Intramuscular opioids for maternal pain relief in labour: a randomised controlled trial comparing pethidine with diamorphine


  • F. M. Fairlie,

    Consultant Obstetrician and Gynaecologist , Corresponding author
    1. University Department of Obstetrics and Gynaecology, Glasgow Royal Maternity Hospital
      Correspondence: Dr F. Fairlie, The Jessop Hospital, Leavygreave Rd, Sheffield S3 7RE, UK.
    Search for more papers by this author
  • L. Marshall,

    Labour Ward Sister/Midwife
    1. University Department of Obstetrics and Gynaecology, Glasgow Royal Maternity Hospital
    Search for more papers by this author
  • J. J. Walker,

    Professor (Obstetrics and Gynaecology)
    1. University Department of Obstetrics and Gynaecology, Glasgow Royal Maternity Hospital
    Search for more papers by this author
  • D. Elbourne

    Senior Lecturer (Medical Statistics)
    1. Medical Statistics Unit, London School of Hygiene and Tropical Medicine
    Search for more papers by this author

Correspondence: Dr F. Fairlie, The Jessop Hospital, Leavygreave Rd, Sheffield S3 7RE, UK.


Objective To compare the pain relief and side effects of intramuscular pethidine with intramuscular diamorphine in labour.

Design Double-blind randomised controlled trial.

Setting The labour ward in a UK teaching hospital.

Participants Sixty-nine nulliparous women and 64 multiparous women in labour who requested narcotic analgesia and remained undelivered one hour after trial entry.

Methods Nulliparous women were randomised to receive either 150 mg intramuscular pethidine or 7–5 mg intramuscular diamorphine. Multiparous women were randomised to receive either 100 mg intramuscular pethidine or 5 mg intramuscular diamorphine. All participants received the anti-emetic prochloroperazine at the same time as the trial drugs.

Main outcome measures Maternal analgesia assessed by a visual analogue score and verbal scales of pain intensity and pain relief, maternal sedation and vomiting, neonatal outcome assessed by Apgar scores and the need for resuscitation.

Results More women allocated to receiving pethidine than to diamorphine reported slight or no pain relief at 60 minutes after administration of these drugs (P= 0.03). This trend was repeated in most of the other measures for maternal analgesia. There was no difference in maternal sedation, but the incidence of vomiting within 60 minutes was lower for women who received diamorphine (P= 0.02). Pethidine was associated with lower Apgar scores at 1 minute (P < 0.05).

Conclusion Intramuscular diamorphine in labour appears to have some benefits, compared with intramuscular pethidine, but the trial was small and further research, particularly into alternative opioids and long term effects on the infants is still needed.


In the UK intramuscular pethidine is widely used for analgesia in labour1. It was first used for intrapartum pain relief in 1947, and is the only opioid allowed to be ordered by midwives2. There are, however, concerns that pethidine may not provide very effective pain relief.3,4. Some studies have suggested that fewer than 20% of women may expect a pain-free labour with this opioid5,6, and little or no intrapartum pain relief was reported by 75% of women receiving either 50 mg or 100 mg intramuscular pethidine7. Not only is pethidine often ineffective as an analgesic, but it has disagreeable side effects. It delays gastric emptying8. Nausea and vomiting, amnesia and dysphoria have been reported to occur in at least 25% of women receiving this opioid in labour9. Pethidine rapidly crosses the placenta and maximum concentrations are found in fetal scalp blood and in umbilical arterial blood between one and five hours after an intramuscular maternal dose10. Pethidine may be associated with a change in the fetal electroencephalogram and decreased variability of the fetal heart rate11. Reported neonatal effects include abnormal behavioural patterns, depression of reflexes and impaired sucking reflex for up to 48 hours after birth12,13.

These concerns have led to an exploration of the relative advantages and disadvantages of an alternative intramuscular opioid for maternal pain relief in labour. In 1943 a report concluded that diamorphine provided a rapid and intense analgesia without adverse maternal side effects14. An observational study of the benefits and side effects of analgesia available to labouring women admitted to the Jessop Hospital in Sheffield in 1967 led Beazley et al.15 to recommend intramuscular diamorphine. In their experience this drug provided effective analgesia for up to four hours, and the incidence of maternal vomiting was negligible. In addition, it was associated with euphoria rather than the dysphoria and sedation often associated with opiates. More recently, it has been suggested that diamorphine, compared with pethidine, is more rapidly eliminated from maternal and neonatal p1asma16. Despite these positive findings, there has been a reluctance to administer diamorphine to pregnant women. This may be based on the potential danger of addiction17, although addiction resulting from the use of diamorphine in labour has never been proven.

As we were unable to find any trials in the literature which compared pethidine with diamorphine in labour, we designed a double-blind, randomised controlled trial to compare pain relief and side effects of intramuscular diamorphine with intramuscular pethidine administered in labour.


The trial was conducted at Glasgow Royal Maternity Hospital between May 1990 and February 1992. The women studied were in active labour, with a gestational age of between 37 and 42 weeks. Active labour was defined as regular uterine contractions, at least two in 10 minutes, and cervical dilatation of at least 3 cm. Labour was either spontaneous in onset or induced by amniotomy and intravenous oxytocin. All women recruited to the trial gave informed consent, and the trial design was approved by the University of Glasgow ethics committee. Although women were invited to enter the trial when they were in early labour, actual recruitment occurred only when analgesia was requested for the first time. Once recruited, women were randomly allocated to receive either intramuscular pethidine or intramuscular diamorphine. Pethidine was administered according to the standard dose regime used in the labour ward at that time (i.e. nulliparous women received 150 mg and multiparous women (parity 1–4) received 100 mg). Women randomised to receive diamorphine were given 7.5 mg if they were nulliparous and 5 mg if they were multiparous.

Randomisation was achieved by packing each woman's medication individually in a box which was identified by a trial number. A trial population of 200 (100 nulliparous and 100 multiparous women) was planned based on pragmatic considerations. A total of 216 numbered boxes were packaged; 108 with doses suitable for nulliparous women (54 pethidine and 54 diamorphine) and 108 with doses suitable for multiparous women (54 pethidine and 54 diamorphine). The boxes were packaged in blocks of six to ensure similar numbers in the two analgesic groups. To maintain the blindness of the midwife managing the woman's labour to the medication received, the drug was administered by a midwife not involved in the trial. The randomisation code was not broken until completion of the trial.

All women received an intramuscular injection of 12.5 mg prochloroperazine as an anti-emetic at the same time as the trial drug. Both the trial drug and prochloroperazine were administered into the gluteus muscle. If a woman requested additional analgesia before delivery, she was given the choice of a second (open) dose of narcotic analgesia or lumbar epidural analgesia.

In all labours studied the fetal heart rate was monitored continuously. If meconium staining of the amniotic fluid occurred, the time this sign first appeared in relation to the time of administration of the trial drug was noted. Neonatal condition was assessed by 1 minute and 5 minute Apgar scores; the need for resuscitation (indicated by either the administration of the opioid antagonist naloxone and/or intubation and intermittent positive pressure ventilation); admission to the special care baby unit; and significant neonatal morbidity. The latter included seizures, intraventricular haemorrhage, necrotising enterocolitis and respiratory distress syndrome.

Maternal pain severity was assessed by using a visual analogue score and verbal scales of pain intensity and pain relief18,19. Assessments were made when a woman first requested analgesia and at intervals of 30 minutes after receiving the trial drug for a maximum of three hours or until delivery occurred or additional analgesia was requested. The visual analogue score ranged from 0 (representing no pain) to a maximum of 100 (representing excruciating pain). Women marked the visual analogue score without reference to previously marked scores. Measurement of the pain at each recording time referred to the peak intensity during the most recent contraction. The verbal pain intensity and pain relief were scored by the women themselves using the following scales:

Verbal pain intensity was scored as follows: 0 = no pain; 1 = mild pain; 2 = moderate pain; 3 = severe pain. Pain relief was scored as: 0 = none; 1 = slight; 2 = moderate; 3 = good; 4 = complete. Maternal sedation was assessed by the midwife on a 4-point scale as: 0 = alert; 1 = mildly drowsy; 2 = moderately drowsy; and 3 = asleep.

The incidence of maternal vomiting within one hour of receiving the trial drug was recorded. All women participating in the trial were interviewed by the midwife co-ordinating data collection (L. M.) 24 hours after the birth of the infant. They were asked to give a global assessment of the intrapartum analgesia provided by the trial drug they received, scoring it as either good or poor.

Statistical analysis

The primary outcome was maternal pain one hour after randomisation. Women who were delivered within one hour of randomisation were therefore excluded from further study. For all other women in the trial, outcomes were analysed for the total group and for nulliparous and multiparous women separately.

The χ2 test (with Yates' correction) and Fisher's exact test were used to compare categorical variables. The Student's t test and the Mann-Whitney U test were used to compare continuous outcome. Differences were considered statistically significant at the 5% level. Differences of means and relative risks (RR) and are given with 95% confidence intervals (CI) as appropriate.


All women initially recruited to the trial received the drug to which they were randomised. After recruitment of 150 women, enthusiasm for the trial waned and a preliminary analysis was undertaken. This revealed statistical differences in both pain relief and side effects and the trial was terminated. By then 75 nulliparous women and 86 multiparous women had been recruited. Women in the pethidine and diamorphine groups had similar characteristics at trial entry (Table 1).

Table 1.  Maternal characteristics at trial entry. VAS = visual analogue score; VPI = verbal pain intensity. Values are given as n (%) or mean [SD].
 Diamorphine (n=65)Pethidine (n=68)Diamorphine (n=34)Pethidine (n=35)Diamorphine (n=31)Pethidine (n=33)
Spontaneous labour43 (66)44 (65)22 (65)21 (60)21 (68)23 (70)
Maternal weight (kg)73 [13]74 [13]72 [13]73 [10]14 [13]75 [15]
Gestational age (weeks)40 [1]40 [1]40 [1]40.5 [1]40 [1]40 111
Bishop score pre-analgesia8.4 [1.6]8.0 [1.8]8.4 [1.6]7.9 [1.7]8.5 [1.8]8.1 [1.7]
Contraction frequency of geqslant R: gt-or-equal, slanted3 in 10 mins47 (72)50 (74)19 (56)23 (66)28 (90)21 (82)
VAS71 [22]65 [22]72 [20]67 [21]69 [24]63 [24]
Moderate or severe VPI65 (100)67 (98.5)34 (100)35 (100)31 (100)32 (97)

As labour progressed the number of women remaining undelivered declined. Figure 1 shows the number of women remaining within each analgesic and parity group as time progressed. Of the 161 women recruited to the trial, 28 (17%) were excluded because they were delivered within one hour of receiving the trial drug. This proportion did not differ significantly between the two trial arms. Data from the remaining 69 nulliparous women and 64 multiparous women were available for analysis.

Figure 1.

Number of women remaining undelivered as labour progressed, by allocation and parity. *One woman delivered at 60 mins and is therefore in the trial but does not contribute to the primary outcome.

Table 2 shows the outcomes of labour and delivery and the incidence of maternal sedation and vomiting. None of the parameters compared showed a statistically significant difference between pethidine and diamorphine, except fewer women in the diamorphine group vomited within one hour of receiving the trial drug compared with the pethidine group (RR 0.39, 95% CI 0.17 to 0.86; P= 0.02). The effect was similar (although not statistically significant) for the nulliparous and multiparous groups separately.

Table 2.  Labour and delivery outcomes. Values are given as n (a) or mean [SEMI
 Diamorphine (n=65)Pethidine (n=68)Diamorphine (n=34)Pethidine (n=35)Diamorphine (n=31)Pethidine (n=33)
  1. *P= 0.02.

Moderately drowsy or asleep at 60 mins post drug administration16 (25)18 (26)13 (38)10 (29)3 (10)8 (24)
Maternal vomiting7 (11)*19 (28)3 (9)9 (26)4 (13)10 (30)
Trial drug administration to delivery interval (h)4.9 [0.4]4.5 [0.4]6.0 [0.6]5.8 [0.6]3.8 [0.3]3.2 [0.5]
Spontaneous vaginal delivery52 (80)52 (76)22 (65)20 (57)30 (97)32 (97)
Instrumental delivery11 (17)12 (18)10 (29)11 (32)1 (3)1 (3)
Caesarean section2 (3)4 (6)2 (6)4 (11)00
Operative delivery for fetal distress4 (6)3 (4)3 (9)3 (9)1 (3)0
Meconium staining after trial drug administration9 (14)13 (19)4 (12)9 (26)5 (16)4 (12)

Table 3 shows the effects of pethidine and diamorphine on maternal pain and pain relief. There was no statistically significant difference between the trial drugs in visual analogue score 60 minutes after trial entry, but there was an overall trend for more women who received pethidine to report a verbal pain intensity of moderate or severe, compared with those who received diamorphine, 60 minutes after receiving the trial drugs. This reached statistical significance for multiparous women (RR 0.84, 95% CI 0.72 to 0.98; P= 0.02– Fisher's exact test), but not for nulliparous and multiparous groups combined (RR 0.92, 95% CI 0.79 to 1.08; P= 0.43). This trend was reflected in the smaller proportion of women in the diamorphine group who said they had slight or no pain relief at this time (RR 0.63, 95% CI 0.43 to 0.94; P= 0.03). The effect was similar, although not statistically significant, for the nulliparous and multiparous groups separately. Neither administration of additional pain relief (whether narcotic or epidural), nor global assessment of pain relief as poor, differed significantly between the two randomised groups. By 90 minutes after administration of the drugs, no differences were evident for any of the outcomes considered (data available but not shown).

Table 3.  Maternal pain assessment and analgesia requirements. VAS = visual analogue score; VPI = verbal pain intensity. Values are given as n (%) or mean [SEM].
 Diamorphine (n=65)Pethidine (n=68)Diamorphine (n=34)Pethidine (n=35)Diamorphine (n=31)Pethidine (n=33)
  1. *P= 0.02 (Fisher's exact test).

  2. P= 0.03.

VAS at 60 mins post trial drug administration58 [4]67 [3]52 [5]62 [5]64 [5]71 [4]
Moderate or severe VPI at 60 mins post trial drug administration48 (75)59 (87)22 (67)26 (74)26 (84)*33 (100)
None or slight pain relief at 60 mins post trial drug administration23 (36)38 (56)8 (24)17 (49)15 (48)21 (64)
Second dose of narcotic analgesia administered9 (14)7 (10)3 (9)3 (9)6 (19)4 (12)
Epidural analgesia21 (32)18 (26)17 (50)16 (46)4 (13)2 (6)
Global assessment of pain relief as poor or fair37 (57)44 (65)16 (47)18 (51)21 (70)26 (79)

The benefit of diamorphine was evident in fewer babies with an Apgar score c 7 at 1 minute (RR 0.41, 95% CI 0.18 to 0.91; P= 0.04), but there were no other significant differences in neonatal outcomes between the groups (Table 4). One baby, who was born to a woman in the pethidine arm of the trial by spontaneous vaginal delivery, and who did not need resuscitation at birth, developed seizures 48 hours after birth. Computerised axial tomography showed a left-sided, temporal-frontal cerebral infarct of unknown origin.

Table 4.  Neonatal outcomes. Values are given as n (%).
 Diamorphine (n=65)Pethidine (n=68)Diamorphine (n=34)Pethidine (n=35)Diamorphine (n=31)Pethidine (n=33)
  1. *P= 0.04.

  2. P= 0.04 (Fisher's exact test).

Apgar < 7 at 1 min7 (11)*18 (26)3 (9)10 (29)4 (13)8 (24)
Apgar < 7 at 5 mins1 (1.5)3 (4)01 (3)1 (3)2 (6)
Neonatal resuscitation22 (34)19 (28)10 (29)10 (29)12 (39)9 (27)
Admission to special care baby unit5 (8)9 (13)4 (12)6 (17)1 (3)3 (9)
Neonatal morbidity01 (1.5)01 (3)00


Labour pain is unusual, compared with other causes of pain, in that it is not always considered undesirable. Morgan et al. found that 45% of women felt that pain in labour was an essential part of the emotional experience of childbirth20 and that effective pain relief does not ensure a satisfactory birth experience21. Nevertheless, many women find the contractions of labour painful and most request pain relief. Lumbar epidural anaesthesia provides excellent intrapartum analgesia in 50% to 80% of recipients5,6, and may have fewer maternal and neonatal side effects22–24, but in a survey of 1000 women postpartum, Morgan et al.21 observed that those who received epidural anaesthesia were more likely to be dissatisfied, compared with women who received pethidine or entonox. An unsatisfactory experience was more likely to be associated with forceps delivery and a long labour, but not ineffective pain relief. Epidural anaesthesia may be associated with both uncommon but serious acute complications of dural tap, spinal injection and epidural haematoma, as well as more frequent longer term side effects such as backache25. Moreover, the experienced anaesthetic staff required to perform the technique may not always be available. There is therefore a continuing need for intramuscular narcotic analgesia in labour. Our desire to obtain an effective and safe alternative to intramuscular pethidine led us to re-examine the role of intramuscular diamorphine in labour. To our knowledge this is the first randomised trial comparing the benefits and side effects of intramuscular diamorphine in labour with intramuscular pethidine.

A method of pain measurement is required to compare the analgesia provided by two narcotic drugs. The method must account for increases in pain intensity as labour advances and the variation in women's expectation and tolerance of pain. In our trial we used pain scores to quantify pain relief. Pain scores are easy to use and have proved to be a sensitive, reliable method for the measurement of pain19,20. It might be anticipated that pain scores should all show the same effect. However, this is not necessarily so since the score may assess different aspects of the pain experience. In our study visual analogue score did not detect statistically significant differences between the drugs, whereas there was a statistically significant benefit of intramuscular diamorphine in terms of greater pain relief (P= 0.03) and less verbal scales of pain intensity for multiparous women (P= 0.02), compared with pethidine. Regardless of the measure we used, the direction of effect showed similar trends.

Almost 50% of the women in the trial reported the trial drug provided poor pain relief. This was true for both diamorphme and pethidine. The incidence of requests for ‘second line analgesia’1—either epidural analgesia or a second dose of a narcotic—was not different for the two narcotics. Approximately 40% of women in each analgesic group requested second line analgesia. More nulliparous women requested second line analgesia than multiparous women, reflecting the shorter length of labour in the latter group.

Published studies of analgesia in labour have not always analysed nulliparous and multiparous women separately. This distinction is important, as previous experience of labour may affect pain expectations and parity influences the level of uterine activity and contraction frequency in labour26. In our trial the effect on verbal pain intensity seemed more marked for multiparous women while the effect on pain relief seemed more marked in nulliparous women. This may be due to the different aspects of pain that verbal pain intensity and pain relief assess or due to differences in the perception of pain by nulliparous and multiparous women. The differences were, however, compatible with chance.

We chose a dose of 150 mg pethidine for nulliparous women and 100 mg pethidine for multiparous women because these were the standard doses used in our labour ward. The doses of diamorphine were selected as being equivalent to 150 mg and 100 mg pethidine. Although the duration of labour may be shorter in multiparous women there is no evidence to indicate that it is less painful or that less analgesia is required. We suggest that future studies comparing pethidine with diamorphine should use the same analgesic dose regime for both parity groups.

It has been suggested that the site of injection may influence the efficacy of pethidine in labour. In our trial both pethidine and diamorphine were administered intramuscularly into the gluteus muscle. In theory it is possible that the absorption and metabolism of the two drugs were sufficiently different to explain the difference in analgesia and side effects. Although this is unlikely, assays of plasma levels of the drugs and their metabolites would be necessary to address this question.

Evaluation of analgesia also includes consideration of side effects. We found that the incidence of vomiting was significantly less after diamorphine compared with pethidine when all women in the trial were analysed. Beazley et al.15 also observed less maternal vomiting with diamorphine compared with pethidine. Neonatal respiratory depression associated with pethidine is likely to be most severe if delivery takes place between 2.5 and 3.5 hours after an intramuscular injection28. We observed a significantly higher incidence of low 1 minute Apgar scores after pethidine compared with diamorphine. It may be argued that a difference in 1 minute Apgar score does not represent significant neonatal sedation. To determine neonatal effects in detail would require a much larger study and long term follow up of infant behaviour and neurological development.

There remains a demand for narcotic analgesia in labour even though most women have access to a 24-hour service for regional analgesia. Awareness of the problems of pethidine has led to an increasing trend in the use of alternative opioids for pain relief in labour in UK maternity units. This is reflected in the results from a recent survey of 117 midwives working in 63 maternity units in the England and Wales undertaken at the annual Midwives Research Conference in April 1997 showing that although pethidine was used in all the units, 21% now reported using diamorphine and 27% were using meptazinol (C. Littler, personal communication). Asked which was the standard opioid used in the first stage of labour, 84% said pethidine, but a further 5% said meptazinol. In seven hospitals, no single opioid was identified as standard, but a mixture of either pethidine or diamorphine (2%) or pethidine or meptazinol (10%) was given. In Scotland, diamorphine was used in 42% of obstetric units29.

Our findings have shown for the first time in a randomised controlled trial that there may be benefits of intramuscular diamorphine, compared with intramuscular pethidine, although the size of effect may have been biased as the trial was stopped early because of these results. Further studies are indicated to define the most effective dose regime, route and frequency of drug administration and neonatal side effects.

The trial also confirms that for women requesting intramuscular narcotic analgesia, neither diamorphine nor pethidine provide good pain relief and so there is a need to consider alternative opioids. In fact, it was preparation for a systematic review comparing different types of intramuscular opioids30 that prompted the belated retrieval of the data of the trial reported here. The review compared pethidine to tramadol, to meptazinol, and to pentazocine as well as to diamorphine, and concluded that as many of the trials had methodological problems, policy recommendations were premature. There is a need for further good quality trials comparing the most commonly used treatments to the promising alternatives, with enough power to consider neonatal outcomes, so that the many thousands of women worldwide who request injected narcotic opioids in labour can be assured of the effectiveness and safety of pain relief they are offered.