Evoked Human Oesophageal Hyperalgesia: A Potential Tool for Analgesic Evaluation?


  • Anne Estrup Olesen,

    1. Mech-Sense, Department of Gastroenterology, Aalborg Hospital, Aalborg, Denmark
    2. Center for Sensory-Motor Interactions, Department of Health Science and Technology, Aalborg University, Aalborg, Denmark
    Search for more papers by this author
  • Camilla Staahl,

    1. Mech-Sense, Department of Gastroenterology, Aalborg Hospital, Aalborg, Denmark
    2. Center for Sensory-Motor Interactions, Department of Health Science and Technology, Aalborg University, Aalborg, Denmark
    Search for more papers by this author
  • Christina Brock,

    1. Mech-Sense, Department of Gastroenterology, Aalborg Hospital, Aalborg, Denmark
    Search for more papers by this author
  • Lars Arendt-Nielsen,

    1. Center for Sensory-Motor Interactions, Department of Health Science and Technology, Aalborg University, Aalborg, Denmark
    Search for more papers by this author
  • Asbjørn Mohr Drewes

    1. Mech-Sense, Department of Gastroenterology, Aalborg Hospital, Aalborg, Denmark
    2. Center for Sensory-Motor Interactions, Department of Health Science and Technology, Aalborg University, Aalborg, Denmark
    Search for more papers by this author

Author for correspondence: Asbjørn Mohr Drewes, Mech-Sense, Department of Gastroenterology, Aalborg Hospital, Hobrovej 18-22, 9000, Aalborg, Denmark (fax +45 99326507, e-mail drewes@smi.auc.dk).


Abstract:  Hypersensitivity is a common finding in visceral disorders. Therefore, in the development and testing of analgesics for the treatment of visceral pain, it is important to establish an experimental pain model of visceral hypersensitivity. Such a model will mimic the clinical situation to a higher degree than pain models where the receptors and peripheral afferents are briefly activated as with, for example, electrical, thermal, and mechanical stimulations. In this study, a model to evoke experimental hyperalgesia of the oesophagus with a combination of acid and capsaicin was introduced. The study was a randomised, double-blind, cross-over study. Fifteen healthy volunteers were included. Sensory assessments to mechanical, heat, and electrical stimulations were done in the distal oesophagus, before and after perfusion with a 200 ml solution of acid+capsaicin (180 ml HCL 0.1 M and 2 mg capsaicin in 20 ml solvent) or saline. Oesophageal pain assessment and referred pain areas were evaluated. There were reproducible pain assessments between repetitions within the same day and between days (all P > 0.05). Acid+capsaicin perfusion induced 56% reduction of the pain threshold to heat (P = 0.04), 19% reduction of the pain threshold to electrical stimuli (P < 0.001), 78% increase of the referred pain areas to mechanical stimulation (P < 0.001) and 52% increase of the referred pain areas to electrical stimulus (P = 0.045). All volunteers were sensitised to one or more modalities by acid+capsaicin. The model was able to evoke consistent hyperalgesia and may be useful in future pharmacological studies.

Visceral disorders are often associated with pain, which can be difficult to investigate due to heterogeneity of patients. Thus, human being's experimental pain models have been developed to induce standardised pain in the viscera. Hereby, the stimulus can be controlled and the response assessed both qualitatively and quantitatively [1,2]. Such models have successfully been used to evaluate mechanisms and effect of analgesics [3–6]. We have previously demonstrated several differences between the sensations evoked by the different stimulus modalities in the viscera (thermal, electrical, and mechanical), indicating activation of different visceral nerve pathways. This offers the possibility for multi-modal stimuli activating the superficial and deeper layers of the human being's gut and hereby mimicking experiences present in painful visceral diseases [7].

Hypersensitivity is a common finding in visceral disorders [8,9]. Therefore, in the development and testing of analgesics for treatments of visceral pain, it is important to establish an experimental pain model of visceral hypersensitivity. Such a model will mimic the clinical situation to a higher degree than pain models where the receptors and peripheral afferents are briefly activated as with e.g. electrical, thermal, and mechanical stimulations. Thereby, it could bridge experimental pain to clinical pain [2,10]. Experimental pain models of hyperalgesia have been developed [8,11–13] and proved their value in the testing of analgesics [14–17]. However, human being's visceral pain models with chemical stimulation of the gastrointestinal tract are not investigated in detail and some studies show conflicting results regarding central sensitisation [11,12,18]. Moreover, the response to acid perfusion of oesophagus is highly individual with only 75% developing sensitisation [14]. As tissue injury generates release of multiple molecules acting synergistic, it may be necessary to use a mixture of chemical substances with diverse effects on the tissue to mimic hyperalgesia. Such different cellular interaction sites of acid+capsaicin have been proposed, as acid targets the transient receptor potential vanilloid type 1 (TRPV1) extracellulary, while the capsaicin targets TRPV1 predominantly intracellulary [19].

Capsaicin binds to the TRPV1, which plays an important role in activation of the pain system and tissue inflammation [20]. Recently, studies have tested the analgesic effect of TRPV1 antagonists on capsaicin-induced acute pain [21,22]. Therefore, it could be important to include capsaicin in a multiple pain-producing stimuli to mimic the clinical situation, which includes tissue inflammation. Acid alters the activation of capsaicin responses in vitro [23], and capsaicin responses were found to be about 10-fold longer lasting than acid responses in an animal model [23]. We therefore hypothesised that the combination of acid+capsaicin would evoke stable oesophageal hyperalgesia in human beings.

The aims were: (i) to explore the effect of acid+capsaicin in induction of oesophageal hyperalgesia in a multi-modal visceral pain approach, assessing modality-specific hyperalgesia after induced sensitisation and (ii) to evaluate referred pain areas to experimental visceral stimuli before and after sensitisation, reflecting sensitisation of the central nervous system (CNS).


Subjects.  The study was conducted according to the Declaration of Helsinki. The local ethical committee approved the study protocol (VN 2003/120). Oral and written informed consents were obtained from all volunteers. None of the volunteers consumed chilli daily and none reported chest pain, reflux, regurgitation or other symptoms. The study included 15 healthy volunteers – seven females, eight males; mean age 34 years, range 22–63 years.

Protocol.  To improve the psychophysical assessment of the intensity of visceral pain and discomfort, learning sessions are important before the experiments [1]. Therefore, each volunteer participated in a screening session where all pain stimulations were performed.

The study was a randomised, double-blind, placebo-controlled, cross-over study. The volunteer fasted for at least 4 hr before the probe was inserted through the mouth and placed in the stomach. The bag was inflated and the probe was then carefully retracted until resistance was met, in order to identify the lower oesophageal sphincter. The probe was then retracted 8 cm to ensure placement above lower oesophageal sphincter and taped to the chin to ensure the position. The exact length from the tip of the probe to the teeth was registered and the probe was also placed with this distance at the second visit to ensure the same position at both visits. Initially, two mechanical distensions were applied to pre-condition the tissue and to train the -volunteers in assessing sensations in the oesophagus. These distensions were stopped at mild pain intensity (visual analogue scale [VAS] = 6) and have previously been shown to facilitate discrimination of the different VAS ratings [1,7]. Hereafter, pain tests were initiated. Each test battery was carried out in the same order: heat, mechanical, and electrical stimulation. As the stimulated area of oesophagus was re-activated with warm water before each test battery, it was most practical firstly to stimulate with heat with the same stimulation devices. Thereafter, mechanical stimulation was done as some of the same equipment was used for water running into the probe. Finally, electrical stimulation using another computer and stimulator was applied. The flowchart illustrating the protocol is shown in fig. 1. The duration of each test battery was approximately 10 min. depending on the individual volunteer's pain threshold. First, a baseline test was done. Second, the perfusion with acid+capsaicin or placebo (saline) was done over 28 min. After the perfusion, at time = 0, the first pain test was done.

Figure 1.

 Flowchart that illustrates each experiment. Each test battery was done in the same order, heat stimulation → mechanical stimulation → electrical stimulation. First, a baseline test was done. Then, the perfusion with acid+capsaicin was done over 28 min. After the perfusion, at time = 0, the first pain test was done. The second pain test was started at t = 28 min. with rekindling by circulating hot water in the balloon in the distal part of the oesophagus (45° for 2 min.), followed by the pain test (t = 30 min.). The third pain test was started at t = 58 min. with rekindling, followed by the last pain test (t = 60 min.).

Sensory assessment. Prior to the study, all volunteers were instructed to use the 0–10 electronic VAS, where 0 = no perception, 1 = vague perception of mild sensation, 2 = definite perception of mild sensation. 3 = vague perception of moderate sensation, 4 = definite perception of moderate sensation, 5 = the pain threshold, 6 = mild pain. 7 = moderate pain, 8 = pain of medium intensity, 9 = intense pain, and 10 = unbearable pain. This scale has been described in detail previously [1] and has been shown to be robust and valid in assessment of experimental oesophageal pain [24]. The volunteers were instructed and trained to score sensation and pain, and to differentiate this from the unpleasantness in the throat induced by the probe. They were asked to quantify referred pain areas corresponding to the maximal pain intensity by drawing the area and location on a transparent paper. The size of the referred pain area was later measured and digitised (Trust, 1200 wireless tablet, Trust International BV, Dordrecht, The Netherlands).

Probe design.  An oesophageal probe was designed for multi-modal stimulation including a non-conducting polyurethane bag for mechanical and thermal stimuli, and electrodes for electrical stimuli. The bag was placed 1.5 cm above the tip of the probe. Stainless steel electrodes for electrical stimuli were mounted on the probe 10 cm proximal to the bag. A channel for acid+capsaicin perfusion was placed inside the probe with an outlet 1 cm above the bag. Detailed information of the probe and stimulation system has been described in reference [7]. The system allowed sequential stimulation of individual stimuli with control of site, duration, and intensity.

Induction of hyperalgesia.  Each volunteer was investigated twice. At least, one week separated the two randomised study days. On one day, the volunteers underwent an acid+capsaicin perfusion test with a 200 ml solution consisting of 180 ml HCL 0.1 M (Bie & Berntsen, Rødovre, Denmark) and 2 mg capsaicin in 20 ml solvent (polyoxyethylene sorbitan monooleate) (Hospital Pharmacy, Aalborg Hospital, Aalborg, Denmark). Preliminary pilot studies led to the conclusion that 2 mg of capsaicin was useful, as it was tolerable and induced sensitisation. The total amount of pure capsaicin has been derived from 2.5 ml Tabasco and found to be 0.42 mg [25]. Therefore, the 2 mg of capsaicin used in this study correspond to approximately 12 ml Tabasco sauce. On the other study day, the volunteers underwent placebo perfusion (isotonic saline). The 200 ml of liquid was administered through a perfusion channel in the probe at a rate of 7 ml/min. for 28 min. If the volunteer reached the pain detection threshold, the perfusion was stopped for 1 min. and then continued until 200 ml was infused or until it became too unpleasant (pain intensity higher than pain detection threshold for more than 1 min.). This was done to use the maximum tolerated amount of acid+capsaicin. It has been demonstrated in the skin that ongoing peripheral nociceptive stimulation with heat is necessary for maintaining central sensitisation [26]. Consequently, before the sensory test battery was performed, a re-activation (re-kindling) of the stimulated area was done (fig. 1) by re-circulating innocuous heat (45° for 2 min. or until the volunteers reached the pain detection threshold).

Heat stimuli.  Heat stimulation was carried out by a peristaltic pump system (Type 110, Ole Dich Instrument Makers Aps, Hvidovre, Denmark) that re-circulated water (60°) through two channels in the probe with a rate of 180 ml/min. The feeding channel ended inside the bag where a sensor monitored the temperature. The stimulation has been described in detail previously [5]. As a proxy for the thermal energy, the stimulus intensity was calculated as the area under the curve (AUC, [time vs. temperature]) from start to end of the stimulation [24,27].

Mechanical stimuli.  Mechanical distension was done by inflating the bag with water (37°) at a rate of 15 ml/min. by a pump (Type 110, Ole Dich Instrument Makers Aps) [5]. The bag was distended to a VAS rating of 7 (moderate pain) after which the bag was deflated. The pressure in the balloon was continuously registered and the volume in the balloon at moderate pain was used for analysis.

Electrical stimuli.  Stainless steel electrodes were mounted on the probe 10 cm proximal to the bag. As we aimed at investigating the sensitivity locally in the perfused area, the probe was pushed 10 cm in distal direction before stimulation, so the electrodes were in contact with the perfused area. After electrical stimulation, the probe again was retracted 10 cm proximal. A computer-controlled constant current stimulator (University of Aalborg, Aalborg, Denmark) delivered the single electrical stimulations to the oesophagus [5]. The current intensity was increased in 1 mA in increments with random sham stimulations having the same or lower intensity until the pain detection threshold was found [5].

Overall sensitivity.  The overall effect of the acid+capsaicin perfusion on pain threshold and referred pain area was evaluated. Calculation was done taking the placebo infusion and the two baseline investigations into account. All values were corrected for baseline values when comparing the two treatment arms. The following equation was used to calculate relative sensitisation (S):


This equation has previously been used to define whether volunteers are sensitised or not to a given stimulus [28].

X has different units corresponding to different modalities. For heat stimulation, X denotes AUC at pain detection threshold or referred pain (cm2), for mechanical stimulation, X denotes volume (ml) in the balloon when the volunteer scores 7 on the VAS or referred pain (cm2), and for electrical stimulation, X denotes mA at pain detection threshold or referred pain (cm2).

For the assessments, which were supposed to decrease after sensitisation (ml and mA), a volunteer was defined as being sensitised when S < –0.1 and desensitised when S > 0.1. ‘No change’ was indicated when S was between –0.1 and 0.1 [28].

For the assessments, which were supposed to increase after sensitisation (referred pain and AUC), a volunteer was defined as being sensitised when S > 0.1 and desensitised when S < –0.1. ‘No change’ was indicated when S was between −0.1 and 0.1.

Statistical analysis.  The results are expressed as mean ± S.D. unless otherwise specified. To eliminate errors relating to differences in baseline pain recordings, the change in stimulus intensity relative to baseline was used in the calculations [24]. Two-way anova was used. The software package SigmaStat 3.0 (Aspire Software International, Ashburn, VA, USA) was used. Three types of statistical analyses were applied to evaluate the reproducibility of the pain model. For assessment of systematic bias over time, the levels of mean thresholds were analysed using one-way anova. Significant P-values would indicate that the scores had changed due to systematic causes (bias relating to the method). For measurement of intraindividual variance, the intraclass correlation was calculated. Intraclass correlation evaluates each volunteer's ability to reproduce a response across sessions. There is no standard acceptable level of reliability using the intraclass correlation, ranging from 0 to 1. Values closer to one represent higher reliability, and thus it is recommended that any measure should have an intraclass correlation coefficient of at least 0.6 to be useful [29]. For determination of the overall interindividual variance, the coefficient of variation was calculated. Coefficient of variation expresses the S.D. as a proportion of the mean [29].


As the potency between capsaicin and acid was unknown, a comparative study was not performed. Based on pilot studies and results from animal studies, we chose to make an explorative study of the effect of acid+capsaicin.

The volunteers reported heartburn and nausea during perfusions, but 15 min. after the perfusion was stopped, there were no spontaneous sensations and none of the volunteers felt any symptoms related to the perfusion the day after the experiment. Hence, none of the chemical stimuli induced longer lasting pain. The symptoms experienced by the volunteers mimicked those seen in patients with gastro-oesophageal reflux disease suffering from heartburn and regurgitation.


Data from the pain assessments after perfusion of the oesophagus with saline (placebo) are summarised in table 1. The reproducibility of the pain recordings are summarised in table 2. The pain assessments were reproducible between repetitions within the same day and between days (all P > 0.05). The overall variability (coefficient of variation) was 41.6–135.7% and close to or above 100% for the referred pain areas.

Table 1. 
Results from sensory testing before and after perfusion of the oesophagus with isotonic saline (placebo).
StimuliUnitBaseline0 min.30 min.60 min.
  1. AUC is area under the temperature curve, mA is milliamp. Data were obtained before (baseline), right after, 30 min., after and 60 min. after placebo perfusion of the oesophagus.

HeatAUC at pain threshold329.5 ± 313.6278.1 ± 331.3184.2 ± 177.4173.8 ± 131.4
Referred pain (cm2) 46.7 ± 60.3 46.7 ± 56.6 42.3 ± 41.8 45.3 ± 50.8
MechanicalVolume (ml) at moderate pain 20.1 ± 8.1 20.2 ± 10.7 19.9 ± 7.4 19.8 ± 8.3
Referred pain (cm2) 37.3 ± 50.6 40.0 ± 52.2 40.0 ± 52.2 39.0 ± 57.0
ElectricalmA at pain threshold 10.8 ± 5.0 12.0 ± 6.9 12.1 ± 7.2 13.4 ± 7.2
Referred pain (cm2) 22.4 ± 23 22.6 ± 21.6 24.4 ± 23.9 23.1 ± 22.3
Table 2. 
Reproducibility of the pain recordings.
  Within dayBetween days
  1. The left side of the table is data obtained on four subsequent tests obtained on the placebo day (before, 0, 30 and 60 min. after perfusion with isotonic saline). The right side of the table is baseline data obtained on two study days separated by at least one week. The P value is the probability of equality between baseline recordings. As all P-values were above 0.05, there was no difference in baseline recordings between days or in four subsequent tests obtained on the placebo day. ICC (intra-class correlation) is intra-individual variance; CV (coefficient of variation) is the overall inter-individual variance. AUC is area under the temperature curve, mA is milliamp.

 AUC at pain threshold0.560.76 96.60.610.78 84.6
HeatReferred pain (cm2)0.980.90116.00.440.90134.6
 Volume (ml) at moderate pain0.960.82 43.10.720.81 42.6
MechanicalReferred pain (cm2)0.970.99135.70.360.77126.2
 mA at pain threshold0.800.89 41.6
ElectricalReferred pain (cm2)0.850.88 98.10.920.71104.4

Induction of hyperalgesia.

On average, the volunteers tolerated 95 ± 40 ml acid+capsaicin. Thus, a mean of 0.95 mg capsaicin was given corresponding to 5.6 ml Tabasco sauce. One of 15 volunteers tolerated 200 ml. Two volunteers were unable to tolerate 200 ml of isotonic saline (tolerated 157 and 158 ml).

The results from sensory testing after perfusion of the oesophagus with acid+capsaicin are shown in table 3.

Table 3. 
Results from sensory testing after perfusion of the oesophagus with acid+capsaicin.
StimuliUnitBaseline mean ± S.D.Time (60 min.) mean ± S.D.F- and P-values
  1. Results are given as mean ± S.D. (standard deviation). AUC is area under the temperature curve. There was hyperalgesia to heat and electrical stimulation and an expansion of referred pain areas (mechanical and electrical stimulation). The F- and P-values are calculated from multiple comparisons, including assessment at 30, 60, and 90 min. 1Significant effect of acid+capsaicin perfusion compared to placebo.

 AUC at pain threshold344.8 ± 254.1152.4 ± 202.8F = 4.4, P = 0.041
HeatReferred pain (cm2) 34.3 ± 48.159.11 ± 42.1F = 1.1, P = 0.3
 Volume (ml) at moderate pain 20.3 ± 8.7 28.1 ± 31.5F = 0.8, P = 0.4
MechanicalReferred pain (cm2) 24.9 ± 29.0 44.2 ± 33.7F = 12.6, P < 0.0011
 mA at pain threshold 13.0 ± 4.8 10.5 ± 3.8F = 14.5, P < 0.0011
ElectricalReferred pain (cm2) 20.4 ± 21.7 30.9 ± 28.3F = 3.7, P = 0.0451

Sensitivity to heat stimulation.

For heat stimulation, the AUC after acid+capsaicin perfusion was smaller than at baseline compared to placebo perfusion (F = 4.4, P = 0.04) indicating sensitisation. The referred pain area did not change after acid+capsaicin perfusion compared to placebo (F = 1.1, P = 0.3).

Sensitivity to mechanical stimulation.

No difference was seen for the maximal tolerated volume after acid+capsaicin compared to placebo (F = 0.8, P = 0.4). In contrast, an increase in the referred pain area after acid+capsaicin perfusion compared to placebo was seen (F = 12.6, P < 0.001) (fig. 2A + B).

Figure 2.

 Change in referred pain area to mechanically evoked pain. (A) Comparison of acid+capsaicin and saline perfusions of the distal oesophagus, illustrated as changes from baseline values (e.g. the referred pain area to mechanical stimulation increases 12 cm2 right after perfusion of oesophagus with acid+capsaicin compared to baseline, whereas it only increases 4 cm2 after perfusion with saline). Overall, the referred pain area increased after acid+capsaicin perfusion (P < 0.001). The error bars represent S.E.M. (standard error of mean). (B) Illustration of the referred pain area ‘before’ acid+capsaicin perfusion (baseline) and 60 min. ‘after’ acid+capsaicin perfusions of the distal oesophagus.

Sensitivity to electrical stimulation.

During electrical stimulation, a decrease in pain detection threshold was seen after acid+capsaicin perfusion compared to placebo (F = 14.5, P < 0.001). An increase in the referred pain area after acid+capsaicin perfusion compared to placebo was found (F = 3.7, P = 0.045).

Overall sensitisation.

The relative sensitisation for all volunteers was calculated at all times: right after perfusion with acid+capsaicin, 15 volunteers were sensitised in 49 cases, after 30 min. in 52 cases and after 60 min. in 42 cases. As there were more cases of sensitisation 30 min. after perfusion of the oesophagus with acid+capsaicin, the relative sensitisation for all volunteers 30 min. after perfusion is illustrated in table 4. As baseline values were not significantly different for any modality (table 2), all volunteers showed increased sensitivity to two or more stimulation modalities.

Table 4. 
The distribution of sensitisations to oesophageal stimulation, 30 min. after perfusion with acid+capsaicin. Thumbnail image of


For the testing and evaluation of analgesics, it is important to have experimental pain models of hyperalgesia mimicking clinical pain. Visceral hyperalgesia can be induced by acid perfusion of oesophagus, but in vitro data have indicated that the model can be improved due to a synergistic effect of acid+capsaicin [23,30]. Thus, the current study explored a human being's oesophageal model of hyperalgesia utilising two different chemicals. The combined perfusion of acid+capsaicin resulted in hyperalgesia to heat and electrical stimulation, and an expansion of referred pain areas to mechanical and electrical stimulation. Hence, the model mimics clinical visceral pain conditions involving hyperalgesia with a strong CNS component.

Methodological considerations.

Previously, the analgesic effect of drugs has mostly been tested in acute experimental skin models. Visceral pain differs to a major degree from the somatic counterpart due to the different peripheral and central organisation of the nervous systems together with differences in pain biochemistry and mechanisms [1]. Therefore, models of visceral pain are necessary to understand the basic functions as well as the effects of analgesics [1].

Changes in the CNS have been shown in animals after peripheral-induced hyperalgesia [31] and in human being studies [32,33]. These changes may also include altered expression of specific opioid receptors and thereby affect the response to analgesics [34,35]. In healthy volunteers, experimentally induced peripheral hyperalgesia in muscle and viscera has also evoked both peripheral and central sensitisation as a result of increased excitability of the CNS [13,36–38]. Thus, chemical perfusion of the oesophagus can induce generalised sensitisation and act as a translational bridge to the clinical situation. Such a model could be used in the future to test the effect and mechanisms of both opioids and other analgesics, for example TRPV1 antagonists, which are up-coming drugs for the treatment of inflammatory pain.

However, experimental pain methods have limitations, for example, it will never be possible to induce manifest clinical inflammation due to the potential risks for the volunteers. Another shortcoming could be that capsaicin can increase amplitudes, propagation velocity of pressure waves, and delay gastric emptying [25]. The present study was not designed to evaluate the possible prokinetic effect of acid+capsaicin. Therefore, if the model is applied in future drug studies, it is important to analyse both pharmacokinetics and pharmacodynamics of the administered drug. Thereby, it can be concluded whether the chemical procedure will influence the pharmacokinetics of the drug or not.

The placebo arm, including perfusion with isotonic saline, was essential to test for increased sensitivity or tolerance over time causing psychological factors like habituation. It has furthermore been shown that when tests were repeated with an interval of 1 week or more, the pain scores were less reproducible mainly because of a higher intra-volunteer variance, necessitating baseline recordings on all study days [24]. This was seen in the present study as slight difference in baseline recordings from the placebo period (table 1) and the acid+capsaicin period (table 2). Therefore, when applying experimental pain tests, the study design should be cross-over and include a placebo arm. Moreover, all values should be corrected for baseline values when comparing the two treatment arms. This ensures that the data analysis considers the increasing sensitivity to the stimulation [24].

Several reasons were taken into account when perfusion with isotonic saline was chosen instead of active control perfusion; (i) a control perfusion creating an equally intense sensation as acid+capsaicin perfusion, which simultaneously did not sensitise the nervous system, was non-existing, (ii) saline perfusion caused symptoms in two volunteers (13%). Previous pilot studies in our group have shown that not all volunteers report symptoms from acid perfusion and (iii) constructing the study as a three-way cross-over study including placebo, acid, and acid+capsaicin perfusion was not done, as two active perfusions could induce a long-term sensitisation, and thus affect one another depending on the time interval. Hence, this confounder was avoided in the current design. Moreover, the potency between acid and capsaicin has not been revealed. Thus, it was not possible to construct a reasonable study comparing equipotent doses. Furthermore, the study was a methodological experiment carried out to explore the effect of acid+capsaicin, and not whether acid+capsaicin was superior to either acid or capsaicin. To make stronger conclusions on superiority, future studies in a large number of healthy volunteers and patients should firstly investigate the effect of different doses of both acid and capsaicin, and then investigate the difference of combined perfusion and the single chemicals in detail.

It has been demonstrated that duration and magnitude of hypersensitivity is related to exposure or dose of the chemicals [39–41]. However, a recent study from Matthews et al. reported that short duration (5 min.) sequential infusions of acid in the distal oesophagus as well as weaker concentrations of acid (0.075 M) could induce and maintain hypersensitivity [42]. The acid-concentration was chosen because it is used in the Bernstein test and as it has been used in previous studies to induce oesophageal hyperalgesia [37,43]. The concentration of capsaicin was chosen as a result of previous experience in this group [44]. On average, 95 ± 40 ml of the acid+capsaicin was tolerated. Two volunteers were unable to tolerate the saline perfusion. This indicates that psychological parameters may influence the results or it could be due to normal variability in sensation. The increased sensibility after placebo could be a result of acid+capsaicin perfusion causing long-term sensitisation. This would affect the experiment one week later by increased sensibility to placebo perfusion. This was, however, not the case as these two volunteers underwent the placebo perfusion on the first of the two days.

Experimental pain research has limitations as the order and magnitude of the stimulations could affect the results. However, performing the study as a two-way cross-over study accounted for these effects. The order of stimulations was the same and so was the magnitude. Coefficient of variation and intraclass correlation values were calculated to investigate if these biases differentially affected the outcome. The pain stimulations provided stable pain assessments between repetitions the same day and between days. As expected, the intraclass correlation values within the day were better than those between days [24]. The large coefficient of variation values illustrates the interindividual variation in pain thresholds. The coefficient of variation values for referred pain areas indicate a larger variation, possibly caused by the difficulties of reporting this phenomenon. Some volunteers reported a small distinct area of 1 cm2, others as larger areas up to 220 cm2. Consequently, the coefficient of variation values was expected and the approach has previously been validated for the measurement of referred pain area [24].

Concerning ethical aspects, none had any spontaneous perfusion-related sensations 15 min. after perfusionstop, and none felt any symptoms the following day. The acid pH was 1, comparable to stomach juice, and chilli is a known spice for cooking. Investigations of 24-hr pH measurements in the distal oesophagus of volunteers have shown that frequent pH drops of same magnitude are present [45]. Thus, the chemicals in combination are not regarded as problematic. The exact duration of the induced hyperalgesia is unknown. Sarkar et al. showed that electrical pain threshold in the upper oesophagus decreased for 5 hrs, following lower oesophageal acid infusion [39], and consistent with this observation none of our volunteers experienced symptoms the day after the experiment.

Pain stimulations after the induced hyperalgesia.

Heat stimuli.  In the placebo arm, the pain detection threshold decreased over time indicating sensitisation (table 1). However, the acid+capsaicin perfusion reduced pain threshold to heat more than the saline perfusion. Capsaicin targets the TRPV1, which is also sensitive to protons and heat [30,46]. Heat-induced oesophageal pain is also mediated by the TRPV1 receptor [47]. Thus, a peripherally induced hyperalgesia caused by sensitisation of peripheral TRPV1 receptors could explain the reduced pain detection threshold to heat stimulation [27,48].

Mechanical stimuli.  The mechanosensitive afferents are mainly located in the deeper muscle layers and are not directly exposed to the acid+capsaicin [49,50]. This could explain equality between acid+capsaicin and placebo perfusion to distension in this study. Several studies have shown increased sensitivity to mechanical stimuli after oesophageal perfusion with acid and with capsaicin in ileum and jejunum [37,51,52]. Other studies support our findings as capsaicin did not sensitise mechanosensation neither in oesophagus nor in jejunum [25,53]. However, the lack of effect in this study could possibly be caused by chemically evoked secondary contractions and muscle hyper-responsiveness in the distal oesophagus interfering with the response to mechanical stimulation [49]. This would clearly influence the tolerated volume. Hence, if contractions result in the bag conforming to another shape after acid+capsaicin, the volume needed for pain detection will be different and affect the result [48]. This bag conformation could possibly explain the increased bag volume at moderate pain, 60 min. after perfusion with acid+capsaicin compared to baseline (table 2).

An increase in referred pain areas to the mechanical stimuli following acid+capsaicin perfusions was demonstrated. Visceral primary afferents terminates in the dorsal horn where it converges with sensory somatic input. Hence, the referred pain areas represent a central component of visceral pain [54]. Previous studies have also demonstrated central sensitisation in human being models of acid exposure [8]. This makes the model interesting in assessment of centrally acting mechanisms to mechanical stimulation.

Electrical stimuli.  Increased pain and referred pain areas to electrical stimulation were demonstrated. This was probably due to both a peripheral and a central effect. The central effect could be caused by a direct stimulation of the nerve.

Overall sensitisation.  Previously, oesophageal perfusion with 200 ml 0.15 M HCl induced proximal sensitisation in 75% of healthy volunteers [14]. Some healthy volunteers normally desensitise to perfusion of the oesophagus with acid for unknown reasons [49]. Therefore, it was expected that some volunteers tolerated more painful stimulations after perfusion with acid+capsaicin. Analysing the data, 0.1 seemed to be the best cut-off value to indicate whether a volunteer was sensitised or not. It cannot be excluded that the diffuse noxious inhibitory control influenced the results. However, this study was not designed to investigate the effect of diffuse noxious inhibitory control. This can be done in experimental pain studies by evoking diffuse noxious inhibitory control with e.g. the cold pressor test. Therefore, future studies should aim at investigating the effect of central control mechanisms on this model of hyperalgesia.

All volunteers showed increased sensitivity to two or more modalities. As the percentage of sensitised volunteers was not reported in all previous studies, a direct comparison was not possible. On the other hand, the tolerated volume of acid+capsaicin was lower than the tolerated volume of pure acid in another study [28]. Moreover, it appeared that the capsaicin was less tolerable than expected. Thus, the tolerated amount in combination with acid was only 1 mg. However, 1 mg was enough to induce hyperalgesia lasting for at least 60 min. This, indirectly, indicates a synergistic effect of capsaicin on acid.

As the multi-modal stimuli affect different oesophageal layers, we expected heterogeneous results as illustrated in fig. 3. Heat stimuli are mainly affecting the mucosa, thus the peripheral effect was more pronounced than the central (measured indirectly as referred pain area). The effect of mechanical stimulation was primarily central caused by the deeper location of the mechanosensitive afferents unaffected by the perfusion. Electrostimulation affects several layers unspecifically, which was demonstrated by both peripheral and central effect.

Figure 3.

 Schematic illustration of the proposed gut layers. The gut layers are preferentially affected with: (T) Thermal stimuli (mucosa, and submucosa, dark grey); (M) Mechanical stimuli (circular and longitudinal muscle – light and hatched grey); (E) Electrical stimuli (all layers depending on stimulus intensity). Perfusion of the oesophagus with acid+capsaicin (curved arrows) gives the possibility to evoke peripheral and central sensitisation (illustrated with stars). Referred pain in the somatic tissues is believed to be generated by central mechanisms, where visceral and somatic nerves converge on nerves in the same area of the spinal cord or at supraspinal centres. Hence, an increase in the referred pain after sensitisation with acid+capsaicin reflects central neuronal hyperexcitability as illustrated with the size of the arrows. Electrical stimulation results in peripheral and central sensitisation, thermal stimulations with heat only resulted in peripheral sensitisation, whereas the mechanical stimulations (mainly affecting deep neurons that are not affected by the peripheral sensitisation) only resulted in an increase in the referred pain area.

The study was not designed to investigate differences in women's pain perception during their menstrual cycle. Menstrual cycle can influence pain thresholds, but no or minor relation between the phase of the menstrual cycle and the pain perception has been found [55–57]. Vignolo et al. concluded that the intake of oral contraceptive is associated with decreased levels of reported pain [57]. However, this would not influence the results, as women would be on oral contraceptive throughout the study.


Oesophageal sensitisation and hyperalgesia can be induced experimentally with acid+capsaicin perfusion of the oesophagus and assessed quantitatively by the multi-modal pain assessment approach. Evaluation of the referred pain areas to experimental stimuli before and after the sensitisation can be used to elucidate central component of the hyperalgesia. The evoked phenomena mimic clinical disorders of the oesophagus to a higher degree than previously and can be applied in future clinical and pharmacological studies.


The Research Initiative of Aarhus University Hospital, ‘Det Obelske Familiefond,’ and the Spar Nord Foundation are acknowledged for funding this project.