Optimum pain management following Caesarean section remains unresolved. Intramuscular (i.m.) opioids, patient controlled analgesia systems and epidural/intrathecal opioids all have their risks and benefits. Intrathecal midazolam hydrochloride has been demonstrated to have antinociceptive properties in both laboratory animals [1,2] and in man [3-5]. Doses up to 2 mg have been described without adverse effects. However, there have been no randomized controlled studies in which the efficacy of intrathecal midazolam in the management of acute post-operative pain in man has been examined. It would therefore be benefical to undertake a prospective, randomized, double-blind pilot study to assess the analgesic properties of intrathecal midazolam when given alone and in conjunction with spinal diamorphine.
Patients and methods
The study was approved by the local Ethics Committee. Written informed consent was obtained from 52 ASA I or II patients due to undergo elective Caesarean section with regional anaesthesia. Patients were excluded from the study if there was a contraindication to regional anaesthesia, known sensitivity to any of the drugs to be used, a history of chronic pain or use of medications known to modify pain perception. Fetal prematurity (less than 36-weeks gestation) was also considered a contraindication.
All patients received premedication with cimetidine 400 mg at 22.00 hours the night before, and at 07.00 hours on the day of surgery. Sodium citrate 30 mL (0.3 M) was given immediately prior to anaesthesia. Patients were randomly allocated to one of four groups. Active and placebo solutions were prepared by a second anaesthetist otherwise uninvolved with the case. The anaesthetist performing the block and post-operative assessments was blinded to the solution administered. Group B were randomized to receive intrathecal heavy bupivacaine 15 mg, 0.2 mL saline 0.9% and 1 mL saline 0.9%; group BD to receive intrathecal heavy bupivacaine 15 mg, diamorphine 0.2 mg (0.2 mL) and 1 mL saline 0.9%; group BM intrathecal heavy bupivacaine 15 mg, midazolam 1 mg (1 mL) and 0.2 mL saline 0.9%; and group BMD intrathecal heavy bupivacaine 15 mg, diamorphine 0.2 mg (0.2 mL) and midazolam 1 mg (1 mL). All agents administered were preservative-free; diamorphine was reconstituted from the powder immediately prior to injection. Midazolam was diluted from its commercial formulation of 5 mg mL−1 (Roche, Welwyn Garden City, UK) to a dose of 1 mg mL−1 in normal saline.
Monitoring was established with electrocardiography, pulse oximetry and non-invasive blood pressure. A 16-gauge intravenous (i.v.) cannula was sited and a pre-load of 500 mL modified fluid gelatin (MFG) administered. Regional anaesthesia was performed in the sitting position using a two-site combined epi-spinal technique  with a 19-gauge epidural catheter through 17-gauge Tuohy needle. This epidural was not used except where subsequent subarachnoid blockade (SAB) proved inadequate. SAB was effected via a 24-gauge Sprotte spinal needle. A further 500 mL MFG with ephedrine 30 mg was commenced immediately prior to intrathecal injection to prevent hypotension [7,8].
SAB was assessed by loss of sensation to cold and pinprick, a dermatomal sensory loss from T6 to S4 being considered adequate. Bolus doses of i.v. ephedrine 3 mg were given as required to maintain arterial blood pressure. Lower-segment Caesarean delivery was performed through a Pfannenstiel incision. Syntocinon (10 units) was given at the time of umbilical cord clamping. Intravenous volume replacement with MFG was given as clinically indicated. Dextrose 4% in saline 0.18% was infused at the end of the procedure and for 24 h at 100 mL h−1. No patient required blood transfusion or supplementary epidural analgesia.
After skin closure, all patients received diclofenac 100 mg as a rectal suppository. Morphine via PCAS was commenced (Graseby Medical Ltd, Watford, UK). Droperidol 5 mg was added to each 50 mL syringe of morphine . No other analgesic drugs were given, but additional anti-emetics were administered as required. Patients were assessed at 1, 6 and 24 h post-operatively. Visual analogue pain score (VAS) 0-100 mm and PCAS morphine usage in milligrams were recorded. Sedation (unsedated vs. drowsy), pruritus and PONV were recorded.
Statistical analysis of continuous variables (PCAS usage and VAS) were performed using one-way analysis of variance (Anova) with post-tests adjusted by Bonferroni's correction. Discrete variables (sedation score, pruritus and PONV) were assessed using χ2 or Fisher's exact test. The null hypothesis was rejected at the 5% level.
Caesarean sections were performed by one of eight obstetric registrars. No operation exceeded 1 hour's duration. Sixteen patients received BMD, 11 received B, 12 received BD and 13 received BM. There were no differences in demographic data (age, weight, height) between groups (Table 1), nor were there any differences in VAS (Table 2), sedation or PONV at 1, 6 and 24 h post-operatively. The number of patients who experienced nausea/vomiting at 1, 6 and 24 h post-operatively were 3, 2 and 1 in group BD; 1, 1, 0 in group B; 4, 3, 2 in group BMD; and 2, 1, 3 in group BM. Numbers of patients who complained of drowsiness or who were asleep (sedation score 1 or 2) at the same time points were 0, 3, 0 (BD); 2, 3, 1 (B); 0, 2, 2 (BMD) and 1, 3, 5 (BM).
Morphine requirements (Table 3) significantly differed between groups at 1 h (P=0.004), at 6 h (P=0.01) and at 24 h (P=0.046) (ANOVA). Further analysis of the four groups taking into account multiple comparisons (Bonferroni adjustment) revealed reduced PCAS morphine usage in groups BD, BM and BMD when compared to group B at 1 h (P<0.05). At 6 h morphine requirements were less in group BMD compared with group B (P<0.05). At 24 h the differences between groups could not be attributed to specific pairs of groups after Bonferroni correction.
The incidence of pruritus differed between the groups at 1 h (P<0.05), at 6 h (P<0.007) and at 24 h (P<0.004) post-operatively, pruritus being most common in patients receiving diamorphine. No adverse effects attributable to intrathecal midazolam hydrochloride were seen (0-10.3%, 95% confidence limit).
In this study we have shown post-operative analgesic requirement following Caesarean section to be reduced by intrathecal diamorphine. There was a supplementary effect of shorter duration with the addition of midazolam. In the absence of diamorphine, bupivacaine with midazolam showed better analgesia than bupivacaine alone only at 1 h. The clinical significance of this observation is unclear. As this was a small study, analgesic effects of intrathecal midazolam at later stages may have remained undetected (type 2 error). Differences between groups may have been further obscured by coadministration of diclofenac.
The best management for post-operative pain following Caesarean section remains controversial. There are well documented problems associated with intramuscular opioids; their use is suboptimal . Patient controlled analgesia systems provide excellent pain control but they are relatively expensive, not always available and associated with a high incidence of sedation, nausea and vomiting . Over recent years, our use of intrathecal opioids has steadily increased. Low cost, ease of administration and effectiveness of epidural and spinal opioids makes them an attractive option. However, sedation, itching, urinary retention, nausea and vomiting and the risks of late respiratory depression have perhaps limited widespread use of these routes of administration . As yet the correct dosage for intrathecal opioids has not been clearly defined. Low doses are associated with fewer side effects without reducing the analgesic potency, however, duration of action appears to be significantly reduced .
Midazolam hydrochloride is a potent, short acting imidazobenzodiazepine presented as an aqueous solution buffered to pH 3.5 to attain water solubility. It is preservative-free and becomes highly lipophilic when injected into the body at neutral pH. Intrathecal midazolam was originally shown to have antinociceptive properties in studies performed in animals in the early 1980s . Neurotoxicity studies have been performed and contradictory results published with at least two studies showing midazolam to be free of neurotoxicity [15,16]. However, one study performed with rabbits demonstrated a significant incidence of reactions on histological examination . Intrathecal midazolam has been used in man since 1986 and doses of up to 2 mg have been described, neither acute nor chronic side effects have been reported in the literature. Midazolam has been demonstrated to be effective in the treatment of chronic low back pain , acute post-operative pain and to control the cardiovascular response to surgery in man . Failure to find an analgesic benefit from spinal midazolam in man has also been described .
Benzodiazepine receptors have been implicated in an analgesic response elicited in male laboratory animals by potentially threatening olfactory stimuli . Analgesia produced by intrathecal midazolam can be reversed by the benzodiazepine antagonist flumazenil and the gamma-aminobutyric acid A (GABA A) antagonist bicuculline suggesting that antinociceptive actions are mediated through a typical benzodiazepine/GABA A receptor complex . Midazolam induced analgesia is also reversed by the opioid antagonist naloxone  and has been linked to a nonmu, possibly kappa, opioid pathway. In man a dermatomal level to pin-prick and injection of intradermal saline has been demonstrated, supporting evidence from laboratory animals that midazolam produces cord-level analgesia . There is no direct effect on nerve conduction and reversal of antinociceptive effects can be achieved by specific antagonists; this suggests that local anaesthetic action is unlikely. Motor neurone blockade has been demonstrated but only at extremely high dosage .
GABA is synthesized from glutamate in the presynaptic nerve terminal. Most GABA-minergic neurones are short interneurones with local connections and their actions are generally inhibitory. GABA, on binding to GABA A receptors, opens ligand gated chloride channels. Chloride conductance is increased leading to hyperpolarization and pre-synaptic inhibition of afferent terminals in the spinal cord. This results in less excitatory neurotransmitter being released onto the post-synaptic neurone and decreased central propagation of the action potential .
The effects of GABA are terminated by reuptake of neurotransmitter into the surrounding glial cells and into the nerve terminal. Post-synaptic GABA A receptors, to which the antagonists picrotoxin and bicuculin bind, are also described. High doses of these antagonists are known to cause convulsions suggesting GABA produces similar effects at pre- and post-synaptic receptors . Bicuculin has also been shown to attenuate the antinociceptive effects of midazolam .
Benzodiazepines alone lack GABA-mimetic effects. Acting through their own binding sites they enhance the effects of GABA at many sites within the central nervous system. The greatest density of binding sites in spinal cord are found within lamina 2 of the dorsal horn , a region strongly associated with the processing of noxious stimuli.
The present study is the first randomized controlled trial into the clinical efficacy of intrathecal midazolam as an analgesic for the early post-operative period. We have shown that midazolam, at a dose of 1 mg, can produce clinically detectable, beneficial effects on post-operative pain. However, this effect was not pronounced, nor was it as prolonged as previously described. Goodchild and Noble  demonstrated analgesia and a sensory block lasting for as long as 72 h into the post-operative period. In the present study, the duration of analgesia attributable to midazolam alone appeared less than 6 h although if combined with spinal opioids the effect was prolonged to beyond 6 h. The difference between the present results and those of Goodchild may have been because the patients additionally received diclofenac.
In conclusion, there is good evidence from the literature that midazolam has an antinociceptive effect at spinal level. This is borne out in the present pilot study which demonstrates analgesic benefit in the early post-operative period following Caesarean section. The effect while statistically significant was clinically marginal in this patient group. Further studies in other patient populations are required.
1 Edwards M, Serrao JM, Gent JP, Goodchild CS. On the mechanism by which midazolam causes spinally mediated analgesia. Anesthesiology
2 Serrao JM, Stubbs SC, Goodchild CS, Gent JP. Intrathecal midazolam
and fentanyl in the rat: Evidence for different spinal antinociceptive effects. Anesthesiology
3 Serrao J, Marks RL, Morley SJ, Goodchild CS. Intrathecal midazolam
for the treatment of chronic mechanical low back pain: a controlled comparison with epidural steroid in a pilot study. Pain
4 Goodchild CS, Noble J. The effects of intrathecal midazolam
on sympathetic nervous system reflexes in man - a pilot study. Br J Clin Pharmacol
5 Yanez A, Peleteiro R, Cambra MA. Intrathecal administration of morphine, midazolam and their combination in 4 patients with chronic pain. Rev Esp Anestesiol Reanim
6 Brownridge P. Epidural and subarachnoid analgesia for elective caesarean section
7 Kang YG, Abouleish E, Caritis S. Prophylactic intravenous ephedrine infusion during spinal anesthesia for caesarean section
. Anesth Analg
8 Frazer RS, Edward GM. Prophylactic ephedrine infusion in obstetric anaesthesia. Br J Anaesth
9 Williams OA, Clarke FL, Harris RW, Smith P, Peacock JE. Addition of droperidol to patient controlled analgesia: Effect on nausea and vomiting. Br J Anaesth
10 The Royal College of Surgeons of England and The College of Anaesthetists. Report of the Working Party on Pain after Surgery.
London: HMSO, 1990.
11 Eisenach JC, Grice SC, Dewan DM. Patient controlled analgesia following Ceasarean section: a comparison with epidural and intramuscular narcotics. Anesthesiology
12 Chrubasik S, Chrubasik J. Selection of the optimum opioid for extradural administration in the treatment of post-operative pain
. Br J Anaesth
13 Morgan M. The rational use of intrathecal and extradural opioids. Br J Anaesth
14 Niv D, Whitwam JG, Loh L. Depression of nociceptive reflexes by intrathecal administration of midazolam. Br J Anaesth
15 Serrao JM, Mackenzie JM, Goodchild CS, Gent JP. Intrathecal midazolam
in the rat: an investigation of possible neurotoxic effects. Eur J Anaesthesiol
16 Schwieger IM, Jorge-Costa M, Pizzolato GP, Forster A, Morel DR. Intrathecal midazolam
reduces isoflurane MAC and increases apnoeic thresholds in rats. Can J Anaesth
17 Malinovsky JM, Cozian A, Lepage JY, Mussini JM, Pinaud M, Souron R. Ketamine and midazolam neurotoxicity in the rabbit. Anesthesiology
18 Baaijens PFJ, Van Dongen RTM, Crul BJP. Intrathecal midazolam
for the treatment of chronic mechanical low back pain: a randomised double-blind placebo-controlled study. Br J Anaesth
Suppl 1: A470.
19 Kavaliers M, Innes DG. Male scent-induced analgesia in the deer mouse, Peromyscus maniculatus: involvement of benzodiazepine systems. Physiol Behav
20 Serrao JM, Gent JP, Goodchild CS. Naloxone reverses the spinal analgesic effects of midazolam. Br J Anaesth
21 Crawford ME, Molke Jensen F, Toftdahl DB, Madsen JB. Direct spinal effect of intrathecal and extradural midazolam on visceral noxious stimulation in rabbits. Br J Anaesth
22 Goodchild CS. GABA receptors and benzodiazepines. Br J Anaesth
23 Faull RLM, Villiger JW. Benzodiazepine receptors in the human spinal cord; a detailed anatomical and pharmacological study. Neurosci