- Top of page
Objective The aim of this study was to investigate the efficacy and safety of recombinant human relaxin (rhRIx) as a cervical ripening agent in women with an unfavourable cervix before induction of labour at term.
Design A multi-centre, double-blind, placebo-controlled trial performed in Edinburgh, Glasgow and Oxford. Women were treated with 0, 1, 2 or 4 mg of rhRIx in a gel vehicle administered intravaginally. Analysis of variance tests were performed on all continuous variables, and Cochran Mantel-Haenszel tests employed for all discrete variables.
Participants Ninety-six women at 37 to 42 weeks of gestation with a singleton pregnancy and a modified Bishop score of 4 were recruited.
Results There was no significant difference in the change in modified Bishop score between the four treatment groups. The lengths of the first and second stages of labour were similar in all 4 groups. PGE2 and oxytocin requirements were similar in all groups, as was the mode of delivery. There was no evidence that relaxin was absorbed systemically when given in this way.
Conclusion Recombinant human relaxin 1 to 4 mg, administered as an intravaginal gel, has no effect as a cervical ripening agent before induction of labour at term.
- Top of page
The success of labour induction is influenced by the state of the cervix; induction of labour in the presence of an unripe cervix is associated with an increase in both maternal and neonatal morbidity1. Cervical ripening occurs during the phase known as pre-labour. Its exact mechanism remains unclear but the process results in structural changes within the cervix making it more compliant. These include a decrease in collagen concentration, an increase in water content and an alteration in proteoglycan/glycosaminoglycan ratio within the tissue. It is this process that we are striving to mimic when ripening the cervix pharmacologically. The ideal cervical ripening agent should have a selective effect on the cervix without any additional effect on uterine contractility, and in theory the polypeptide hormone relaxin would meet these requirements.
Relaxin was first identified in 1926 by Hisaw2 who demonstrated its capacity to promote separation of the pubic symphysis in guinea-pigs. Relaxin facilitates connective tissue remodelling, and in some animals also inhibits myometrial contractility. In most species, cervical ripening before the onset of labour is associated with an increase in serum relaxin concentrations. However, this is not the case in humans and the exact role of this hormone in human parturition remains uncertain. Early clinical trials3,4 employing porcine relaxin for cervical ripening in human subjects reported conflicting results which probably reflected the impurity of the agent at that time. The subsequent purification of porcine relaxin5 led to renewed interest in this hormone. A number of studies showed that porcine relaxin had some therapeutic benefit as a cervical ripening agent in women6,7, despite the primary peptide structure of porcine relaxin having only about 50% homology with that of human relaxin. The development of recombinant human relaxin by Genentech Inc (San Francisco, California, USA) has provided an agent that could have superior cervical ripening properties in women and has stimulated further research in this field.
Phase I studies conducted by Genentech demonstrated that recombinant human relaxin was safe and not associated with serious adverse effects, and that there was no maternal development of antibody to rhRIx following treatment8,9. Phase II studies were therefore initiated in Australia and the UK. This paper reports the results of the UK double-blind, three-centre study investigating the effect of recombinant human relaxin, administered as an intravaginal gel, on cervical ripening in pregnant women at term with an unfavourable cervix.
- Top of page
Ethical approval for this study was obtained from the local ethics committees of the three participating centres.
Ninety-six women, comprising nulliparous and parous patients, were recruited to the study. All women had a singleton pregnancy of at least 37 weeks of gestation with a cephalic presentation and a modified Bishop score10 of 4 or less. Women with a uterine scar, ruptured membranes or evidence of placental abruption or placenta praevia, were not recruited to the study. Other exclusion criteria were significant systemic disease, recent ingestion of aspirin or other nonsteroidal anti-inflammatory drugs, fetal malformation, growth retardation or macrosomia, and oligo- or polyhydramnios. The decision to induce labour was made by the woman's attending obstetrician and written informed consent was obtained before recruitment. The indication for induction in the majority of cases was either pregnancy induced hypertension or prolonged pregnancy.
The study involved a randomised, double-blind comparison of four treatment regimens: 0 (placebo), 1, 2 or 4 mg of recombinant human relaxin (rhRIx) in a gel vehicle. The randomisation sequence was generated in blocks of four by the Genentech Biostatistical Department and the code was stratified according to parity. A randomisation list accompanied each drug shipment. Primiparous women were assigned sequential numbers starting from the top of the list, and multiparous women were assigned sequential numbers starting from the bottom of the list. The study medication was supplied by Genentech, Inc., as a two-part administration set: a stoppered 20 cm3 glass vial of lyophilised rhRIx or placebo and a 5 cm3 syringe of gel. The vial contained 0 mg or 12 mg rhRIx in isotonic citrate buffer. The syringe delivered 3 mL of sterile 4% methylcellulose gel in isotonic citrate buffer. The active doses were made by reconstituting 12 mg of lyophilised study medication in either 12, 6 or 3 mL of sterile water and combining this with the fixed 3 mL volume of methyl-cellulose gel.
Women were admitted to hospital in the afternoon before the day of induction. A medical and obstetric history was taken and a general examination performed. Fetal weight and amniotic fluid volume were estimated by ultrasound scan and a cardiotocograph performed. Blood was taken for haematology, coagulation profile, biochemistry, serum relaxin level and serum antibody to relaxin. Upon randomisation, the study medication was reconstituted using sterile water by one of the hospital pharmacy staff and delivered to the labour ward. The baseline modified Bishop score was recorded and the study gel administered intravaginally to the posterior fornix that evening. The woman remained recumbent for 1 hour following gel application. Blood pressure, pulse, respiration rate, uterine activity and fetal heart rate were monitored hourly for 4 hours post-treatment, and then every 4 hours for at least 24 hours or until delivery. Observations were suspended overnight if the woman was asleep. Blood was taken at 1,4, 15 and 24 hours after gel administration for serum relaxin concentrations.
The following morning, at approximately 15 hours post-treatment, a vaginal examination was performed by the same investigator and the modified Bishop score recorded after which prostaglandin E2 (PGE2) gel 2 mg was instilled intravaginally. Further PGE2 gel, 1 mg or 2 mg, was given after a 6-hour interval as required. Amniotomy was performed when the cervix was at least 3 cm dilated and fully effaced, and oxytocin administered according to hospital protocol. Uterine activity and fetal heart rate were monitored continuously throughout labour. If labour occurred before the planned 15 hour assessment, the modified Bishop score was recorded at the onset of labour and the woman managed as above. The onset of labour was defined as the onset of regular painful contractions with evidence of progressive cervical dilatation of the cervix. For women delivered by caesarean section before the second stage of labour, the length of the first stage was taken to be the total length of labour.
At delivery, umbilical cord blood was taken to assay relaxin concentration and blood gas analysis, and maternal blood for serum relaxin concentration. Apgar scores were recorded at 1 and 5 minutes post-delivery. Twenty four hours after delivery, maternal blood was collected for repeat haematology, coagulation profile, routine biochemistry and serum relaxin concentration. A general physical examination was also performed. Women were reviewed 6 weeks post-delivery when serum relaxin levels were measured and any postnatal problems documented.
The primary outcome measure was the change in modified Bishop score between baseline and 15 hours post-treatment. Secondary outcome measures included duration of first and second stage of labour, prevalence of ‘spontaneous’ labour and caesarean section, and oxytocin and total PGE2 requirement. Maternal safety measures included vital signs, haematology and biochemistry screening. Fetal safety measures included stillbirth, neonatal death, fetal heart rate disturbances, Apgar scores, cord blood gases, need for resuscitation and incidence of neonatal morbidity.
Sample size and power consideration
Ninety-six women were recruited to the study with 24 women in each of the four treatment groups. The primary outcome measure, change in modified Bishop score, was first to be analysed by ANOVA to take account of any effect of dose influencing modified Bishop score. If a significant effect were found subsequent pairwise comparisons were to have been performed. The determination of sample size was based on such pairwise comparisons. Assuming that the standard deviation and the between-treatment difference with respect to the change in cervical score are equal, a two-sided t test with a significance level of 1.67% will have 84% power to detect such a difference.
All data were recorded on standardised case record forms and analysed at Genentech Inc in California. Analysis of variance tests were performed on all continuous variables, and Cochran Mantel-Haenszel tests employed for discrete variables.
- Top of page
Numbers of women in the four treatment groups were as follows: placebo: n= 23, 1 mg: n= 23, 2 mg: n= 25 and 4 mg: n= 25. There were no statistically significant differences in the characteristics of the women in the four groups (Table 1).
Table 1. Subject characteristics at recruitment by treatment group. Age and gestation analysed by ANOVA; parity and baseline modified Bishop score analysed by Cochran Mantel-Haenszel test. Values are given as mean (SD).
| ||Treatment group|| |
|Characteristic||Placebo (n= 23)||1 mg (n= 23)||2 mg (n= 25)||4 mg (n= 25)||P|
|Age (years)||27.0 (21.3-32.7)||26.8 (21.3-32.3)||26.7 (22.1-31.3)||25.8 (21.4-30.2)||0.796|
|Modified Bishop score||2.9 (2.0-3.8)||3.0 (2.3-3.7)||2.8 (1.9-3.7)||2.5 (1.4-3.6)||0.076|
|Gestation (weeks)||40.0 (38.5-41.5)||39.6 (38.2-41.0)||39.9 (38.8-41.0)||40.1 (38.9-41.3)||0.694|
|Parity|| || || || || |
The mean changes in modified Bishop score were 1.64 for placebo, 1.35 for 1 mg relaxin, 1.76 for 2 mg relaxin and 1.32 for 4 mg relaxin. There was no statistically significant difference in the increase in modified Bishop score between the four groups (Table 2). In addition, there was no significant difference between the four groups when nulliparae and parous women were analysed separately (data not shown). The length of the first and second stages of labour was similar in all treatment groups. There were no differences in the time from treatment to first stage of labour or from treatment to delivery between the groups (Table 2). Excluding the patients delivered by caesarean section from the analysis did not alter the results for any of the above variables (data not shown).
Table 2. Labour characteristics by treatment group. All parameters analysed by ANOVA. Values are given as mean (95% CI).
| ||Treatment group|| |
| ||Placebo||1 mg||2 mg||4 mg||P|
|Change in Bishop score||1.64 (1.0-2.3)||1.35 (0.7-2.0)||1.76 (0.5-3.0)||1.32 (0.8-1.9)||0.853|
|Duration of 1st stage (h)||4.9 (3.9-5.9)||5.0 (3.7-6.2)||5.3 (4.1-6.5)||6.5 (5.1-7.9)||0.222|
|Duration of 2nd stage (h)||0.7 (0.4-1.1)||1.3 (0.8-1.7)||1.0 (0.5-1.5)||1.1 (0.6-1.5)||0.387|
|Total duration of labour (h)||5.6 (4.6-6.6)||6.2 (4.7-7.8)||6.3 (4.9-7.7)||7.6 (5.9-9.2)||0.286|
|Treatment to 1st stage (h)||22.3 (19.2-25.3)||23.7 (21.0-26.3)||26.9 (22.3-31.6)||23.1 (20.2-25.9)||0.202|
|Treatment to delivery (h)||28.0 (24.8-31.2)||29.9 (26.7-33.2)||39.3 (26.4-52.2)||36.7 (23.8-49.6)||0.306|
The majority of women required induction 15 hours after study medication and only a few women laboured before this time (Table 3). The use of PGE2 and oxytocin is also illustrated in Table 3. There was no statistical difference in requirement for either drug between groups, and the doses employed were similar in each treatment group. There was no significant difference in the mode of delivery between the groups with the majority of women achieving a spontaneous vaginal delivery. The overall caesarean section rate for the study was approximately 20% and the apparent increase in section rate in the 2 mg group did not reach statistical significance. Again, there was no statistical difference in the variables displayed in Table 3 when primigravid and parous women were analysed separately, or when women delivered by caesarean section were excluded from the analysis (data not shown).
Table 3. Delivery characteristics, oxytocin and PGE, requirements, by treatment group. Total oxytocin and PGE, use analsyed by ANOVA; mode of labour and delivery analysed by Cochran Mantel-Haenszel test. Values are given as n and mean (95% CI).
| ||Treatment group|| |
| ||Placebo (n= 23)||1 mg (n= 23)||2 mg (n= 25)||4 mg (n= 25)||P|
|Spontaneous labour||*|| || || || |
| Yes||2||1||5||2|| |
|Women requiring PGE2||18||18||21||20*||0.928|
|Total PGE2 dose (mg)||2.1 (1.5-2.6)||2.1 (1.5-2.6)||2.6 (2.0-3.1)||2.3 (1.8-2.8)||0.553|
|Women requiring oxytocin||10||13||12||13||0.916|
|Total oxytocin dose (U)||2.4 (0.4-4.4)||2.7 (1.2-4.1)||2.6 (0.8-4.5)||2.4 (1.1-3.7)||0.993|
|Mode of delivery:||*|| || || || |
|Operative vaginal||2||6||4||5|| |
|Caesarean section||4||3||8||4|| |
Maternal and neonatal safety outcomes
There was no significant difference in pre- and post-treatment measurements of pulse and respirations between the four groups. There was a tendency to lower blood pressures in the treatment groups when compared with placebo and this was statistically significant for diastolic blood pressure: placebo −83 mmHg; 1, 2 and 4 mg rhRIx −77, −77 and 70 mmHg, respectively (P < 0.012). This effect was seen over the first 14 hours following the start of treatment. Haematological and biochemical profiles were within the normal range for pregnancy. There was no association between treatment group and maternal outcome measures such as post-partum haemorrhage, infection and urinary retention.
There were significantly higher fetal heart rates in the treatment groups: placebo: 133 beats per minute (bpm); 1,2 and 4 mg rhRIx: 136,141 and 137 bpm, respectively (P 0.011). This effect was evident for approximately 24 hours post-treatment. There were no significant differences in 1 and 5 min Apgar scores, or cord blood gas results between the groups. Neonatal outcome measures such as meconium stained amniotic fluid, hypoglycaemia and hyperbilirubinaemia occurred with the same frequency in all treatment groups. There were no stillbirths or neonatal deaths in the study.
Maternal serum relaxin concentrations were assayed by Genentech. Concentrations following intravaginal administration of 1,2 and 4 mg rhRIx were no different from the endogenous relaxin levels measured in the placebo group (Fig. 1). There was no correlation between baseline serum relaxin concentration and base-line modified Bishop score, or maximum serum relaxin concentration and change in modified Bishop score. Most cord blood relaxin concentrations were below the level of detection of the assay irrespective of treatment group. Details of the relaxin assay are ‘in house’ at Genentech.
Figure 1. Maternal serum relaxin concentrations (pg/mL): ▪= placebo, □= 1 mg, = 2 mg, = 4 mg, pre- and post treatment (mean, 95% CI).
Download figure to PowerPoint
- Top of page
A precise role for relaxin in the physiology of human parturition has yet to be defined. In a number of animals (eg. pig and rat), the onset of labour is clearly preceded by a surge in serum relaxin concentrations and this appears to facilitate connective tissue remodelling in the reproductive tract. In human pregnancy serum relaxin concentrations peak at approximately 10 weeks of gestation and thereafter concentrations fall progressively towards term11. Labour itself is not associated with any further change in relaxin concentrations11. If relaxin plays a key role in cervical ripening it would be during the time of pre-labour that changes in serum relaxin concentrations might occur.
The corpus luteum is the primary source of relaxin during pregnancy, although extra-ovarian sites of production have been identified12. The fact that relaxin is not a pre-requisite for cervical dilatation has been demonstrated by the successful induction of labour in a woman with premature ovarian failure in whom serum relaxin levels were too low to record13. This finding also highlights the fact that relaxin production by other tissues is very small compared with ovarian production.
This study did not show any therapeutic effect of recombinant human relaxin on cervical ripening when administered as an intravaginal gel. There was no significant difference in mean change in modified Bishop score following treatment with 1, 2 or 4 mg of rhRIx compared with placebo. This is not surprising since pre-and post-treatment serum relaxin concentrations were the same in the placebo and active groups suggesting that relaxin was not absorbed from the preparation when administered via this route or that the dosages were too small. There was no association between endogenous serum relaxin levels and baseline modified Bishop score, or maximum serum relaxin concentrations and change in modified Bishop score. In addition, there was no change in endogenous levels throughout induction, labour and delivery which confirms previous findings. RhRIx appears to be safe and not associated with any clinically significant side effects. In view of the lack of evidence for absorption of rhRIx, which is reflected in the absence of clinical effect, the changes in maternal diastolic blood pressure and fetal heart rate, which were small and clinically irrelevant, may well have been spurious.
These findings are in keeping with a smaller study14 employing recombinant human relaxin in a dose of 1.5 mg which also failed to demonstrate a significant effect on cervical ripening. Similarly, this dose of recombinant human relaxin was not associated with any adverse maternal or fetal complications. There could be a role for relaxin in cervical ripening without any parallel increase in circulating relaxin concentrations if it were acting at receptor level. However, to date, a receptor for relaxin in humans has not been identified.
The reduction in cervical collagen concentration necessary for cervical ripening can, in part, be explained by an increase in enzymatic collagen degradation. Labour is associated with an increase in circulating collagenase levels15 and collagenase activity within cervical tissue. It has been reported that relaxin increases collagenase activity16. Relaxin receptors have been identified on human fibroblasts17, which along with leucocytes are capable of collagenase production. A number of clinical trials have demonstrated a cervical ripening effect of porcine relaxin in pregnant women6,7. However, these trials were small and the women were of mixed parity. In addition, in one study6, women in the treatment group tended towards higher baseline Bishop scores limiting the conclusions that can be drawn. Cervical ripening has been likened to an inflammatory response18 and one potential mechanism whereby porcine relaxin, although purified, could stimulate cervical ripening is by provoking an immunological reaction resulting in neutrophil degranulation and release of collagenase, perhaps mediated by inter-leukin-8 which has recently been shown to be produced by the human cervix19.
The effect of relaxin on myometrial contractility in different species is varied. An inhibitory effect of purified porcine relaxin on the spontaneous contractility of nonpregnant human myometrium has been demonstrated in vitro20. This effect has been confirmed on nonpregnant and pregnant myometrium from rats and pigs, but porcine relaxin had little, if any, effect on spontaneous or induced, pregnant or nonpregnant, human myometrial contractility in vitro21. More recently, human relaxin has been shown to have only a minor effect on human myometrial contractility in late pregnancy22. The data for uterine activity in this clinical trial were not assessed, but as there was no difference in duration of labour or need for augmentation between groups it is likely that contractility was unaffected.
Prostaglandins are currently the most successful agents for ripening the cervix pharmacologically23 but they have the disadvantage of simultaneously stimulating uterine activity. Although theoretically relaxin meets the requirements for the ideal cervical ripening agent, that is ripening without contractility, this study has not demonstrated any cervical ripening effect of recombinant human relaxin. This may simply be due to incorrect choice of dosage. However, it seems more likely that the route of administration for this large polypeptide hormone was inappropriate. The intravenous route may be more suitable and deserves to be investigated before we can discard relaxin as a ripening agent.
We are grateful to Genentech, Inc. for their support in the coordination and financing of this trial. Dr M. M. Chou was awarded a Research Fellowship from the Taichung Veterans General Hospital.