Klinger2017 - Placebo Effects LBP

PAIN Publish Ahead of Print
DOI: 10.1097/j.pain.0000000000000977
Placebo effects of a sham opioid solution: a randomized controlled study

in patients with chronic low back pain
Regine Klinger a *, Ralph Kothe b, Julia Schmitz a, Sandra Kamping, a, Herta

Flor c
D
a) Department of Anesthesiology, University Medical Center Hamburg –
Eppendorf, Hamburg, Germany
TE
b) Schön Klink Hamburg Eilbek, Hamburg, Germany; Department of Ortho-
pedics, University Medical Center Hamburg-Eppendorf, Hamburg, Germany
c) Department of Clinical and Cognitive Neuroscience, Central Institute of
Mental Health, Medical Faculty Mannheim, University of Heidelberg and
EP
Department of Psychology, University of Mannheim, Mannheim, Germany
Corresponding author: Regine Klinger, PhD, University Medical Center

Hamburg-Eppendorf (UKE), Center for Anesthesiology and Intensive Care
C
Medicine, Department of Anesthesiology, Pain Therapy and Pain Psychol-

ogy, Martinistraße 52, 20246 Hamburg, Germany, E-mail:
r.klinger@uke.de, phone: +49 40 7410 52837, Fax: +49 40 7410 44963,
C
URL: www.uke.de
A
Abstract
This study tested the experimental placebo effect in a group of chronic pain
patients. Forty-eight patients suffering from chronic back pain participated
in a randomized clinical trial that tested the efficacy of a sham opioid solu-
tion (NaCl) compared to an alleged neutral, completely inactive solution
(NaCl). We shaped the placebo effect by two interventions: sugges-
1
Copyright Ó 2017 by the International Association for the Study of Pain. Unauthorized reproduction of this article is prohibited.
tion/instruction and conditioning. The patients were either told that the
“solution reduces pain/improves physical capacity” or the “solution is neu-
tral, a placebo”. Half of each group was additionally conditioned (coupling
solution with reduced experimental pain), yielding 4 subgroups with 12
participants each. Outcome measures were: the patients’ clinical back pain
ratings and acute pain ratings (both examined by numerical rating scale 0-
D
10), and self-rated functional capacity (0-100%; time required for the exer-
cise). Expected pain relief before and after solution intake was also as-
TE
sessed. The inactive solution (NaCl), when presented as an effective treat-
ment (sham “opioid” solution), i duced placebo analgesia as evident in
lower ratings of the patients’ clinical back pain (F (3.12, 144.21) = 25.05, p<
.001), acute pain ratings (F (1.99, 87.40) = 18.12, p < .01) and time needed
to complete a series of daily activities exercises (F (1, 44) = 8.51, p < .01) as
EP
well as increased functional capacity (F (1, 44.00) 19.42, p < 0.001). The
two manipulations (instruction and conditioning) changed pain expecta-
tions and they were maintained in both sham opioid groups. The results
suggest that it may be clinically useful to explicitly integrate placebo anal-
C
gesia responses into pain management.

C
Keywords: Placebo analgesia; chronic back pain; sham opioid solution; pla-
cebo effects; placebo responses; randomized controlled clinical trial; classi-
cal conditioning; expectancy theory.
A
1. Introduction
The good evidence for the efficacy of analgesic placebo effects in
healthy samples13, 15, 31 raises the question if they could also be experimen-
2
tally induced in clinical pain treatment.26 Clinical use of placebo effects
could induce additional pain relief and diminish the amount of analgesics
and related side-effects in these patients.
So far there has been little research on the effects of experimental place-
bos on the patients’ clinical pain. Most studies conducted on patients (e.g.
D
those with irritable bowel syndrome 22,23, headache 21, neuropathic pain
TE
32
or back pain 7,16,17) used experimental or acute pain and obtained signifi-
cant placebo effects.31,39,18,19 Only a few studies also examined underlying
placebo mechanisms such as experiential (e.g. partial reinforcement and
classical conditioning), social (e.g. vicarious learning and modeling) and

EP
verbal (e.g. suggestion-induced expectations) learning 21,22,23,32. Only one
study has compared the patients’ placebo response with that of healthy
participants 27 and no studies have examined the effects of analgesic pla-

C
cebo mechanisms on both clinical pain and disease-related function.
Until now, there have been no studies conducted on patients with

C
chronic pain investigating placebo effects on clinical pain using sham opioid
A
solutions. However, studies of this kind would be of great interest because
many neurobiological studies have documented great similarities between
the molecular basis of the mode of action of an opioid and the related pla-
cebo response, showing high effect sizes of an opioid placebo and of the
placebo component of an opioid.1, 4, 10, 33
3
An important source of the response variability to analgesics appears to be
the difference in the placebo-related activation of endogenous opioid sys-
tems. Clinical application of placebo effects requires that they can be relia-
bly triggered.26 It is, therefore, essential to know the underlying core
mechanisms: expectancy34, 40 and learning processes such as classical con-
D
ditioning or social learning.8, 11
TE
The aim of this study was to show that it is possible to manipulate an in-
ert solution via instruction and additional classical conditioning so that the
solution can reduce the perceived intensity of both an experimental pain
stimulus and clinical pain, and improve the physical capacity of chronic
EP
back pain patients. For this purpose, we examined the analgesic placebo
effect in a short-term (2 hours) experiment studying 48 patients with
chronic back pain using standardized daily activities. They were given a
C
pharmacologically neutral solution, which was introduced either as a pain-
reducing opioid solution that also improved physical function and move-
C
ment, or as a neutral solution, a placebo. The central questions were: (1)

A
Can an analgesic placebo response (on experimental acute as well as clini-
cal pain) be attained in chronic back pain patients by manipulating an inert
neutral solution via instruction and additional conditioning? (2) Can this
placebo response also reduce pain behavior and increase perceived func-
tional capacity in daily activities? (3) Does this placebo effect change ex-
4
pectancies of pain reduction? In addition, we hypothesized that the pa-
tients who received the supposed ”opioid solution“ would exhibit a
stronger placebo response than the patients who were told that the solu-
tion was a placebo.
2 Method
D
2.1 Participants
TE
The participants were 48 patients with chronic back pain who were re-
cruited from local multidisciplinary pain centers, back pain management
centers, and the Outpatient Clinic of Behavior Therapy at the Department
of Psychology, University of Hamburg, Germany. We estimated that a sam-

EP
ple size of 48 was needed for between- and within-subject comparisons to
attain 80% power and a 5% significance level against the backdrop of a
previous study in which similar outcome measures (numerical rating scales

C
for pain intensity, see below 2.3.1) were used.14 Our sample was recruited
within a larger project investigating placebo effects with back pain patients
C
.38
A
The sample included 36 (9 per group) women and 12 (3 per group)
men with a mean age of 49.97 years (SD = 13.64).
Inclusion criteria were: chronic, persistent or recurrent (≥ 1 per week)
back pain for a minimum of 6 months duration. The patients were diag-
nosed by the referring physician who excluded patients exhibiting “red
5
flags”. Red flags included, for example, significant trauma, unexplained
weight loss and widespread, neurological changes.28 The patients were
also excluded if they had inflammatory, neuropathic or tumor-related back
pain, and severe acute or chronic somatic diseases, such as acute radicular
syndrome, rheumatic disorders, cancer, severe heart diseases, which
D
would not allow experiment participation. Only patients between 20 and
TE
55 years of age were included. All somatic exclusion and inclusion criteria
were likewise assessed by the referring physicians. Mental health problems
were diagnosed by psychotherapists employing a psychological assessment
(interview) and additional questionnaire (depression: General Depression

EP
20 30
Scale (ADS) ; anxiety: State-Trait Anxiety Inventory (STAI) , excluding
persons with severe acute or chronic mental disorders in accordance with
the International Classification of Diseases (ICD). However, patients with

C
anxiety or depressive disorders secondary to the back pain diagnosis were
included. Further exclusion criteria were: Cognitive impairment, preg-

C
nancy, symptoms or diseases which prohibit the application of electrical

A
impulses (e.g. cardiac insufficiency and/or need of cardiac pacemakers or
implanted spinal cord stimulation devices). Patients were asked not to use
any acute pain medication (rescue medication) on the day of their partici-
pation in the experiment. They were allowed to take their long-term medi-
6
cation (medication to be taken regularly), for example, for pain or other
disorders.
On the day of the experiment the following long-term medication was
taken: 11 patients (22,92%; see Table 1) had taken opioids (hydromor-
phon, tramadol, tilidin phosphate/hydrochloride, morphine sulfate, fen-
D
tanyl, oxycodon) or 16 patients (33.33%) had taken non-opioids (ibupro-
TE
fen, flupirtine, paracetamol with/without coffein, carbamazepine, di-
clofenac, acetylsalicylic acid, metamizole, gabapentin). During the week
before the experiment 15 patients (31,25%) took opioids (long-term medi-
cation). Patients with medication overuse or impairments due to the

EP
medication (e.g. opioid intolerance) were also excluded from the experi-
ment.
The patients were randomly assigned to the four experimental groups

C
(1. instruction “solution is neutral, a placebo” / no conditioning, 2. instruc-
tion solution is neutral, a placebo” / conditioning, 3. instruction “solution is

C
an opioid, reduces pain/ improves physical capacity” / no conditioning, 4.

A
instruction “solution is an opioid, reduces pain/ improves physical capac-
ity” / conditioning). The study adhered to the guidelines of the Declaration
of Helsinki and informed consent was obtained from all subjects. The study
was approved by the local institutional review board (No. 2275) and regis-
tered (German Clinical Trials Register: DRKS00011201).
7
2.2 Setting and materials
For the experiment, an intracutaneous electrical painful stimulus was
applied.5, 6 An electrical stimulator especially adapted for intracutaneous
electrical nerve stimulation (G 43, Fa. Bentronic, Munich, Germany) was
used and painful stimulation was introduced through an intracutaneous
D
platinum electrode placed on the epidermis of the finger pad of the left
TE
index finger after the hard surface of the skin had been removed with a
curette. The electrode was then connected to the power supply system.
The device for the application of the painful stimulus was controlled
through a LabVIEW-based (National Instruments Corporation, Austin,

EP
Texas) computer program. The placebo was a saline solution, colored
lightly red with a bitter taste of quinine. The patients were given a cotton
swab with this solution which they were asked to put in their mouth. The 2
C
bottles containing the solution were labeled with the fictional terms 1.
“Opioid KP / 310“ or 2. “KP / 236 (Placebo)“, in which “KP“ stood for

C
“Klinische Prüfung“ (“Clinical trial“). The patients in all groups received the
A
pharmacologically inert neutral solution; none of the patients received a
verum.
The experiment was set in a treatment room at the Department of Or-
thopedics at the University Hospital Hamburg-Eppendorf. This clinical situa-
8
tion was chosen to improve the credibility of the experiment and the po-
tential effectiveness of the placebo.
2.3 Procedure
According to the authorized informed consent, the patients were in-
D
formed about the fact that they deliberately received altered information
about the true processes in the experiment and that a detailed debriefing
TE
would be provided after the experiment. All information the patients re-
ceived was specifically authorized by the local ethics committee. The pa-
tients were told that they were taking part in a clinical study to test the
EP
effects of a new highly effective pain-reducing solution for CLBP treatment,
which also improves physical function and movement (see Supplement 1).
This solution would be tested against a solution without an active compo-

C
nent, a placebo. The participants would be randomly assigned to the group
treated with the solution or the control group with the placebo. All partici-
C
pants were informed about the application of painful stimuli according to
ethical guidelines and could leave the experiment at any point in time.
A
2.3.1 Measures / calibration trials
The magnitude of the experimental pain stimulus was individually tai-
lored.6, 27, 42 Stimuli with rising intensity were applied through the electrode
beginning with 0.12 miliamperes (mA) in a sequential up and down proce-
dure with stimulus intensity increments of 0.02 mA (in the range of 0.12
9
mA to 0.20 mA), 0.04 mA (in the range of 0.20 mA to 1.00 mA) and 0.10 mA
(in the range of 1.00 mA to 3.00 mA). Each stimulus had to be rated on a 9-
point scale according to its pain intensity (0= not noticeable, 1= just about
perceivable, 2=clearly perceivable, 3= strongly perceivable but not painful,
4=strongly perceivable, noticeably painful, 5= clearly painful, 6= strongly
D
painful, 7= very strongly painful, 8= immensely painful). When the partici-
TE
pant first scored a “7” or had reached the maximum current of 3 mA, the
stimuli then were applied again in reverse order and had to be reassessed.
This procedure was repeated twice, resulting in six sequences of pain stim-
uli. Detection threshold was defined as the mean value of the mA between
EP
a value of 0 and 1 on the rating scale. Pain threshold was defined as the
mean value of the mA values between a value of 3 and 4 on the rating
scale. This value was then doubled and employed to assess the placebo
C
effect in the experiment.6
For the entire sample (N=48), the detection threshold (Dt) was M= 0.37
C
mA (SD= 0.16), the pain threshold was (Pt) M= 0.83 mA (SD= 0.39), and the
A
painful stimulus that was employed in the experiment was (Pst) M= 1.67
mA (SD= 0.77). There were no significant differences between the groups..
Experimental design
Prior to the experiment all participants were informed that they would
have to execute standardized daily activities (4 exercises) twice, first as a
10
baseline condition and a second time after the repeated intake of the solu-
tion. They were also informed that painful stimuli would be applied as a
baseline condition and then before and after the intake of the solution for
the first and the second time. They also were told that half of the partici-
pants would be given the opioid, the other half would receive a placebo.
D
Then the patients had to draw a lot indicating which group they had been
TE
allocated to (independent variable “instruction”: “solution reduces
pain/improves physical capacity” / “solution is neutral, a placebo”). Half of
each group was additionally conditioned (independent variable “classical
conditioning”: yes/no) resulting in 4 subgroups with 12 participants in each

EP
(see Table 1 for the experimental design).
Conditioning was performed by reducing the intensity of the experimen-
tal electrical painful stimulus by 50% after the intake of the solution (ex-
C
perimental phase “post 1st solution – conditioning”), i.e. the participants
thus inadvertently experienced a reduction of the experimental pain stimu-

C
lus not knowing that this was manipulated by the experimenter.

A
Dependent variables/measurements
Assessment of clinical back pain
Before and after the interventions (exercise, intake of the solution) the
patients had to rate the intensity of their clinical back pain on an 11-point
11
numerical rating scale (NRS: 0= no pain, 10 = worst pain imaginable), yield-
ing a total of 8 ratings across all conditions (see Table 2).
Assessment of acute pain
The experimental pain stimuli were rated for intensity on an 11-point
NRS (0= no pain, 10 = worst pain imaginable) for a total of 6 times before
D
and after the intake of the solution and the exercise (experimental pain
TE
stimuli were only given at baseline, before and after the 1st and 2nd solution
and for manipulation (conditioning) of the 1st solution).
Assessment of pain behaviors and disability
Functional disability was assessed using a 12-item scoring system (Han-

EP
nover Activities of Daily Living questionnaire [ADL] 29 29 29 29 29 28 28 29, 37
.
The Hannover ADL instrument has excellent psychometric properties and it
has been shown to correlate highly with the Roland–Morris Scale.35, 36 The
C
ADL-questions are answered by using the response options ‘‘able to per-
form without difficulties’’, ‘‘able to perform with some difficulties’’, and

C
‘‘unable to perform or only with help’’. Raw scores were rescaled to values
A
between 0% and 100 % of functional capacity. The questionnaire was
completed before and after the experiment.
In addition, the patients had to perform 4 exercise which are also de-
scribed as items of the ADL (1. bending forward to pick up a small object
from the floor, 2. rising from a prone position, 3. putting on socks, 4. bend-
12
ing sideways from a seated position to pick up a small object on the floor
just beside the chair). Each exercise performance of the patient was re-
corded on camera before and after taking the solution. Both the experi-
menter and the patients rated the functional capacity of each performed
exercise according to the categories of the ADL.
D
Pain behavior was operationalized by using the time the patients
TE
needed to perform their exercise. This was measured by an independent
investigator who analyzed the videos and was blind to the group assign-
ment and the time points at which the exercise was being performed.
Assessment of expectations on the pain-reducing properties of the so-

EP
lution
Before the first and second administration of the solution, the patients’
expectations of the pain-reducing effect of the solution was measured by

C
the question “What do you expect will happen after taking the solution?”.
The answer was based on a 6-point verbal rating scale ranging from 1= no
C
pain reduction to 6=pain free. On the same scale, the patients’ expecta-
A
tions were measured again after the intake of the second solution and the
experimental testing of the painful experience. They were asked about
their expectations on prospective pain reduction by imaging a future intake
of the solution (“What do you expect will happen after taking the solution
in the future?”).
13
Time course of the experiment
The entire experiment lasted about 2 hours. The participants began with
the exercise and after a pause of 5 minutes they received 5 successive
painful stimuli to adapt to pain perception and again subsequent 5 succes-
D
sive pain stimuli to determine the baseline ratings (phase 1 “baseline”). The
TE
first solution was administered, followed by a time period of 5 minutes
rest declared as time for “medication release”. Then, 10 stimuli were
given. In the “conditioning groups”, the intensity of the painful stimulus
was reduced by 50%, which means that each of the 10 stimuli was given by
EP
half of the intensity as in phase 1 “baseline”. The non-conditioned partici-
pants received the exact same painful stimuli as in the first baseline. After
this procedure, a post- baseline assessment of 5 painful stimuli with the

C
same intensity as in the first baseline phase was administered to test the
results of the first solution. Then, 5 stimuli were given as a new baseline for
C
the second solution. Afterwards, the second solution was applied. Again,
A
the patients had 5 minutes rest for “medication release”. Finally, the ef-
fects of the second solution were tested by applying 5 painful stimuli., of
the same intensity as in the baseline phase. In this phase the conditioning
was not repeated, so we did not apply 10 stimuli again.
14
2.3.2 Statistical analysis
To reduce the data for statistical analysis, the mean scores of the self-
reported data of the acute pain stimuli in every block of 5 pain stimuli were
determined. For these as well as for the other outcome variables, we esti-
mated means and standard deviations to provide descriptive data for every
D
group. Preliminary assumption testing was conducted to check for normal-
TE
ity, linearity and sphericity. All data were normally distributed. In the case
of violation of sphericity, Greenhouse-Geisser corrections were used.
Primary intervention effects were estimated by employing generalized
linear modeling (GLM) with repeated measure analyses of variance

EP
(ANOVAs) containing the within-subject-factor “time-point” (back pain - 8
time points: pre and post 1st exercise, pre and post 1st solu-
tion/conditioning, pre and post 2nd exercise, pre and post 2nd solution;
C
acute pain - 6 time points: baseline, pre and post 1st solution/conditioning,
post 1st solution, pre and post 2nd solution); pain behavior - 2 time-points:
C
pre and post 1st exercise, no repetition was planned in the study design;
expectation – 3 time-points: pre 1st solution/conditioning, pre and post 2nd

A
solution) and the between-subject factor group (group “instruction –
‘opioid’”, group “instruction ‘placebo’”, group “instruction ‘opioid’ with
conditioning”, group “instruction – ‘placebo’ with conditioning”). We pre-
sent the results of the F-Tests with Greenhouse-Geisser corrections and an
15
adjusted model (only interaction effects). To determine the course of every
group, we calculated a repeated measures ANOVA and Cohen’s d to assess
effect sizes (first and last time point) for every group. To determine differ-
ences between the groups, we calculated the ANOVAs for every time point
and the mean differences and 95% confidence interval (CI), p-values and
D
Cohen’s d between the groups. All analyses and post hoc tests were Bon-
TE
ferroni-corrected.
Analyses were conducted using SPSS 22.0 (IBM Corporation, Armonk,
NY, USA) for Mac.

EP
3 Results
3.1 Clinical back pain
Table 2 presents the means and standard deviations of the back pain
C
ratings for every group and for each of the 8 time points of the experiment.
It also presents the F values and effect sizes (Cohen's d) calculated for
C
every group with regard to the difference between “pre 1st exercise” and
“post 2nd solution”.

A
The results revealed a significant main effect of “time-course” , F (3.12,
144.21) = 25.05, p < .001. Compared to the starting point of the experiment
after the 1st exercise, back pain ratings increased significantly (p < .01).
After the 1st solution intake, back pain ratings decreased significantly (p <
16
.02; p <.01) and again after 2nd solution intake (p < .01), in spite of the 2nd
exercise.
However, this main effect cannot be interpreted without considering
the significant overall interaction effect between “instruction”, “condi-
tioning”, and “time-course”, F (9.36, 143.23) = 8.29, p < .001. The patients’
D
back pain ratings differed as a consequence of the intervention “instruction
TE
* conditioning” (type of treatment group). Post hoc ANOVAs showed that
the groups differed significantly at all time points after conditioning and
first solution intake: pre 2nd exercise: F(3, 47) = 3.42, p= .025; post 2nd exer-
cise: F(3,47) = 6.27, p= .001; pre 2nd solution: F(3,47) = 3.77, p= .017; post
EP
2nd solution: F(3,47) = 7.29, p< .001 (cf. Figure 1A).
Significant reductions of clinical back pain were only present for the
groups that had received the opioid instruction (group “opioid-info / with-
C
out conditioning”: F (1,11) = 24.00, p < .01, d= .83; group “opioid-info / with
conditioning”: F (1,11) = 30.31, p < .01, d=1.83). The group “placebo instruc-
C
tion / without conditioning” by contrast showed a significant increase of

A
back pain (F (1,11) = 10.17, p < .01, d= - .64), the group “placebo instruction /
with conditioning” showed no significant change in pain (F (1,11) = 1.44, p <
.26, d= .32) over time (cf. Table 2).
Scores of anxiety, depression and functional capacity at the time point
before the experiment were used as a covariate to control the placebo ef-
17
fect (time-points “pre 1st exercise” vs. “post 2nd solution”) for individual
differences. There was no confounding influence of these variables on the
results (anxiety: F(1,43) = 0.91, p = 0.35; depression: F(1,43) = 1.82, p = 0.18;
functional capacity: F(1,43) = 0.84, p = 0.36).
Post hoc comparisons between the groups at time point “post 2nd solu-
D
tion” (see Supplement 2 and 3) revealed that the group “opioid-instruction
TE
and conditioning” displayed a significantly stronger pain reduction than
both of the placebo-informed groups. The group “opioid-instruction”
(without conditioning) did not display a significant pain reduction com-
pared to the group “Placebo-instruction and conditioning”.

EP
Pain increased in the placebo instructed group without conditioning
with a high effect size (cf. Table 2).
3.2. Acute pain

C
Table 3 shows the means and standard deviations of the acute pain rat-
ings for every group and each of the 6 acute pain time points as well as the
C
F values and effect sizes (Cohen's d) calculated for every group with regard
to the difference between “pre 1st exercise“ and “post 2nd solution”.
A
We found a significant main effect of “time-course” , F (1.99, 87.40) =
18.12, p < .01. Bonferroni-corrected contrasts showed that, compared to
baseline, acute pain ratings were significantly decreased after conditioning
and the 1st solution intake (p < .01), as well as after the 2nd solution intake
18
(p < .01). In addition, a significant interaction effect was found between
type of group and “time-course”, F (5.96, 87.40) = 4.40, p < .001, indicating
different group trends of the patients' acute pain ratings as a consequence
of the intervention “instruction and conditioning” (type of treatment
group), cf. Figure 1B. Post hoc ANOVAs showed that the groups differed
D
significantly at the time point “post 2nd solution”: F (3,44) = 4.24, p= .01
TE
after conditioning and intake of solution twice. Significant acute pain re-
duction over time was only achieved for both opioid informed groups: in
the group “opioid instruction without conditioning”: F (1,11) = 8.19, p= .02;
and the group “opioid instruction with conditioning: F (1,11) = 41.27, p <
EP
.01; see Table 3). At the time point “post 2nd solution” the groups “placebo
without conditioning” and “opioid with conditioning” differed significantly
from each other (cf. Supplement 2 and 3).

C
The effect sizes for the acute pain reductions in both “opioid instruc-
tion” groups were high (“without conditioning”: d=1.11, “with condition-

C
ing”: d= 1.79, see Table 3).

A
3.3. Pain Behavior: Time needed for exercise
Table 4 presents the means and standard deviations of the amount of
time the patients required for their 1st and 2nd exercise, carried out before
and after intake of the 1st solution (see Figure 1C).
19
We observed a significant main effect of “time-course” for pain behav-
ior F (1, 44) = 8.51, p < .01, indicating a significant decrease in the time
needed to complete exercise (cf. Figure 1C). A significant interaction ef-
fect between “instruction”, “conditioning”, and “time-course”, F (3, 44) =
3.15, p < .03 indicated that the patients' pain behavior differed over time
D
as a consequence of the intervention “instruction * conditioning” (type of
TE
treatment group). Except for group 3 (“placebo instruction without condi-
tioning”) all other groups (both opioid instructed groups and the placebo
instructed group with conditioning) showed a decrease in the amount of
time needed to complete the daily activities after intake of the solution,
EP
yielding significant reductions for group 1 (“opioid instruction – without
conditioning”: p < .04) and group 4 (“placebo instruction – conditioning”: p
< .03) . Effect sizes were small (see Table 4).

C
For the point of time “post exercise”, post hoc comparisons (see Sup-
plement 2) were conducted via one-way ANOVAs (Bonferroni-corrected).

C
When examining the time taken to complete the daily activities exercises
A
there only was a trend of significant differences (LSD = p< .05) between the
group “opioid” instruction with classical conditioning and “placebo” in-
struction without classical conditioning.
20
3.4 Pain Behavior: Self-ratings of functional capacity
Table 4 (cf. also Figure 1D) presents the means and standard deviations
of the patients' self-ratings of their functional capacity directly after their
1st and 2nd exercise, carried out before and after 1st solution intake.
D
There was a significant main effect of “time-course” on perceived func-
tional capacity, F(1, 44.00) 19.42, p < 0.001. In addition, a significant inter-
TE
action effect for “time-course * conditioning * instruction”, F (3.00, 44.00)
10.83, p < 0.001 indicated that the patients' self-ratings differed over time
as a consequence of the intervention “instruction * conditioning” (type of

EP
treatment group). With the exception of group 3 (“placebo instruction
without conditioning”), all groups exhibited an increase of functional ca-
pacity, which was significant for group 1 (“opioid instruction – without

C
conditioning”: p < .01) and group 2 (“opioid instruction – with condition-
ing”: p < .01), and failed to reach significance for group 4 (“placebo instruc-
C
tion with conditioning”) and group 3 (“placebo instruction – without con-
ditioning”) . The effect sizes were medium to high (cf. Table 4).
A
Analyses relating to associations between placebo effects on pain and
effects on physical functioning are outside the scope of the current paper
and will be reported elsewhere.
21
3.5 Expectation of pain reduction
Table 5 presents the means and standard deviations of the patients' ex-
pectations of the pain-reducing effects of the solution before they used
each of the solutions. After taking the second solution and the experimen-
tal testing of pain experience, the participants were asked about their ex-
D
pectation of prospective pain reduction caused by the solution after the
TE
end of the experiment.
There was only a significant effect of “instruction” F(1, 44) 92.03, p <
.001, indicating a difference for placebo” vs “opioid” at each time point. In
the group “opioid instruction – without conditioning” a significant decrease

EP
of expectation was found (F (1,11) 8.51 p< .01), however, the expectation
of pain reduction still prevailed. In the group “opioid instruction – with
conditioning”, expectations of pain reduction were maintained (F(1,11)

C
1.55, p= .24). In the group “placebo-instruction – without conditioning”, no
expectation for pain relief existed before and after the manipulation, how-
C
ever, in the group “placebo-instruction – with conditioning” there was a

A
significant increase of expectations after conditioning (F (1,11) 5.21, p< .04,
d=-.66).
4 Discussion
This experimental study examined the short term influence of a phar-
macologically neutral (placebo) solution shaped via instruction and condi-
22
tioning on the perception of clinical and acute pain, pain behavior, and
functional capacity in patients with chronic back pain. Furthermore, we
investigated the patients' expectations after receiving the instructions and
their intake of the solution, which seems to play an essential part in shap-
ing the placebo response.
D
Our study indicated that pain perception and pain behaviors are highly
TE
affected by the instructions and the learning experience the patients
gained during the experiment. First, the instruction “the solution you will
receive is an opioid, reduces pain/ improves physical capacity” produced a
higher placebo analgesia to clinical and acute pain and to pain behaviors
EP
than the instruction “the solution you will receive is neutral, a placebo”
with high effect sizes. The former instruction also produced higher expecta-
tions of pain reduction. Moreover, the “opioid-instructed“ group with addi-

C
tional conditioning showed more pronounced pain reductions in clinical
pain (Cohen’s d: 1.83 vs. .83), acute pain (Cohen’s d: acute pain: 1.79 vs.
C
1.11), the time to perform certain behaviors (Cohen’s d: .58 vs. .36), and
A
functional capacity (Cohen’s d: -.92 vs. -.59). After conditioning, a nearly
comparable placebo response developed in the placebo instructed group
in comparison to the non-conditioned “opioid instructed group“. The pla-
cebo-instructed group without conditioning showed no significant decrease
in time required for the exercise. In conclusion, this study demonstrates
23
that clinically relevant, albeit short term ,placebo effects were shaped via
instructions and conditioning in a sample of patients with chronic back
pain.
Our sample, even though small, consisted of patients characterized by
high back pain chronicity. Prior to the experiment, the patients' average
D
back pain ratings were at NRS 5 points (range 0-10). After the intervention,
TE
the opioid-instructed group with conditioning (pre NRS: 5.08) attained NRS
ratings of 1.92 points (Cohen’s d: 1.83) indicating a mean reduction of 3.16
points on an 11-point scale. Compared to other medication treatments for
back pain 28, this result points to high short-term efficacy of placebo for this
EP
chronic group.
Our study primarily targeted the motivational component of pain. We
were able to shape a placebo response that increased mobility in back pain
C
patients as measured during the experiment. This is an important point
because most of these patients tend to exhibit fear-avoidance behavior,41

C
following a former experience of pain triggered by a movement. Patients

A
may maintain this avoidance of mobility because they are convinced that
they will experience new and more pain when moving. This prevents them
from having a new experience. This maladaptive avoidance behavior often
causes subsequent musculoskeletal disorders and, is hence, a cause of new
pain.41 As a result, the treatment of back pain places special emphasis on
24
guiding patients into movement, allowing pain-free experiences to occur
and, in turn, fostering increased mobility. The instructions in our study
stated that the solution would enable participants to raise their levels of
pain-free mobility. As a result, the perception of physical functionality in-
creased and pain behaviors decreased. Hence, the management of instruc-
D
tions could be an important way to treat fear-avoidance behavior in pa-
TE
tients. At the same time, the same ethical principles as in other placebo
instructions must be followed. Based on the principles of classical condi-
tioning, patients with chronic back pain could, for example, be encouraged
to use analgesics when restarting exercising after a period of avoidance of

EP
movement to associate mobility with the absence of increased pain.
Even though we only examined short-term effects related to an experi-
mental setting and did not collect data over extended periods of time, our
C
results are important in the context of clinical applications of analgesic
placebo effects. The results support the claim that placebo research is
C
transferable to patients and that placebo effects they can be deliberately

A
applied. This can boost the efficacy of pain treatment, thereby meeting
demands for clinical application of placebo research.26 Our study with back
pain patients shows that placebo effects can be elicited in patients in ways
comparable to the results obtained in healthy participants (4, 12). By apply-
ing placebo mechanisms (instruction and conditioning) both of these
25
mechanisms, pain reducing expectations occurred which led to the percep-
tion of higher functional capacity and less pain behavior. These changes in
function and pain behavior are the result of a completely inactive sub-
stance being solely shaped by instructions and conditioning.
Our data suggest that this approach may add an additional placebo
D
component (placebo response) to the pharmacological component of anal-
TE
gesics. Comparisons have revealed that randomized controlled trials where
the subjects are certain that they receive active treatment tend to have up
to five times higher efficacy than studies where the subjects are informed
39
that they have only a 50% chance of obtaining verum Benedetti et al. 3
EP
and Colloca et al. 8, 11 suggested that the effect of analgesic medication has
both a psychological and a pharmacological component. In this sense, the
analgesic placebo effect can be considered to be ‘‘additive,’’ supplement-

C
ing pain management and enhancing analgesic medication beyond its
purely pharmacological effect, within ethical borders. In their open-hidden

C
paradigm, Benedetti et al.3 and Colloca et al.11 developed a model to use

A
the placebo response as an additive to highly effective analgesics. In the
open administration condition, participants could consciously perceive the
analgesics, whereas in the hidden condition they were unaware of receiv-
ing the drug. The pain reducing effect was up to one third higher in the
26
“open” administration of opioids and NSAIDS than in the “hidden” admini-
stration.1, 2
Bingel et al. 4also used the open-hidden model to investigate how diver-
gent expectancies alter the analgesic efficacy of a potent opioid in healthy
volunteer. By using an active opioid, they attained larger effect sizes com-
D
pared to our results: hidden application of remifentanil decreased pain
TE
intensity ratings on a numerical rating scale (0-100) from 66 (SD=2) to 55
(SD=3) points, with a medium effect size of d= 4.4 (11%), open application
with positive instruction resulted in a decrease of pain ratings from 66
(SD=2) to 39 (SD=3) points, with a very large effect size of d= 10.8; 27%,
EP
and open application with negative instruction increased pain ratings from
39 (SD=3) to 64 (SD=3) points, also with a very large effect size of d= -8.3
(25%). This underlines that the boosting of pharmacological effectiveness

C
via instruction is of great importance. It can improve the outcome of pain
therapy and should not be withheld from patients from an ethical perspec-
C
tive.
A
Another possibility for clinical use of these findings on placebo effects in
patients is to extend the effects of drug treatments by interspersing place-
bos. Colloca et al. 9 reviewed a database of placebo studies, searching for
studies indicating that placebos given after repeated administration of ac-
tive treatments acquire medication-like effects. This is a special case of
27
intermittent reinforcement of the inactive placebo substance because it
followed the analgesic, thereby coupling it with an active (pharmacological)
substance.
Limitations
One limitation of our study is that differential pain medication history
D
was not included. We only assessed medication intake of the last week and
TE
on the day of the experiment. On the day of the experiment, 22.92% of the
sample took opioids, and 31.25% took opioids during the last week. We
didn’t control for opioid intake in our randomisation.
Differentiating learning history and attitudes towards opioids would

EP
have helped to investigate individual placebo effects. Reactivating prior
experiences can have both positive and negative consequences, for exam-
ple, positive associations can channel the experience with a new analgesic
C
or a new placebo into a positive direction. Verbal instructions can be em-
ployed to have patients recall and reactivate previously learned associa-

C
tions that are otherwise implicit and not accessible by verbal report.25 New
A
medication experiences occur on the basis of this learning history and they
are influenced by it.24 Colloca and Benedetti 8

found that the prior experi-
ence of a beneficial effect of a drug led to a higher placebo response than
the experience of a drug being ineffective, and these effects lasted for sev-
28
eral days. Our sample was too small to draw any robust conclusions on the
influence of prior opioid intake and placebo effects.
Another limitation is that our results reflect only short-term effects re-
lated to the experimental setting. Clinical applications of analgesic placebo
effects require long-term studies to assess more extended effects.
D
Outlook
TE
This study supports the practical value of the use of placebo mecha-
nisms in a clinical context. However, placebo analgesia does not need de-
ception. There are many ways within ethical borders8, 9, 15, 26 to exploit pla-
cebo effects. By using the targeted application of instruction and condition-

EP
ing, thus shaping expectancies towards pain reduction, the patients experi-
enced less pain and were able to increase their mobility. The goal would be
to integrate the placebo response in pain management to optimize pain

C
treatment outcomes, not only pharmacological pain treatment. Following
the concept of open medication, it is advisable to enhance expectations,

C
for example, by placing emphasis on positive drug effects, avoiding an

A
overemphasis on side effects, and to enhance learning components by ap-
plying analgesics in an open manner and to include many sensory aspects
to increase analgesic effectiveness .26 Provision of well-balanced informa-
tion about analgesics in a valued and trusting atmosphere to facilitate pa-
tient-physician communication may increase efficacy. These principles
29
should be included in medical education to facilitate use of placebo effects.
More research in clinical settings is needed to confirm this framework.
Acknowledgements
This research was funded by grants by the Deutsche Forschungsgemein-
D
schaft DFG (FOR 1328/1) to R.K. (Kl 1350/3-1) and H.F. (Fl 156/33-1) and
TE
SFB1158/B03 to HF.
Conflict of interests: No conflicts of interest to be claimed.
References:
EP
1. Amanzio M, Pollo, A., Maggi, G., & Benedetti, F. Response
variability to analgesics: A role for non-specific activation of
endogenous opioids. J Pain. 90:205-215, 2001
2. Benedetti F, Amanzio, M., & Maggi G. Potentiation of placebo
analgesia by proglumide. Lancet. 346(8984):1231, 1995
3. Benedetti F, Pollo, A., Lopiano, L., Lanotte, M., Vighetti, S., &
C
Rainero, I. Conscious expectation and unconscious conditioning in

analgesic, motor, and hormonal placebo/nocebo responses. J
Neurosci. 23:4315–4323, 2003
4. Bingel U, Colloca, L., & Vase, L. Mechanisms and clinical
C
implications of the placebo effect: Is there a potential for the elderly?

A mini-review. Gerontology. 57:354–363, 2011
5. Bromm B: Pain and the brain: From nociception to cognition, Raven
Press, New York, 1995.
6. Bromm B, & Meier, W. The intracutaneous stimulus: A new pain
A
model for algesimetric studies. Methods Find Exp Clin Pharmacol.

6:405 - 410, 1984
7. Charron J, Rainville, P., & Marchand, S. Direct comparison of
placebo effects on clinical and experimental pain. Clin J Pain.
22:204-211, 2006
8. Colloca L, & Benedetti, F. Placebos and painkillers: Is mind as real
as matter? Nat Rev Neurosc. 6:545–552, 2005
9. Colloca L, Enck, P., & De Grazia, D. Relieving pain using dose-
extending placebos: A scoping review. J Pain. 157:1590-1598, 2016
30
10. Colloca L, Klinger, R., Flor, H., & Bingel, U. Placebo analgesia:
Psychological and neurobiological mechanisms. J Pain. 154:511,
2013
11. Colloca L, Lopiano, L., Lanotte, M., & Benedetti, F. Overt versus
covert treatment for pain, anxiety, and Parkinson’s disease. Lancet
Neurol. 3:679–684, 2004
12. Elsenbruch S, Kotsis, V., Benson, S., Rosenberger, C., Reidick, D.,
Schedlowski, M., Bingel, U., Theysohn, N., Forsting, M. & Gizewski,
E. R. Neural mechanisms mediating the effects of expectation in
visceral placebo analgesia: An fMRI study in healthy placebo
D
responders and nonresponders. J Pain. 153:382-390, 2012
13. Enck P, Bingel, U., Schedlowski, M., & Rief, W. The placebo
response in medicine: Minimize, maximize or personalize? Nature
reviews Drug discovery. 12:191-204, 2013
TE
14. Faul F, Erdfelder, E., Lang, A. & Buchner, A. *Power 3: A flexible
statistical power analysis program for the social, behavioral, and
biomedical sciences. Behav Res Meth. 39:175-191, 2007
15. Finniss DG, Kaptchuk, T. J., Miller, F., & Benedetti, F. Biological,
clinical, and ethical advances of placebo effects. Lancet. 375:686-
695, 2010
16. Flor H, Haag, G., & Turk, D. C. Long-term efficacy of EMG
EP
biofeedback for chronic rheumatic back pain. J Pain. 27:195-202,
1986
17. Flor H, Haag, G., Turk, D. C., & Köhler, H. Efficacy of EMG
biofeedback, pseudotherapy, and conventional medical treatment
for chronic rheumatic back pain. J Pain. 17:21-31, 1983
18. Haake M, Müller, H-H., Schade-Brittinger, C., Basler, H. D.,
Schäfer, H., Maier, C., Endres, H., Trampisch, H. J., & Molsberger,
A. German Acupuncture Trials (GERAC) for chronic low back pain:
C
Randomized, multicenter, blinded, parallel-group trial with 3 groups.

Arch Intern Med. 167:1892-1898, 2007
19. Haake M, Müller, HH., Schade-Brittinger, C., Prinz, H., Basler, H.
D., Streitberger, K., Schäfer, H., & Molsberger, A. The German
C
multicenter, randomized, partially blinded, prospective trial of

acupuncture for chronic low-back pain: a preliminary report on the
rationale and design of the trial. J Altern Complement Med. 9:763-
770, 2003
A
20. Hautzinger M, Bailer M: Allgemeine Depressions Skala (ADS).

Deutsche Form der “Center for Epidemiologic Studies Depression
Scale (CES-D). [General Depression Scale (ADS). German version
of the Center for Epidemiologigic Studies Depression Scale].”, Beltz
Test GmbH, Göttingen, 1993.
21. Kam-Hansen S, Jakubowski, M., Kelley, J. M., Kirsch, I., Hoaglin, D.
C., Kap-tchuk, T. J., & Burstein, R. Labeling of medication and
placebo alters the outcome of episodic migraine attacks. Sci Transl
Med. 6:218ra215, 2014
22. Kaptchuk TJ, Kelley JM, Conboy LA, Davis RB, Kerr CE, Jacobson
EE, Kirsch I, Schyner RN, Nam BH, Nguyen LT, Park M, Rivers AL,
31
McManus C, Kokkotou E, Drossman DA, Goldman P, Lembo AJ.
Components of placebo effect: randomised controlled trial in
patients with irritable bowel syndrome. BMJ. 336:999-1003, 2008
23. Kelley JM, Lembo, A. J., & Ablon, J. S. Patient and practitioner
influences on the placebo effect in irritable bowel syndrome.
Psychosom Med. 71:789-797, 2009
24. Kessner S, Wiech, K., Forkmann, K., Ploner, M., & Bingel, U. The
effect of treatment history on therapeutic outcome: An experimental
approach. JAMA Intern Med. 173:1468–1469, 2013
25. Kirsch I, Lynn, S. J., Vigorito, M., & Miller, R. R. The role of
D
cognition in classical and operant conditioning. J Clin Psychol.
60:369–392, 2004
26. Klinger R, Colloca, L., Bingel, U., & Flor, H. Placebo analgesia:
Clinical applications. Pain. 155:1055, 2014
TE
27. Klinger R, Soost, S., Flor, H., & Worm, M. Classical conditioning
and expectancy in placebo hypoalgesia: A randomized controlled
study in patients with atopic dermatitis and persons with healthy
skin. J Pain. 128:31-39, 2007
28. Koes BW, van Tulder, M., Lin, C. W. C., Macedo, L. G., & McAuley,
J., Maher, C. An updated overview of clinical guidelines for the
management of non-specific low back pain in primary care. Eur
EP
Spine J. 19:2075–2094 2010
29. Kohlmann T, & Raspe, H. Der Funktionsfragebogen Hannover zur
alltagsnahen Diagnostik der Funktionsbeeinträchtigung durch
Rückenschmerzen (FFbH-R) [Hannover functional questionnaire in
ambulatory diagnosis of functional disability caused by back-ache].
Die Rehabilitation. 35:I-VIII, 1996
30. Laux L, Spielberger CD: Das State-Trait-Angstinventar: STAI. [The
State-Trait Anxiety Inventory: STAI]. Beltz Test, Göttingen, 2001.
C
31. Peerdeman KJ, van Laarhoven, A. I., Keij, S. M., Vase, L., Rovers,
M. M., Peters, M. L., & Evers, A. W. Relieving patients' pain with
expectation interventions: A meta-analysis. J Pain. 157:1179-1191,
2016
C
32. Petersen GL, Finnerup, N. B., Nørskov, K. N., Grosen, K.,

Pilegaard, H. K., Benedetti, F., Price, D. D., Jensen, T. S., & Vase,
L. Placebo manipulations reduce hyperalgesia in neuropathic pain.
J Pain. 153:1292-1300, 2012
A
33. Pollo A, Amanzio, M., Arslanian, A., Casadio, C., Maggi, G., &
Benedetti, F. Response expectancies in placebo analgesia and their
clinical relevance. J Pain. 93:77-84, 2001
34. Price DD, Milling, L. S., Kirsch, I., Duff, A., Montgomery, G. H., &
Nicholls, S. S. An analysis of factors that contribute to the
magnitude of placebo analgesia in an experimental paradigm. J
Pain. 83:147–156, 1999
35. Roese I, Kohlmann, T., & Raspe, H. Measuring functional capacity
in backache patients in rehabilitation: A comparison of standardized
ques-tionnaires. Rehabilitation. 35:103-108, 1996
32
36. Roland M, & Morris, R. A study of the natural history of back pain.
Part I: Development of a reliable and sensitive measure of disability
in low-back pain. Spine. 8:141-144, 1983
37. Silman A, Hochberg M: Epidemiology of the rheumatic diseases. In:
Back pain.(Raspe, H., Ed.), Oxford University press, Oxford, 2001,
pp. 309-338.
38. Tretrop J: Der Placeboeffekt bei chronischen
Rückenschmerzpatienten: eine klinisch-experimentelle Studie.
Dissertation. , University of Hamburg, 2007.
39. Vase L, Petersen, G. L., Riley, J. L., & Price, D. D. Factors
D
contributing to large analgesic effects in placebo mechanism studies
conducted between 2002 and 2007. J Pain. 145:36-44, 2009
40. Vase L, Robinson ME, Verne GN, Price DD. The contributions of
suggestion, desire, and expectation to placebo effects in irritable
TE
bowel syndrome patients. An empirical investigation. Pain. 105:17-
25, 2003
41. Vlaeyen JW, & Linton, S. J. Fear‐avoidance model of chronic
musculoskeletal pain: 12 years on. J Pain. 153:1144-1147, 2012

EP
42. Voudouris NJ, Peck, C. L., & Coleman, G. S. The role of
conditioning and verbal expectancy in the placebo response. J Pain.
43:121-128, 1990
C
Figure legend
C
Figure 1
Patients‘ pain ratings and pain behaviors: (A) acute pain ratings (B) pain
behavior and (C) self-ratings of functional capacity (D) as a function of in-

A
struction, conditioning, and pre and post intake of solutions. Means, stan-
dard deviations, effect sizes are presented in the corresponding Tables 2 –
4.
33
Table 1
1)
Baseline demographic, clinical and pain characteristics
Opioid Instruction – Opioid instruction – Placebo instruction – Placebo instruction –

no conditioning conditioning no conditioning conditioning
(n=12) (n=12) (n=12) (n=12)
Demographic characteristics
Age (years) 49.99 (13.26) 53.04 (13.24) 50.83 (17.01) 50.33 (15.17)
Sex: Male 3 (6.25%) 3 (6.25%) 3 (6.25%) 3 (6.25%)
Sex: Female 9 (18.75%) 9 (18.75%) 9 (18.75%) 9 (18.75%)
D
2)
BMI 24.33 (4.08) 25.52 (4.57) 27.22 (6.22) 24.27 (4.67)
Clinical characteristics
Back pain duration in

119.33 (75.23) 158.83 (138.96) 165.67 (117.17) 162.17 (106.86)
TE
total (month)
Back pain duration of

current intensity 2.92 (0.29) 2.83 (0.39) 2.75 (0.45) 2.92 (0.29)
(month)
Disability through by
4.33 (3.23) 4.25 (2.05) 5.25 (1.49) 5.83 (3.16)
back pain
EP
Back pain intensity
5.58 (2.47) 6.42 (1.62) 5.67 (1.23) 6.83 (1.95)
(average last 6 month)
Back pain intensity

8.67 (1.44) 7.92 (1.68) 8.67 (0.99) 9.00 (1.13)
(strongest pain last 6 month)
Opioids today yes 2 4 2 3

Opioids today no 10 8 10 9
Non-opioids today yes 6 4 1 5

C
Non-pioids today no 6 8 11 7
Opioids last week yes 3 4 3 5

Opioids last week no 9 8 9 7
C
Non-opioids last week yes 11 9 12 9

Non-opioids last week no 1 3 0 3
Experimental characteristics (Milliampere mA) experimental stimmt nicht ganz würde es pain characteristics nennen pain stimulus ist unklar meinst du
A
stimulus strength applied during the experiment dann so sagen zB in der Fussnote
Detection threshold 0.39 (0.17) 0.39 (0.14) 0.32 (0.17) 0.39 (0.16)
Pain threshold 0.90 (0.34) 0.79 (0.22) 0.72 (0.39) 0.92 (0.54)
Pain stimulus 1.80 (0.68) 1.58 (0.44) 1.43 (0.77) 1.84 (1.08)
Pain stimulus, condi.3) - 0.79 (0.22) - 0.92 (0.54)
1)
Data are mean values (SD), n (%) or n. None of the variables revealed significant differences between groups.
2)
BMI = Body Mass Index.
3)
Pain stimulus, condi. = Conditional pain stimulus during phase of conditioning; individual pain stimuli were reduced by 50% in both conditional groups. The
non-conditional groups received their original pain stimulus (100%).
Table 2
Patients' back pain ratings: Mean values, standard deviations, and effect sizes of the placebo responses according to the different instructions and conditioning conditions coupled with the solution
Experimental manipulation 1st exercise 1st solution and conditioning 2nd exercise 2nd solution
pre post pre post pre post pre post Difference
st 1) st st 2) st nd 4) nd nd 5) nd 6) st
1 exercise 1 exercise 1 solution 1 solution - 2 exercise 2 exercise 2 solution 2 solution pre 1 exercise -
3) nd
conditoning post 2 solution
7) 8)
Instruction Conditioning Group M SD M SD M SD M SD M SD M SD M SD M SD F (1,11) p d
D
Opioid No 1 5.00 2.09 5.75 2.30 5.42 2.19 4.00 2.13 3.75 2.34 3.67 2.54 3.58 2.71 3.00 2.73 24.00 < .01 0.83
Yes 2 5.08 1.73 5.75 2.30 5.33 1.92 3.58 1.56 2.83 1.59 2.75 1.66 3.08 1.88 1.92 1.73 30.31 < .01 1.83
Placebo No 3 4.67 1.67 5.67 1.72 5.00 1.86 5.00 1.76 5.25 1.82 6.08 1.93 5.58 1.78 5.83 1.95 10.17 < .01 - 0.64
TE
Yes 4 5.25 1.87 6.00 1.48 5.33 1.61 5.25 1.55 4.50 1.93 5.17 2.04 5.00 1.86 4.58 2.31 1.44 < .26 0.32.
Differences between groups/time F (3,44)= .21 F (3,44)= .06 F (3,44)= .11 F (3,44)= 2.43 F (3,44) = 3.42 F (3,44) = 6.27 F (3,44) = 3.77 F (3,44) = 7.29
point p = .89 p =.98 p =.95 p =.08 p =.03 p=.001 p =.02 p = < .001
1)
Baseline. Patients' rating of back pain, numerical rating scale from 0 (no pain) to 10 (worst pain imaginable).
2)
After resting and before 1st solution intake.
3)
After receiving the 1st solution (opioid/ placebo briefings) and experience (conditioning/no-conditioning).
EP
4)
After resting and before 2nd exercise.
5)
After resting and before 2nd solution intake.
6)
After receiving the 2nd solution (opioid/ placebo briefings).
7)
Analysis of variance: pre 1st exercise - post 2nd solution.
8)
Cohen’s d: difference between time point before 1st exercise and time point after 2nd solution intake.
C
C
A
Table 3
Patients' acute pain ratings: Mean values and standard deviations of the placebo responses relating to the different instructions and conditioning conditions associated with the solution
1st solution and conditioning 2nd solution

pre post post pre post Difference
st st st nd 3) nd 4) st
Baseline 1 solution 1 solution - 1 solution - 2 solution 2 solution 1 baseline -
1) nd
conditoning after post 2 solution
2)
conditoning
D
excluded in VA
5) 6)
Instruction Conditioning Group M SD M SD M SD M SD M SD M SD F (1,11) sign d
Opioid No 1 6.95 1.20 6.88 1.11 6.32 1.38 5.63 1.76 5.72 1.51 5.09 2.16 8.19 .02 1.11
TE
Yes 2 6.83 0.86 6.93 0.96 2.95 2.01 5.30 1.75 5.92 1.34 4.62 1.62 41.27 < .01 1.79
Placebo No 3 6.67 0.90 6.88 1.05 6.81 1.09 6.53 0.97 6.82 1.05 6.95 0.98 1.57 .24 - 0.30
Yes 4 6.12 1.48 6.35 1.33 3.14 1.50 5.97 1.66 5.98 1.82 5.66 1.82 0.91 .36 0.27
Differences between groups/time F (3,44) = 1.26 F (3,44) = 0.73 F (3,44) = 14.42 F (3,44) = 1.48 F (3,44) = 1.34 F (3,44) = 4.24
point
EP
1)
1st solution intake had occurred followed by conditioning or no conditioning: Both conditioned groups received stimuli of only 50% intensity of their individual pain stimulus. This phase was not included in the
analyses of variance because the reduced experience of pain was caused by reduced experimental pain stimulation and not by subjective experience.
2)
After 1st solution.
3)
Before 2nd solution.
4)
After 2nd solution.
5)
Significance.
6)
Cohen’s d: Difference between 1st baseline – time point after 2nd solution.
C
C
A
Table 4
Patients' pain behavior: Mean time needed for the exercise and self-ratings of functional capacity (0-100%)
Time needed for exercise Self-ratings of functional capacity
pre post pre post

st st 1) st st 1)
1 solution 1 solution Difference 1 solution 1 solution Difference
D
2) 2)
Instruction Conditioning Group M SD M SD F (1,11) p d M SD M SD F (1,11) p d
Opioid No 1 43.58 32.49 35.33 24.35 5.44 .04 0.36 52.52 22.89 67.78 29.24 25.07 < .01 -0.59
Yes 2 31.02 12.88 24.93 08.00 3.81 .07 0.58 60.56 20.78 77.22 15.43 17.55 < .01 -0.92
TE
Placebo No 3 37.10 13.04 39.74 18.14 1.97 .19 -0.17 51.11 15.13 44.44 15.66 4.40 = .06 0.43
Yes 4 39.70 19.85 35.92 16.46 6.13 .03 0.21 50.00 20.99 53.89 24.03 1.67 = .22 -0.17
F (3,44) = 0.75 F (3,44) = 1.54 F (3,44) = 0.68 F (3,44) = 5.29
p = .53 p = .22 p = .57 p = .003
1)
After 1st solution intake, followed by conditioning or no conditioning.
EP
2)
Cohen’s d: difference between pre – post.
C
C
A
Table 5
Patients' expectation of the pain-reducing effects of the solution
Expectation
pre - post - post - Difference

st 1) st nd 3)
1 solution 1 solution / 2 solution
D
nd 2)
pre 2 solution
4)
Instruction Conditioning Group M SD M SD M SD F (1,11) p d
Opioid No 1 4.58 1.51 3.58 1.62 3.33 1.72 8.51 .01 .77
TE
Yes 2 4.25 1.49 3.58 1.08 3.67 .89 1.55 .24 .74
Placebo No 3 1.00 .00 1.00 .00 1.00 .00 - - -
Yes 4 1.00 .00 1.83 1.12 1.75 1.14 5.21 .04 -.66
Differences between groups/time point F (3,44) = 41.99 F (3,44) = 15.98 F (3,44) = 15.37
p = .001 p = .001 p = .001
EP
1)
Before 1st solution intake: Expectation of pain reduction.
2)
After 1st solution intake, followed by conditioning or no conditioning: Expectation of pain reduction after 2nd solution.
3)
After 2nd solution intake: Expectation of pain reduction of prospective pain reduction after end of experiment.
4)
Cohen’s d: difference between pre 1st solution - post 2nd solution.
C
C
A
Placebo effect on back pain Placebo effect on acute pain ratings
10 Patients' acute pain ratings (NRS 0-10) 10
Patients' back pain rating (0-10)
7.5 7.5
D
5 5
TE
2.5
2.5
1st solution / 2nd
1st solution / 2nd solution
conditioning solution
conditioning
EP
0
A B
Pain behavior Placebo effect on pain behavior:

(time needed for exercise) Self-ratings of functional capacity 0-100%
50
C 100
C
Functional capacity 0-100 %
40 75
A
30
50
Time (seconds)
20
25
10
0 0
C D

Klinger2017 - Placebo Effects LBP

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Klinger2017 - Placebo Effects LBP

Uploaded by

Copyright:

Available Formats

PAIN Publish Ahead of Print

Placebo effects of a sham opioid solution: a randomized controlled study

Regine Klinger a *, Ralph Kothe b, Julia Schmitz a, Sandra Kamping, a, Herta

Corresponding author: Regine Klinger, PhD, University Medical Center

Medicine, Department of Anesthesiology, Pain Therapy and Pain Psychol-

gesia responses into pain management.

The good evidence for the efficacy of analgesic placebo effects in

and related side-effects in these patients.

cant placebo effects.31,39,18,19 Only a few studies also examined underlying

placebo mechanisms such as experiential (e.g. partial reinforcement and

classical conditioning), social (e.g. vicarious learning and modeling) and

participants 27 and no studies have examined the effects of analgesic pla-

cebo mechanisms on both clinical pain and disease-related function.

Until now, there have been no studies conducted on patients with

solutions. However, studies of this kind would be of great interest because

many neurobiological studies have documented great similarities between

placebo component of an opioid.1, 4, 10, 33

the difference in the placebo-related activation of endogenous opioid sys-

bly triggered.26 It is, therefore, essential to know the underlying core

mechanisms: expectancy34, 40 and learning processes such as classical con-

solution can reduce the perceived intensity of both an experimental pain

effect in a short-term (2 hours) experiment studying 48 patients with

pharmacologically neutral solution, which was introduced either as a pain-

ment, or as a neutral solution, a placebo. The central questions were: (1)

Can an analgesic placebo response (on experimental acute as well as clini-

cal pain) be attained in chronic back pain patients by manipulating an inert

tients who received the supposed ”opioid solution“ would exhibit a

tion was a placebo.

cruited from local multidisciplinary pain centers, back pain management

centers, and the Outpatient Clinic of Behavior Therapy at the Department

of Psychology, University of Hamburg, Germany. We estimated that a sam-

attain 80% power and a 5% significance level against the backdrop of a

previous study in which similar outcome measures (numerical rating scales

The sample included 36 (9 per group) women and 12 (3 per group)

men with a mean age of 49.97 years (SD = 13.64).

Inclusion criteria were: chronic, persistent or recurrent (≥ 1 per week)

nosed by the referring physician who excluded patients exhibiting “red

weight loss and widespread, neurological changes.28 The patients were

also excluded if they had inflammatory, neuropathic or tumor-related back

syndrome, rheumatic disorders, cancer, severe heart diseases, which

were likewise assessed by the referring physicians. Mental health problems

were diagnosed by psychotherapists employing a psychological assessment

(interview) and additional questionnaire (depression: General Depression

persons with severe acute or chronic mental disorders in accordance with

the International Classification of Diseases (ICD). However, patients with

anxiety or depressive disorders secondary to the back pain diagnosis were

included. Further exclusion criteria were: Cognitive impairment, preg-

nancy, symptoms or diseases which prohibit the application of electrical

impulses (e.g. cardiac insufficiency and/or need of cardiac pacemakers or

On the day of the experiment the following long-term medication was

taken: 11 patients (22,92%; see Table 1) had taken opioids (hydromor-

phon, tramadol, tilidin phosphate/hydrochloride, morphine sulfate, fen-

clofenac, acetylsalicylic acid, metamizole, gabapentin). During the week

before the experiment 15 patients (31,25%) took opioids (long-term medi-

cation). Patients with medication overuse or impairments due to the

The patients were randomly assigned to the four experimental groups

(1. instruction “solution is neutral, a placebo” / no conditioning, 2. instruc-

tion solution is neutral, a placebo” / conditioning, 3. instruction “solution is

an opioid, reduces pain/ improves physical capacity” / no conditioning, 4.

instruction “solution is an opioid, reduces pain/ improves physical capac-

ity” / conditioning). The study adhered to the guidelines of the Declaration

tered (German Clinical Trials Register: DRKS00011201).

For the experiment, an intracutaneous electrical painful stimulus was