Article Text

Download PDFPDF

The use of hypnosis in gastroscopy: a comparison with intravenous sedation
  1. Philip Conlong,
  2. Wynne Rees
  1. Hope Hospital, Eccles Old Road, Salford, Manchester M6 8HD, UK
  1. P Conlong, 14 Lidgate Grove, Didsbury, Manchester M20 6TS, UK


A total of 124 subjects who were undergoing routine endoscopy were randomly assigned to one of three groups. All three groups received lignocaine throat spray. The first group additionally received midazolam, the second received hypnosis, whilst the third only received lignocaine throat spray. Although hypnotised patients were deemed by an independent observer to be less agitated than the other two groups (p<0.03), they reported the gastroscopy to be significantly more uncomfortable (p<0.042) and scored higher in their memory for the procedure (p<0.001). They also took slightly longer to induce than the midazolam group. The midazolam group on the other hand rated the procedure as significantly more comfortable although paradoxically were seen by an independent observer as being more agitated. They were also significantly more amnesic. The endoscopist encountered more procedural difficulties with this group but this did not reach levels of significance. Hypnosis was not shown to be an effective alternative to intravenous sedation in gastroscopy.

  • hypnosis
  • gastroscopy
  • sedation
  • midazolam

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

Short-acting benzodiazepines can cause significant morbidity and mortality. Cardiorespiratory problems account for 50% of the morbidity and 60% of deaths following endoscopy.1-9 Most endoscopists use sedation2 although an increasing number offer the alternative option of avoiding it altogether with obvious benefits to the patient, including quicker recovery times and reduced time off work. With these thoughts in mind we wanted to investigate the possibility of using a different approach to sedation.

Although hypnosis has been used in dental practice, obstetrics and surgery,10-15 it has not been used in gastroscopy. We wanted to find out firstly whether hypnosis could be used in endoscopy to provide the required level of sedation and amnesia without some of the problems that can be associated with midazolam, and secondly whether it offered any apparent advantage over and above a lignocaine throat spray.


A total of 124 patients were recruited from patients needing a diagnostic gastroscopy at the Royal Oldham Hospital in Manchester. The study group comprised 72 males and 52 females with an average age of 58 years (range 19–92). The study was approved by the local ethics committee.

Patients were randomised to one of three groups, all of which received lignocaine throat spray. Group 1 also received midazolam, group 2 received hypnosis, and group 3 only had the lignocaine throat spray. Nine patients were excluded; seven who could not understand the instructions or complete the questionnaire and two with unstable angina. Of the remaining patients, 45 had cardiorespiratory problems requiring medication; 18 of these had chronic obstructive pulmonary disease, 16 had angina, five had suffered myocardial infarcts more than 6 months previously and six had asthma. There were 46 patients in the sedation group (24 men), 45 in the lignocaine group (20 men) and 33 in the hypnosis group (13 men).

Following randomisation to these groups the patients were then randomly assigned to positions on a busy gastroscopy list with an average of 10 patients per session. The procedure was explained to them by a nurse who also explained the nature of the grouping to which they had been allocated.

The doctor carrying out the hypnosis was fully trained in the technique. All patients were taken individually into a separate anteroom next to the endoscopy suite. They all had a venflon inserted and lignocaine spray to the hypopharynx. The midazolam group were then sedated by titrating the dose to produce ptosis and drowsiness. Those having hypnosis were induced by one of the authors, using a fixation technique. This involved looking at a blue light from a spotlight directed onto a wall opposite the patient. The hypnotist was allowed 5 minutes from the time the venflon was inserted to adequately hypnotise the patient. This was assessed by a nurse present in the room. A breathing technique was then used to induce a progressively deeper hypnotic state. This involved giving suggestions to the patient of increasing drowsiness and heaviness of the eyes whilst slowly counting to 20 in time with inspiration. The patient was then asked to close their eyes. Further suggestions were given to the patient to produce increasing drowsiness. The hypnotist used the onset of a deeper respiratory pattern and the development of changes in muscular tone to deduce that the patient was hypnotised. At this point a suggestion was given to produce anaesthesia of the throat and oesophagus.

Patients were then individually brought into the endoscopy suite. An endoscopist and nurses present were completely unaware of the grouping into which the patients had been assigned. Both endoscopists used in the study had been trained in gastroscopy.

At this point readings were taken of the patients' pulse and oxygen saturation and the gastroscopy was carried out. Further readings were taken at intubation and upon completion of the procedure. The endoscopists instructed the patient to swallow the endoscope on intubation.

After completion of the gastroscopy a nurse filled in a ‘level of agitation’ score from 1 to 5 using a visual analogue scale: the value 1 being assessed as ‘quiet throughout’; 3 as ‘agitated’ and 5 as ‘extreme agitation’. The endoscopists also completed a form describing any difficulty they might have had with the procedure. Hypnotised patients were ‘reversed’ with a standard counting method, which involved the hypnotist counting from 20 to 1 in time with the patient's inspiratory rate; and then finally commanding them to awaken. After arrival in the recovery area, patients were given a questionnaire which asked them to assess the comfort of the procedure on a visual analogue scale from 1 ‘very comfortable’, to 5 ‘very uncomfortable’. Once the patients were fully orientated, they were asked a series of questions to determine what they could remember about the procedure (including whether they could recall being taken into the endoscopy unit, the insertion of the venflon, intubation, and being taken back to the recovery room).


The hypnosis group had a significantly lower level of agitation with a mean score of 1.57, compared to the sedation group (2.14) and the lignocaine group (2.30) (p < 0.03, ANOVAR) (table1).

Table 1

 Comparison of group scores

There were no significant differences observed in the time taken to induce the patient. After inserting the venflon, the midazolam group took an average of 4 min compared with 4 min 52 s for the hypnosis group.

There were no significant differences in the total time taken for the gastroscopy; with the hypnotised group completed in an average of 9 min 25 s, the sedated group 8 min 27 s, and the lignocaine group 8 min 36 s.

No significant differences were found in oxygen saturation. It was satisfactory (ie, at least 95%) in 27% of the hypnotised group, 32% of the sedated group and 34% of the lignocaine group. Most of the patients who had an unsatisfactory oxygen saturation developed this at intubation. Although there were differences between the groups in the number of difficult endoscopies, these did not reach levels of significance. There were problems with three patients in the hypnotised group; 12 in the sedated group, and 10 in the non-sedated group. The vast majority of these problems were intubation-related.

There was a significant difference in levels of comfort ratings between the three groups. The sedation group scoring significantly better (2.90) than the hypnosis group (3.57) or the lignocaine throat spray group (3.65) (p<0.042, ANOVAR). There were no significant differences between the latter two groups.

A significantly smaller number of sedated patients (27) had full recall of the gastroscopy compared to the hypnotised group (28) and the lignocaine group (45) (p<0.001, χ2). There were no significant differences in the abnormalities found at gastroscopy between the three groups (table 2).

Table 2



The group receiving midazolam outscored the hypnosis group in reported levels of comfort for the procedure, the degree of amnesia, and the time taken for induction, the first two items reaching statistical significance. They only did poorly on two measures: the number of patients having a less than satisfactory oxygen saturation, and those that had a difficult endoscopy, although neither of these reached statistical significance.

The hypnosis group performed similarly to the lignocaine throat spray group, only managing to score significantly better in the nurse's score of levels of agitation during the procedure. This seems at odds with most of the patients' reports of poor levels of comfort for the procedure. It also took slightly longer for the hypnotist to achieve what he thought was a hypnotised state in comparison to midazolam. The endoscopist, however, reported less difficulty with the endoscopy in this group, although there was no statistically significant difference compared to the other two groups. The hypnotised group scored only slightly better that the lignocaine throat spray group on amnesia for the procedure, and considerably worse than those receiving midazolam.

There would seem to be two ways to interpret the apparent failure of the hypnosis group. Firstly, it is possible that most of the patients were not in the required hypnotic state. In the limited time available to hypnotise the patient prior to gastroscopy, some may simply have been relaxed. This would explain the very good scoring on the agitation scale assessed by the nurses in the endoscopy unit, the relative ease of the gastroscopies in this group, the fact that only some of the patients achieved amnesia for the procedure, and the higher subjective levels of discomfort. An alternative hypothesis is that hypnosis simply does not work at all and that patients in this group were all just acting what they assumed would be a hypnotic state, therefore producing only marginally better scores than the group receiving only lignocaine.

Reviewing the evidence on hypnosis, there are several studies that suggest it is of use across a broad range of specialties.10-15 If one accepts the premise that hypnosis works, albeit that the mechanisms involved are unclear, then one has to ask what happened in this study. The most likely explanation is that the hypnotist did not have enough time to produce a hypnotic state in all the patients. Judging depth of hypnosis is difficult and one weakness of this study is that we did not ask the hypnotist afterwards to assess his satisfaction with this in each subject prior to the gastroscopy. One improvement to the study design would be to allow more time for hypnosis. This would make it easier to carry out accepted tests on the adequacy of hypnosis such as changes in sensory perception. This of course begs the question of how useful hypnosis would be on a busy endoscopy list. If one could conclusively show that it worked, it might well have a use in patients with cardiorespiratory disease in whom the risks of using benzodiazepines are higher.

In conclusion, we have not found hypnosis to be of value in gastroscopy, although this may have been because the hypnotist did not have enough time available to induce the patient fully.