SEARCH

SEARCH BY CITATION

Keywords:

  • Cervical ripening;
  • induction of labour;
  • isosorbide mononitrate;
  • outpatient

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Aims
  5. Methods
  6. Results
  7. Discussion
  8. Conclusion
  9. Disclosure of interests
  10. Contribution to authorship
  11. Details of ethical approval
  12. Funding
  13. Acknowledgements
  14. References

Objective  To determine whether isosorbide mononitrate (IMN), self-administered vaginally by women at home, improves the process of induction of labour.

Design  Randomised double blind placebo-controlled trial.

Setting  Large UK maternity hospital.

Population or Sample  Nulliparous women with a singleton pregnancy, cephalic presentation ≥37 weeks gestation, requiring cervical ripening prior to induction of labour.

Methods  IMN (= 177) or placebo (= 173) self-administered vaginally at home at 48, 32 and 16 hours prior to the scheduled time of admission for induction.

Main outcome measures  Admission to delivery interval and women’s experience of induction of labour.

Results  IMN did not shorten the admission to delivery interval as compared with placebo [mean difference of −1.6 hours (95% CI −5.1,1.9, P = 0.37)], despite being more effective than placebo in inducing a change in Bishop score [mean difference of 0.65 (95% CI 0.14,1.17, P = 0.013)]. While both groups found the overall experience of home treatment to be positive, (mean score of 3.8/10 ± 2.3/10 for the IMN group, where 1 = extremely good and 10 = not at all good) women in the placebo group found it marginally more positive than those in the IMN group (just over half a unit on a 10-point scale, P = 0.043). There were no differences between the groups in the pain or anxiety experienced or willingness to take the treatment in a subsequent pregnancy.

Conclusions  IMN self-administered vaginally at home does not shorten admission to delivery interval despite a significant effect on cervical ripeness assessed using the Bishop score. However, women report positive views on cervical ripening at home, and the setting deserves further investigation.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Aims
  5. Methods
  6. Results
  7. Discussion
  8. Conclusion
  9. Disclosure of interests
  10. Contribution to authorship
  11. Details of ethical approval
  12. Funding
  13. Acknowledgements
  14. References

Approximately 20% of pregnant women undergo induction of labour in the UK.1 In primigravidae, the mean time taken from induction to delivery is 27 hours, of which up to 18 hours is spent in the cervical ripening phase before labour itself starts.2 An ideal agent for cervical ripening would induce adequate cervical ripening without causing uterine contractions. No fetal monitoring would then be required, and such an agent could be given at home. Nitric oxide, a free radical gas with a short half life, is thought to be a fundamental mediator of cervical ripening.3,4 We and others have hypothesised that nitric oxide (NO) donors, such as isosorbide mononitrate (IMN), nitroglycerin and sodium nitroprusside may prove to be agents that can be administered at home without causing uterine contractions or other adverse effects of clinical importance during the ripening process.2,3,5–10

There is an increasing interest in outpatient cervical ripening. Preliminary studies suggest that outpatient cervical ripening with prostaglandins (PG) is associated with significantly shorter admission to delivery times,11,12 reduced healthcare costs13 and higher satisfaction rates in the women compared with inpatient cervical ripening.14 However, there are concerns about the use of prostaglandins in the outpatient setting, because of the potential for uterine contractions to cause fetal hypoxia, which would remain undiagnosed in the absence of fetal monitoring.15,16 The use of PG for outpatient cervical ripening has, therefore, been declared unsafe by many authorities15–17 and the National Centre for Clinical Excellence and Royal College of Obstetricians and Gynaecologists have stated that they cannot make any recommendations on the safety of outpatient treatment.

We report here a randomised placebo-controlled trial of cervical ripening with IMN on an outpatient basis at home prior to induction of labour. Previous trials of cervical ripening in the ‘outpatient’ setting have involved administration of the cervical ripening agent by the investigators in the hospital environment and a period of observation with fetal monitoring before being discharged home.12,18–20 In contrast, in the study described here women self-administered the trial drugs at home in the absence of any fetal monitoring. Only one previous study has reported the use of IMN for cervical ripening at home.21

Aims

  1. Top of page
  2. Abstract
  3. Introduction
  4. Aims
  5. Methods
  6. Results
  7. Discussion
  8. Conclusion
  9. Disclosure of interests
  10. Contribution to authorship
  11. Details of ethical approval
  12. Funding
  13. Acknowledgements
  14. References

To test the hypothesis that IMN (40 mg), given every 16 hours up to a total of three doses improves the process of induction of labour as compared with placebo. Our prior hypotheses were that outpatient pre-induction cervical ripening with IMN would reduce the elapsed time from hospital admission to delivery and improve women’s experience of induction of labour. We also aimed to estimate the cost of IMN versus placebo with the view to establishing its cost-effectiveness. Costs to the health services are reported in detail in our companion paper and will not be discussed further here.22

Methods

  1. Top of page
  2. Abstract
  3. Introduction
  4. Aims
  5. Methods
  6. Results
  7. Discussion
  8. Conclusion
  9. Disclosure of interests
  10. Contribution to authorship
  11. Details of ethical approval
  12. Funding
  13. Acknowledgements
  14. References

Design

Double blind randomised placebo-controlled trial (ISRCTN39772441) for which we have previously published the protocol.23

Participants

Between March 2005 and December 2006, women scheduled for admission to the Princess Royal Maternity Hospital, Glasgow for cervical ripening and labour induction were approached by the clinical researcher (SSB). The inclusion criteria included all of the following: nulliparity, singleton fetus, cephalic presentation, ≥37 completed weeks gestation, modified Bishop score24 <7 and willingness to self administer vaginal tablets. We excluded patients who were <16 years of age, who needed delivery within the next 48 hours in the fetal or maternal interest or who had ruptured membranes.

Ethical approval

The study was approved by the North Glasgow University NHS Trust ethics committee and written informed consent was gained prior to randomisation.

Recruitment

Once a decision to induce labour was made, women who wished to participate in the study were recruited and randomised at an antenatal visit by SSB prior to scheduled admission for induction of labour. At this visit, randomised study medication (either 40 mg IMN tablets or placebo) was issued, with instructions to self-administer the tablets vaginally at home at 48, 32 and 16 hours prior to the scheduled time of admission for induction of labour. After randomisation, women did not attend the hospital until admission for inpatient induction of labour, unless required for any medical reason.

After admission to hospital, treatment reverted to the protocol for induction of labour for our institution, whereby prostaglandins were given until cervical ripening (defined as Bishop score >6) was achieved or a maximum of three doses of prostaglandin tablets (3 mg each) were administered. Once the cervix was ripe the fetal membranes were ruptured and oxytocin administered if required.25 The women’s inpatient cervical ripening and intrapartum labour management was managed independently of the study team.

Outcomes

The primary clinical outcome measure was elapsed time interval from hospital admission to delivery (defined as the time from admission for inpatient induction or admission in labour to delivery – either vaginally or by caesarean section).

The primary economic outcome was cost to the health service of IMN versus placebo. The companion economic paper presents evidence on the cost-effectiveness of outpatient cervical ripening with IMN.22

The primary acceptability outcome was women’s experience of induction of labour. The secondary outcomes were the component parts of the women’s labour and delivery together with the neonatal outcomes.23

  • 1
    Primary clinical outcome (elapsed time from admission to delivery) for the subgroup of women who deliver vaginally.
  • 2
    Operative delivery rates.
  • 3
    Incidence of unscheduled admission for reasons other than labour commencing.
  • 4
    Duration and frequency of neonatal admissions to special care.
  • 5
    Incidence of adverse maternal and fetal outcomes such as uterine hypercontractility, postpartum haemorrhage (maternal outcomes) and meconium stained liquor, 5 minute Apgar of less than seven (fetal outcomes).
  • 6
    Length of labour.
  • 7
    Oxytocin augmentation rates.
  • 8
    Epidural usage.
  • 9
    Proportion with unfavourable cervix at 24 hours after admission.
  • 10
    Requirement for additional inpatient cervical ripening agent.

A cardiotocograph (CTG) was defined as a normal, suspicious or pathological as per the National Institute for Health and Clinical Excellence guideline on Intrapartum care.26

Collection of outcome data

Clinical data

The clinical outcomes were recorded by SSB onto a dedicated Case Report Form. Compliance was assessed, as far as possible, by asking subjects to complete a diary card indicating when they took study medication.

Maternal satisfaction/acceptability data

Women’s experience and satisfaction with labour was assessed in two ways; first, through a series of short questionnaires and secondly, interviews with a sub-sample (to be reported elsewhere). At recruitment women were given short questionnaires to be completed before administering each tablet, the first acting as a baseline. The questionnaire asked about physical symptoms. It also asked them to indicate on a 10-point scale how, respectively, happy, tense and anxious they felt, and whether they were experiencing abdominal pain; in the first questionnaire, women were also asked to complete a standardised personality scale (the Eysenck Personality Scale).27 The final questionnaire, completed after birth and before discharge from the hospital, ascertained women’s’ views (also using a 10-point scale) on the experience, the treatment at home and whether they would repeat the experience and advise a friend to do the same (Table 3). Only those who had taken at least one dose of the trial treatment were asked to complete the questionnaire. The acceptability part of the study was conducted during a consecutive 20-month period (May 2005 to December 2006), during which 220 of the 240 women who could have completed the questionnaires returned the data.

Table 3.   Maternal satisfaction (Likert scales 1–10)
QuestionIMN (= 112–116) Mean (SD)Placebo (= 106–111) Mean (SD)IMN–placebo (95% confidence interval)P-value
  1. Data shown are for number of women (n) by randomised group, with mean difference and 95% confidence interval and associated P-value.

How do you think your labour went? (1 = very easy…10 = very difficult)6.18 (2.46)6.52 (2.16)−0.35 (−0.96, 0.26)0.26
Thinking back, how do you feel about the experience of taking the tablets for the home treatment?  (1 = Extremely good…10 = Not at all good) 3.84 (2.30)3.23 (2.15)0.61 (0.02, 1.21)0.043
How painful do you think treatment at home was?  (1 = Not at all painful…10 = Very painful)2.76 (2.30)2.18 (2.18)0.57 (−0.02, 1.16)0.056
How anxious were you about being at home while taking the treatment? (1 = Not at all anxious…10 = Very anxious)2.50 (1.96)2.39 (1.88)0.11 (−0.40,0.61)0.67
Would you have the same treatment at home in your next pregnancy? (1 = Definitely…10 = Definitely not)3.39 (2.74)2.77 (2.19)0.62 (−0.04, 1.27)0.063
Would you advise a friend to have the same treatment at home? (1 = Definitely…10 = Definitely not)3.10 (2.38)2.69 (2.07)0.41 (−0.18, 0.99)0.17

Randomisation and blinding

Randomisation to IMN or placebo was in the ratio 1:1, using randomised permuted blocks of four. The randomisation schedule was generated by the Study Data Centre at the Centre for Healthcare Randomised Trials (CHaRT) at Aberdeen University. The randomisation schedule was used by the Clinical Trials section of the pharmacy at the Western Infirmary, Glasgow, to prepare treatment packs of identical appearance (either active or placebo) for each patient, each labelled with the relevant unique study number and then allocated to each subject during a telephone call by SSB to an automated interactive telephone response (IVR) randomisation service at CHaRT in Aberdeen.

Failure to initiate on randomised study medication

Recruited women who went into spontaneous labour in the interval between being randomised and initiating on their randomised study medication – usually 7 days after being recruited and 48 hours before their scheduled admission, would not have started on study drug. These women, and any women who withdrew from the study before being eligible to start on their randomised medication were included in the primary analysis (with their consent), but excluded from a supporting secondary analysis.

Statistical issues

The required sample size used the primary outcome of time from hospital admission to delivery. In a previous study of women undergoing induction of labour with Bishop scores 0–4 and 5–8, the mean (SD) induction to delivery intervals were 19.7 (10.8) hours and 15.3 (6.8) hours respectively (JE Norman, data on file). With 150 women in each of the IMN and placebo groups (300 in total) the study had a 96% power at a 5% level of significance to detect a difference in mean time from admission to delivery of 4 hours, assuming a common standard deviation of 9 hours, using a two sample two-sided t-test. We had anticipated some lack of initation on treatment in our published protocol,23 and reported power calculations based on estimates of 12.5% and 25% not initiating on treatment. The study still had a 93% power to detect a difference of 3 hours and 30 minutes with 250 women and 81% power with 200 women, as we reported.23 However, the sample size was increased to 350 after 200 women were recruited, because of more women than anticipated going into spontaneous labour during the 7-day gap between being randomised and starting the study medication, and the resultant increased variability in admission to delivery interval. An increase in sample size of 50 was agreed without unblinding or examination of any treatment effect and approved prospectively by the Ethics committee and MHRA.

For the primary outcome of patient satisfaction, we assumed the results would be similar to the PRIM study (that those pretreated with IMN will give a mean (SD) score of satisfaction of 7.0, and those who have prostaglandins only as the active ripening agent will give a mean (SD) score of satisfaction of 5.8, with a common standard deviation of 3.0). The study, therefore, had 93% power to detect a significant difference in satisfaction rates, given a sample size of 150 in each group.

Statistical analysis

The primary endpoint of time from admission to delivery was compared between the two groups using a two sample two-sided t-test, and then adjusted for covariates pre-specified as important using a linear model. Analysis of secondary outcomes was performed using t-tests (or Wilcoxon tests) or chi-squared tests as appropriate to the distribution of the data, again with adjustment for covariates using appropriate generalised linear models. All primary analyses were performed according to the intention to treat principle. No interim analyses were performed, other than that requested at a single meeting of the Data Monitoring Committee when about half the primary outcome data were mature. All data were analysed in SAS 9.02 for Windows (SAS Institute, Cary, NC, USA). No adjustments have been made for any multiple comparisons.

Data and safety monitoring

No concerns were raised by the independent Data and Safety Monitoring Committee.

Role of funding source

The funder (WellBeing) had no input into the study design, data collection, data analysis, data interpretation or writing of the final report.

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Aims
  5. Methods
  6. Results
  7. Discussion
  8. Conclusion
  9. Disclosure of interests
  10. Contribution to authorship
  11. Details of ethical approval
  12. Funding
  13. Acknowledgements
  14. References

The participant flow diagram is shown in Figure 1. Of the 350 women randomised, 23% women in each group did not have the opportunity to initiate treatment as they went into spontaneous labour before the scheduled time of starting the trial medication (48 hours prior to scheduled IOL). Three percent of women in each group (5 IMN, 6 placebo) withdrew after randomisation and before initiating treatment. One woman in each group was diagnosed to have the baby in breech presentation after randomisation and before initiating treatment. Both chose to have an elective cesarean section and were advised to withdraw from the study. All 350 randomised women, were included in the primary analysis as analysis was by intention to treat. The women randomised to the active group were similar to those randomised to the placebo group (Table 1). The study was conducted according to the published protocol23 with one exception: women were recruited into the study up to 9 days before scheduled induction of labour instead of up to 7 days as we initially planned.

image

Figure 1.  Flow diagram of participants in the IMOP study.

Download figure to PowerPoint

Table 1.   Baseline characterisctics
CharacteristicsIMN (n = 177)Placebo (n = 173)
  1. Data are number of patients (%), unless otherwise stated.

  2. *All women were between 37 and 42 completed weeks gestation. Mean gestational age was at randomisation.

  3. **All previous pregnancies ended at <24 weeks gestation.

Mean age in years (SD)27.1 (6.2)27.4 (6.4)
Mean gestational age in days (SD)*290.0 (2.96)290.1 (2.93)
Any previous pregnancies**38 (21.5)47 (27.2)
Indication for labour induction
Post dates171 (96.6)168 (97.1)
Other6 (3.4)5 (2.9)
Baseline modified Bishop score
Mean (SD)3.18 (1.61)3.12 (1.73)
011 (6)20 (11)
116 (9)15 (9)
235 (20)20 (11)
338 (22)40 (23)
436 (20)39 (23)
528 (16)26 (15)
613 (7)13 (8)
0–3100 (56)95 (55)
4–677 (44)78 (45)

Primary outcomes

Primary clinical outcome

IMN did not significantly shorten the admission to delivery interval as compared with placebo; mean difference IMN–Placebo −1.59 hours (95% confidence interval −5.08 to 1.89, P = 0.37) (Table 2).

Table 2.   Primary outcome – admission to delivery interval in hours
Admission to delivery intervalIMNPlaceboIMN–placeboP-value
nMean (SD)nMean (SD)(95% CI)
All17725.1 (18.1)17326.7 (16.7)−1.59 (−5.08, 1.89) 0.37
Secondary analysis based on the amount of trial treatment taken
Didn’t take treatment4714.2 (10.2)4615.0 (11.3)−0.79 (−5.23,3.64) 0.72
Took treatment (all)13029.0 (18.8)12730.9 (17.7)−1.89 (−6.37, 2.59) 0.41
Took 1 dose2424.0 (18.7)521.1 (15.2)2.9 (−16, 21) 0.75
Took 2 doses413.3 (3.3)1020.1 (18.4)−7 (−28, 14) 0.48
Took 3 doses10230.8 (18.7)11232.3 (17.4)−1.48 (−6.35, 3.39) 0.60
Primary acceptability outcome

There was no difference between the two groups in either abdominal pain or anxiety levels when taking the treatment at home or on admission to hospital for induction of labour (Figures 2 and 3). Women in both groups become more anxious on admission to the hospital. While both groups found the overall experience of home treatment to be positive, women in the placebo group found it marginally more positive than those in the IMN group (just over half a unit on a 10-point scale, P = 0.043) (Table 3). However, there was a clear relationship within the IMN group between those women who experienced strong headaches and their views about the experience. Perhaps not surprisingly, women in the IMN group with strong headaches reported significantly less positive views in relation to views 2–6 inclusive in Table 3 than those who did not (data not shown). Women in both groups reported little anxiety and pain with the home treatment and reported that they were willing to take the same treatment in the next pregnancy and to advise a friend to take the treatment (Table 3).

image

Figure 2.  The X axis indicates the time points at which women completed the questionnaire. This was just before administering each tablet and on admission to the hospital (before the start of inpatient induction of labour). The Y axis indicates the response of the women on a 10-point scale on how anxious they felt.

Download figure to PowerPoint

image

Figure 3.  The X axis indicates the time points at which women completed the questionnaire. This was just before administering each tablet and on admission to the hospital (before the start of inpatient induction of labour). The Y axis indicates the response of the women on a 10-point scale on whether they were experiencing any abdominal or pelvic pain.

Download figure to PowerPoint

Secondary outcomes (Table 4)
Table 4.   Secondary outcomes
OutcomeIMNPlaceboIMN–placeboP value
nMean (SD) or n (%)nMean (SD) or n (%)Mean difference or odds ratio (95% confidence interval)
  1. *Unripe cervix defined as Bishop’s score <7.

  2. **P-value from Wilcoxon test.

Effect on outcome
Mean (SD) change in Bishop’s score1272.29 (2.10)1281.64 (2.06)0.65 (0.14, 1.17)0.013
Proportion women with unripe cervix >48 hours of home treatment*13083 (64%)12798 (77%)0.52 (0.30, 0.90)0.020
Proportion women requiring further prostaglandins13083 (64%)12794 (74%)0.62 (0.36, 1.06)0.079
Proportion women unfavourable cervix >24 hours after admission13016 (12%)12720 (16%)0.75 (0.37, 1.53)0.43
Mean (SD) admission to vaginal delivery interval (hours)6517.4 (15.9)6318.1 (15.3)−1.03 (−6.00, 3.93) 0.68
Vaginal delivery not achieved in 24 hours11781 (69%)12186 (71%)0.92 (0.53, 1.60)0.76
Delivery and neonatal outcomes
Spontaneous vaginal delivery17765 (37%)17363 (36%)1.01 (0.66, 1.57)0.95
Instrumental vaginal birth17747 (27%)17354 (31%)0.80 (0.50, 1.57)0.34
Caesarean section17765 (37%)17356 (32%)1.21 (0.78, 1.89)0.39
Mean (SD) Birthweight (kg)1773.64 (0.44)1733.62 (0.41)0.02 (−0.07, 0.11)0.60
Mean (SD) Apgar score at 5 minutes1779.41 (0.76)1739.33 (0.99)0.08 (−0.11, 0.26)0.41
Labour and birth complications
Suspicious or pathological CTG on admission13001270N/AN/A
Suspicious or pathological CTG after prostaglandin administration819 (11.2%)875 (5.8%)2.05 (0.66, 6.40)0.22
Epidural usage177116 (66%)173120 (69%)0.84 (0.54, 1.31)0.45
Meconium stained liquor17751 (30%) 17353 (31%)0.92 (0.58, 1.46)0.73
Need for oxytocin12780 (63%)12775 (59%)1.18 (0.71, 1.96)0.81
Mean (SD) length of labour (hours)1778.38 (4.74)1738.14 (4.67)0.24 (−0.74, 1.23)0.63
Mean (SD) blood loss (ml)177667 (1318)173563 (506)104 (−107, 315)0.33
Blood loss >500 ml17759 (33%)17347 (27%)1.34 (0.85, 2.12)0.21
Neonatal complications
Stillbirth or neonatal death1770 (0%)1731 (1.73%)N/AN/A
Apgar score <7 at 5 minutes1773 (1.7%)1732 (1.2%)N/AN/A
Admission to neonatal unit17718 (10.2%)17316 (9.3%)1.11 (0.55, 2.26)0.77
Mean (SD) length stay in neonatal unit for admitted babies (hrs)1825 (37)165 (8)20 (1, 39)0.0034**
Maternal complications
Headache – mild11250 (45%)10821 (19%)3.34 (1.83, 6.12)<0.0001
Headache – medium11236 (32%)1081 (1%)51 (7, 378)0.0001
Headache – strong11222 (20%)1081 (1%)26 (3, 198)0.0016
Any headache11274 (66%)10822 (20%)7.61 (4.14, 14.0)<0.0001
Nausea11219 (17%)10813 (12%)1.49 (0.70, 3.20)0.30
Flushes11210 (9%)10812 (11%)0.78 (0.32, 1.90)0.59
Faintness1122 (2%)1083 (3%)0.64 (0.10, 3.89)0.62
Drowsiness11218 (16%)10831 (29%)0.48 (0.25, 0.92)0.026
Vaginal soreness1127 (6%)10813 (12%)0.49 (0.19, 1.27)0.14
Indigestion8933 (37%)8732 (37%)1.01 (0.55, 1.87)0.97
Any side effect (non-headache)11254 (48%)10859 (55%)0.77 (0.46, 1.31)0.34
Any side effect (including headache)11283 (74%)10861 (56%)2.21 (1.25, 3.89)0.0064
Admission (reasons other than labour during treatment at home)13012 (9%)12710 (8%)1.19 (0.50, 2.86)0.70

The mean change in modified Bishop score from recruitment to hospital admission was significantly greater in the IMN group as compared with the placebo group [mean difference of 0.65 (95% CI 0.14, 1.17, P = 0.013)]. The proportion of women with an unripe cervix (defined as Bishop score of ≤6) after 48 hours of outpatient treatment was significantly less in the IMN group as compared with the placebo group [64% versus 77%; odds ratio 0.52 (95% CI 0.30, 0.90, P = 0.020)]. The proportion of women requiring prostaglandins was also lower in the IMN group as compared with placebo group (64% versus 74%), but this difference did not reach statistical significance. By 24 hours after admission, the proportion of women with an unripe cervix was similar in both groups. There was no difference in the admission to vaginal delivery interval, the induction to delivery interval or the proportion of women not achieving vaginal delivery in 24 hours between the two groups.

There was no difference in cardiotocographic abnormalities, use of epidural analgesia in labour, oxytocin requirements during labour, length of labour, mode of delivery and blood loss at delivery between the two groups.

There was no difference in the neonatal outcomes with respect to birthweight, 5 minute Apgar score, proportion of babies with Apgar score of <7 at 5 minutes and the proportion of babies admitted to the neonatal unit between the two groups. However, of those babies admitted to the neonatal unit, babies in the IMN group spent on average a longer time in the unit than those in the placebo group: 20 hours (95% CI 1, 39, t-test P-value 0.0030, Wilcoxon Test P = 0.0034).

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Aims
  5. Methods
  6. Results
  7. Discussion
  8. Conclusion
  9. Disclosure of interests
  10. Contribution to authorship
  11. Details of ethical approval
  12. Funding
  13. Acknowledgements
  14. References

The findings of this study refute our primary clinical hypothesis that IMN would significantly reduce the admission to delivery interval. However, we confirmed our own work and the work of others demonstrating that IMN is an effective cervical ripening agent in each of the first and third trimesters when measured either by change in Bishop score (assessed clinically) or in a reduction of cervical resistance (assessed by using a tonometer).5,8 This combination of demonstrable cervical ripening efficacy, but lack of a substantial improvement in the process of induction of labour, is perhaps surprising. There is evidence in the literature to suggest that ripening of cervix prior to induction of labour in women either pharmacologically or physiologically is associated with a shorter latent and active phase of labour and a decreased rate of cesarean section.28,29 We were unable to demonstrate this following cervical ripening with IMN, leading us to conclude that the clinical utility of IMN for pre-induction cervical ripening is minimal, at least in the doses, formulations and frequency used here and for the patient population in whom it was used.

Only two other studies have used IMN on an outpatient basis for cervical ripening.18,21 Bullarbo et al.18 found that more of the women who had IMN (all of whom had a single dose) compared with those who had placebo went into labour within 24 hours of treatment [22% (22/100) versus 8% (8/100), P = 0.01)]. In contrast, we found that similar proportions of women in both groups went into labour within 24 hours [9% following IMN (12/130) versus 4% following placebo (5/127), P = 0.09]. Differences in findings between the two studies may possibly be explained by the differences in the parity of patients. Bullarbo et al.23 recruited both nulliparous and multiparous women, while our study was confined to nulliparous women. There are also differences in the mean gestational age at delivery between the two studies (42.2 weeks in the study by Bullarbo et al.18 as compared with 41.4 weeks in our study). Indeed, subgroup analysis of the primigravid women in the Bullarbo et al. study shows (as did ours) that there was no significant difference in the proportion of women with a ripe cervix following outpatient care [17% (12/72) versus 8% (6/74), P = 0.11 in Bullarbo et al.’s study; 9% (12/130) versus 4% (5/127), P = 0.09 in our study]. Somewhat surprisingly, Bullarbo et al. failed to demonstrate an improvement in the mean Bishop score following IMN despite showing clinical effectiveness in shortening labour, in contrast to our own data.

Habib et al.21 found that the admission to delivery interval was significantly shorter in the IMN group as compared with the placebo, in contrast to our own results. Differences in the findings could possibly be explained by the differences in parity of patients, mean gestational age at delivery and the indication for the induction of labour. Habib et al.21 recruited both nulliparous and multiparous women, while our study was confined to nulliparous women. Habib et al. have not given the subgroup analysis for the nulliparous and multiparous women, hence it is not possible to know if their findings are applicable to both groups. There are also differences in the mean gestational age at delivery between the two studies (39.2 weeks in the study by Habib et al.21 as compared with 41.4 weeks in our study) and the indication for induction of labour (only 35% had induction of labour for postdates in the study by Habib et al. as compared with 97% in our study). It has been shown that nitric oxide metabolites in the cervical fluid are lower in women going beyond term as compared with those delivering at term, suggesting a possible inherent cervical ripening incapacity in women with postterm pregnancies.30 Also, the nitric oxide metabolites are more in parous women as compared with nulliparous women, suggesting that cervical ripening with IMN may be more successful with parous women as compared with nulliparous women.

A difficulty with the study reported here is that the relatively long interval between the decision to induce labour and randomisation on the one hand, and the actual date of induction of labour on the other, meant that over 20% of participants did not initiate on randomised treatment, thus weakening the power of the study. However, even when analysis is restricted to women who actually initiated treatment (257 women), admission to delivery interval was only shortened by 1.89 hour (95% CI shortened by 6.37 hours to lengthened by 2.59 hours).

As IMN was self administered by the women in our trial in the home setting, there was a potential concern about compliance and suboptimal delivery of the assigned treatment which could account for the lack of efficacy. However, given that a greater proportion of women had a ripe cervix after active treatment, together with the finding that 66% of the women in the IMN group in our trial had headaches (the most common side effect of IMN), it seems likely that most of the women were complying with the treatment. It is, however, possible that the results may have been different (i.e. the treatment more effective) if the tablets had been inserted by a healthcare professional.

Putting these findings together, we believe that IMN does ripen the cervix, but that its effects on clinically important parameters such as admission to delivery interval are insufficient to recommend its use on either an inpatient or an outpatient basis in postterm nulliparous women. We have explained the rationale for the dosage, dosage interval and the duration of treatment of IMN in our protocol,23 but it may well be that the dosage or the formulation that we used were inadequate. It is also possible that the reduction in the induction to delivery interval with other cervical ripening agents such as prostaglandins relates more to stimulation of uterine contractions rather than to cervical ripening. In addition, prostaglandins have a recognised role in inducing the presence of uterine oxytocin receptors, which is not a feature of IMN. Moreover, NO donors are potent tocolytic agents. Vaginally administered IMN may have systemic effects by vaginal absorption thus leading to myometrial relaxant effects which may influence outcome measures, in particular the admission to delivery interval.

In addition to lack of reduction in the admission to delivery interval, our study raises possible concerns about safety aspects of nitric oxide donors. We have previously shown that NO donors, in the doses and formulations shown here, have no significant adverse effects on maternal or fetal haemodynamics6 or fetal heart rate patterns.2 In this study, although the proportion of babies admitted to the neonatal unit was similar in both groups, the length of the stay in the neonatal unit was greater in the IMN group compared with the placebo group. The greater stay in the IMN group is of potential concern although there is no obvious mechanism by which IMN could cause this. The majority of the babies who were admitted to the neonatal unit were admitted for a short period (from 2–24 hours). There were seven babies who spent longer than a day in the neonatal unit. The reasons for their admission were meconium aspiration (2), Group B streptococcal infection (2) and one each of respiratory distress, neonatal pneumonia and sepsis, none of which have biological plausibility for attribution to IMN. While other trials have found similar neonatal admission rates in the two arms,2,18,31 none have looked at the length of stay in the neonatal unit.

Regarding acceptability, maternal enthusiasm for home cervical ripening was evident and women reported very positive views on home treatment with low anxiety and pain levels in both the IMN and placebo groups. When compared with IMN, women found the overall experience with placebo more positive, but this is not surprising as IMN caused headaches but did not improve the process of labour. The women in both study arms had similar scores in terms of psychological characteristics. Anxiety levels of both groups of women rose significantly on admission to the hospital suggesting that women find hospital admission stressful rather than reassuring. Furthermore, when questioned before hospital discharge, women in both group were very willing to try the same treatment at home in their next pregnancy and advise it to a friend. This is a very strong endorsement by the women of outpatient cervical ripening as part of the process of induction of labour. It is possible that those who agreed to participate in the study were a self-selected group with low base-line anxiety, but their mean scores for neuroticism (which includes a strong emotional reaction of anxiety) were similar between the two groups [5.10 (3.19) and 5.13 (3.24) for IMN and placebo group] and when compared with normative values given in the EPQ-R manual of 5.93 (2.89).27 Additionally, 65% of the women (350/541) that were approached and eligible were willing to participate in the study suggesting that their views could be generalisable to the entire population.

We have previously highlighted the difficulties in measuring the effects of cervical ripening agents.2 This is partly because some women go into spontaneous labour and this has led to inconsistent reporting between trials. Additionally, it is difficult to determine appropriate outcome measures in clinical practice. It has been suggested that time to (vaginal) delivery and vaginal delivery not achieved in 24 hours should be used as a benchmark to measure the efficacy of a chosen induction method.1 While we report these outcomes, we believe that they are not the only factors that induction of labour should be measured by, and that women’s views are also important. While speed with which an agent achieves induction is the most commonly reported outcome measure, it is not necessarily the most important effect of a cervical ripening or induction agent. The outcome considered most important will vary depending on the indication for labour induction and it is likely that a variety of outcomes are important. Where induction of labour is applied for post dates pregnancy, our own belief is that patient satisfaction, achieving vaginal (rather than caesarean) delivery and minimising adverse events in mother and baby are all important outcomes. We suggest that there should be discussion and consensus among healthcare providers as to what the best outcome measures might be.

Conclusion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Aims
  5. Methods
  6. Results
  7. Discussion
  8. Conclusion
  9. Disclosure of interests
  10. Contribution to authorship
  11. Details of ethical approval
  12. Funding
  13. Acknowledgements
  14. References

In summary, we have shown that IMN given at home ripens the cervix but is not efficacious enough to shorten the admission to delivery interval or improve maternal satisfaction. Thus, the clinical utility of IMN in this setting is likely to be limited. In contrast women reported very positive views on cervical ripening at home. IMN does not appear to be a useful ripening agent when used at home, but the strong endorsement for the principle of cervical ripening at home should encourage researchers to continue to evaluate other agents in this setting.

Contribution to authorship

  1. Top of page
  2. Abstract
  3. Introduction
  4. Aims
  5. Methods
  6. Results
  7. Discussion
  8. Conclusion
  9. Disclosure of interests
  10. Contribution to authorship
  11. Details of ethical approval
  12. Funding
  13. Acknowledgements
  14. References

Shrikant Bollapragada was responsible for recruitment of patients, collection of data, execution of the study and prepared the first draft of the paper. Fiona Mackenzie was responsible for the proposal, design and supervision of the study. John Norrie was responsible for the study proposal, design and analysis of data. Stavros Petrou was responsible for the proposal and design of the study and analysis of the health economics data. Oya Eddama was responsible for the collection and analysis of health economics data. Margaret Reid was responsible of study proposal and design of the study and for the interviews with the subjects and holding focus groups with the midwives. Jane Norman was responsible for the study proposal, design, analysis of data, supervised the work, was the Chief Investigator and is the corresponding author. All authors contributed to drafting the final manuscript.

Details of ethical approval

  1. Top of page
  2. Abstract
  3. Introduction
  4. Aims
  5. Methods
  6. Results
  7. Discussion
  8. Conclusion
  9. Disclosure of interests
  10. Contribution to authorship
  11. Details of ethical approval
  12. Funding
  13. Acknowledgements
  14. References

The study was approved by the North Glasgow University NHS Trust ethics committee and written informed consent was gained prior to randomisation.

Funding

  1. Top of page
  2. Abstract
  3. Introduction
  4. Aims
  5. Methods
  6. Results
  7. Discussion
  8. Conclusion
  9. Disclosure of interests
  10. Contribution to authorship
  11. Details of ethical approval
  12. Funding
  13. Acknowledgements
  14. References

The IMOP study was funded by WellBeing. http://www.wellbeingofwomen.org.uk/The funder had no input into the study design, data collection, data analysis, data interpretation or writing of the final report.

Acknowledgements

  1. Top of page
  2. Abstract
  3. Introduction
  4. Aims
  5. Methods
  6. Results
  7. Discussion
  8. Conclusion
  9. Disclosure of interests
  10. Contribution to authorship
  11. Details of ethical approval
  12. Funding
  13. Acknowledgements
  14. References

This work was supported by a grant from the Wellbeing for which we are grateful. We thank all the patients who participated in the study and all the doctors and midwives at the Princess Royal Maternity Hospital. We are grateful to the chair (PO) and members of the Data and Safety Monitoring committee: Dr Philip Owen, Dr Lena Macara, Professor Deirdre Murphy and Dr Chris Weir. We are grateful to Dr Andrew Thomson for being the independent chair of the Trial Management Committee. We are grateful to Dr Inass Osman and Professor Ian Greer for their comments on the protocol and their help in securing grant funding. We are grateful to Susan McKechnie, Liz Terrace, and June Grant for their help with collection of the health economics data and Liz Shroeder who helped with the analysis of the health economics data. We are very grateful to Alison McDonald (trial management) and Gladys McPherson (database management) and Graeme Maclennan (statistics for the DMC), all of CHaRT in Aberdeen, for their work on the study.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Aims
  5. Methods
  6. Results
  7. Discussion
  8. Conclusion
  9. Disclosure of interests
  10. Contribution to authorship
  11. Details of ethical approval
  12. Funding
  13. Acknowledgements
  14. References