An endoscopic method for thermal and chemical stimulation of the human oesophagus


  • S. S. Olesen,

    1. Mech-Sense, Department of Gastroenterology, Aalborg Hospital, Aarhus University, Denmark
    Search for more papers by this author
  • A. E. Olesen,

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

    1. Mech-Sense, Department of Gastroenterology, Aalborg Hospital, Aarhus University, Denmark
    2. Center for Sensory-Motor Interactions (SMI), Department of Health Science and Technology, Aalborg University, Denmark
    Search for more papers by this author
  • J. L. Poulsen,

    1. Mech-Sense, Department of Gastroenterology, Aalborg Hospital, Aarhus University, Denmark
    Search for more papers by this author
  • P. Funch-jensen,

    1. Department of Surgical Gastroenterology L, Aarhus Hospital, Aarhus University, Denmark
    Search for more papers by this author
  • H. Gregersen,

    1. Mech-Sense, Department of Gastroenterology, Aalborg Hospital, Aarhus University, Denmark
    Search for more papers by this author
  • A. M. Drewes

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

Professor Asbjørn Mohr Drewes, MD, PhD, DMSc, Mech-Sense, Department of Gastroenterology, Aalborg Hospital, Aarhus University, Hobrovej 18-22, 9000 Aalborg, Denmark.
Tel: +45 99326243; fax: +45 99326507; e-mail:


Abstract  Previous methods for visceral thermal stimulation have lacked control of the temperature rate and visual inspection of the organ. The aims of this study was to develop a method for linear control of heat stimulation in the human oesophagus combined with endoscopy, to assess the reproducibility of this method and to investigate sensitivity to thermal stimulation of the distal oesophagus before and after acid perfusion. A probe with a 2.8 mm endoscope inside was constructed permitting heat and chemical stimulation. Three different temperature ramps were applied in the distal oesophagus in 12 healthy subjects by recirculation of heated water in a bag. Endoscopy of the oesophageal mucosa was performed prior to experimental stimulation. The temperature, the time of stimulation and the area under the temperature curve (AUC) were measured at the pain detection threshold. Thermal stimulation was repeated after perfusion of the oesophagus with acid. The method was tested on two subsequent days to assess reproducibility. All subjects had a normal endoscopic examination. Day-to-day reproducibility was good for the three temperature ramps (intra-class correlations >0.6). The subjects tolerated less heat stimulation, a decrease in AUC (= 0.0003), a decrease in time to pain detection threshold (= 0.005) and decreased temperature at pain detection threshold (= 0.0001) after acid perfusion. The slow ramp was the most sensitive, showing a decrease in AUC of 29%. The present method was easily implemented and showed good reproducibility. It can potentially be used in basic experiments, drug and clinical studies as it provides a controllable thermal stimulus.

Painful sensations from the human oesophagus are common symptoms in the clinic but the underlying aetiology is often difficult to identify. In clinical practise, it is most likely that this is due to heterogeneity of the patient groups and the many confounding factors such as sedation, nausea and general malaise that all can affect the pain experience.1 In addition, the relationship between disease severity and symptoms are ambiguous.2 Experimental human models have been developed to increase our understanding of oesophageal pain mechanisms. These models permit the investigator to control the stimulus and assess the evoked response quantitatively and qualitatively whereby selectively pain modalities can be studied.1

Previous methods for heat stimulation in the human gastrointestinal (GI) tract have lacked control of the temperature rate and visual inspection of the organ. The temperature rate is an important factor in experimental pain models. In skin, C-fibres were activated using rates below 1 °C s−1, while higher rates of heating activated thinly myelinated Aδ-fibres.3 Hence, control of the temperature rate provides information about fibre and nociceptor activation.4 In most experiments with heat stimulations of the oesophagus the subjects have been healthy volunteers. However, oesophageal diseases such as gastro-oesophageal reflux syndrome GORD are common and seen in up to 20% of the population dependent on age and sex.5 As these diseases have variable clinical presentations and not necessarily present clinical symptoms an endoscopic examination of the oesophagus during the experiments will increase the validity.6

A new method recently developed by our group provides a controllable and linear increasing temperature stimulus of the oesophagus. We hypothesized that this method was valid and reliable. Hence, the aims of the study were: (i) to develop a probe permitting thermal and chemical stimulation together with endoscopy; (ii) to develop a computer controlled system applying heat stimulations with different linear ramps; (iii) to asses reproducibility of the method and (iv) to investigate sensitivity to thermal stimulation of the distal oesophagus before and after acid perfusion.

Material and methods


Twelve healthy volunteers (six females mean age 28 ± 5.2 years) participated in the study. They were recruited from hospital and university staff. Volunteers had normal physical examinations and were all healthy without pain complaints. In particular, they denied having any heartburn, regurgitation, chest pain or other upper GI symptoms. Written informed consent was given prior to participation and the protocol was approved by the local ethics committee (No VN 2003/120 mch).

Probe design

A custom made probe measuring 60 cm in length with a diameter of 6.2 mm was fitted with a bag near the tip (Ditens A/S, Aalborg, Denmark). The bag was made of 25 μm thick polyestherurethane and had a maximum volume of 80 mL. An inlet and an outlet channel permitted recirculation of water in the bag. A temperature wire with a tip sensor was placed in the inlet channel for continuous temperature assessment during stimulation (TC Ltd., Uxbridge, England). A side hole for acid perfusion was placed 5 mm above the bag. A fiberscope was positioned (Olympus BF-XP160F, Ballerup, Denmark) in the central channel of the probe (Fig. 1).

Figure 1.

 Schematic illustration of the experimental setup. The computer (PC) was connected to the syringe pump controlling the flow rate as a push/pull system. The water bath was designed to heat water to a predefined temperature. Water running through the pipes in the water bath was not in direct contact with the heated water. Changing the flow rate changed the temperature in the bag. The arrows indicate the flow direction of the water. The temperature sensor is positioned inside the bag.

Thermal stimulation

Heat stimuli were applied by recirculation of heated water in the bag. Water was heated to 63 °C in a special designed temperature controlled water bath. A software program (Openlab; GMC APS, Hornslet, Denmark) was used to control an infusion/withdrawal syringe pump (PHD 22/2000; Harvard Apparatus, Holliston, MA, USA) through a RS232 interface (Fig. 1). The pump held two 140cc syringes for infusion and withdrawal of water from the bag and the volume in the bag was kept constant by withdrawing the same amount of fluid as infused.

In vitro pilot studies confirmed that water leaving the water bath was heated to 63 °C at the maximum flow rate of the syringe pump (200 mL min−1). The tube from the water bath to the probe was 20 cm and due to cooling from the surroundings the temperature of the water entering the bag depended on the flow rate. In vitro tests were used to find the optimum flow rate functions providing linear temperature incline at different rates. To mimic the in vivo conditions, the probe was submerged into a water bath heated to 37 °C. A constant flow rate resulted in non-linear temperature changes. To create a linear temperature change an exponential software function controlled the flow rate


A, B and C were constants determining the flow rate and hence the temperature ramp. The constants resulting in linear temperature inclines of ∼0.2 °C s−1 (slow), 0.5 °C s−1 (medium) and 2 °C s−1 (fast) in the bag were determined in vitro. An upper temperature limit of 60 °C was chosen following a series of pilot studies in healthy volunteers. All subjects tolerated heat stimulations at 60 °C without reporting any adverse effects.

Endoscopic examination

For each subject the oesophageal mucosa was examined before experimental testing and the oesophago-gastric junction was identified to ensure exact positioning of the probe. Three separate video sequences were recorded before and after acid sensitization and stored for later documentation.

Sensory assessment

Prior to the study, all subjects were instructed to use the 0–10 electronic modified visual analogue scale (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. For painful sensations, the subjects used the scale from 5 to 10 anchored at 5 = pain detection threshold; 6 = slight pain; 7 = moderate pain; 8 = medium pain intensity; 9 = intense pain and 10 = unbearable pain. The scale has been described in details previously proving to be reliable in discriminating sensations in the oesophagus with high accuracy 1,7,8 The subjects were instructed to score the evoked sensations during the thermal stimulations and to differentiate this from unpleasantness in the mouth and throat.

Study protocol

A schematic illustration of the study protocol is shown in Fig. 2. Subjects fasted at least 6 h prior to experimental testing. The probe was inserted through the mouth. The oesophageal mucosa was inspected with the fibre-video endoscope to ensure that no pathological conditions were present. To maintain an exact probe position the oesophago-gastric junction was identified and the bag was positioned 10 cm above the junction. The distance to teeth from the tip of the fiberscope was noted and the probe was taped to the chin. Three video sequences of the oesophageal mucosa were recorded for documentation. The bag was filled with 5 mL of water at a temperature of 37 °C to ensure sufficient mucosa contact, but still avoiding mechanical stimulation as it was not felt by any subjects. Heat stimulation was started with preconditioning using the medium ramp 0.5 °C s−1. This was performed twice to familiarize the subjects with the stimulation. After preconditioning, the subjects were stimulated with the three different ramps of increasing temperature rate: 0.2 °C s−1 (slow), 0.5 °C s−1 (medium) and 2 °C s−1 (fast). Subjects were blinded to the sequence of the ramps, but to ensure continuous VAS registrations, subjects were informed when heat stimulations were started. The stimulations were interrupted when subjects reached the pain detection threshold (PDT). Baseline temperature was re-established between stimulations.

Figure 2.

 The protocol used for thermal stimulation of the human oesophagus before and after sensitisation with acid. Endoscopy of the oesophageal mucosa was done prior to experimental stimulation by using a small endoscope in the centre channel of the probe. The heat stimulations were applied as a total of five subsequent stimulations separated by a pause to revert to baseline temperature. The first two stimulations were preconditioning stimuli (0.5 °C s−1) made to familiarize subjects to the method. Thereafter, a slow ramp stimulation (0.2 °C s−1) followed by a medium ramp stimulation (0.5 °C s−1) and a fast ramp stimulation (2 °C s−1) was applied. Stimulations were terminated at the pain threshold. After the first series of stimulations the subjects underwent a modified acid perfusion test. Here the distal oesophagus was perfused by 0.1 mol L−1 hydrochloric acid for 30 min and the endoscopy and the thermal stimulations were repeated. To assess day-to-day reproducibility all subjects underwent the protocol on two subsequent days separated by at least two weeks. Numbers are minutes from the beginning of the experiment.

Acid perfusion of the distal oesophagus was performed after the first series of thermal ramps. Hydrochloric acid (HCL 0.1 mol L−1, Bie & Berntsen, Rødovre, Denmark) was infused through one of the perfusion channels in the probe at a rate of ∼7 mL min−1 for 30 min, in total 200 mL.9 If any painful sensations due to acid stimulation were reported, the perfusion was temporarily stopped and the subjects were allowed to drink 10–20 mL water. Following acid perfusion, video recordings and the thermal stimulations were repeated. In total, the study period lasted for 2 h.

To assess day-to-day reproducibility of the method subjects completed the study two times with a minimum of 2 weeks apart.


Descriptive statistics are reported as mean ± SD. The responses to heat stimulations were compared with two-way anova. Normality was checked by QQ-plots and the assumption of variance homogeneity by Levene’s test. < 0.05 was considered statistically significant.

Intra-class correlation (ICC) was calculated for measurement of intra-individual variance. Intra-class correlation evaluates each subject’s ability to reproduce a response between days of study and illustrates the difference between a subjects score at repeated experimental sessions. The ICC was calculated as


where σwithin is the variance within individuals and σtotal is the sum of variance within individuals and between the repetitions of heat stimulations. The parameter ranges from 0 to 1, with values closest to one indicating the highest reproducibility. Acceptable levels of ICC was set to >0.6. The software package stata version 10.1 (StataCorp LP, College Station, Texas, USA) was used for the statistical calculations.


All subjects completed the experiment. Two reported slight soreness on swallowing the following day. One subject reported heartburn on the following day. Otherwise no adverse effects were observed.

In vitro experiment

Flow rates above 150 mL min−1 changed the volume in the bag during stimulations. To avoid this, the maximum flow rate was set to 130 mL min−1. The experiments showed that the flow rate equation for the slow ramp was: exp (0.025 × t) × 1 + 8, for the medium ramp: exp (0.07 × t) × 1.5 + 12 and for the fast ramp: exp (0.15 × t) × 10 + 20. These equations resulted in linearly increasing temperature ramps. In a subsequent experiment the rate of temperature increase measured in vitro using these equations was 0.18 °C s−1 ± 0.08 for the slow ramp, 0.51 °C s−1 ± 0.10 for the medium ramp and 2.40 °C s−1 ± 0.52 for the fast ramp.

Endoscopic examination

Normal mucosa was found in all volunteers prior to experimental stimulation and after the acid perfusion. Especially, no obvious changes were (e.g. redness) seen after the acid perfusion.

Reproducibility experiment

The overall reproducibility of the method is reported in Table 1. All baseline assessments showed good reproducibility between days (ICC >0.6).

Table 1.   Intra-class correlation coefficient between days
RampBaselineAfter acid
  1. AUC, Area under temperature curve.


Heat stimulation and acid sensitization

It was possible to control the temperature rates for the three different heat stimuli and subjects were able to rate the evoked pain simultaneously on the VAS without knowing the predefined rates of temperature change. All subjects reported a hot and burning sensation during the heat stimulations. Fig. 3 shows representative tracings from a single subject.

Figure 3.

 A representative stimulation trace from a single subject. The temperature rate for the three different temperature ramps and the simultaneously recorded pain rating on the visual analogue scale are illustrated.

The average areas under the temperature curve (AUC), time to PDT and temperature at PDT before and after acid perfusion are reported in Table 2. The first day a decrease in AUC (= 5.17; = 0.03) was observed after acid perfusion. The decrease in time to PDT (= 2.47; = 0.12) and temperature at PDT (= 2.66; = 0.11) were not different on day one. The second day acid perfusion resulted in a decrease in AUC (= 14.38; = 0.0003) and in time to PDT (= 8.51; = 0.005) and temperature at PDT (= 18.55; = 0.0001). The responses to acid sensitization are illustrated in Fig. 4. The slow ramp showed the widest dynamic range with a decrease in AUC of 29% following acid perfusion. For the medium and fast ramps the range was 20% and 19% respectively.

Table 2.   Heat stimulation of the human oesophagus by three different temperature ramps before and after acid perfusion
DayRampBaselineAfter acid
AUC (°C*s)Time at PDT (s)Temperature at PDT (ºC)AUC (°C*s)Time at PDT (s)Temperature at PDT (°C)
  1. Data from day one and two are reported separately as mean ± SD.

  2. AUC, Area under the temperature curve; PDT, Pain detection threshold.

  3. *Statistically significant difference after acid perfusion compared with baseline (P < 0.05).

1Slow859 ± 417125 ± 2353 ± 4768 ± 374*121 ± 2452 ± 3
Medium382 ± 7550 ± 456 ± 1318 ± 133*46 ± 854 ± 3
Fast200 ± 4219 ± 259 ± 1150 ± 50*18 ± 358 ± 2
2Slow1121 ± 20139 ± 1356 ± 2795 ± 235*121 ± 16*53 ± 2*
Medium431 ± 10153 ± 557 ± 1343 ± 129*49 ± 7*55 ± 2*
Fast185 ± 6119 ± 358 ± 1150 ± 5018 ± 2*58 ± 1*
Figure 4.

 The response to acid sensitisation in healthy subjects for the first (top) and second (bottom) stimulation experiment. On the first day the response to acid was variable with some subjects being sensitized whilst other subjects were desensitized. At the second day a uniform response was seen with sensitisation in all subjects. Data from the slow ramp stimulations (0.2 °C s−1) are illustrated.


A new method for linear increasing heat stimulation in the human oesophagus was developed and tested in 12 healthy subjects with a multimodal probe combined with endoscopy. The mucosa was normal in all subjects and no difference in baseline recordings between days was found. The response to acid was highly variable on the first day, but on the second day all subjects were sensitized. The slow ramp was the most sensitive, showing a decrease in AUC of 29% following acid sensitization.

Previous methods for heat stimulation in the GI tract

The earliest methods for heat stimulation in the human GI tract were based on recirculation of heated water in a bag. The stimulation was performed by using a push-pull system controlled manually by the researchers. This resulted in varying temperature rates and fluctuations in the temperature curve.10 Later methods were designed to recirculate the water using a peristaltic pump with measurement of the temperature of the water inside the bag. As the temperature was fluctuating during the experiment the AUC was used as a proxy for the mean thermal energy load.9 In the rectum, another method for thermal stimulation was based on the peltier device heater. It allowed control of the temperature and location and avoided heat stimulation of the anal canal. However, problems to tolerate the stimulation due to the stiff and uncomfortable instrument were reported and it was difficult to ensure continuous mucosa contact during stimulation.11 Although reliable and reproducible, these methods did not allow control of the temperature rate or visual inspection of the organ prior to stimulation. Taken together with the above mentioned practical problems better control of visceral thermal stimulations are needed.

Methodological considerations

The method described in this study was able to control the temperature rate. The pain following slow ramp stimulation was described as hot and burning. In the skin, C-fibres typically encode such sensations reflecting that the majority of the fibres in the oesophageal mucosa are unmyelinated.12

The baseline recordings were reproducible for the three different temperature ramps. With respect to the intra-individual variability, the most reproducible stimulus was the slow ramp. This finding supports previous published data obtained from heat stimulation in the skin were high rates of temperature changes (2 °C s−1) showed a 75% larger variability than low rates (1 °C s−1).4 Furthermore, from a physiologic perspective the slow ramp is the most appealing as it has the widest dynamic range following acid perfusion and evokes a burning pain quality like that reported by patients with GORD.5

In previous studies, painful heat levels could not always be reached. As the temperature was rapidly increased a thermal limit of 50 °C was used not to harm the mucosa. Therefore, the exact temperature at the pain threshold could not be assessed as the stimulation was stopped before reaching the pain threshold.9,13 In this study, the slow ramp allowed a linear increasing heat stimulus with only minor fluctuations. The subjects could feel the temperature increase during the whole stimulation and thus stop the stimulus at the pain threshold. Three different parameters was measured for each stimulation; AUC, time to PDT and temperature at PDT. The AUC could be used as a measure of the subjects’ heat tolerance and the temperature at the pain threshold could be assessed. The AUC combine information on the temperature and time of stimulation and therefore represent a very responsive parameter. Furthermore, the statistical properties of the AUC are compelling.14 For these reasons, we suggest using the AUC as the primary efficacy parameter for upcoming studies.

The response to acid perfusion of the oesophagus is highly individual and typically only 75% of the subjects are sensitized to experimental stimulation after acid perfusion.15,16 In this study, the response to acid was highly variable at the first day. Sensitization was seen in most subjects, but some subjects were desensitized following the acid perfusion. Consequently, only changes in the AUC reached statistically significant levels. Further, the responses to heat stimulations were more variable on the first day compared with the second day. Recently, Paine et al. reported a relationship between personality and pain-related changes in autonomic variables following distal oesophageal distensions.17 These findings could explain the diverse response to acid perfusion on the first day, although this study was not designed to evaluate this. Interestingly, the response to acid sensitization was more uniform at the second day with all subjects being sensitized after acid perfusion. This was also depicted in the statistical analysis were all parameters reached statistically significant changes after acid perfusion. An explanation for the difference in sensitization between days could be a ‘carry over effect’ due to long term sensitization induced by acid perfusion at the first day. However, the two experiments were done at least 2 weeks apart and no significant difference in baseline values between the first and second day was observed. Another explanation could be a training effect in the individual subject. It is most likely that the individual differences in personality, anxiety and other confounding variables would be counterbalanced by training the subjects, making the pain scoring at the second day more consistent. Therefore, we suggest doing learning experiments in upcoming studies to familiarize the subjects to the stimulations. Although not addressed in this study, such learning experiments are normally used in experimental pain experiments to ensure uniform pain ratings in the subsequent study sessions and limit the intra-individual variation.13

Endoscopic perspectives

Oesophageal pathology in terms of erosions and ulcerations are a very common finding in asymptomatic patients undergoing upper GI endoscopy for other indications.2 In previous visceral pain models, the mucosa could not be examined prior to stimulation. Hence, otherwise healthy subjects could have mucosal pathology without overt symptoms. In this study, an endoscopic examination was applied in all subjects prior to heat stimulation. The mucosa was inspected for pathology and the position of the probe could be verified during the experimental testing. The endoscopic procedure was robust and easy to implement although it required some experience in upper GI endoscopy. For upcoming studies, the endoscopic method could be further developed. Previously, a probe was designed to study mucosa blood flow by using a laser probe18 and cross-sectional ultrasound was used for biomechanical imaging.19 In addition, biopsies could make material available for histological evaluation. Combining these techniques with the present method would offer a new and comprehensive setup for future studies of the upper GI tract.

Clinical and pharmacological perspectives

The novel method for heat stimulation could be used to study pathophysiology or pharmacological interventions in healthy volunteers. It may also have applicability in patients with diseases involving the upper GI tract. For example, patients suffering from GORD show hypersensitivity to experimental heat stimulation of the oesophagus.20,21 This is considered to be due to peripheral and central sensitization of the nervous system.22 The response to heat stimulation could serve as a test to discriminate between peripheral and central pain components. Hence, patients showing a predominant response to heat stimuli in a specific range would likely suffer from peripheral hypersensitivity due to acid sensitization of the vanilloid receptor (TRPV1 receptor). By contrast, it is most likely that patients with central sensitization would be hypersensitive to all stimuli and not specifically to heat pain (>43 °C).23 The present method would make it possible to test such hypothesis in a controlled study design; although the application of the method in oesophageal diseases will require further studies.

The TRPV1 receptor is widely discussed and is believed to have a major influence in visceral hypersensitivity and pain.24,25 Pharmacological blockade of the TRPV1 receptor represents a new strategy in pain treatment and controllable thermal stimulation is essential when the effects of new potential TRPV1 antagonist are studied.26 The method presented in the present paper provides such a controllable thermal stimulus usable in the gut and could be used in forthcoming drug studies of TRPV1 antagonism.

In conclusion, the new endoscopic method for linear control of heat stimulation was controllable and reproducible. For upcoming studies, we suggest to use the slow temperature ramp and to do a training experiment to familiarize subjects to the stimulations. The method should be used in future basic experiments, drug and clinical studies as it provides a differentiated and controllable stimulation with the possibility to take biopsies and to activate different receptor populations and fibre types.


This study was supported from Nordjyllands Lægekredsforenings forskningsfond, Karen Elise Jensens Fond, Det Obelske Familiefond and EU FP7 Diamark grant agreement #223630. The authors would also like to thank Olympus Denmark for providing the endoscope.