European Journal of Pain (1998) 2: 25-34

Effects of an opioid antagonist on pain intensity

and withdrawal reflexes during induction of
hypnotic analgesia in high- and low-hypnotizable
volunteers
R. Zachariae”, 0. K. Andersenb, P. Bjerring”, M. M. Jargensen” and
L. Arendt-Nielsenb

alnstitute of Psychology, University of Aarhus, Denmark

bCenter for Sensory- Motor Interaction, Aalborg University, Aalborg, Denmark
“Department of Dermatology, Aarhus University Hospital, Aarhus, Denqark

The aim of this investigation was to study the effect of suggestions of hypnotic analgesia on spinal pain
transmission and processing. Pain intensity and amplitude of nociceptive withdrawal reflexes to electrical
stimuli were measured in 10 high- and 10 low-hypnotizable subjects during two sessions taking place at least
24 h apart under five conditions of: (1) pre-hypnosis; (2) neutral hypnotic relaxation; (3) suggestions of hypnotic
analgesia; (4) suggestions of hypnotic analgesia after injections of either naloxone (1 ml, 1 mg/ml) or saline
(1 ml) under double-blinded conditions; and (5) post-hypnosis. The conditions of naloxone or saline were
allocated at random to either Day 1 or Day 2 in a double-blinded fashion. Results showed significant reductions
of pain intensity during hypnotic analgesia, and a significant reduction in nociceptive reflexes during hypnotic
analgesia on Day 1 in the highly hypnotizable group. No differences were found for low-hypnotizable subjects.
The results support previous findings that pain intensity as well as the nociceptive reflex can be modulated by
suggestions of hypnotic analgesia. While no effect of naloxone on pain intensity was found during hypnotic
analgesia, naloxone significantly reversed the suppressive effect of suggestions of hypnotic analgesia on reflexes
in high-hypnotizable subjects. Subsequent analysis showed that the effect of naloxone was associated with the
intensity of the stimulus needed to elicit a reflex, and was unrelated to hypnotic susceptibility when controlling
for stimulus intensity. These results suggest that the effect of naloxone was related to the greater stimulus
intensities needed to elicit a reflex in the high-hypnotizable group, rather than to hypnosis or hypnotic
susceptibility in itself. It is unclear why greater stimulus intensities were needed in high-hypnotizable subjects
and further studies are needed.

INTRODUCTION evoked by painful stimulation, can be modified

by suggestions of hypnotic analgesia (Arendt-
Numerous investigations and case reports have Nielsen et aI., 1990; Zachariae & Bjerring, 1990;
shown that hypnotic suggestions of analgesia are 1994; De Pascalis et al., 1996), and it has recently
able to reduce both experimental and clinical been shown that hypnotic analgesia corresponds
pain (Ome, 1983; Hilgard, 1986; Dahlgren et al., with localized increases in cerebral blood flow
1995). Studies have also indicated that physio- in high- but not in low-hypnotizable subjects
logical correlates of pain, such as brain potentials (Crawford et al., 1993). These and other findings
indicate that the pain reduction experienced by
Paper received 17 August 1997 and accepted in revised form high-hypnotizable subjects cannot, as has been
4 December 1997.
Correspondence to: R. Zachariae, Institute of Psychology, suggested (Spanos et al., 1985-86; Spanos et al.,
Aarhus University, Asylvej 4, DK8240 Risskov, Denmark. 1988), be explained exclusively in terms of social

1090-3801/98/010025+ 10 $12.00/O
0 1998 European Federation of Chapters of the International Association for the Study of Pain
26 R. ZACHARIAE ETAL.

role enactments and responding to contextual be modulated by severalfactors, including muscle

demands. Although results of recent studies sug- tension (Desmedt & Godaux, 1976), attention
gest that attentional processesmay be important (Bathien & Hugelin, 1969;Arendt-Nielsen et al.,
in the experienceof hypnosis and hypnotic phe- 1991) and stress (Willer et al., 1981), and any
nomena (Crawford, 1994),the possiblecognitive, attempt to determine whether the ability to mod-
neurophysiological and biochemical mechanisms ulate pain is specifically related to hypnotically
underlying hypnotic analgesia are still generally induced analgesia must therefore include high
unknown. hypnotically responsivesubjectsas well ascontrol
Results from animal studies have indicated the subjectswith low-hypnotic susceptibility. Only if
existenceof several analgesic mechanisms, some it can be shown that the responses of high-
of which involve processing at the spinal level hypnotizable subjects differ substantially from
and may be mediated by endogenous opioids. the responsesof low-hypnotizable subjects, may
Others may involve processingat a higher cortical one conclude that the observed results may be
level and seem to be opioid independent (Frenk specifically related to the process of hypnosis.
et al., 1986; Hamon et al., 1988). It has been One possibility is that high-hypnotizable subjects
shown that while certain forms of stress-induced would be able to reduce their spinal reflexes,
analgesiaare reversibleby injections of the opioid while low-hypnotizable subjects would not. One
antagonist naloxone, others are not (Willer et al., could also speculatethat the possible reductions
1981;Cannon et al., 1983).It has been speculated of spinal reflexeswere influenced by opioid-me-
that hypnotic analgesia could be mediated by diated mechanisms in high- but not in low-hyp-
opioid-dependent mechanisms, but all previous notizable subjects. Since a possible effect of
investigations have shown that hypnotic anal- hypnosis on the reflexes could be related to the
gesia, as measured by the reduction of reported relaxed condition of hypnosis itself rather than
pain intensity, is not reversible by the opiate the specific suggestions of altering pain per-
antagonist naloxone (Goldstein & Hilgard, 1975; ception, a neutral hypnotic control condition
Barber & Mayer, 1977;Moret et al., 1991).These omitting specific suggestions should also be in-
results could suggest that hypnotic analgesia cluded.
primarily involves processing at the higher cor- On this background, the aim of the present
tical rather than at the spinal level. investigation was: (1) to study the effect of hyp-
It hasbeen reported, however,that suggestions notic suggestion of analgesia on electrically in-
of hypnotic analgesiamay inhibit the nociceptive duced pain intensity and spinal nociceptive
reflexes; (2) to investigate whether a possible
withdrawal reflex induced by needle pricks on
the foot (Finer, 1970).This reflex has beenshown effect was opioid dependent by including a con-
dition of hypnotic analgesiaafter an injection of
to be inhibited by morphine (Willer et al., 1985);
either naloxone, an opioid antagonist, or placebo;
an inhibition which is thought to take place
and (3) to investigate whether the effect could be
primarily at the spinal level. An attempt to mod-
specifically related to hypnotic analgesia by: (i)
ulate this spinal nociceptive reflex by suggestions including high-hypnotizable as well as low-hyp-
of hypnotic analgesia therefore seems to be a notizable subjects in the investigation, and (ii)
potentially useful model for investigating possible including a relaxed neutral hypnotic control con-
spinal mechanisms involved in this process. A dition.
recent study has been able to show reductions
of this reflex following suggestionsof hypnotic
analgesia (Kiernan et al., 1995). Unfortunately, MATERIALS AND METHODS
the researchersdid not report the hypnotic sus-
ceptibility level of the individual subjectsincluded Subjects
in the study. This makes it difficult to determine
Selection of volunteers
to which degree the reductions of the reflexes
were specifically related to the suggestions of A group of 103 subjects (37 males, 69 females,
hypnotic analgesia. The spinal reflex may in fact mean age: 31.3k 9.9 years) volunteered to have

European Journal of Pain (1998). 2

HYPNOTIC MODULATION OF SPINAL PAIN TRANSMISSION 27

their hypnotizability tested using a Danish ad- amplified (20.000 or 50.000 times), sampled
aptation of the Harvard Group Scaleof Hypnotic (2 kHz), monitored on a computer storage os-
Susceptibility, Form A (Shor & Orne, 1962; Za- cilloscope, and stored on disk for later analysis.
chariae et al., 1996). This scale measures hyp- EMG signals for 1 s were stored from 200 ms
notizability on a scale from 0 to 12, with 12 prior to the electrical stimulation. The EMG
representingthe highestpossible score.The mean recordings were rectified and averaged (eight
hypnotizability score of this group was 7.2k2.7 stimulations). The onset latency and duration
compared to the mean score of the original US of the reflexes following the stimulations were
sample (7.39) (Shot-,1963)and the original Dan- measured from the averaged reflex by visual
ish sample (7.53) (Zachariae et al., 1996). Ten inspection using cursors on a computer screen.
low- and 10 high hypnotizable volunteers were The criterion for determination of the onset and
selected for the study. The low hypnotizable the end of the reflex was presenceof non-zero
group consisted of four males and six females activity (i.e. reflex),as no background activity was
with a mean hypnotizability score of 2.5 (+ 1.4) present.The root-mean-square(RMS) amplitude
and a mean age of 32.7 (A 11.1 years), and the was measuredin the interval 70-200 ms after the
high-hypnotizable group consistedof three males electrical stimulus.
and seven females with a mean hypnotizability
score of 11.0 (f0.9) and a mean age of 26.6
(2 3.1 years). The study was approved by the Investigation procedure
local Ethics Committee.
All subjectswereaskedto participate individually
in two sessions(Day 1 and Day 2) taking place
Measurement of nociceptive reflexes at least 24 h apart. On Day 1, the subjects were
given a brief description of the aims and methods
Stimulation
of the investigation. They then had their systolic
The nociceptive electrical stimuli were delivered blood pressuremeasured as a simple assessment
percutaneously to the sural nerve behind the of arousal prior to the experiment, and venous
lateral malleolus (Hugon, 1973; Willer, 1977) accesswas obtained using a Venflon catheter.The
by a constant current rectangular pulse train, subjects had their nociceptive-reflex threshold
consisting of five pulses, each of 1 ms and de- determined and their nociceptive reflexes meas-
livered at 200 Hz.The stimulus intensity was 1.5 ured in the waking state. The subjectswere then
times the individual reflex threshold, which was given a brief hypnotic induction lasting 10min
defined as the lowest stimulus intensity that and their reflexeswere measuredin the following
evoked reflexeson three successivetrials. A reflex three hypnotic conditions: (1) neutral, relaxed
was defined as electromyogram (EMG) activity hypnosis; (2) suggestionsof hypnotic analgesia;
above 30 PV for a period of at least 5 ms. The and (3) suggestions of hypnotic analgesia after
stimulation intensity of 1.5times the reflex thresh- an injection of either 1 ml saline or 1ml (1 mg/
old was above the subjects’ pain threshold and ml) naloxone (Narcanti, Du Pont Pharm). The
ensured a nociceptive reflex in all subjects order of Conditions 1 and 2 was determined
throughout the experiment. beforehand in a randomized fashion. The ad-
ministration of either saline or naloxone was
done in a double-blinded, randomized fashion.
Electrom yographic recordings
The subjects were then instructed to return to
The EMG was recordedfrom the ipsilateral thigh their normal waking state, after which the fifth
muscles (one electrode on biceps femoris and and final measurement of their reflexeswas com-
one electrode on rectus femoris) using Ag-AgCl pleted. The procedure on Day 2 was similar
surfaceelectrodes (diameter: 7 mm). Prior to the to the procedure on Day 1 with the following
electrode attachment, the skin was lightly ab- exceptions: (1) the order of Conditions 1 and 2
raded and cleaned with isopropyl alcohol. The was reversed from the order on Day 1; and (2)
EMG signals were filtered (20-500 Hz; 4. order), the subjects received an injection of the drug or

European Journal of Pain 11998), 2

28 R. ZACHARIAE ETAL.

placebo they had not receivedon Day 1, also in hypnotic conditions, and finally after the last
a double-blinded fashion. measurement in the post-hypnotic condition.

Hypnotic suggestions Statistical analysis

Manuscripts had been prepared beforehand to Baseline pain intensity, reflexesand stimulus in-
ensurethe uniformity of the suggestionsgiven to tensities were analysed using two-factor [day (1
each subject. and 2) and ‘hypnotizability’ (high and low)] ana-
lysis of variance (ANOVA). The data for pain
intensity and reflexes were then analysed using
Hypnotic induction
repeated-measuresANOVAs with ‘day’ ( x 2) and
Both high- and low-hypnotizable subjects were ‘condition’ ( x 4) as within-subjects factors, and
subjectedto the same lo-min hypnotic induction hypnotizability ( x 2) as a between-subjectsfac-
and deepening procedure. tor. When appropriate, the result for high- and
low-hypnotizable subjects were analysed sep-
arately using one-way ANOVA’s, with post-hoc
Neutral hypnosis
multiple comparisons performed using the Tukey
In the condition of neutral hypnosis, the subjects test. Pearson correlation coefficients were cal-
were given continuous suggestionsto experience culated to further describepossible relationships
a pleasantfeeling of relaxation while experiencing between selectedvariables.
normal sensitivity in the leg and foot receiving
the electrical stimulation.
RESULTS
Suggestions of hypnotic analgesia
Data are missing for one high-hypnotizable sub-
In the analgesiccondition, the subjectsweregiven ject who did not attend Day 2. Data are also
suggestions to imagine that their foot had re- missing for analgesiawith naloxone for one low-
ceived an injection of a ‘powerful local hypnotizable subject, who did not wish to receive
anaesthetic’, and continuous suggestionsto ima- the naloxone-placebo injections.
gine their foot as if it ‘was made out of wood
or rubber’. There were no differences in the
suggestionsgiven to high- and low-hypnotizable Baseline measures
subjects in the three hypnotic conditions, except
that the words ‘hypnosis’ and ‘trance’ were sub- The mean systolic blood pressureat the beginning
stituted with the word ‘relaxation’ in the in- of the sessionwas 127f 19on Day 1 and 127f 17
duction procedure given to low-hypnotizable on Day 2 for the low, hypnotizable group, and
subjects so as to minimize possible effects due to 126f18onDay 1 and 119+17onDay2forthe
negative expectations regarding the effectiveness high-hypnotizable group. The differences were
of hypnosis. not statistically significant. As shown in Table 1,
greater mean stimulus intensities were needed
elicit reflexesin high-hypnotizable subjects.Ana-
Subjective measures lysis of variance(ANOVA) showedthis difference
to be statistically significant [F( 1,37)= 6.3;
The subjects were askedto rate the pain intensity p<O.O2], while no differenceswere found between
on visual analogue scales (VAS, 100mm), with Days 1 and 2. Likewise, ANOVA showed a sig-
the left endpoint (0) representing ‘no pain at all’, nificant effect of hypnotizability [F( 1,38)= 6.0;
and the right endpoint (100) the ‘greatest pain p<O.O2] on baseline pain-intensity levels, with
possible’. This was done after the first baseline high-hypnotizable subjects displaying greater
measurement,after hypnosis for each of the three pain intensities than low-hypnotizable subjects.

European Journal of fain (1%X3, 2

HYPNOTIC MODULATION OF SPINAL PAIN TRANSMISSION 29

TABLE 1. Mean stimulus intensity (mA), mean pain intensity (visual analogue scale mm +SD) for baseline and
conditions of neutral hypnosis, hypnotic analgesia and post-hypnosis for high- and low-hypnotizable on Days
1 and 2

Ww Day ti Stim. (mA) Pre (mm) Neutral (mm) Analgesia (mm) Post (mm)
suscept.

High 1 10 33.6f22.9 59.8+ 16.9 49.8 k 23.0 29.9+ 14.1 58.6rfr 17.4
High 2 10 26.2 f 18.2 63.05 17.9 54.7 k 18.1 35.4f 16.4 55.9 & 16.0
Low 1 10 17.1 k5.2 49.0+21.1 43.1 &- 22.7 38.0 + 14.6 45.6 k 18.7
Low 2 10 18.1 27.7 44.6& 18.0 41.0f17.1 37.1 k 14.6 44.8& 19.2
High If2 20 29.9 * 20.9b 61.3f16.9” 52.1 k20.4” 32.5+ 15.1 57.3 f 16.3”
Low 1+2 20 17.6k6.5 46.8+ 19.2 42.1 k 19.6 37.6* 14.2 45.2 f 18.5

a ~0.05 [Post-hoc tests (Tukey) compared to analgesia].

bComparing high- and low-hypnotizable subjects at baseline.

No effect of day was found. Results are shown The data for high- and low-hypnotizability sub-
in Table 1. jects were then analysed for Days 1 and 2 sep-
arately using one-way ANOVAs. A significant
effect of ‘condition’ [F(3,32)= 3.4; p<O.O5] was
Pain intensity WAS) found for high-hypnotizable subjects on Day
1. Post-hoc tests (Tukey) showed significantly
A statistically significant effect was found for (pcO.05) lower reflexes in the condition of hyp-
‘condition’ [F(3,51- = 15.3; p<O.OOl], but there notic analgesiacompared to baseline. The differ-
were no effects of ‘day’ [F(1,17)=0.01; p=O.94] encesbetweenconditions did not reach statistical
and of ‘hypnotizability’ [F( 1,17)= 1.2; p = 0.281. significancefor low-hypnotizable subjectson Day
A significant interaction was seenbetween ‘con- 1, or for high- and low-hypnotizable subjects on
dition’ and ‘hypnotizability’ [F(3,51) = 6.9; Day 2. The results are shown in Table 2.
p<O.Ol]. Since no effect or interaction was seen
for ‘day’, the data for Days 1 and 2 were pooled.
The data for high- and low-hypnotizable subjects Comparing the effect of naloxone and
were then analysed separately using one-way placebo
ANOVAs. A significant effect of ‘condition’ was
found for high-hypnotizable subjects [F(3,72) = To evaluate a possible effect of naloxone on
10.3;p~O.O001],withpost-hoc tests(Tukey) show- hypnotic analgesia,the relative percentdifference
ing a significant difference between ‘hypnotic for both VAS scoresand nociceptive reflexeswere
analgesia’ and the three remaining conditions. calculated between: (1) the hypnotic analgesia
No effect of ‘condition’ was seen in the low condition and the subsequentcondition of hyp-
hypnotizable group. The results are shown in notic analgesia with naloxone; and (2) between
Table 1. the hypnotic analgesic condition and the sub-
sequent condition of hypnotic analgesia with a
saline injection (placebo). The results are shown
Reflexes in Table 3. The results were analysed using two-
factor (‘condition’ x ‘group’) ANOVAs. While
A statistically significant effect was found for there was no effect for pain intensity [F(2,32)=
‘condition’ [F(3,51)=7.31; p<O.OOl], but no 1.1;p = 0.34), an effect of condition was found for
effects of ‘day’ [F(l, 17)= 0.07; p = 0.791 or ‘hyp- reflexes[F’(1,36)= 4.3; p<O.O5]. When the relative
notizability’ [F( 1,17)= 0.13;p = 0.721were found. difference reflexesof high- and low-hypnotizable
A significant interaction was found between‘day’ subjects were compared, a significant difference
and ‘hypnotizability’ [F( 1,17)= 5.19; p<O.O5]. (p<O.O5, t-test for independent samples) was

European Journal of Pain (1998). 2

30 R. ZACHARIAE ETAL.

TABLE 2. Mean nociceptive reflexes (@*SD) for baseline and conditions of neutral hypnosis, hypnotic analgesia
and post-hypnosis for high- and low-hypnotizable subjects on Days 1 and 2

Hypnotic Day n Baseline ($f) Neutral ($/) Analgesia (VW Post (/W
susceptibility

High 1 9 47.7k25.1 28.0& 12.4 21.2~15.1” 32.9 * 15.8

High 2 9 62.8f30.5 41.Ok23.4 37.9 + 20.4 44.7 k 23.1
Low 1 10 66.6 f 43.3 46.7 + 28.4 34.8* 16.0 43.4f30.6
Low 2 10 41.0* 19.1 36.0k21.4 28.4& 16.8 41.4k29.7

“p<O.O5 [Post-hoc test (Tukey) compared to baseline].

TABLE 3. Mean changes (%&SD) of pain intensity and nociceptive reflexes in

high- and low-hypnotizable subjects during conditions of: (1) hypnotic analgesia
after an injection of naloxone; and (2) hypnotic analgesia after a placebo
injection (saline) compared to the condition of hypnotic analgesia with no
injection

Measure Hypnosis Naloxone Placebo (saline)

susceptibility

VAS% High -14.Ok35.4% +3.1 +37.6%

VAS% Low -0.4+11.8% +5.4+ 16.0%
EMG% High +95.4f 155.4% -1.4f22.1%
a
EMG% Low +3.6*21.5% - 5.7 k35.9%

“p~o.05 (one-tailed). VAS, visual analogue scale; EMG, electromyogram.

found for the analgesiawith naloxone condition, coefficient of correlation between stimulus in-
while no difference was found for the placebo tensity and baselinepain intensity, no correlations
condition. were found for either the high- (R = -0.16, n.s.)
or the low-hypnotizable groups (R =0.12, n.s.).
Stimulus intensity for naloxone
Correlations between changes in pain
The mean stimulus intensity needed to elicit a
intensity (VAS) and reflexes
reflex in the session which included naloxone
injection was 17.1 ( + 6.4)mA for the low-hyp- The relative changesin pain intensity and reflexes
notizable group and 31.2 (+ 19.0)mA for the were calculated as the percent change,compared
high-hypnotizable group. This difference was to baseline. Significant correlations were found
found to be statistically significant (~~0.05, t-test between relative changes in pain intensity and
for independent samples). relative changesin reflexesin both the high- (R =
0.32, ~~0.05) and the low-hypnotizable groups
(R=0.40, ~~0.01) for the three conditions of
Correlation analyses neutral hypnosis, analgesia and post-hypnosis
Correlations between baseline measures taken as a whole.
To evaluate whether general arousal as rep-
resented by blood pressure had an influence on Correlations between stimulus intensity
and the relative effect of naloxone on
pain intensity, the correlation between blood
reflexes
pressure and baseline pain intensity was cal-
culated for both Day 1 and Day 2. No cor- Significant correlations were found betweenhyp-
relations were found. When calculating the notizability and stimulus intensity (R = 0.49,

European Journal of Pain (19981, 2

HYPNOTIC MODULATION OF SPINAL PAIN TRANSMISSION 31

p<O.O5), between hypnotizability and relative Effects on pain intensity

changesin reflexesduring suggestionsof hypnotic
analgesia with naloxone (R =0.42, p-=0.05), and Theseresults showedthat suggestionsof hypnotic
between stimulus intensity and the relative analgesia produced a considerable reduction
changesin reflexesduring suggestionsof hypnotic (-46.9%) in pain intensity in high-hypnotizable
analgesiawith naloxone (R= 0.88,p~O.Ol).When subjects, while the differences in the low-hyp-
calculating the partial correlation coefficient be- notizable subjects did not reach statistical sig-
tween stimulus intensity and relative changesin nificance. These findings are consistent with the
reflexes, controlling for hypnotizability, a sig- results of previous investigations (Hilgard, 1986;
nificant correlation of 0.85 ($0.01) was found. Evans, 1990;Zachariae & Bjerring, 1994),which
When controlling for stimulus intensity, no cor- indicate that high-hypnotizable subjects are gen-
relation was found between relative changes in erally able to produce greater reductions in pain
reflexesduring analgesiawith naloxone and hyp- intensity under the condition of hypnotic anal-
notizability (R = 0.01). gesia than low-hypnotizable subjects, regardless
of the pain stimulus and the criteria of pain
sensation used.
DISCUSSION

The influence of hypnotizability on Effects on nociceptive reflexes

baseline measures
When analysing the results for nociceptive re-
At the pre-hypnotic baseline, high-hypnotizable flexes, high-hypnotizable subjects showed a sig-
subjects were found to experience significantly nificant reduction (- 55.6%)during the condition
greater pain than low-hypnotizable subjects did. of hypnotic analgesiaon Day 1, when compared
Although significantly greaterstimulus intensities to baseline,whereasthe difference between base-
were required to elicit spinal reflexes in high- line and the condition of hypnotic analgesia
hypnotizable subjects than in low-hypnotizable (- 39.6%) d’d 1 not reach statistical significance
subjects, no correlations were found between on Day 2. No effectsof any of the conditions were
pain intensity and stimulus intensity. Also, the seenin the group of low-hypnotizable subjects.
difference in stimulus intensity was not reflected Theseresults generally support the previous find-
in significant differences in baseline reflexes be- ings of Finer (1970) and Kiernan et al. (1995),
tween the two groups. The authors have no ready that it is possible to reduce the nociceptive spinal
explanation for these observations. It has been reflex by suggestionsof hypnotic analgesia, and
observed that increased arousal may augment suggestthat hypnotizability may be of import-
the spinal reflex (Willer et al., 1979) as well as ance.The lack, however, of a clear-cut difference
other physiological responses,including evoked between the neutral hypnotic and the hypnotic
brain potentials (Eason et al., 1969; Arendt- analgesic conditions could indicate that, in ad-
Nielsen, 1994). One possible explanation could dition to the specific suggestions of analgesia,
therefore be that low-hypnotizable subjects ex- other factors, including the general element of
perienced higher levels of anxiety with regard relaxation found in the hypnotic conditions,
to the electrical stimuli, and the experimental might be of importance. This finding is in agree-
situation in general, than high-hypnotizable sub- ment with the results of Kiernan et al. (1995),
jects did, so that lower stimulus intensities were who did not find any difference betweenhypnotic
required to elicit a nociceptive reflex in this group. suggestionsof comfort and relaxation, and spe-
Anticipating this possibility, systolic blood pres- cific suggestionsof analgesia.Although the pres-
sure had been measured prior to the baseline as ent authors, contrary to Kiernan et al. (1995),
a crude assessment of the degree of arousal. did find moderate correlations between the
However, no differenceswere found between the changesin pain intensity and nociceptive reflexes
two groups. in both high- (R=0.32) and low-hypnotizable

European Journal of Pain (1998), 2

32 R. ZACHARIAE HAL.

groups (R =0.40), the present results, taken to- when controlling for hypnotic susceptibility, sug-
gether, suggest that the importance of hyp- gestsanother explanation. Although the authors
notizability is less clear-cut with regard to the also found a significant correlation between the
ability to suppressthe reflexesthan with regard relative changesin reflexes and hypnotizability,
to the reduction of subjective pain responses. there was no correlation when we controlled for
This has also been shown to be the case with stimulus intensity. These results suggestthat the
other physiological correlatesof nociception, e.g. higher stimulus intensity levels found in the high-
pain evoked brain potentials (Zachariae & Bj- hypnotizable group are responsible for the sup-
erring, 1994;Dahlgren et al., 1995)3,5).The au- pressiveeffect of naloxone on hypnotic analgesia,
thors have no ready explanation for the reduced and that hypnotizability in itself may be of less
effect of suggestions of hypnotic analgesia on importance. It could be that, as a result of these
the reflexes on Day 2 in the group of high- greater intensities, higher levels of endogenous
hypnotizable subjects. One possible explanation opioids were releasedin this group. This hypo-
could be that the subjects were less stressedby thesis finds support in studies which have dem-
the painful stimuli on Day 2, thus reducing their onstrated release of endogenous opioids
motivation to respond to hypnotic suggestions following certain types of painful stressors in
of analgesia. both laboratory animals (Frenk et al., 1986)and
in humans (Willer et al., 1981).It has also been
shown that the analgesic effect of morphine is
Effect of naloxone on hypnotic analgesia dependent on the intensity of the pain stimulus
(Buchsbaum et al., 1981).It may also be that the
When the authors calculated and compared the higher stimulus intensities needed to elicit the
relative effect of suggestionsof hypnotic analgesia reflexes in the high-hypnotizable group resulted
with naloxone to the relative effect of suggestions in activation of more C-fibres, which are known
of hypnotic analgesia with saline, no significant to respond to opioids, in this group. It is therefore
differences were found for pain intensity. This possiblethat the reflexesof the high-hypnotizable
finding is in agreementwith studies showing that subjects were generally inhibited by a pain- or
naloxone does not influence the perception of stress-relatedreleaseof endogenousopioids, and
pain induced by electrical or thermal stimuli that the observedsuppressiveeffect of opioids is
(Stacheret al., 1988),as well as with the previous unrelated to both the specific condition of hyp-
studiesthat found that injections of naloxone are notic analgesiaand the level of hypnotizability.
unable to reverse the effect of suggestions of
hypnotic analgesia on pain intensity (Goldstein
& Hilgard, 1975; Barber & Mayer, 1977;Moret CONCLUSION
et al., 1991). However, when the effects of na-
loxone on the nociceptive reflexeswere analysed, The present results confirm earlier observations
a significant reduction in the effect of suggestions demonstrating the ability of high-hypnotizable
of hypnotic analgesia was found in the high- subjects to reduce their sensation of pain sig-
hypnotizable group, while there was no ob- nificantly. These results also support a previous
servable effect in the low-hypnosis-susceptible finding that nociceptive spinal reflexes may be
group. modulated by psychological intervention, and
One explanation could be that the psycho- suggest that the ability to reduce nociceptive
physiological and biochemical mechanisms that reflexes is dependent on hypnotizability. While
mediate the effect of hypnotic suggestionon pain these results are in agreement with earlier in-
processing of high-hypnotizable subjects differ vestigations that all failed to show an effect of
from those that mediate the response of low- naloxone on reported pain intensity, naloxone
hypnotizable subjects. However, the finding of a was found to reverse the suppressive effect of
significant correlation betweenstimulus intensity suggestionsof hypnotic analgesia on the reflexes
and the relative changes in reflexes during the in the group of high-hypnotizable subjects.These
condition of hypnotic analgesia with naloxone results suggestthat this effect was related to the

European Journal of Pain (1998). 2

HYPNOTIC MODULATION OF SPINAL PAIN TRANSMISSION 33

higher stimulus intensities generally needed to Dahlgren LA, Kurtz RM, Strube MJ, Malone MD. Differ-
ential effects of hypnotic suggestion on multiple dimen-
elicit a reflex in this group, rather than hypnotic sions of pain. J Pain Symptom Managem 1995; 10:
susceptibility in itself. The authors are unable to 464-470.
explain the greater stimulus intensities neededin De-PascalisV, Perrone M. EEG asymmetry and heart rate
during experience of hypnotic analgesia in high and low
high-hypnotizable subjects, and further in- hypnotizables. Znt J Psychophysiol 1996; 21: 163-175.
vestigations are clearly needed. Desmedt JE, Godaux E. Habituation of exteroceptive sup-
pression and of extero-ceptive reflexes in man as in-
fluenced by voluntary contraction. Brain Res 1976; 106:
21-29.
ACKNOWLEDGEMENTS Eason RG, Harter MR, White CT. Effects of attention and
arousal on visually evoked cortical potentials and reaction
time in man. Physiol Behav 1969; 4: 283-289.
This study was in part supported by the Danish Evans FJ. Hypnosis and pain control. Austral J Clin Exp
Pain ResearchCenter and The Danish National Hypn 1990; 18: 21-33.
ResearchFoundation. Finer B. Physiological Studies of the Relationship Between
Doctor and the Patient in Pain. Dissertation, Uppsala,
1970.
Frenk H, Cannon T, Lewis JW, Liebeskind JC. Neural
and neurochemical mechanisms of pain inhibition. In:
REFERENCES Sternbach RA, editor. The Psychology ofPain. New York:
Raven Press, 1986: 2547.
Arendt-Nielsen L, Zachariae R, Bjerring P. Quantitative Goldstein A, Hilgard ER. Failure of the opiate antagonist
evaluation of hypnotically suggested hyperaesthesia and naloxone to modify hypnotic analgesia. Proc Nat Acad
analgesia by painful laser stimulation. Pain 1990; 42: Sci 1975; 72: 2041-2043.
243-25 1. Hamon M, Bourgoin S, Le Bars D, Cesselin F. In vivo and
Arendt-Nielsen L, Nielsen J, Brennum JR. Effect of at- in vitro release of central neurotransmitters in relation to
tention and inattention upon the nociceptive flexor reflex pain and analgesia. Progr Brain Res 1988; 77: 431-445.
(RII). Abstract Book, The 15th Annual Meeting, Scan- Hilgard ER. Hypnosis and pain. In: Sternbach RA, editor.
divanian Association,for the Study of Pain, Ebeltoft, 1991: The Psychology of Pain, New York: Raven Pressm 1986.
40. Hugon M. Exteroceptive reflexes to stimulation of the
Arendt-Nielsen L, Brennum J, Sindrup S, Bak P. Elec- sural nerve in normal man. In: Desmedt JE, editor. New
trophysiological and psychophysical quantification of Developments in Electromyography and Clinical Neuro-
temporal summation in the human nociceptive system. physiofogy. Basel: Karger, 1973: 713-729.
Eur J Appl Physiol 1994; 68: 266273. Kiernan BD, Dane JR, Phillips LH, Price DD. Hypnotic
Arendt-Nielsen L. Characteristics, detection, and mod- analgesia reduces R-III nociceptive reflex: further evi-
ulation of laser-evoked vertex potentials. Acta Anaes- dence concerning the multifactorial nature of hypnotic
thesiol Stand 1994; 38 (Suppl. 98): l-44. analgesia. Pain 1995; 60: 3947.
Barber J, Mayer D. Evaluation of the efficacy and neural Moret V, Forster A, Laverriere MC, Lambert H, Gaillard
mechanics of a hypnotic analgesia procedure in ex- RC, Bourgois P, Haynal A, Gemperle M, Buscher E.
perimental and clinical dental pain. Pain 1977; 4: 4148. Mechanism of analgesia induced by hypnosis and acu-
Bathien N, Hugelin A. Reflexes monosynaptiques et poly- puncture: is there a difference? Pain 1991; 45: 135-140.
synaptiques de l’homme au tours de l’attention. Elec- Ome MT. Hypnotic methods for managing pain. In: Bonica
troenceph Clin Neurophysiol 1969; 26: 604-612. JJ, editor. Advances in Pain Research and Therapy, Vol. 5.
Buchsbaum MS, Davis GC, Coppola R, Naber D. Opiate New York: Raven Press, 1983: 847-862.
pharmacology and individual differences. II. Somato- Shor RE, Ome EC. Harvard Group Scale of Hypnotic
sensory evoked potentials. Pain 1981; 10: 367-377. Susceptibility, Form A. Palo Alto: Consulting Psy-
Cannon JT, Lewis JW, Weinerg VE, Liebeskind JC. Evidence chologists Press, 1962.
for the independence of brainstem mechanisms mediating Shor EC. Norms on the Harvard Group Scaleof Hypnotic
analgesia induced by morphine and two forms of stress. Susceptibility, From A. Znt J Clin Exp Hypn 1963; 11:
Brain Res 1983; 269: 231-236. 3947.
Crawford HJ, Gruzelier JH. A midstream view of the Spanos NP, Ollerhead VG, Gwynn MI. The effects of three
neuropsychophysiology of hypnosis: recent research and instructional treatments on pain magnitude and pain
future directions. In: From E, Nash MR, editors. Con- tolerance: implications for theories of hypnotic analgesia.
temporary Hypnosis Research. New York: Guildford Zmag Cogn Pers 1985-1986; 5: 321-337.
Press, 1992: 227-268. Spanos NP, MacDonald DK, Gwynn MI. Instructional set
Crawford HJ, Gur C, Skolnick B, Gur RE, Benson DM. and the relative efficacy of hypnotic and waking analgesia.
Effects of hypnosis on regional cerebral blood flow during Canad J Behav Sci 1988; 20: 64-72.
ischemic pain with and without suggested hypnotic anal- Stacher G, Abatzi T-A, Schulte F, Schneider C, Stacher-
gesia. Znt J Psychophysiol 1993; 15: 181-195. Janotta G, Gaupmann G, Mittelbach G, Steinringer H.
Crawford HJ. Brain dynamics and hypnosis: attentional Naloxone does not alter the perception of pain induced
and disattentional processes.Znt J Clin Exp Hypn 1994; by electrical and thermal stimulation of the skin in healthy
42: 204-232. humans. Pain 1988: 34: 271-276.

European Journal of Pain (1998), 2

34 R. ZACHARIAE ETAL.

Willer J-C. Comparative study of perceived pain and no- with postoperative pain. Anesthesiology 1985; 63: 6755680.
ciceptive flexion reflex in man. Pain 1977; 3: 69-80. Zachariae R, Bjerring P. The effect of hypnotically induced
Willer JC, Boureau F, Albe-Fessard D. Supraspinal in- analgesia on flare reaction to cutaneous histamine prick
fluences on nociceptive flexion reflexes and pain sensation test. Arch Dermatol Res 1990; 282: 539-543.
in man. Bruin Res 1979; 179: 6168. Zachariae R, Bjerring P. Laser-induced pain-related brain
Willer JC, Albe-Fessard D. Electrophysiological evidence potentials and pain ratings in high and low hypnotic
for a release of endogenous opiates in stress-induced susceptible subjects during hypnotic suggestions of re-
analgesia in man. Bruin Res 1980; 198: 419426. laxation, dissociated imagery, focused analgesia, and
Wilier JC, Dehen H, Cambier J. Stress-induced analgesia placebo. Int J Clin Exp Hypn 1994; 42: 5680.
in humans: endogenous opioids and Naloxone-reversible Zachariae R, Jorgensen MM, Sommerlund B, Molay F.
depression of pain reflexes. Science 1981; 212: 689-691. Danish norms on the Harvard Group Scale of Hypnotic
Willer J-C, Bergeret S, Gaudy J-H. Epidural morphine Susceptibility, Form A. Int J Clin Exp Hypn 1996; 44:
strongly depresses nociceptive flexion reflexes in patients 140-151.

European Journal of Pain (IX%), 2