SEARCH

SEARCH BY CITATION

Abstract

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. METHODS
  5. RESULTS
  6. DISCUSSION
  7. CONCLUSION
  8. Acknowledgements
  9. References

Objective  It has recently been suggested that systemic pethidine is ineffective in relieving labour pain. This study aims to evaluate the analgesic efficacy of pethidine in labour.

Design  This is a double blinded randomised placebo-controlled study.

Setting  Labour suite, Prince of Wales Hospital, Hong Kong.

Population  Fifty normal Chinese pregnant women in early labour.

Methods  We planned to randomise 112 women in early labour to receive either intramuscular pethidine 100 mg or normal saline when they requested analgesia. An interim analysis was performed at sample size 50.

Main outcome measures  The primary outcome assessed was the visual analogue scale (VAS) pain score after 30 minutes.

Results  The study was terminated after recruitment of the first 50 women when planned interim analysis showed a significantly greater reduction of VAS pain score in the pethidine group (P= 0.009). The median difference in VAS pain score between the two groups at 30 minutes was −17 [95% CI −30 to −4]. Mean time to first subsequent request for analgesia was greater in the pethidine group (232 minutes, 95% CI 135 to 329 minutes) compared with the control group (75 minutes, 95% CI 54 to 95 minutes). Eight women (32%) in the pethidine group needed no further analgesia vs one woman (4%) in the control group (P= 0.011). Women in the pethidine group gave greater scores for sedation and satisfaction. Neonatal outcome was similar.

Conclusions  Systemic pethidine was more effective at relieving labour pain than placebo. Its analgesic effect, however, was modest.


INTRODUCTION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. METHODS
  5. RESULTS
  6. DISCUSSION
  7. CONCLUSION
  8. Acknowledgements
  9. References

Systemic opioids are widely used for relief of labour pain. They are cheap, simple to use and readily available. Of available opioids, the most commonly used is pethidine, usually given intramuscularly in doses of 50–100 mg. In a recent survey in the United States, the use of parental opioids was found to be between 39% and 56% in various hospital obstetrics units.1 Yet, surprisingly, the efficacy of systemic pethidine for relieving labour pain has not been substantiated. On the contrary, many studies have suggested that intramuscular pethidine may be ineffective at relieving labour pain.2–4 In a randomised double-blinded study, Olofsson et al.5 administered pethidine 1.5 mg/kg or morphine 0.15 mg/kg in three divided intravenous doses to labouring women and found that there was no reduction in pain scores over 30 minutes after either drug. However, there was a marked increase in maternal sedation. This apparent lack of analgesic effect of systemic opioids has led to considerable controversy6,7 and it has been suggested that their use may even be unethical and medically incorrect.5 A recent review article by Bricker and Lavender8 has also shown that convincing evidence concerning the efficacy of pethidine in labour was limited.

Systemic opioids are associated with adverse effects on both mother and baby. For the mother, these include dysphoria, sedation, respiratory depression, nausea and vomiting and delayed gastric emptying.9 Pethidine readily crosses the placenta and may accumulate in the fetal circulation, especially in the presence of fetal compromise.10 Early neonatal respiratory depression may occur and behavioural and feeding problems for up to six weeks after delivery have been shown to be associated with the intrapartum use of pethidine.11–13 Moreover, administration of an ineffective therapy delays the administration of therapy that is known to be effective, such as epidural analgesia. Therefore, proper evaluation of the analgesic efficacy of pethidine and other systemic opioids is warranted.

Available data on the efficacy of systemic opioids for labour analgesia are limited because of the absence of randomised controlled blinded trials comparing opioids with placebo. Although ethical considerations may previously have contributed to this lack of placebo-controlled studies, Reynolds recently commented that this should no longer be considered the case,6 as there was great uncertainty whether pethidine had any analgesic effect in labouring women. Therefore, we designed this prospective, randomised double-blinded placebo controlled study to evaluate the efficacy of intramuscular pethidine given for pain relief in the first stage of labour.

METHODS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. METHODS
  5. RESULTS
  6. DISCUSSION
  7. CONCLUSION
  8. Acknowledgements
  9. References

The study was approved by the Clinical Research Ethics Committee of the Chinese University of Hong Kong with the proviso that the initial study period (after which rescue analgesia would be available) would be limited to 30 minutes to minimise the period during which patients would be exposed to a potentially inferior treatment (placebo). Data were collected in the labour ward of the Prince of Wales Hospital, Shatin, Hong Kong, China between September 2000 and May 2001. Healthy American Society of Anesthesiologists physical status 1 and 2 Chinese women in early labour with uncomplicated singleton term pregnancies and cephalic presentations were invited to participate in the study after they were admitted to the labour ward. Women were excluded if they made immediate request for analgesia or had already decided to have epidural analgesia, if they had obstetric complications (such as pre-eclampsia, antepartum haemorrhage or known fetal abnormality), had any contraindication to vaginal delivery or contraindication to pethidine or intramuscular injections, or if they had a history of drug abuse. Women undergoing induction of labour were included. Advanced written informed consent was obtained from eligible subjects in early labour. Use of a 100-mm visual analogue scale (VAS) for the measurement of pain (0 mm = no pain, 100 mm = worst pain imaginable) was explained. Subsequently, at each woman's first request for analgesia, consent was confirmed and she was entered into the study.

After request for analgesia, each woman was randomly allocated to receive a single intramuscular injection of pethidine 100 mg in 2 mL (pethidine group) or saline 2 mL (control group). Randomisation was stratified by parity and was performed using computer-generated random numbers in blocks of 10. Randomisation codes were placed in sequentially numbered, opaque, sealed envelopes. When each woman requested pain relief, the next numbered envelope was opened by a midwife who prepared the study drug and had no further involvement with the woman's care or assessment. A second midwife who was blinded to the contents of the syringe injected the drug into the gluteus maximus muscle.

VAS pain scores were recorded immediately before (baseline) and 15 and 30 minutes after injection by a blinded investigator. Women were monitored using continuous maternal pulse oximetry (Ohmeda Biox 3700e, BOC Health Care, Louisville, Kentucky, USA), continuous fetal cardiotocography (CTG), and non-invasive arterial pressure measurement. Maternal sedation as judged by the woman was measured at the time of injection, and then subsequently at 15 and 30 minutes using a separate 100-mm VAS (0 = totally alert, 100 = almost asleep). In addition, the presence or absence of maternal sedation as judged by the attending midwife was also recorded. Intrapartum management was according to our usual standardised labour ward protocol, which included fetal scalp blood sampling for cases with suboptimal CTG tracings. The incidence of vomiting and the incidence of any episodes of arterial oxyhaemoglobin desaturation, which is defined by a pulse oximetry reading <95% for 5 minutes or more, were recorded.

At 30 minutes, maternal satisfaction was recorded using a five-point numerical scale, with 1 being totally dissatisfied and 5 being very satisfied. The randomisation code was revealed to the attending staff (but not the woman), and further analgesia was then made available on the woman's request. Women in the pethidine group were offered the options of epidural analgesia or Entonox (50% nitrous oxide + 50% oxygen), and women in the control group were offered intramuscular pethidine, Entonox or epidural analgesia. The time from initial injection of the study drug to the first subsequent request for analgesia was recorded for each case.

For women who had received pethidine and were agreeable to venipuncture, a venous blood sample was taken at 30 minutes for measurement of plasma concentration of pethidine. Samples were immediately centrifuged and the plasma was stored at −70°C before assay using a modified gas chromatography method.14,15 The calibration curves for the assay were linear over the range 10–3000 ng/mL with coefficients of variation for within-day variation of 9.0% and 9.9% at 240 and 1000 ng/mL, and coefficients of variation for between-day variation of 10.3% and 11.2% at 225 and 980 ng/mL. The lower limit of detection was 10 ng/mL.

At delivery, Apgar scores were recorded at 1 and 5 minutes and arterial and venous blood samples were taken from a double-clamped segment of umbilical cord for immediate measurement of blood gases and pH using a Ciba-Coring 278 Blood Gas System (Ciba-Corning, Medfield, Massachusetts, USA) blood gas analyser. The requirements for resuscitative measures, including oxygen insufflation, bag and facemask ventilation and admission to the neonatal intensive care unit, were recorded.

Prospective power analysis was based on pooled data from 400 labouring parturients from published16 and unpublished studies from our unit. In these studies, the mean VAS pain score at first request for analgesia was 73.7 mm with an SD of 20.8 mm. Using these data, we calculated that a sample size of 56 subjects per group would have 90% power at the 5% significant level to detect a mean difference of 13 mm in VAS pain score between groups. This magnitude of difference was based on previously published data by Todd et al.,17 who reported that the minimum change on a standard 100 mm VAS pain scale that should be considered clinically significant was a mean of 13 mm or a median of 10 mm.

For ethical considerations, we planned interim analysis after recruitment of the first 50 women with a predetermined statistical stopping rule. An a priori decision was made to terminate the study at this point should the active treatment be more effective than control. For this analysis, to account for the possibility that the baseline VAS pain score might be different between the two groups, we chose to use differences in pain score reduction between the two groups as the outcome variable. For the interim analysis, a P value of <0.01 was required to terminate the study.18 Differences in VAS pain scores and differences in pain score reductions between the two groups were analysed using the Mann–Whitney U test for the median difference. The Student's t test, Mann–Whitney U test and χ2 test were used for between-group comparison of other outcome variables as appropriate. Kaplan–Meier survival analysis was performed to compare the difference in incidence and timing of further analgesic requirements between the two treatment arms. ‘Survival time’ was defined as the time from injection of the study drug to the time of first subsequent request for analgesia. Women who delivered without the need for further analgesia were considered ‘censored’. All analyses were performed using the Statistical Package for Social Sciences for Windows version 10.0 (SPSS, Illinois, USA) and Confidence Interval Analysis version 2.0.0 (T. Bryant, University of Southampton). Values of P < 0.05 were considered statistically significant.

RESULTS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. METHODS
  5. RESULTS
  6. DISCUSSION
  7. CONCLUSION
  8. Acknowledgements
  9. References

Figure 1 shows the trial profile. Recruitment was difficult as anticipated. Out of 198 eligible women approached, only 93 (47%) gave consent to be entered into the study. Of these, a further 43 (46%) were not randomised subsequently. Of the 50 women who were randomised, 25 (15 nulliparous and 10 multiparous) were randomised to each group. All 50 women completed the study with no crossovers during the study period. The characteristics of these women are summarised in Table 1. The study was terminated after recruitment of the first 50 women according to the predefined schedule when the interim analysis showed that the reduction in VAS pain scores from baseline to 30 minutes was significantly greater (P= 0.009) in the pethidine group (median change: decrease by 11 mm) compared with the control group (median change: increase by 4 mm).

image

Figure 1. Trial profile.

Download figure to PowerPoint

Table 1.  Maternal characteristics of randomised subjects. Values were expressed in median {interquartile range}, mean [SD] or percentage as appropriate.
 Control group (n= 25)Pethidine group (n= 25)
Age (years)28.4 [5.1]28.4 [4.4]
Nulliparous/multiparous ratio3:23:2
Labour induction (%)4880
Augmentation (%)4020
Baseline VAS pain score (mm)73 {59–86}73 {60–87}
Cervical dilation (cm)2 {1–3}1 {1–2}
Injection to delivery interval (minutes)344 [237]319 [210]

For the primary outcome of this study, the median difference in VAS pain score at 30 minutes between the pethidine and the control groups was −17 mm (95% CI −30 to −4). As shown in Table 2, the median baseline VAS scores were similar for both groups (73 vs 73 mm, P= 0.07). At 30 minutes, the median VAS score was significantly lower in the pethidine group compared with the control group (54 vs 78 mm, P= 0.01). In the pethidine group, 12 patients (24%) achieved a reduction in VAS of at least 40 mm (2 SD difference) compared with no patient in the control group (P= 0.01).

Table 2.  Changes in VAS pain scores. Values are median and {interquartile range} or [95% CI].
 Control groupPethidine groupMedian difference [95% CI]**
  • *

    Statistically significant.

  • **

    Estimation of median difference and corresponding CI were calculation using Confidence Interval Analysis version 2.0.0 (T. Bryant, University of Southampton), available with Atman DG, Machin D, Bryant TN, Gardner MJ. Statistics with Confidence, 2nd edition. Bristol: BMJ Books, 2000.

Baseline VAS pain score73 {59–86}73 {60–87}2 [−8 to 13]
VAS pain score at 15 minutes76 {59–85}62 {48–87}−5 [−21 to 6]
VAS pain score at 30 minutes78 {61–91}54 {41–75}−17 [−30 to −4]*

The requirement for and timing of subsequent analgesia is shown in the survival curve in Fig. 2. The mean time to first subsequent request for analgesia was greater in the pethidine group (232 minutes, 95% CI 135 to 329 minutes) compared with the control group (75 minutes, 95% CI 54 to 95 minutes). Eight women (32%) of the pethidine group needed no further analgesia compared with one woman (4%) in the control group (P= 0.011). Figure 3 shows the subsequent analgesics administered in each group. In the control group, six women (24%) eventually required epidural analgesia, compared with three women (12%) in the pethidine group (P= 0.27).

image

Figure 2. Kaplan–Meier survival plot showing the need for further analgesia before delivery. Survival time was defined as the time from initial injection to first subsequent request for analgesia. Cases were considered censored (represented by circles) if they delivered without the need for further analgesia. The pethidine group is represented by the full line, the control group is represented by the dotted line.

Download figure to PowerPoint

image

Figure 3. Profile of subsequent analgesics requested.

Download figure to PowerPoint

Satisfaction scores at 30 minutes (Table 3) were greater in the pethidine group (median 2, interquartile range 2–3) compared with the control group (median 1, interquartile range 1–2, P < 0.001). Of note, 8% of women in the pethidine group were totally dissatisfied (score = 1), compared with 60% in the control group and no woman in either group gave a score of 5 (very satisfied).

Table 3.  Satisfaction score at 30 minutes. Values are n (%).
Satisfaction scoreControl groupPethidine group
115 (60)2 (8)
29 (36)15 (60)
31 (4)5 (20)
403 (12)
500

The changes in VAS sedation score from baseline to 30 minutes are shown in Table 4. The median sedation score was similar between the control and pethidine groups at the time of request for analgesia (21 mm vs 26 mm, P= 0.82). Sedation scores increased in both groups. However, median sedation scores were significantly higher in the pethidine group compared with the control group at 15 minutes (68 vs 41 mm, P= 0.002) and at 30 minutes (78 vs 47 mm, P= 0.005).

Table 4.  Changes in VAS sedation scores. Values are median and {interquartile range} or [95% CI].
 Control groupPethidine groupMedian difference [95% CI]**
  • *

    Statistically significant.

  • **

    Estimation of median difference and corresponding CI were calculation using Confidence Interval Analysis version 2.0.0 (T. Bryant, University of Southampton), available with Atman DG, Machin D, Bryant TN, Gardner MJ. Statistics with Confidence, 2nd edition. Bristol: BMJ Books, 2000.

Baseline VAS sedation score21 {3–56}26 {6–46}1 [−16 to 14]
VAS sedation score at 15 minutes41 {14–61}68 {48–75}24 [8 to 43]*
VAS sedation score at 30 minutes47 {18–69}78 {51–88}26 [8 to 41]*

Other maternal side effects and neonatal outcome were similar between groups (Table 5).

Table 5.  Maternal and neonatal side effects and outcome. Data are number (%) or mean [SD].
 Pethidine groupControl groupRelative risk/mean difference (95% CI)
  • NICU = neonatal intensive care unit; CTG = cardiotocogram; BE = base excess.

  • *

    Statistically significant.

Maternal
Sedation (observed by midwife)16 (64)3 (12)5.33 (1.77 to 16.05)*
Vomiting2 (8.3)0 (0)Not significant
Arterial oxyhaemoglobin saturation <95%8 (32)6 (24)1.33 (0.54 to 3.29)
Operative vaginal delivery6 (24)7 (28)0.86 (0.34 to 2.19)
Caesarean section5 (20)6 (24)0.83 (0.29 to 2.38)
 
Fetal/neonatal
Suboptimal CTG requiring fetal blood sampling3 (12.0)5 (20)0.60 (0.16 to 2.25)
Bag-mask ventilation4 (16.0)2 (8)2.00 (0.40 to 9.95)
Oxygen given10 (40)8 (32)1.25 (0.59 to 2.64)
Tracheal intubation1 (4)1 (4)1.00 (0.07 to 15.12)
Apgar score 1 minute ≤73 (12)4 (16)0.75 (0.19 to 3.01)
Apgar score 5 minutes ≤700Not significant
Umbilical arterial pH7.26 [0.09]7.27 [0.09]0.01 (−0.04 to 0.06)
Umbilical arterial BE (n= 26)−6.51 [2.76]−6.57 [2.99]−0.52 (−2.38 to 2.27)
Admission to NICU1 (4)1 (4)1.00 (0.07 to 15.12)

Blood samples were obtained from 12 women in the pethidine group. Of these, the mean plasma concentration of pethidine at 30 minutes was 373 (SD 76) ng/mL, which was within 1 SD of the median effective concentration of pethidine (mean 460, SD 180 ng/mL) previously reported for post-operative analgesia.19 There was no significant correlation between the plasma pethidine concentration and the reduction in VAS score at 30 minutes (P= 0.620).

DISCUSSION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. METHODS
  5. RESULTS
  6. DISCUSSION
  7. CONCLUSION
  8. Acknowledgements
  9. References

Our study showed that intramuscular pethidine 100 mg was significantly more effective than saline placebo as an analgesic in the first stage of labour. This was evidenced by the finding that patients in the pethidine groups had VAS pain scores at 30 minutes that were lower than the corresponding values in the control group and were lower than the baseline pain scores and also by the finding that the pethidine group had a lower requirement for subsequent analgesia compared with the control group. Women's satisfaction with their analgesia was also better in the pethidine group compared with the placebo group. These results suggest that, contrary to other reports, pethidine does relieve labour pain and it is therefore both appropriate and ethical to retain pethidine in the list of therapeutic options available to women in the labour ward.

However, it should be noted that the analgesic efficacy of intramuscular pethidine was modest. The median change in VAS pain score at 30 minutes compared with baseline in the pethidine group was only 11 mm (95% CI 2 to 26 mm). Moreover, pethidine was associated with maternal sedation, nausea and vomiting. In comparison, a much greater degree of analgesia is achievable with regional techniques, such as epidural or combined spinal–epidural analgesia. For example, studies of regional analgesia have typically used criteria to define adequate analgesia, such as a reduction of VAS pain scores to 10 mm or less20 or by more than 90% from baseline.21 Regional anaesthesia is also associated with a low incidence of nausea and vomiting and does not cause sedation. On the other hand, regional anaesthesia is more invasive and requires more resources and expertise compared with intramuscular opioids. Not every obstetric unit is able to provide an epidural service and some patients may not want an epidural. Although pethidine is clearly not as effective for analgesia as epidurals, nonetheless, the degree of analgesia obtained may be sufficient for a large proportion of patients. This is evident in audit data from our unit. Despite having an epidural on-demand service, 38% of parturients in labour chose intramuscular pethidine for analgesia compared with an overall rate of epidural analgesia of 17.8% in the year 2000. In our current study, we found that only three women (12%) of those randomised to receive pethidine eventually requested epidural analgesia.

Why has our study shown pethidine to be effective at providing analgesia in labour whereas other studies have not? The reason for this may be related to study design and dosage. Previous studies have not compared pethidine with a placebo group. Because labour pain may increase in intensity as labour progresses, failure of opioids to decrease pain scores to below baseline does not necessarily equate to lack of effect. In our study, although the median decrease in VAS pain score from baseline was only 11 mm at 30 minutes, this has more clinical significance when one takes into account the finding that pain scores in the placebo group trended towards an increase at 30 minutes. Individual response to opioids is very variable and failure of analgesia may be related to inadequate dosing. As an example, Ramin et al.22 compared intravenous pethidine (up to 200 mg over 4 h) with epidural analgesia and reported that a large proportion of women (103/666) who received pethidine found it ineffective and requested epidural anaesthesia. However, when the same authors subsequently repeated the study, this time administering pethidine by a more liberal patient-controlled analgesia regimen, pethidine was much more effective, as evidenced by a reduction of VAS pain scores, relatively high patient satisfaction and only a few patients (5/357) requesting epidural analgesia because of inadequate analgesia.23 Another study performed by Vella et al.24 also found a 22–26% decrease in VAS scores after pethidine administration when up to 150 mg was used.

All of the women in our study were Chinese, and their mean body weight (64.1 kg) was likely to have been lower than that of Caucasian parturients. We chose to give women a standard dose of 100 mg pethidine, the dose used in most obstetrics departments in Hong Kong, which was equivalent to a mean dose of 1.56 mg/kg. This is greater than that used in many other studies.3,5,25 Despite this, we found that the mean plasma concentration of pethidine 30 minutes after injection was slightly less than the previously reported mean effective analgesic concentration for pethidine.19 This suggests that we may have seen better efficacy had we used a more liberal dosing regimen. Finally, racial differences in analgesic response may be important. Previous published work has shown that pethidine requirement for post-operative analgesia was less in Asian patients compared with European patients,26 and in our current study, this may partially explain why there may have been a greater response in our patients compared with studies of European patients.

In our study, the primary outcome variable on which we based the power analysis was the between-group difference in VAS pain scores at 30 minutes. However, the interim analysis and the criterion for terminating the study were based on the difference between groups in the within-group pain score reduction; this was done in an attempt to account for potential differences in baseline pain scores. Because the criterion used in the interim analysis differed from the primary outcome, this constitutes a methodological flaw in the study design. Nevertheless, because both variables were significantly different between groups, this does not affect the validity of our results.

Our study also has a number of limitations. First, there may have been selection bias. We excluded women who had already decided to have epidural analgesia. In addition, of all the women who matched the criteria for inclusion into the study, only a quarter was subsequently randomised. More than half of the women initially approached would not agree to randomisation and a further half of the women who were agreeable did not enter the study for various reasons. Thus, our cases were highly selected and it may be argued that their pain threshold and response to therapy may not necessarily have been representative of the general population. Women in the pethidine group and control group had comparable baseline VAS pain score at the time of first request for analgesia, indicating that the women in the groups had similar pain threshold. However, women in the pethidine group had lower baseline cervical dilatation and there was a greater incidence of induction of labour in the pethidine group compared with the control group. These are potential confounding variables in our analysis. However, because induction of labour may be associated with more painful and arduous labour,27 the greater proportion of women in the pethidine group who had labour induced if anything would most likely have reduced the study's ability to demonstrate an effect of pethidine. The small difference in cervical dilatation was probably of little clinical significance.

Our result confirmed previous findings that pethidine causes maternal sedation and vomiting.4,11–13 We did not observe any difference in neonatal outcome between groups although our study was not powered to do this.

CONCLUSION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. METHODS
  5. RESULTS
  6. DISCUSSION
  7. CONCLUSION
  8. Acknowledgements
  9. References

In conclusion, this is the first randomised double-blinded controlled trial to show that intramuscular pethidine is more effective than placebo at providing analgesia in labour in a group of normal Chinese pregnant women. Our findings support the continued use of pethidine as a simple and cheap therapeutic option in the management of labour pain, particularly in units where access to epidural analgesia is limited. Because the analgesic effect of pethidine 100 mg was modest, further research should be directed at optimising dose and methods of administration, such as patient-controlled analgesia.

Acknowledgements

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. METHODS
  5. RESULTS
  6. DISCUSSION
  7. CONCLUSION
  8. Acknowledgements
  9. References

We would like to thank the midwives of the Labour Ward, Prince of Wales Hospital for their assistance in this study and Ms Pet Tan for performing the pethidine assays.

References

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. METHODS
  5. RESULTS
  6. DISCUSSION
  7. CONCLUSION
  8. Acknowledgements
  9. References

Accepted 27 February 2004