/j.jash.2016.10.009Get rights and content

Highlights

  • There are significant placebo effects on systolic and diastolic blood pressure.

  • Placebo effects on diastolic blood pressure accounted for 47% of the drug effect.

  • Placebo effects on systolic blood pressure accounted for 34% of the drug effect.

  • No placebo effect on heart rate was found.

  • Placebo mechanisms could be utilized to optimize the pharmacological treatment.

Abstract

Hypertension is often treated pharmacologically. Since there is evidence that the cardiovascular system is sensitive to placebo mechanisms, our aim was to conduct an effect size analysis of placebo groups in double-blinded randomized controlled parallel-group drug trials using beta-blockers to treat hypertensive patients. A comprehensive literature search via PubMed, PsycINFO, PSYNDEX, PQDT OPEN, OpenGREY, ISI Web of Knowledge, and the WHO International Clinical Trials Registry Platform provided the basis of our meta-analysis. Effect sizes were estimated using a random-effects model based on 23 studies covering a total of 11,067 participants. Main outcomes were systolic blood pressure (sBP) and diastolic blood pressure (dBP). Blood pressure was lowered in placebo groups with significant and robust effect sizes (Hedges’ g). The estimates for sBP (−0.27, P < .001) and dBP (−0.49, P < .001) can be interpreted as small to moderate. The placebo response accounted for 34% of the drug response for sBP and 47% of the drug response for dBP. Our moderator analyses indicated that a higher study quality and more study site visits were marginally associated with a higher placebo response. In light of these strong placebo responses, placebo mechanisms need to be considered in order to improve antihypertensive treatment.

Introduction

Placebo groups are important in randomized controlled drug trials to differentiate the specific effects of the drug from nonspecific symptom change. The physiological change in patients allocated to placebo groups, the so-called placebo effect,1 includes symptom change caused by the natural course of the disease, various statistical artifacts (ie, regression to the mean, reporting biases, etc.), and the true placebo-induced change.2 Positive treatment expectations, learning, and the doctor–patient interaction are major mechanisms driving the true placebo-induced change.1 Randomized placebo controlled trials are used to contrast the specific medication effect with the overall placebo effect, assuming an additive model of both effects.3 The effect in the drug group needs to be substantial in comparison to the effects in the placebo group in order to demonstrate the drug's efficacy. Research, however, demonstrates that not only statistically, but also clinically significant symptom improvement is seen in placebo groups in drug trials.4

But how much of the positive symptom change in the drug group can be observed in the placebo group? Research on antidepressant medication suggests that about 75% of the positive effect of the medication can also be observed in placebo groups.5 For pain and generalized anxiety disorder, the proportion is around 50%.6, 7 These results are mostly based on subjective outcomes, such as self-reported scales. Therefore, it is important to look at diseases where the drug response can be quantified by objective outcomes. Winkler and Rief8 conducted a meta-analysis and showed in insomnia patients that based on objective, polysomnographic data (combined data from EEG, ECG, etc.), over 60% of the positive effects of the drug were also achieved in placebo groups. Thus, placebo responses are more than reporting biases in self-reported scales. They do not only occur in psychiatric disorders, but are evident in various physiological systems and a broad range of disorders, for example, in respiratory functions, immune functions, motor functions, and others.9

The neuro-bio-behavioral mechanisms of the placebo response are currently being investigated (for review, see Schedlowski et al10). In placebo analgesia, the neurobiological mechanism could be traced down to the spinal cord level.11 Less is known about the neurobiological mechanism of the cardiovascular placebo response, but reduction of β-adrenergic receptor activity in the heart and changes in coronary diameter have been discussed.12, 13 Efferents of the autonomous nervous system seem to contribute to this effect,14 though the detailed neurobiological pathways have not yet been fully elucidated. This evidence, however, stems mostly from experimental approaches, and we know little about the placebo effects in the cardiovascular system in long-term treatments.

Hypertension is a meaningful example of a disease requiring permanent medication, since it is a major health issue worldwide.15 Beta-blockers are a common permanent medication in hypertension that lowers blood pressure significantly.16 Blood pressure reduction is important to decrease the morbidity and mortality of hypertension and ensuing cardiovascular diseases.17 Beta-blockers were introduced in the 1960s and were recognized by most guidelines for the treatment of hypertension.18, 19 Even though they are not considered as first-line treatment for hypertension anymore,20 beta-blockers are still being investigated in clinical trials.21

The cardiovascular system is sensitive to placebo mechanisms.22 Experimental research suggests that placebo interventions can decrease systolic blood pressure (sBP).23, 24 In a double-blind randomized clinical trial, Grenfell et al25 described a blood pressure-lowering effect in hypertensive patients through parenteral placebo administration. These placebo effects are different from the regression to the mean phenomenon.26 It describes that if a variable is extreme on its first measurement, it tends to be closer to the average on its second measurement. Thus, hypertensive patients with a very high blood pressure on baseline measurement might have a lower blood pressure after placebo treatment, which could be confused with a placebo effect. Previous research, however, has already demonstrated in mild-to-moderate hypertension that the placebo response can be reliably distinguished from the regression to the mean phenomenon.27 A randomized controlled trial showed that many hypertensive patients (31%) reach their goal blood pressure even in the placebo group.28 However, the study included a small sample of only male hypertensive patients with mild-to-moderate hypertension.

At this point in time, there has been no meta-analysis that specifically investigated the relative effect of placebo groups compared to beta-blocker groups in hypertension. The existing meta-analyses16 are not well suited to answer this research question. To adequately estimate placebo effects in RCTs, the characteristics of the trial design need to be considered. The number of trial arms and different drug types are hypothesized to have a substantial influence on the placebo effect.29 Parallel-group designs are necessary, as crossover designs lead to an underestimation of the placebo effects.9 Additionally, the effect sizes in placebo groups are highly correlated with those in the respective drug groups.5 Therefore, we focused on beta-blockers and their respective placebo groups, in order to achieve a homogenous estimator of the placebo effect. Beta-blockers have been used for over 50 years30 and have often been compared to placebo groups in randomized controlled trials.31 In the light of this large amount of research data, a meta-analysis of beta-blocker trials seems to be well suited to investigate placebo responses in hypertension in a large sample. Thus, our meta-analysis focused on parallel-group designs only, considered only one type of medication and examined the effects of trial design (probability of receiving placebo) on the placebo effect.

Section snippets

Methods

We followed PRISMA guidelines for this meta-analysis.32

Study Selection

As shown in Figure 1, we identified 1,284 articles. Additionally, we found 22 abstracts on the World Health Organization International Clinical trials registry. We then removed duplicates and scanned a total of 692 abstracts, of which 113 were selected for full text reading (see Figure 1 for reasons for exclusion). Of those 113 studies, 39 studies were removed because they used a crossover design, 35 because they did not report exact values for blood pressure, 6 because they were not randomized

Discussion

Hypertensive patients in randomized controlled beta-blocker trials respond to placebo interventions. This means that even the intake of an inert pill lowers blood pressure significantly with robust small-to-medium effect sizes. The blood pressure change was quantified through objective measures of sBP and dBP. Almost half of the blood pressure-lowering effect that was observed in the drug groups was also observed in the placebo groups. For sBP, no substantial moderators of the placebo effect

References (87)

  • J.W. Black et al.

    A new adrenergic beta-receptor antagonist

    Lancet

    (1964)
  • A. Liberati et al.

    The PRISMA statement for reporting systematic reviews and meta-analyses of studies that evaluate health care interventions: explanation and elaboration

    J Clin Epidemiol

    (2009)
  • J.A. Glombiewski et al.

    Psychological treatments for fibromyalgia: a meta-analysis

    Pain

    (2010)
  • R. Rosenthal

    The file drawer problem and tolerance for null results

    Psychol Bull

    (1979)
  • C. Su et al.

    A meta-analysis of the placebo rates of remission and response in clinical trials of active ulcerative colitis

    Gastroenterology

    (2007)
  • S. Leucht et al.

    Antipsychotic drugs versus placebo for relapse prevention in schizophrenia: a systematic review and meta-analysis

    Lancet

    (2012)
  • B.R. Rutherford et al.

    Clinic visits in late-life depression trials: effects on signal detection and therapeutic outcome

    Am J Geriatr Psychiatry

    (2014)
  • P.A. Ades et al.

    Cardiac and skeletal muscle adaptations to training in systemic hypertension and effect of beta blockade (metoprolol or propranolol)

    Am J Cardiol

    (1990)
  • M.C. Houston et al.

    The effects of clonidine hydrochloride versus atenolol monotherapy on serum lipids, lipid subfractions, and apolipoproteins in mild hypertension

    Am Heart J

    (1990)
  • V. Papademetriou et al.

    Factorial antihypertensive study of an extended-release metoprolol and hydrochlorothiazide combination

    Am J Hypertens

    (2006)
  • E.J. Pérez-Stable et al.

    The effects of propranolol on cognitive function and quality of life: a randomized trial among patients with diastolic hypertension

    Am J Med

    (2000)
  • G. Mancia et al.

    Lack of placebo effect on ambulatory blood pressure

    Am J Hypertens

    (1995)
  • B.A. Gould et al.

    Does placebo lower blood pressure?

    Lancet

    (1981)
  • W. Rief et al.

    Rethinking psychopharmacotherapy: The role of treatment context and brain plasticity in antidepressant and antipsychotic interventions

    Neurosci Biobehav Rev

    (2016)
  • P. Easterbrook et al.

    Publication bias in clinical research

    Lancet

    (1991)
  • M.C. Juergens et al.

    Illness beliefs before cardiac surgery predict disability, quality of life, and depression 3 months later

    J Psychosom Res

    (2010)
  • J.A.C. Laferton et al.

    Enhancing the efficacy of heart surgery by optimizing patients’ preoperative expectations: study protocol of a randomized controlled trial

    Am Heart J

    (2013)
  • B.K. Doering et al.

    Utilizing placebo mechanisms for dose reduction in pharmacotherapy

    Trends Pharmacol Sci

    (2012)
  • P. Enck et al.

    The placebo response in medicine: minimize, maximize or personalize?

    Nat Rev Drug Discov

    (2013)
  • W. Rief et al.

    Mechanisms involved in placebo and nocebo responses and implications for drug trials

    Clin Pharmacol Ther

    (2011)
  • I. Kirsch et al.

    Initial severity and antidepressant benefits: a meta-analysis of data submitted to the food and drug administration

    PLoS Med

    (2008)
  • M. Shedden-Mora et al.

    Lessons learned from placebo groups in antidepressant trials

    Philos Trans R Soc Lond B Biol Sci

    (2011)
  • L. Colloca et al.

    Placebos and painkillers: is mind as real as matter?

    Nat Rev Neurosci

    (2005)
  • G. Andrews

    Placebo response in depression: bane of research, boon to therapy

    Br J Psychiatry

    (2001)
  • A. Winkler et al.

    Effect of placebo conditions on polysomnographic parameters in primary insomnia: a meta-analysis

    Sleep

    (2015)
  • M. Schedlowski et al.

    Neuro-bio-behavioral mechanisms of placebo and nocebo responses: implications for clinical trials and clinical practice

    Pharmacol Rev

    (2015)
  • F. Eippert et al.

    Direct evidence for spinal cord involvement in placebo analgesia

    Science

    (2009)
  • M. Law et al.

    Lowering blood pressure to prevent myocardial infarction and stroke: a new preventive strategy

    Health Technol Assess

    (2003)
  • M. Ezzati et al.

    Rethinking the “diseases of affluence” paradigm: global patterns of nutritional risks in relation to economic development

    PLoS Med

    (2005)
  • P. Boutouyrie et al.

    Local pulse pressure and regression of arterial wall hypertrophy during long-term antihypertensive treatment

    Circulation

    (2000)
  • G. Pucci et al.

    Effects of β-blockers with and without vasodilating properties on central blood pressure: systematic review and meta-analysis of randomized trials in hypertension

    Hypertension

    (2016)
  • K. Meissner

    The placebo effect and the autonomic nervous system: evidence for an intimate relationship

    Philos Trans R Soc Lond B Biol Sci

    (2011)
  • S. Agras et al.

    Expectation and the blood-pressure-lowering effects of relaxation

    Psychosom Med

    (1982)
  • Conflict of interest: M.W., A.W., W.R., and B.K.D. have no conflicts of interest including any financial, personal, or other relationships with other people or organizations to declare that could inappropriately influence, or be perceived to influence, the present work. This study did not require ethical approval.

    View full text