Article Text

Download PDFPDF
Midazolam in conjunction with local anaesthesia is superior to Entonox in providing pain relief during bone marrow aspirate and trephine biopsy
  1. G Chakupurakal,
  2. J Delgado,
  3. E Nikolousis,
  4. S Pitchapillai,
  5. D Allotey,
  6. K Holder,
  7. L Bratby,
  8. J de la Rue,
  9. D W Milligan
  1. Department of Haematology, Heart of England Hospital, Birmingham, UK
  1. Dr D W Milligan, Department of Haematology, Birmingham Heartlands Hospital, Birmingham B9 5SS, UK; d.w.milligan{at}


Aim: To compare intravenous titrated midazolam 5–10 mg and inhaled Entonox in addition to local anaesthesia in order to identify which agent provides optimum pain relief.

Methods: Randomised, controlled trial. 49 patients were recruited, of which 46 were evaluable. 24 and 22 patients were recruited into the Entonox and midazolam arms, respectively. Patient experiences as well as staff observations were recorded with questionnaires after recovery from the procedure and 24 hours later.

Results: 45% and 59% of the patients in the midazolam arm could recollect the procedure after 15 minutes and 24 hours, respectively, compared to 96% and 88% who received Entonox. Midazolam provided a more comfortable experience (p<0.01) and improved pain relief (p = 0.01) compared to Entonox immediately after the procedure; this further improved when recalled 24 hours later. Nausea, dizziness and hallucinations were observed with both treatments, but dizziness was significantly more frequent with Entonox (p = 0.048). Clinically relevant respiratory depression (O2 saturation <90%) occurred in 19% of patients in the midazolam arm; sedation was reversed with flumazenil.

Conclusion: Midazolam in conjunction with local anaesthesia provides rapid and reversible sedation as well as effective pain relief during bone marrow biopsy, and is superior to Entonox; however, care must be taken to monitor respiratory function.

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.

Take-home messages

  • Bone marrow biopsy is viewed with trepidation by many patients; local anaesthesia alone may provide inadequate analgesia.

  • The addition of either Entonox or a benzodiazepine reduces the pain, but it is not clear which is the best agent.

  • In this randomised trial, midazolam was significantly superior to Entonox, but care must be taken to monitor oxygen levels.

The technique of bone marrow aspiration and trephine biopsy (BMAT) has a central role in the diagnosis of haematological disorders. The procedure involves infiltration of the skin and periosteal lining over the posterior iliac crest with local anaesthetic, followed by the insertion of a needle into the bone cavity. In a bone marrow aspirate, about 5 ml of liquid marrow is aspirated, but a trephine biopsy requires the use of a larger needle and the removal of a 2–3 cm core of bone containing marrow. Although most patients tolerate an aspirate procedure with local anaesthetic alone, many find the biopsy test painful.1 Studies using a combination of local anaesthetic agents and benzodiazepine sedatives2 have demonstrated that this makes the procedure more tolerable, and for a number of years intravenous midazolam in addition to local anaesthesia has been offered to patients to reduce distress as routine practice in many centres in the UK.

Midazolam acts mainly as an amnesic and sedative rather than an analgesic. Hence although pain is experienced during the procedure, there is limited recall afterwards. This relieves patient anxiety and apprehension, particularly when the test needs to be repeated. Its ease of use, safety and efficacy, together with availability of flumazenil as an antidote, render it a useful agent to be used as sedative prior to BMAT.3 Despite this profile its major pitfall is possible respiratory depression, highlighted in the guidance by the Royal College of Anaesthetists.4 To minimise this risk, respiratory function must be monitored during the procedure and full resuscitation facilities must be available. Flumazenil rapidly reverses midazolam induced sedation but its duration of action is short, which may lead to re-sedation after the patient has left hospital. Other side effects of midazolam include dysphoria, hallucinations and agitation during the procedure.

Nitrous oxide was identified by Davy in 1800 and has been used widely in a 50:50 mixture with oxygen (Entonox) since the 1960s in obstetrics to provide analgesia during labour.5 6 Entonox is delivered from a small cylinder through a demand valve using a mouthpiece, permitting patient self-administration. If the patient becomes drowsy, the airtight seal necessary for gas flow between the face and mask is broken, disrupting drug delivery. This also makes respiratory depression an unlikely side effect. Low fat solubility and low anaesthetic potency allow speedy recovery as soon as the patient’s breathing becomes shallow. Its strong analgesic and minimal amnesic effects have been shown to relieve distress in a variety of situations, including postoperative pain relief, crushed chest injuries, colonoscopy and burns dressings.7 8 It is currently used in accident and emergency departments and paediatric departments to provide pain relief, and can be safely administered by midwifes, nurses and paramedical staff. Side effects include feeling drunk or dizzy, nausea and sickness, with prompt recovery from these due to short duration of action. The safety profile and ease of administration make Entonox a good alternative to midazolam, and many units are currently using it as an analgesic for BMAT.

We conducted a randomised controlled trial to identify whether Entonox is superior to midazolam in providing patient comfort when given in conjunction with local anaesthesia. Patient satisfaction was assessed with the help of pain scores immediately after and 24 hours after the procedure. We also assessed staff observations with the help of questionnaires.


Study design

A single centre randomised controlled trial was conducted at our institution. The Northern Ireland Multi Ethics Research Committee and MHRA (Medicine and Health products Regulatory Agency) approved the trial protocol. The power calculations for the sample size were based on a previous study of lorazepam.9 This assumed that it would be difficult to demonstrate a superiority of Entonox over midazolam at 24 hours but relatively easy immediately post-procedure, where midazolam may be less effective. Patients who required BMAT as part of the investigations for a suspected underlying bone marrow disorder were recruited. All gave written informed consent and the trial was conducted in accordance with the Declaration of Helsinki. Randomisation was carried out centrally by a member of the Trust Research and Development Department. The study was powered to have an 80% chance to demonstrate non-superiority and a 90% chance of showing a 30% reduction in mean pain scores between the treatments with 42 analysable patients. The data obtained were analysed with SPSS software, using the Wilcoxon, Mann–Whitney and Fisher exact tests; p values were considered statistically significant when <0.05.

The primary endpoint was to determine which agent produced superior analgesia immediately after the test and 24 hours later, and resulted in the least unpleasant experience for patients. The secondary endpoints were to determine the agent which results in fewest side effects, and required least occupancy of day unit beds and nursing time. Questionnaires were devised for the patients and staff performing the procedure to obtain the information.


Participation in the trial was offered to patients requiring a BMAT. Patients were aged 18 or older, able to give informed signed consent, fully conscious, able to understand the procedure and cooperate with Entonox inhalation, and had no major psychiatric disorders. Patients were excluded if they declined sedation. Other exclusion criteria were: chronic obstructive pulmonary disease; myasthenia gravis; recent head injury; middle ear injury; pregnant or breast feeding; history of recent diving; or unable to give informed written consent.

The nursing staff on the day unit performed the procedures. They were competent in administering midazolam intravenously as well as performing bone marrow aspiration and trephine biopsies under local anaesthesia. Midazolam was given by titrated slow intravenous injection; boluses of 2 mg were given every 2 minutes until conscious sedation was achieved. The maximum dose was 10 mg. Oxygen saturation was monitored during the procedure and staff performing the procedure were trained in basic life support. Access to full resuscitation was also available.

For the purpose of this trial, the nursing staff were trained to administer Entonox and trained obstetric staff certified their competence. The patients were asked to complete a questionnaire once they had recovered from the procedure and 24 hours later. This included the estimation of pain on an analogue pain scale. The staff performing the procedure also answered a questionnaire that identified the time taken and recorded whether flumazenil was required to reverse prolonged sedation or hypoxia.


Fifty patients were entered into the trial. One randomisation was unblinded in error, hence data from 49 patients were analysed. We did not receive questionnaires from one patient recruited in each arm, and one patient in the Entonox arm withdrew after randomisation. Of the 46 evaluable patients, 24 and 22 patients were recruited into the Entonox and midazolam arms, respectively. The median age was 61 years (range 30–90); 13 patients were female. Fifty-three per cent of patients had undergone a previous bone marrow biopsy (table 1).

Table 1 Patient demographics


Overall, 13/46 (28%) patients had no recollection of the BMAT 15 minutes later. However, only 1 patient (4%) in the Entonox arm could not recall the procedure compared to 12 patients (55%) receiving midazolam (p<0.001, Fisher’s exact test). This difference was still significant 24 hours after the procedure, when no patient (0%) receiving Entonox failed to recall the procedure compared to 7/20 patients (35%) in the midazolam arm (p = 0.003, Fisher’s exact test).

The subjects’ pain perception was scored on a scale from 0 (no pain) to 10 (extreme pain) at 15 minutes and 24 hours (fig 1). Median pain perception at 15 minutes was 0 (range 0–6.3) in the midazolam arm and 1.9 (range 0–9.2) in the Entonox arm (p<0.01, Mann–Whitney test). At 24 hours, pain perception remained significantly lower (median 0, range 0–7) for patients receiving midazolam compared to patients receiving Entonox (median 1.6, range 0–9; p<0.01, Mann–Whitney test). Indeed, only one patient in the midazolam arm experienced significant pain at both time points (pain scores 6.3 and 7 immediately after and 24 hours later, respectively). There was no significant difference in the pain scores for either modality immediately after the test and 24 hours later. The patients’ subjective experience of how pleasant they perceived the overall experience was scored on a scale from 1 (not unpleasant) to 5 (most unpleasant) (see fig 2 and table 1). At 15 minutes, patients in the Entonox arm perceived the procedure as being significantly more unpleasant than patients in the midazolam arm (p = 0.015, Mann–Whitney test). Furthermore, this difference was still evident 24 hours after the procedure (p = 0.033, Mann–Whitney test).

Figure 1 Pain experienced at 15 minutes and 24 hours in Entonox and midazolam arms, measured on a scale of 0–10.
Figure 2 Unpleasantness of procedure experienced at 15 minutes and 24 hours in Entonox and midazolam arms, measured on a scale of 0–5.

Side effects

The median duration of the procedure was 15 minutes (range 7–30 minutes) in the Entonox arm and 25 minutes (range 10–45 minutes) in the midazolam arm (p = 0.024, Mann–Whitney test). The mean dose of midazolam was 6 mg (range 2–10 mg). The weight of patients in the midazolam arm was 60–90 kg with a mean of 74 kg (data on six patients not available). Variability between operators in the dose administered and time of administration was not recorded.

Three side effects were specifically documented on a scale from 0 to 5: nausea, dizziness and hallucinations. Although all three side effects were observed in both arms, dizziness was significantly increased (fig 3) in patients receiving Entonox (p = 0.048). Moreover, 4/21 patients (19%) receiving midazolam had significant oxygen desaturation (<90%) during the procedure. Flumazenil was administered to 10/21 (48%) patients in the midazolam arm: for oxygen desaturation in 3, and for prolonged sedation in 7. The mean dose of midazolam in these 10 patients was 6.3 mg (range 2.5–10 mg); their mean weight was 74 kg (range 60–90 kg).

Figure 3 Dizziness measured at 15 minutes post-procedure, measured on a scale of 0–5.


Entonox has achieved popularity as an analgesic for a variety of outpatient procedures. Apart from being a potent analgesic, the drug has little risk of respiratory depression or rebound sedation on discharge, and reduces duration of bed occupancy. Its ease of administration and safety make it an attractive analgesic for minor procedures. Two recent studies have identified Entonox to be superior to local anaesthesia alone in improving patient comfort during the procedure.10 11

This is the first randomised controlled trial comparing the efficacy of midazolam to Entonox in conjunction with local anaesthesia for BMAT. Although midazolam is not strictly an analgesic (patients receiving midazolam will still show physiological responses to pain), it is the recall of pain after a procedure that distresses patients. In our hands midazolam was preferred by the patients, both for the reduced pain and for their overall experience. This is contrary to a recent publication,12 comparing pain perception using Entonox to previous experience while using midazolam, which found that 94% of the patients felt that Entonox was better or equal to midazolam. This study was not randomised and not prospective, relying on patients’ recollection of a procedure carried out in the past.

Despite the proven efficacy of Entonox in outpatient procedures, it has some limitations. Patients may not use the apparatus properly, either due to technique or anticipation of pain.13 Studies in obstetrics have revealed that up to 30% of the patients have inadequate pain relief when Entonox is administered to relieve labour pain.14 Midazolam, on the other hand, is a potent sedative.1518 In this study the majority of patients receiving midazolam felt that the procedure was tolerable and, apart from one patient, all patients had adequate pain relief (defined as pain score <2.5 on a scale of 1–10) on recollection. We found that midazolam was superior in alleviating pain or pain recollection during bone marrow biopsy, but its use is not without hazard. Deaths have been reported in other procedures after midazolam treatment, even in patients undergoing conscious sedation.19 In this study clinically important hypoxaemia was reversed by flumazenil in three patients (13%); the use of midazolam sedation requires the full precautions described by the Royal College of Anaesthetists to be taken. We conclude that midazolam in conjunction with local anaesthesia is superior to Entonox in alleviating pain during BMAT and in keeping the patient comfortable. However, it needs to be used with care to avoid the risk of respiratory depression.


We would like to thank all patients who participated in this study and the nurses in the Haematology Day Unit.



  • Competing interests: None.

  • Ethics approval: The Northern Ireland Multi Ethics Research Committee and MHRA (Medicine and Health products Regulatory Agency) approved the trial protocol.