OBJECTIVE: To compare immediate with delayed (4 hours) oxytocin infusion after amniotomy on vaginal delivery within 12 hours and patient satisfaction with the birth process.
METHODS: Parous women with favorable cervixes after amniotomy for labor induction were randomized to immediate titrated oxytocin or placebo intravenous infusion in a double-blind noninferiority trial. After 4 hours, study infusions were stopped, the women were assessed, and open-label oxytocin was started if required. Maternal satisfaction with the birth process was assessed with a 10-point visual numerical rating scale (lower score, greater satisfaction).
RESULTS: Vaginal delivery rates at 12 hours were 91 of 96 (94.8%) compared with 91 of 94 (96.8%) (relative risk 0.98, 95% confidence interval [CI] 0.92–1.04, P=.72), and maternal satisfaction on a visual numerical rating scale (median [interquartile range]) was 3 [3–4] compared with 3 [3–5], P=.36 for immediate compared with delayed arm, respectively). Cesarean delivery, maternal fever, postpartum hemorrhage, uterine hyperactivity, and adverse neonatal outcome rates were similar between arms. The immediate oxytocin arm had a shorter amniotomy-to-delivery interval of 5.3±3.1 compared with 6.9±2.9 hours (P<.001) and lower epidural analgesia rate of 2.9% compared with 9.9% (relative risk 0.3, 95% CI 0.1–1.0, P=.046), but fetal heart rate abnormalities on cardiotocogram were higher, 28.6% compared with 16.8% (relative risk 1.7 95% CI 1.0–2.9, P=.048). In the delayed arm, oxytocin infusion was avoided by 35.6%.
CONCLUSIONS: Immediate or delayed oxytocin infusions are reasonable options after amniotomy for labor induction in parous women with favorable cervixes. The choice should take into account local resources and the woman's wish.
CLINICAL TRIAL REGISTRATION: ISRCTN Register, http://isrctn.org, ISRCTN51476259.
LEVEL OF EVIDENCE: I
After amniotomy for labor induction in parous women, immediate or delayed oxytocin infusion displays a similar vaginal delivery rate within 12 hours and similar patient satisfaction.
Department of Obstetrics & Gynaecology, University of Malaya, Kuala Lumpur, Malaysia.
Corresponding author: Peng Chiong Tan, FRCOG, Department of Obstetrics & Gynaecology, University of Malaya, Lembah Pantai, Kuala Lumpur 50603, Malaysia; e-mail: firstname.lastname@example.org.
Funded by the Department of Obstetrics and Gynaecology, University of Malaya.
Financial Disclosure The authors did not report any potential conflicts of interest.
In contemporary practice within a well-resourced setting, approximately 20% of viable pregnancies have their labor induced.1 Compared with spontaneous labor, labor induction is associated with a higher rate of cesarean delivery of 11.4% compared with 22.0%1 attributable predominantly to an unfavorable Bishop’s score at admission.2
Labor occurs within 24 hours of amniotomy for labor induction in 90.1% of women.3 However, a Cochrane review concludes that data are lacking about the value of amniotomy alone for induction of labor, but there remain clinical scenarios in which amniotomy alone may be desirable and appropriate and this method is worthy of further research.4 Another Cochrane review states that amniotomy and oxytocin infusion can induce labor, but the optimal timing of oxytocin infusion after amniotomy is not known.5 Antepartum use of oxytocin in nulliparas is associated with severe maternal and neonatal morbidity.6 Oxytocin use increases the risk of uterine hyperactivity7 and neonatal jaundice8; trading slight prolongation of induction to delivery interval for fewer complications by reducing oxytocin use can be a prudent strategy.9
Increased parity and a favorable cervix as assessed by a higher Bishop’s score are associated with successful labor induction.10,11 We hypothesize that in parous women with favorable cervixes, after amniotomy for labor induction, a 4-hour period to await labor may reduce antepartum oxytocin use and oxytocin-related adverse events without affecting the proportion of women delivered within 12 hours of amniotomy or their satisfaction with the birth process. We performed a double-blind randomized trial to test our hypothesis.
PATIENTS AND METHODS
The trial was performed in a full-service university hospital with an annual delivery rate of approximately 6,000 women in Kuala Lumpur, Malaysia. Ethical oversight was provided by the University of Malaya Medical Centre medical ethics committee (Approval Reference No. 733.18, dated July 22, 2009). The trial was conducted in accordance with the Declaration of Helsinki (2000) for human studies. Written informed consent was obtained from each participant. The trial was registered with the ISRCTN trial register with the identifier ISRCTN 51476259 before its commencement.
Previous trials have shown that in subanalyses confined to parous women, immediate oxytocin after amniotomy was associated with a 93%12 and a 1-hour delay associated with a 95%13 delivery rate within 12 hours. Taking a 12-hour vaginal delivery rate of 95% for immediate oxytocin and assuming a noninferiority margin of 10% for the 4-hour delay arm, one-to-one ratio, α of 0.05 and power of 90%, and factoring in a dropout of 20%, a total of 205 women were needed.
Primary outcomes were vaginal delivery rate within 12 hours of induction and maternal satisfaction score for the birth process obtained within 24 hours of delivery. Participants were asked within 12 hours of delivery and after their transfer to the postnatal ward to score on a visual numerical rating scale of 1 (denoted as most satisfied) to 10 (denoted as most dissatisfied), their rating of their birth process. Secondary outcomes were amniotomy-to-delivery interval, mode of delivery, opiate or epidural analgesia use in labor, prenatal oxytocin use, uterine hyperactivity, intrapartum and postpartum fever (temperature 38°C or greater), delivery blood loss, maternal antibiotic use, amniotomy-to-hospital discharge interval, and various neonatal outcomes (special care nursery admission, umbilical cord blood pH and base excess, Apgar score, phototherapy for jaundice, and intensive care admission).
The randomization sequence was generated using a computerized random number generator in blocks of 40 by an investigator (P.C.T.). Allocation to treatment arms of immediate or delayed oxytocin was affected by the sequential opening of sealed numbered envelopes.
Potential recruits for the trial were identified by health care providers in our delivery suite when they were admitted for labor induction. Inclusion criteria were parous women (at least one vaginal birth at greater than 24 weeks of gestation), age 16 years or older, singleton pregnancy in cephalic presentation, gestation 37 weeks or greater, Bishop’s score 6 or greater with cervical dilation 2 or greater (suitable for amniotomy), intact membranes, and a reassuring cardiotocogram. Gestational age estimations were all supported by ultrasonographic dating. Exclusion criteria were previous uterine incision or injury (cesarean delivery, myomectomy, or perforation), gross fetal anomaly, and contraindication for vaginal birth. Potential recruits were given the patient information sheet to read and, if agreeable, consented and randomized after their amniotomy. In our center, the standard labor induction technique for women with favorable cervixes was amniotomy followed by immediate oxytocin infusion or a period of up to 4 hours to await labor onset before starting oxytocin infusion as required depending on the health care provider.
The study drug was prepared by a research assistant who was not involved in the subsequent care of the participant. For women assigned to immediate oxytocin infusion, 10 units of oxytocin were added to a standard container of 500 mL Hartmann's solution (oxytocin concentration of 20 milliunits/mL) and the container labeled as “study drug.” For women assigned to delayed oxytocin, an identical container of 500 mL placebo Hartmann's solution was labeled as “study drug.”
For both arms, study drug infusion was started as soon as possible after amniotomy and randomization, started at a rate of 3 mL/h (ie, 1 milliunit/min for the immediate oxytocin arm). The infusion rate was doubled every 30 minutes to a maximum of 48 mL/h (ie, 16 milliunits/min for the oxytocin arm) or until three to four moderate contractions per 10 minutes were achieved at which point the infusion rate was maintained. Continuous electronic fetal heart rate monitoring was maintained throughout the induction and labor.
After 4 hours, infusion of the study drug was stopped for both arms. In the interim, vaginal assessment was performed if clinically indicated (typically for symptom or sign of second stage or in response to nonreassuring cardiotocogram). If the woman was still undelivered at 4 hours, vaginal examination was performed to assess cervical dilation and to determine whether labor was in the active phase (defined as cervical dilation 4 cm or greater and contraction frequency three or more per 10 minutes). In the event of a nonreassuring cardiotocogram associated with uterine hyperactivity, the following actions were suggested: reduce or stop study drug infusion, administer subcutaneous 250 micrograms terbutaline as tocolysis, or expedite operative delivery depending on individual circumstances and the severity of the cardiotocogram abnormality.
At the 4-hour postamniotomy assessment, if labor was not in the active phase, open-label oxytocin infusion was started at 2 milliunits/min and infusion titrated according to our local protocol (infusion rate doubled every 30 minutes as needed to achieve contraction frequency of three to four per 10 minutes or to a maximum infusion rate of 32 milliunits/min). If labor was already in the active phase, the decision was left to the health care provider whether to continue with open-label oxytocin infusion starting at 2 milliunits/min and the rate infusion titrated further according to contraction frequency and labor progress. It has been shown that stopping oxytocin infusion after labor is already in the active phase is not associated with a significantly longer duration of the active phase and the second stage of labor.14
Standard management of labor induction and labor was applied for all participants as previously described for our center15; vaginal assessment was performed at least 4 hourly and the decision for cesarean and instrumental vaginal delivery was made based on the usual obstetric indications.
Participants' characteristics and outcomes were extracted onto a standardized case report form. Case notes and hospital records were scrutinized after delivery to retrieve relevant clinical outcome data. Participants were followed up until hospital discharge.
The cardiotocogram was assessed after delivery by an investigator (M.Z.S.) based on the following criteria for classification as fetal heart rate abnormality: fetal heart rate deceleration (greater than 15 beats per minute [bpm] below baseline for greater than 15 seconds), sustained tachycardia (greater than 160 bpm for more than 15 minutes), baseline bradycardia (less than 100 bpm for more than 15 minutes), or reduced baseline variability (less than 3 bpm for more than 15 minutes) and the following criteria for uterine hyperactivity: tachysystole (six of more contractions in 10 minutes over two consecutive 10-minute periods) or hypertonus (sustained contraction 2 minutes or longer).
Data were entered into SPSS 17. Analysis was based on intention to treat. The Kolmogorov-Smirnov test was used to check for normality of data distribution. Normally distributed continuous data were analyzed with the Student's t test. Two-by-two categorical data sets were analyzed with the Fisher’s exact test and larger categorical data sets with the χ2; ordinal data and nonnormally distributed continuous data were analyzed with the Mann-Whitney U test. Time to vaginal delivery analysis was performed with the Mantel-Cox log rank test16,17 after censoring for cesarean delivery. All tests were two-sided and P<.05 was considered significant.
The trial enrollment period was from February 5, 2010, to May 25, 2012. The latest hospital discharge of a participant or neonate was on May 30, 2012. Trial recruitment was stopped after targeted sample size was achieved. The trial recruitment flow is as shown in Figure 1. Study drug infusion was started for all participants as allocated. There were no withdrawals or dropouts.
Table 1 shows the characteristics of the participants stratified according to their randomization to either immediate or delayed oxytocin arms. There was no significant difference in the participants' basic characteristics, indication for labor induction, presence of meconium at amniotomy, and in their amniotomy-to-commencement-of-study-drug infusion time interval across the trial arms.
Table 2 shows the result for the two primary outcomes. The vaginal delivery rate within 12 hours of amniotomy (after excluding a total of 16 cesarean deliveries) was 91 of 96 (94.8%) compared with 91 of 94 (96.8%) (relative risk 0.98, 95% confidence interval [CI] 0.92–1.04, P=.72) and the medians (interquartile ranges) scores for satisfaction with the birth process were 3 (3–4) compared with 3 (3–5) (P=.36) for the immediate compared with 4-hour delay arms, respectively. Delayed oxytocin infusion was noninferior to immediate oxytocin on the primary outcomes.
Figure 2 depicts the curves for vaginal delivery rate over time for both trial arms (Mantel-Cox test P=.01, cesarean delivery censored) with a shorter amniotomy to vaginal delivery interval for the immediate oxytocin arm (mean±standard deviation amniotomy to delivery 5.3±3.1 compared with 6.9±2.9 hours, P<.001), but by 10 hours after amniotomy, vaginal delivery rates had converged at approximately the 90% mark.
Table 3 shows the result for secondary outcomes. Immediate oxytocin infusion compared with delayed oxytocin was associated with a shorter amniotomy to vaginal delivery interval, delivery before planned review at 4 hours and delivery by 6 hours, labor proceeding to the active phase by 4 hours, and requirement for epidural analgesia (number to treat needed to benefit of 17, 95% CI 8–285). Delayed oxytocin was associated with a lower incidence of fetal heart rate anomalies on cardiotocogram (number to treat needed to benefit of nine, 95% CI 4–232), particularly in the first 4 hours after amniotomy, but no decrease in the incidence of uterine hyperactivity. Cesarean delivery rate, opiate analgesia use, maternal fever and antibiotic use, postpartum blood loss, manual removal of placenta rate, amniotomy to hospital discharge interval, and neonatal outcomes (Apgar score at 5 minutes, umbilical cord arterial blood parameters, antibiotic use, phototherapy for jaundice, intensive care unit admission) were similar across the trial arms. Of the women allocated to delayed oxytocin, 35.6% avoided oxytocin infusion altogether because amniotomy alone was sufficient to induce labor and achieve delivery. Terbutaline was not used because there was no case of uterine hyperstimulation syndrome (concomitant uterine hyperactivity and fetal heart rate abnormality). There were no important unintended harmful events in either arm.
Only a small number of labor induction trials addressing the timing of starting oxytocin infusion after amniotomy have been performed.12,13,18 In our trial of parous women with favorable cervixes, a 4-hour delay compared with immediate oxytocin infusion was associated with a similar vaginal delivery rate at 12 hours after amniotomy and maternal satisfaction with the birth process. In contrast, in an early study of a mixed population of nulliparous and parous women, immediate compared with delayed (24 hours) oxytocin infusion after amniotomy for labor induction results in a vaginal delivery rate at 12 hours of 85.5% compared with 35.5% in favor of the immediate oxytocin arm12; however, the bulk of the deficit in the 12-hour delivery rate was contributed to by nulliparous women.
Although the amniotomy-to-delivery latency was shorter in the immediate oxytocin arm and the epidural rate was also lower in our current trial, maternal satisfaction score with the birth process was not affected. Maternal satisfaction was unaffected probably because the mean increase in the amniotomy-to-delivery interval was only 1.6 hours (mean 5.6 compared with 6.9 hours; Table 3). This time gap probably comprised a fairly comfortable latency with limited uterine activity because by 4 hours, the deficit in the active phase of labor rate was 36.7% (71.4% compared with 34.7%; Table 3) in favor of the immediate oxytocin arm. We did not record the amniotomy-to-labor interval in our trial. An earlier study reported a latency to labor difference of 2 hours (mean 2.25 compared with 4.25 hours) between the early (1-hour delay) oxytocin compared with delayed (up to 24 hours) infusion arms of a mixed study population of nulliparous and parous women after amniotomy for labor induction in favor of early oxytocin infusion.13 An earlier labor induction trial at our center of concurrent dinoprostone and oxytocin infusion compared with dinoprostone in nulliparas with unfavorable cervixes showed greater maternal satisfaction with the birth process despite a higher pain score, increased analgesia use rate, and only a trend toward higher vaginal delivery rate at 24 hours for the concurrent therapy arm19 suggesting that maternal satisfaction in labor induction is a complex outcome as was also shown in our current trial.
Early oxytocin infusion after amniotomy for labor induction in nulliparous women resulted in greater maternal satisfaction, a shorter amniotomy-to-delivery interval, and a better 12-hour vaginal delivery rate (of 77.1% compared with 58.1%).18 In our population of parous women who were more rapidly responsive to amniotomy, a shorter amniotomy-to-delivery interval but with convergent 90% vaginal delivery rates by 10 hours after amniotomy, immediate oxytocin infusion did not affect maternal satisfaction score. Selo-Ojeme et al18 report a borderline result for increased cardiotocogram abnormality for their immediate oxytocin arm (39.3% compared with 29%, relative risk 1.5, 95% CI 0.8–1.7, P=.07) and no increase in uterine hyperactivity. We found a significant increase in minor cardiotocogram abnormalities (predominantly early fetal heart rate decelerations), which did not translate into increased operative delivery rate or adverse neonatal outcome and similarly no increase in uterine hyperactivity. Selo-Ojeme et al did not find a difference in epidural analgesia rates, although their rates were high (52.4% compared with 54.8%) as would be expected for their exclusively nulliparous study population compared with rates of 2.9% compared with 9.9% in our population of parous women (which was significantly different in favor of immediate oxytocin).
We did not find a significant association between immediate and delayed oxytocin on phototherapy for neonatal jaundice: 9.6% compared with 5.0% (P=.28, relative risk 1.9, 95% CI 0.7–5.4). Previous studies have shown that the relationship between intrapartum oxytocin exposure and subsequent neonatal jaundice is tempered by consideration of gestational age,20 use of dextrose solution as a diluent, and neonatal hyponatremia.21 In our trial, gestational age between the arms was similar and we only used Hartmann's solution as a diluent.
Our trial has strengths and limitations. Amniotomy alone and a delay in oxytocin infusion would have the best chance of producing a good outcome in our trial population of parous women with favorable cervixes in contrast to earlier studies of mixed parity women or nulliparas only, which have all demonstrated shorter latency to delivery with early oxytocin.12,13,18 Our findings should be generalizable to any similar population managed in a contemporary delivery suite. Our double-blind design minimized possibility of bias, although it was possible the intervention might be discernible because oxytocin produces contractions sooner than placebo. Our sample size had 90% power of demonstrating noninferiority (margin, difference of 10%) in the 12-hour postamniotomy delivery rate. However, our sample size was not large enough to detect a difference in an important outcome like cesarean delivery. To have 80% power to detect a difference in cesarean delivery rate of 8.6% compared with 6.9%, a future trial requires a sample size of 7,766.
Immediate compared with delayed oxytocin of 4 hours was associated with a similar vaginal delivery rate at 12 hours and similar maternal satisfaction. The induction-to-delivery interval was shorter, and there was less need for epidural analgesia but more cardiotocogram abnormalities with immediate oxytocin. Operative delivery and neonatal adverse outcome rates were not different. More than one-third of women allocated to delayed oxytocin avoided antepartum oxytocin infusion. These mixed findings suggested that the timing of oxytocin infusion after amniotomy for labor induction in a parous woman with a favorable cervix should take into account local resources and the woman's choice.
2. Vrouenraets FP, Roumen FJ, Dehing CJ, van den Akker ES, Aarts MJ, Scheve EJ. Bishop score and risk of cesarean delivery after induction of labor in nulliparous women. Obstet Gynecol 2005;105:690–7.
3. Cooley SM, Geary MP, O'Connell MP, McQuillan K, McParland P, Keane D. How effective is amniotomy as a means of induction of labour? Ir J Med Sci 2010;179:381–3.
4. Bricker L, Luckas M. Amniotomy alone for induction of labour. The Cochrane Database of Systematic Reviews 2000, Issue 4. Art. No.: CD002862. DOI: 10.1002/14651858.CD002862.
5. Howarth GR, Botha DJ. Amniotomy plus intravenous oxytocin for induction of labour. The Cochrane Database of Systematic Reviews 2001, Issue 3. Art. No.: CD003250. DOI: 10.1002/14651858.CD003250.
6. Buchanan SL, Patterson JA, Roberts CL, Morris JM, Ford JB. Trends and morbidity associated with oxytocin use in labour in nulliparas at term. Aust N Z J Obstet Gynaecol 2012;52:173–8.
7. Verspyck E, Sentilhes L. Abnormal fetal heart rate patterns associated with different labour managements and intrauterine resuscitation techniques. J Gynecol Obstet Biol Reprod (Paris) 2008;37(suppl 1):S56–64.
8. Chalmers I, Campbell H, Turnbull AC. Use of oxytocin and incidence of neonatal jaundice. Br Med J 1975;2:116–8.
9. Shyken JM, Petrie RH. The use of oxytocin. Clin Perinatol 1995;22:907–31.
10. Caughey AB, Sundaram V, Kaimal AJ, Cheng YW, Gienger A, Little SE, et al.. Maternal and neonatal outcomes of elective induction of labor. Evid Rep Technol Assess (Full Rep) 2009:1–257.
11. Tan PC, Vallikkannu N, Suguna S, Quek KF, Hassan J. Transvaginal sonographic measurement of cervical length vs Bishop score in labor induction at term: tolerability and prediction of Cesarean delivery. Ultrasound Obstet Gynecol 2007;29:568–73.
12. Patterson WM. Amniotomy, with or without simultaneous oxytocin infusion. A prospective survey. J Obstet Gynaecol Br Commonw 1971;78:310–6.
13. Moldin PG, Sundell G. Induction of labour: a randomised clinical trial of amniotomy versus amniotomy with oxytocin infusion. Br J Obstet Gynaecol 1996;103:306–12.
14. Ustunyurt E, Ugur M, Ustunyurt BO, Iskender TC, Ozkan O, Mollamahmutoglu L. Prospective randomized study of oxytocin discontinuation after the active stage of labor is established. J Obstet Gynaecol Res 2007;33:799–803.
15. Tan PC, Jacob R, Omar SZ. Membrane sweeping at initiation of formal labor induction: a randomized controlled trial. Obstet Gynecol 2006;107:569–77.
16. Mantel N, Haenszel W. Statistical aspects of the analysis of data from retrospective studies of disease. J Natl Cancer Inst 1959;22:719–48.
17. Cox DR. Regression models and life tables. J R Stat Soc 1972;B34:187–220.
18. Selo-Ojeme DO, Pisal P, Lawal O, Rogers C, Shah A, Sinha S. A randomised controlled trial of amniotomy and immediate oxytocin infusion versus amniotomy and delayed oxytocin infusion for induction of labour at term. Arch Gynecol Obstet 2009;279:813–20.
19. Tan PC, Valiapan SD, Tay PY, Omar SZ. Concurrent oxytocin with dinoprostone pessary versus dinoprostone pessary in labour induction of nulliparas with an unfavourable cervix: a randomised placebo-controlled trial. BJOG 2007;114:824–32.
20. Lange AP, Secher NJ, Westergaard JG, Skovgård I. Neonatal jaundice after labour induced or stimulated by prostaglandin E2 or oxytocin. Lancet 1982;1:991–4.
21. Singhi S, Chookang E, Hall JS. Intrapartum infusion of aqueous glucose solution, transplacental hyponatraemia and risk of neonatal jaundice. Br J Obstet Gynaecol 1984;91:1014–8.