SEARCH

SEARCH BY CITATION

Keywords:

  • anxiety;
  • cortisol;
  • expectations;
  • nocebo;
  • placebo;
  • visceral pain

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. Funding
  9. Disclosure
  10. Author contributions
  11. References

Background  In order to elucidate placebo and nocebo effects in visceral pain, we analyzed the effects of positive and negative expectations on rectal pain perception, rectal pain thresholds, state anxiety and cortisol responses in healthy women.

Methods  Painful rectal distensions were delivered at baseline, following application of an inert substance combined with either positive instructions of pain relief (placebo group, = 15), negative instructions of pain increase (nocebo group, = 17), or neutral instructions (control, = 15). Perceived pain intensity, unpleasantness/aversion and urge-to-defecate, state anxiety and serum cortisol were determined at baseline, immediately following group-specific instructions and on a second study day after the same instructions (test day). Rectal pain thresholds were determined at baseline and on the test day.

Key Results  Whereas perceived pain intensity was significantly decreased in the placebo group, the nocebo group revealed significantly increased pain intensity ratings, along with significantly greater anticipatory anxiety on the test day (all < 0.05 vs controls). Cortisol concentrations were significantly increased in the nocebo group following treatment but not on the test day.

Conclusions & Inferences  The experience of abdominal pain can be experimentally increased or decreased by inducing positive or negative expectations. Nocebo effects involve a psychological stress response, characterized by increased anticipatory anxiety. These findings further underscore the role of cognitive and emotional factors in the experience of visceral pain, which has implications for the pathophysiology and treatment of patients with chronic abdominal complaints.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. Funding
  9. Disclosure
  10. Author contributions
  11. References

A better understanding of the neurobiology and neuropsychology of placebo and nocebo effects is of great scientific importance with profound implications for clinical practice.1,2 Placebo analgesia, i.e., a reduction in perceived pain due to positive expectations and/or previous experiences, constitutes one of the most intensely studied phenomena in placebo research. Its ‘negative equivalent’,3 nocebo hyperalgesia, i.e., an increase in perceived pain due to negative expectations and/or previous learning, is less well-understood although its clinical relevance may be just as high.4 Whereas the mechanisms mediating placebo and nocebo effects are increasingly well-characterized in somatic pain models,1 comparatively little placebo research has been conducted in the context of visceral pain (reviewed in Refs 5,6). Indeed, current knowledge about placebo effects in visceral pain comes from only a handful of experimental studies conducted exclusively within irritable bowel syndrome (IBS) patients,7–9 or within healthy individuals.10,11 To this date, no studies exist addressing nocebo effects in visceral pain. Filling this gap is important for two main reasons. First, chronic abdominal pain, such as present in the IBS, is highly prevalent, with detrimental individual and socioeconomic impact.12 Interestingly, IBS patients show high placebo response rates in clinical trials,2 and appear to benefit from placebo interventions.13,14 Second, the same psychological factors assumed to mediate nocebo effects, particularly anxiety and stress,15 have also been implicated in the pathophysiology of IBS16,17 and in altered pain processing in healthy women.18 Hence, the study of placebo/nocebo effects in visceral pain constitutes a model to assess the contribution of psychological factors in the response to pain, relevant to the pathophysiology and treatment of IBS and other conditions associated with visceral pain.19

As an extension of our previous work on the neural mechanisms of placebo analgesia,10 we designed the present pilot study to assess placebo as well as nocebo effects in a clinically relevant visceral pain model in a separate group of healthy subjects. Focussing on healthy subjects at this early stage seems warranted since more preclinical research is needed to discern the mechanisms of placebos and nocebos as well as to derive sophisticated experimental paradigms for application in future patient-oriented research. Given the paucity of existing data specifically on nocebo effects in visceral pain, our primary goal was to implement an experimental nocebo model capable of inducing significantly increased pain perception, i.e., a nocebo response, when compared to a control condition. Using this nocebo paradigm, we then aimed to test the hypothesis that nocebo effects for visceral pain involve a psychological stress response, characterized by increased anxiety along with activation of the hypothalamus–pituitary–adrenal (HPA) axis as has previously been suggested in other pain models.15 Based on findings in somatic placebo analgesia studies, it has been proposed that placebo effects may be mediated by reduced negative emotions including lowered anxiety.20 However, findings remain inconsistent and few studies have been designed to directly contrast placebo and nocebo effects within the same pain model.21–24 Therefore, our second goal was to assess emotional and neuroendocrine responses to placebo and nocebo, compared to a control, within in the same pain model and experimental paradigm, in order to test the hypothesis that placebo analgesia is mediated by reduced negative emotions and stress-related HPA-axis activation. Finally, given previous findings that acute stress can lower rectal pain thresholds in IBS,25 we also aimed to explore possible effects of our placebo and/or nocebo intervention on visceral pain thresholds.

Methods

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. Funding
  9. Disclosure
  10. Author contributions
  11. References

Participants

A total of = 47 healthy women, all naïve to balloon distension studies, were recruited by local advertisement and carefully screened for a number of exclusion criteria. Females were chosen given the well-known female preponderance of functional gastrointestinal disorders including IBS and possible sex differences in visceral pain processing in healthy subjects or animals.26–28 Exclusion criteria included age <18 years and >45 years, body mass index <18 or ≥30, any concurrent medical condition, including gastrointestinal, neurological, psychiatric, cardiovascular, immunological, and endocrine conditions. Only females on hormonal contraceptives were studied to reduce potential confounding by menstrual cycle phase. All participants were evaluated digitally for anal tissue damage (e.g., painful haemorrhoids) which may interfere with balloon placement. A history of psychological conditions (based on self-report) or presently increased scores on the Hospital Anxiety and Depression Inventory (HADS) were also exclusionary.29 Frequency and severity of gastrointestinal complaints suggestive of any functional or organic gastrointestinal condition were assessed in a structured phone interview and subsequently with a standardized in-house questionnaire assessing various relevant gastrointestinal symptoms over the past three months. The study protocol was approved by the local Ethics Committee (protocol number 08-3823). All participants gave written informed consent, and were paid for their participation.

Study design

Women were randomly assigned to a placebo group (= 15), a nocebo group (= 17), or a control group (= 15). The study was comprised of two study days and included three experimental phases, i.e., a baseline, a placebo/nocebo manipulation (‘treatment’) (both day 1) and a test phase (day 2), illustrated in Fig. 1. Rectal sensory and pain detection thresholds were determined initially (day 1) and at the end of the test phase (day 2). On day 1, a baseline was accomplished during which subjects received four identical rectal distensions (duration: 20 s; pauses in-between: 40 s) at pressures just below the individual pain threshold. After a 15-min pause, the treatment (nocebo, placebo, control) was accomplished (placebo/nocebo/control manipulations are explained in detail below). In this phase, four rectal distensions were carried out. In the test phase on day 2, deceptive instructions (depending on group assignment or no instructions in the control group) were repeated, and five distensions at the identical pressure as during baseline were delivered. Subjects rated each distension using three separate visual analogue scales (VAS: 0–100 mm; ends defined as 0: ‘none’ to 100: ‘very much’) assessing perceived pain intensity, unpleasantness/aversion, and urge-to-defecate. Prior to and after each experimental phase, blood samples were taken for analysis of serum cortisol as a marker of HPA-axis activation and state anxiety was assessed with the state version of the State-Trait-Anxiety Inventory.30 To minimize circadian effects, all studies were conducted between 15:00 and 18:00 and both study days were started at the identical time for each subject.

image

Figure 1.  Study design comprised of 2 study days. Pain thresholds were determined on days 1 and 2 with a barostat using a double-random staircase protocol with random pressure increments. At the beginning of the treatment and test phases, all subjects received an intravenous drip that contained sodium chloride, but the placebo group was informed that they received a potent pain killer. The nocebo group received instructions of ‘sensitization’. Blue bars indicate rectal distensions just below the individual pain threshold in all groups and phases, except for the treatment phase in the nocebo group which received slightly increased distension pressures (+2 mmHg) to induce a nocebo response. During test, all subjects received distension pressures identical to those received during the baseline. ‘B’ = blood withdrawal together with a state anxiety questionnaire.

Download figure to PowerPoint

Placebo/nocebo manipulation

All instructions were standardized and provided in writing prior to the experiments following the screening procedure (as part of informed consent). During the experiments, pertinent aspects of the instructions were repeated orally by the investigator and the physician (E.C.) who carried out all medical procedures including placement of rectal catheters. Following the baseline, all subjects received sodium chloride via intravenous (i.v.) drip, but received different instructions to create positive (placebo group), negative (nocebo group) or neutral (control group) expectations. Specifically, in the control group, subjects were truthfully informed that an inert substance, i.e., sodium chloride, was administered. The container used for the infusion drip was clearly marked ‘sodium chloride’. In the placebo group, subjects also received an i.v. drip containing sodium chloride, but were falsely informed that they were receiving a potent pain killer, i.e., Butylscopolaminiumbromid, a commercially available spasmolytic drug. The syringe that was injected into the container used for the continuous infusion drip was labelled with the commercial name of the substance, but in reality contained sodium chloride. We used a similar procedure to induce placebo analgesia in a recently published functional brain imaging study.10 In the nocebo group, subjects also received the i.v. drip, were informed that it contained sodium chloride, but were also told that the aim of the study was to verify our previous observation that pain ratings increased significantly in the majority of subjects over time, i.e., that sensitization occurred, over the course of the 2-day experiment. In addition to inducing these negative expectations by verbal instructions, distension pressures were surreptitiously slightly increased by 2 mmHg during the nocebo manipulation phase. Importantly, on the test day (day 2), all subjects again received an i.v. drip containing sodium chloride, identical instructions as on day 1, and distension pressures that were identical to baseline pressures. Placebo/nocebo responses were quantified by assessing the change in VAS ratings from baseline to the test day.

Rectal distensions

Rectal distensions were carried out with a pressure-controlled barostat system (modified ISOBAR 3 device; G & J Electronics, Toronto, ON, Canada). Perception and pain detection thresholds were determined using a double-random staircase protocol with random pressure increments of 2–10 mmHg as previously described.10,16,17 Subjects were prompted to rate the sensation as follows: 1 = no perception, 2 = doubtful perception, 3 = sure perception, 4 = little discomfort, 5 = severe discomfort, still tolerable, 6 = pain, not tolerable. The threshold for first perception was defined as the pressure when the rating changed from 2 to 3; the pain detection threshold as the pressure at which the rating changed from 5 to 6. The maximal distension pressure was always set at 50 mmHg for safety reasons. For repeated distensions in the baseline, manipulation and test phases, pressures just below the individual pain detection threshold (individual pain threshold – 2 mmHg) were chosen, except for the nocebo group who received pressures at the individual pain threshold during the treatment phase (i.e., 2 mmHg higher than the other groups, see also Fig. 1).

Serum cortisol

Serum cortisol concentrations were analyzed with a commercial enzyme-linked immunosorbent assay (Cortisol ELISA, IBL International, Hamburg, Germany) according to the manufacturer’s instructions. Intra- and interassay variance were 5.6% and 8.8%, respectively.

Statistical analyses

All data were initially assessed for normal distribution using Kolmogorov–Smirnov-tests. Given non-normally distributed data, non-parametric statistics were chosen for all analyses except sociodemographic and psychological characteristics (which were normally distributed). The following analyses were carried out:

  • 1
     Groups were compared with respect to sociodemographic and psychological characteristics using analysis of variance (anova) or chi-square tests for dichotomous variables.
  • 2
     Baseline scores and values (Table 2) for all dependent measures were assessed with non-parametric Kruskal–Wallis tests assessing overall group differences in the baseline. In the event of a significant overall group effect, post hoc group comparisons were accomplished using Mann–Whitney-U tests.
  • 3
     To test group differences induced by our experimental manipulation, non-parametric Kruskal–Wallis tests assessing overall effects of group (i.e., placebo, nocebo, control) in delta scores were conducted for VAS ratings and pain detection threshold. Only in case of a significant overall group effect, specific and directed hypotheses on placebo- and nocebo-related changes in VAS ratings and pain detection thresholds were tested with post hoc non-parametric tests (Mann–Whitney U-tests) comparing delta scores (change from baseline on day 1 to test on day 2) between groups. Given specific and directed hypotheses for VAS ratings, one-tailed tests were used; given explorative analysis of possible changes in pain detection thresholds, these post hoc tests were two-tailed.
  • 4
     To test specific hypotheses regarding anxiety and cortisol, one-tailed post hoc group comparisons were computed on absolute values also using Mann–Whitney U-tests following Kruskal–Wallis tests.

Exact P-values are shown for all post hoc analyses and effects of a Bonferroni correction are reported throughout. In most analyses (exceptions: state anxiety and cortisol), the placebo and nocebo groups, respectively, were separately compared with the control group to avoid inflation of the risk for Type I error by multiple testing. Results are shown as mean ± standard error of the mean (SEM).

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. Funding
  9. Disclosure
  10. Author contributions
  11. References

Sociodemographic characteristics and baseline scores/values

The three groups were comparable with regard to sociodemographic and psychological characteristics, including age, family status, education, smoking behavior, anxiety and depression scores (Table 1). There were no significant differences between groups in any of these measures.

Table 1.   Sociodemographic and psychological characteristics for placebo, nocebo, and control groups
 Placebo (= 15)Nocebo (= 17)Controls (= 15) P
  1. All data are shown as mean ± standard deviation, unless otherwise indicated. HADS, hospital anxiety and depression scale; NS, non-significant (results of univariate anova or chi-square tests).

Age (years)26.2 ± 7.124.7 ± 6.124.4 ± 2.8NS
Body mass index (kg m2)22.5 ± 2.521.0 ± 2.122.3 ± 2.3NS
Married or with a partner % (N) 53.8 (7)52.9 (9)53.8 (7)NS
Education >12 years % (N)92.3 (12)100.0 (17)84.6 (11)NS
Smoker % (N)30.8 (4)23.5 (4)23.1 (3)NS
HADS anxiety score3.5 ± 3.03.9 ± 2.64.2 ± 2.8NS
HADS depression score1.7 ± 2.71.5 ± 1.91.5 ± 1.9NS

Baseline scores and values of all relevant dependent measures (assessed on day 1) are shown in Table 2. These data reveal a significant group effect, assessed with Kruskal–Wallis-test, for VAS ratings of perceived pain intensity (< 0.05); however, post hoc Mann–Whitney-U tests comparing individual groups showed no significant group differences.

Table 2.   Baseline values on day 1 for placebo, nocebo and control groups
 PlaceboNoceboControls P*
  1. All data are shown as mean ± standard error of the mean. VAS, visual analogue scale; *Kruskal–Wallis test assessing an overall group effect; post hoc Mann–Whitney-U tests comparing individual groups revealed no significant effects. NS, non-significant.

VAS pain intensity (mm)64.29 ± 8.0540.34 ± 7.0656.37 ± 7.910.025
VAS unpleasantness/ aversion (mm)76.65 ± 4967.43 ± 5.0874.67 ± 5.09NS
VAS urge-to-defecate (mm)72.08 ± 7.9868.12 6.4170.35 ± 5.34NS
Rectal sensory threshold (mmHg)13.77 ± 1.4213.41 ± 0.7817.46 ± 1.48NS
Rectal pain detection threshold (mmHg)33.92 ± 2.8936.06 ± 2.4738.46 ± 2.68NS
State anxiety (STAI-S)35.15 ± 1.5537.25 ± 1.7932.92 ± 1.76NS
Serum cortisol (nmol/L)301.28 ±  36.85316.28 ±  32.64285.59 ± 28.78NS

VAS ratings of pain intensity, unpleasantness/aversion, and urge-to-defecate

Kruskal–Wallis-tests revealed significant overall group effects for changes in all three VAS ratings, namely in distension-induced pain intensity (< 0.001, Fig. 2A), perceived unpleasantness/aversion (< 0.001, Fig. 2B), and distension-induced urge-to-defecate (< 0.05, data not shown). Post hoc comparisons of group differences in changes in VAS ratings from baseline (day 1) to the test (day 2) showed that the placebo group demonstrated a significantly greater decrease in both perceived pain intensity (= 0.016) and unpleasantness/aversion (= 0.025) when compared to the change observed in the control group (Fig. 2A and B). In contrast, the nocebo group demonstrated a significant increase in both perceived pain intensity (= 0.007) and distension-related unpleasantness/aversion (= 0.01) from baseline to the test when compared to the change in controls (Fig. 2A and B). Applying Bonferroni correction for multiple comparisons did not alter these results except for the effect on unpleasantness/aversion in the placebo group which is then no longer significant. All post hoc tests for urge-to-defecate were non-significant (data not shown).

image

Figure 2.  Changes in pain intensity ratings (A) and distension-induced unpleasantness/aversion (B), assessed with two visual analogue scales (VAS, in mm) from baseline to test in the placebo, nocebo and control groups. The placebo group showed significantly greater decreases in perceived pain intensity (= 0.016) and unpleasantness/aversion (= 0.025) from baseline to test when compared to the change observed in controls. In contrast, the nocebo group revealed a significant increase in both perceived pain intensity (= 0.007) and distension-related unpleasantness / aversion (= 0.01) from baseline to the test when compared to the change in controls (*all post hoc one-tailed Mann–Whitney-U tests). After Bonferroni correction, all group differences remain significant except for the effect on unpleasantness/aversion in the placebo group. Data are shown as mean ± SEM.

Download figure to PowerPoint

Pain detection thresholds

Kruskal–Wallis-tests revealed a significant overall group effect for the change in rectal pain detection threshold from day 1 to day 2 (< 0.05). Post hoc group comparisons showed a significant difference in the change in pain detection threshold (placebo group: 1.54 ± 1.9 mmHg vs control group: −4.23 ± 1.2 mmHg, = 0.014). To follow-up on this decrease in pain threshold from day 1 to day 2 in the control group, we conducted a supplementary within-group comparison, which revealed a significant effect (= 0.009), indicative of pain sensitization in controls.

No significant difference was observed when comparing the change in the nocebo (−2.0 ± 1.6) and the control groups.

State anxiety

For state anxiety, the Kruskal–Wallis-test revealed a significant overall group effect for the pre-treatment (i.e., baseline) assessment on study day 2 (< 0.05, see Fig. 1 for the study design and Fig. 3 for data). Post hoc tests comparing groups at this time point showed that anticipatory state anxiety was significantly increased on the test day in the nocebo group when compared to both the control group (*= 0.01) and the placebo group (*= 0.028) (Fig. 3). After Bonferroni correction for multiple comparisons, the difference between the placebo and nocebo groups is no longer significant.

image

Figure 3.  State anxiety, assessed with the state version of the State-Trait-Anxiety Index,30 in the placebo, nocebo and control groups on both study days (for assessment time points, see study design in Fig. 1). Anticipatory state anxiety was significantly increased on the test day (study day 1) in the nocebo group when compared to both the control group (post hoc Mann–Whitney-U test: *= 0.01) and the placebo group (post hoc Mann–Whitney-U test *= 0.028). Data are shown as mean ± SEM.

Download figure to PowerPoint

Serum cortisol

For serum cortisol concentrations, the Kruskal–Wallis-test revealed a significant overall group effect for the post-treatment assessment on study day 1 (< 0.05, see Fig. 1 for the study design and Fig. 4 for data). Post hoc tests comparing groups at this time point showed that post-treatment serum cortisol concentrations were significantly higher in the nocebo group when compared to the control group (*= 0.007, Fig. 4). Note that since no overall group effects were detected for study day 2, no post hoc tests were carried out on these data.

image

Figure 4.  Serum cortisol concentrations in the placebo, nocebo and control groups on both study days. On study day 1, post-treatment serum cortisol concentrations were significantly higher in the nocebo group when compared to the control group (post hoc Mann–Whitney-U test: *= 0.007). Note that since no overall group effects were detected for study day 2, no post hoc tests were carried out on these data. Data are shown as mean ± SEM.

Download figure to PowerPoint

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. Funding
  9. Disclosure
  10. Author contributions
  11. References

We studied placebo and nocebo responses in a clinically relevant visceral pain model by inducing positive and negative expectations concerning rectal distensions in healthy women. Our first aim was to implement and test an experimental nocebo model capable of inducing significantly increased VAS ratings of perceived pain intensity and distension-induced unpleasantness/aversion, i.e., a nocebo response. We could indeed document for the first time in a visceral pain model significantly increased ratings of rectal distension-induced pain intensity and unpleasantness/aversion in the nocebo group. Of note, our experimental approach to induce a nocebo response differed from some previous studies in the somatic pain field which involved the induction of negative expectations by combining the application of an inert substance (as herein), but together with instructions regarding harmful side effects or worsening of symptoms associated with a specific drug (unlike in our study). Our rationale for choosing this procedure was that we are not aware of a suitable and safe drug that would believably result in a transient but noticeable increase in rectal-distension induced pain and at the same time be acceptable to healthy subjects and possibly also (in future studies) to patients with chronic abdominal pain such as IBS patients. Therefore, we chose to induce negative expectations through a combination of verbal suggestions (i.e., instructions regarding ‘pain sensitization’) and surreptitiously increased distension pressures (i.e., 2 mmHg) during the nocebo manipulation, consistent with many placebos studies implementing ‘preconditioning’ paradigms (e.g., Refs 31,32). Our procedure is somewhat similar to a recently published report23 on placebo and nocebo effects on itch and somatic pain which successfully induced nocebo responses by verbal suggestions also not involving a supposed drug with specific harmful effects and/or adverse side effects. By definition, nocebo and placebo effects require the administration of an inert substance, however, the entire context surrounding medical encounters is now regarded as crucial within this emerging field of research.4 This shift in the definition of placebo/nocebo effects was primarily based on the realization that in daily clinical routine, negative expectations regarding worsening of symptoms can occur through a number of possible factors which may or may not involve an actual drug or medication.4 Therefore, we conclude that our contextual manipulation may constitute a suitable experimental model to assess nocebo effects that is relevant to the pathophysiology of visceral hyperalgesia, is mediated by negative expectations, and applicable to studies including patients. However, we also acknowledge that the present design constitutes a pilot study with several limitations, including a relatively small number of subjects per group combined with considerable variance and group differences in some baseline measures as well as limited statistical power. Clearly, these findings should be regarded as preliminary and proper replication and extension of the present findings, together with a refinement of study protocols and procedures is needed before the initiation of larger trials in patient groups.

In addition to increased perceived pain intensity and aversion ratings (i.e., visceral hyperalgesia), the nocebo group also showed significantly greater anticipatory anxiety on the test day, consistent with our second hypothesis. Serum cortisol results, on the other hand, were less clear. Although we did find increased cortisol concentrations in the nocebo group, this difference was only significant on study day 1 following the treatment phase and no significant effects were observed in the test day. Hence, the present data are not unequivocal with respect to the putative involvement of the HPA stress axis in nocebo responses. The role of anxiety in nocebo effects has previously been documented in other pain models, including experimental ischemic arm pain,33,34 painful mechanical and/or electrical stimulation21–23 and heat pain35 in healthy subjects, as well as in patients with postoperative pain.36 Given our findings, it would appear that nocebo effects are mediated by negative emotions regardless of pain modality. Together, these findings linking anticipatory anxiety and increased pain perception clearly have implications for the pathophysiology of visceral hyperalgesia. In fact, Price et al. noted that ‘…these forms of hyperalgesia are also highly modifiable by placebo and nocebo factors […,] synergistic interactions occur between placebo/nocebo factors and enhanced afferent processing so as to enhance, maintain, or reduce hyperalgesia in IBS’,37 thereby catapulting placebo/nocebo issues at the heart of a multi-factorial disease model of IBS. Based on these considerations, future attempts to improve treatment options in IBS should not only incorporate placebo interventions, as has been elegantly been accomplished by Kaptchuk and colleagues13,14 but also aim to clarify the putative role of nocebo effects in IBS in order to subsequently consider ways to reduce such undesired effects to the benefit of the patient.

In the placebo group, the expectation to receive a potent analgesic drug significantly reduced perceived pain intensity, consistent with our previous fMRI study results employing the same placebo manipulation in a separate group of healthy individuals.10 Together, these findings confirm and complement previous results in IBS patients7,9 as well as results of an esophageal placebo analgesia study in healthy subjects.11 On the other hand, we could not find evidence in support of our second hypothesis testing whether placebo analgesia involves reduced negative emotions and/or reduced stress-related HPA-axis activation. The notion that placebo effects could be mediated at least in part by reduced negative emotions has been put forward20 based on the previously established connection between placebo analgesia and reward processing.38 In healthy humans, findings in support of a role of negative emotions and stress are in fact rather heterogeneous.20 Sex differences as well as personality factors influence the experience of visceral pain in general and appear to play an important role in the intricate relation between negative emotions and visceral pain processing.26,39

For example, a recent study found that placebo administration reduced anticipatory stress only in males but not in females.40 Our findings in females are consistent with this report as they did not provide evidence in support of a relation between negative emotions, cortisol and placebo analgesia in a visceral pain model. Clearly, more research is needed to clarify the intricate connections between emotions, reward processing, and placebo analgesia in different placebo paradigms and study populations.

Finally, we explored possible effects of the placebo or nocebo interventions on visceral pain thresholds as a more objective pain sensitivity measure, in an attempt to differentiate effects on subjective pain evaluation (assessed herein with three VAS scales) and pain detection threshold/pain sensitivity (assessed herein using a double-random random staircase distension protocol). In the nocebo group, we observed no significant effects on pain detection thresholds despite significant effects on VAS ratings. This disparity may indicate that (more subjective) VAS assessments could be more sensitive to nocebo instructions than (more objective) analysis of pain detection thresholds. In the control group, we found a small but significant decrease in pain detection threshold from day 1 to day 2, indicative of a slight sensitization. Interestingly, this was not observed in the placebo group who showed stable pain detection thresholds. One could speculate that placebo instructions may not only decrease VAS pain intensity ratings of pre-defined stimuli, but also reduce visceral pain sensitization (i.e., ‘prevent’ a decrease in pain detection threshold over consecutive study days). Given the clinical relevance of pain sensitization and visceral hyperalgesia together with previous evidence that placebo interventions produce clinically relevant benefits in IBS,13 this finding, given proper replication and further study, may be crucial in the process of translating findings from placebo research into the clinic. Indeed, the broad implications of experimental placebo and nocebo studies involving pain models were elegantly highlighted by recent evidence that positive treatment expectancy substantially enhanced the analgesic effect of a μ-opioid agonist whereas negative treatment expectancy completely abolished the analgesic efficacy.41 Hence, the expectation of a drug’s effect may shape both therapeutic and adverse effects in clinical settings. Given our findings, one could speculate that expectations may also modulate painful experiences in clinical encounters not involving the application of drugs. Clearly, more studies addressing placebo and nocebo responses in patients are urgently needed.

Acknowledgments

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. Funding
  9. Disclosure
  10. Author contributions
  11. References

We thank Kristina Bäumker, Christine Mewes, and Vassilios Kotsis for excellent support in conducting this project.

Funding

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. Funding
  9. Disclosure
  10. Author contributions
  11. References

This project was funded by the German Research Foundation (Deutsche Forschungsgemeinschaft, DFG) (EL 236/8-1). The DFG had no further role in study design; in the collection, analysis and interpretation of data; in the writing of the report; and in the decision to submit the paper for publication.

Author contributions

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. Funding
  9. Disclosure
  10. Author contributions
  11. References

JS, MB and EC performed the research; SE, SB and MS designed the research study; SB and JS analyzed the data; SE and SB wrote the paper; SE acquired funding; all authors contributed to the interpretation of the data and to the final version of the manuscript.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. Funding
  9. Disclosure
  10. Author contributions
  11. References