Critical appraisal of the literature

1. 1. In pregnant term patients (population), does elective induction (intervention) lead to improved fetal or maternal outcomes?

Labor may be induced for either maternal or fetal indications. Induction of labor is undertaken when both of the following criteria are met [3]:

• Continuing the pregnancy is believed to be associated with greater maternal or fetal risk than intervention to deliver the pregnancy, and
  
• There is no contraindication to vaginal birth, including prior classical uterine incision, prior transmural uterine incision entering the uterine cavity, active genital herpes infection, placenta or vasa previa, umbilical cord prolapse, transverse fetal lie.

The magnitude of risk is influenced by factors such as gestational age, fetal lung maturity status, severity of the clinical condition, and cervical status. Appropriately timed induction of women with pregnancy complications can improve maternal‐fetal outcomes (Table 50.1) [4]. This appears particularly true in women with some of the hypertensive complications of pregnancy. In the monitoring for hypertensive disorders of pregnancy between 34 and 37 weeks of gestation (HYPITAT) trial, a small (n = 756) but nonetheless well‐designed study, singleton pregnancies at 36–41 weeks gestation complicated by gestational hypertension or mild pre‐eclampsia were randomized to induction (n = 377) vs. expectant management (n = 379). Of women randomized, 177 (31%) allocated to induction of labor developed poor maternal outcome (maternal mortality or morbidity including eclampsia, hemolysis, elevated liver enzymes and low platelet (HELLP) syndrome, pulmonary edema, thromboembolic disease, or placental abruption), whereas 166 (44%) of women who underwent expectant management suffered from poor outcomes (RR 0.71, [95% CI 0.59–0.86], p = 0.0001). There were no cases of maternal or neonatal death or eclampsia [5]. Based on these findings, induction of labor for gestational hypertension or pre‐eclampsia without severe features is advised at 37 weeks’ gestation (Level of evidence A).

Overall, there is only limited high quality evidence establishing any benefits for specific medical and obstetrical indications for induction [6]. The risks of iatrogenic late preterm birth appear to outweigh any theoretical benefits when the indication for delivery is “soft,” such as suspected macrosomia without maternal diabetes, uncomplicated chronic hypertension, or history of fetal, maternal, or obstetric complication in a previous pregnancy [7, 8].

In comparison to indicated induction where the maternal or fetal health is considered to be in jeopardy, elective induction refers to the iatrogenic stimulation of labor in the absence of maternal or obstetrical indications. The major concerns associated with elective induction of labor at term are the potential for increased rates of cesarean delivery, iatrogenic prematurity, and cost. Another concern is that maternal‐fetal medical benefits, such as reduction in stillbirth, have not been proven. Nevertheless, there are potential advantages to scheduled induction of labor, such as avoiding the risk of delivery en route to the hospital if labor is rapid or the patient lives far away, and avoiding sudden disruption of the patient’s (and provider’s) work and non‐work related responsibilities. There are insufficient data to support a policy of routine elective induction of labor at term. Large, randomized trials with emphasis on maternal and neonatal safety, determination of neonatal benefit as a reflection of reduced unexplained fetal death, and cost‐benefit analyses are needed.

In a 2014 study, Baitlit et al. compared maternal and neonatal outcomes in nulliparous women with non‐medically indicated induction at term vs. expectant management. They concluded that at 39 weeks of gestation, nonmedically indicated induction is associated with lower maternal and neonatal morbidity than expectant management. When induced women were compared with expectantly managed women at the same gestational age, they did not find a substantial increase in the cesarean delivery rate in the induced group [9] (Level of evidence B).

Preventive or risk‐based induction has also been termed Active Management of Risk in Pregnancy at Term (AMOR‐IPAT), or non‐indicated but risk‐based induction. Nicholson et al. performed a meta‐analysis of the associations between the regular use of modeled risk based non‐indicated term labor induction and rate of adverse outcomes. The use of preventive induction, as compared with the standard approach, was associated with a more favorable pattern of birth outcomes. Certain types of non‐indicated induction may be beneficial for maternal‐fetal outcome; however, evidence of the benefits of preventive induction for specific indications is limited [10]. (Level of evidence B).

Currently, there is expert consensus that elective induction should not be performed before 39 weeks gestation; however there is insufficient evidence to recommend for or against induction of labor at ≥39 weeks of gestation [11]. Adequately powered randomized clinical trials are needed to study the risks of non‐indicated term labor induction (Level of evidence C).

The risk of cesarean delivery with elective induction has remained controversial, secondary to differing control groups (either patients undergoing spontaneous labor or expectant management) in the currently published literature. When comparing elective induction outcomes to spontaneous labor, an increased risk of cesarean delivery is noted, as in one of the largest observational studies of elective induction in low risk women, which included 1847 women undergoing elective induction and 35 597 spontaneously laboring women [12]. In this study, the World Health Organization’s Global Survey on Maternal and Perinatal Health in Latin America Study Group observed cesarean rates of 11.7% with elective induction and 8.6% after spontaneous labor (crude RR 1.36, 95% CI 1.19–1.55). The increased risk of operative delivery appears particularly noteworthy in nulliparous women. This is best illustrated by a matched cohort study that compared the fetomaternal outcomes of 7683 women who underwent electively induced labor to 7683 women who experienced spontaneous labor [13]. All of the women were nulliparous with singleton pregnancies in cephalic presentation, gestational age 266–287 days, and birth weight 3000–4000 g. Information on cervical status and use of cervical ripening agents was not available. Elective induction led to statistically significant higher rates of cesarean delivery (10% vs. 7%), instrumental delivery (32% vs. 29%), and use of epidural anesthesia (80% vs. 58%). The higher cesarean rate was attributed to an increased frequency of intervention for failure to progress in the first stage of labor. Similar findings have been reported in multiple cohort studies of nulliparous women with vertex, singleton term pregnancies delivering in the United States [14–20]. These studies consistently showed that the rate of cesarean delivery was increased approximately twofold in women who underwent elective or medical induction of labor compared to those who experienced spontaneous labor.

However, if patients undergoing elective induction are compared to those receiving expectant management rather than those undergoing spontaneous labor, the risk of cesarean delivery appears to be decreased. Women in spontaneous labor, such as in those studies noted above, may not be an appropriate control group for studies evaluating elective induction outcomes. Using them as controls biases differences in cesarean rates to favor the control group because many patients managed expectantly would have gone on to have an indicated cesarean delivery rather than spontaneous labor. Also, in practical terms, a physician cannot decide between spontaneous labor and induction on a given day, but between expectant management and induction. When cesarean delivery rates have been compared for induced labors vs. expectant management at the same gestational age, the differences in cesarean delivery rates were small and favored the induction groups [21, 22]. In a 2012 Cochrane Review of randomized trials in which a policy of induction “at or beyond term” was compared with expectant management, cesarean delivery rate was 11% lower in the induction group (relative risk [RR] 0.89, 95% CI 0.81–0.97). Of note 17 of the 21 trials included in this analysis involved women >41 weeks of gestation [23]. In another systematic review of randomized trials in which induction at “term” was compared with expectant management, the cesarean delivery rate was 13% lower in the induction group (RR 0.87, 95% CI 0.82–0.92). The findings in the latter study were for pregnancies at 37 to <42 weeks gestation [24] (Level of evidence A).

Another issue when evaluating the risk of cesarean delivery is parity. In comparison to nulliparas, most studies in multiparous women have not shown an increased risk of cesarean delivery with induction of labor [25]. Almost all of these reports were retrospective, but one small randomized trial confirmed these findings [26]. One of the largest series was a population‐based cohort study that compared the risk of cesarean delivery in 1775 healthy, low risk multiparous women at term who underwent induction without an identifiable indication to 5785 similar women who entered labor spontaneously [27]. Cervical ripening agents were used in women with unfavorable cervices. The overall cesarean delivery rate was similar for induced and spontaneous labors, 3.8% and 3.6%, respectively (RR 1.07, 95% CI 0.91–1.39). Although the cesarean delivery rate was higher in women with a previous cesarean delivery, the rate did not differ significantly for induced and spontaneous labors (30.5% and 30.7%, respectively).

Even when inductions for medical indications are included, multiparas have a relatively low rate of cesarean delivery. In one retrospective cohort study, the rates of cesarean delivery in multiparas in spontaneous labor (n = 7208), induced with oxytocin (n = 2190), and induced with cervical ripening agents (n = 239) were 4.2%, 6.3%, and 14.2%, respectively [28]. Oxytocin‐induced multiparas were 37% more likely to require cesarean than those with spontaneous labor (OR, 1.37; 95% CI, 1.10–1.71) and nearly three times more likely to undergo cesarean when cervical ripening agents were used (OR, 2.82; 95% CI, 1.84–4.53) (Level of evidence B).

If successful elective induction is defined as achieving a vaginal birth while avoiding excessive costs and admission to a special care nursery, then the best candidates are women (nulliparous or multiparous) with well‐dated pregnancies of at least 39 weeks of gestation and favorable cervices. The excess neonatal morbidity of earlier intervention was illustrated in a prospective observational study that compared the outcome of 790 planned elective inductions at 37–38 weeks of gestation with the outcome of 2004 planned elective inductions at ≥39 weeks of gestation [29]. Earlier induction was associated with a significantly higher risk of neonatal intensive care unit (NICU) admission (7.7% vs. 3.0%) (Level of evidence B).

The major pediatric concerns with regards to elective delivery include neonatal respiratory problems. Respiratory problems can result from inadvertent delivery of a premature infant or transient tachypnea related to cesarean delivery after failed induction. However, several, primarily retrospective, studies have not shown a marked impairment in neonatal outcome when elective induction of labor was undertaken at term in well‐dated pregnancies [30–34]. There may, in fact, be a slight benefit as fewer electively induced infants have meconium passage when compared to spontaneously labored infants. Macrosomia also may be reduced [35] (Level of evidence B).

The risk of respiratory morbidity was illustrated in a retrospective review of infants with respiratory distress or transient tachypnea of the newborn admitted to the NICU following elective delivery at term [36]. The data were stratified by gestational age and route of delivery with a baseline incidence of respiratory distress syndrome of 2.2/1000 deliveries (95% CI 1.7–2.7/1000) and transient tachypnea of 5.7/1000 (95% CI 4.9–6.5/1000) at term. The frequencies of respiratory morbidity following vaginal or cesarean delivery increased with decreasing gestational age, with the highest risk associated with cesarean after labor at 37–38 weeks gestation (57.7/1000 deliveries, 95% CI 26.7–107.1/1000). Delivery by cesarean without preceding labor increased the frequencies of respiratory morbidities even higher across all gestational ages. These data provide support for delaying elective delivery until 39 weeks of gestation (Level of evidence B).

A study using decision analysis analyzed the economic consequences of elective induction of labor at term in a cohort of 100 000 women for whom an initial decision was made to either induce labor at 39 weeks of gestation or follow expectantly through the remainder of pregnancy [37]. All patients in this model underwent elective induction at 42 weeks. Using baseline estimates, the investigators concluded that elective induction would result in more than 12 000 excess cesareans, imposing an annual cost to the medical system of nearly $100 million. A policy of induction at any gestational age, regardless of parity or cervical ripeness, required economic expenditures by the medical system. Although never cost saving, inductions were less expensive at later gestational ages, for multiparous patients, and for those women with a favorable cervix. The inductions most costly to the healthcare system were those performed in nulliparas with unfavorable cervices at 39 weeks. When nulliparous women with favorable cervices undergo labor induction, the estimated cost is approximately halved compared to nulliparas with unfavorable cervices; however, it still resulted in overall added expenditures and additional cesarean deliveries (Level of evidence B).

1. 2. In pregnant patients undergoing induction of labor (population), does transvaginal ultrasound or biochemical examinations (tests) predict labor induction success (outcome) better than cervical examination (comparison)?

Cervical status is one of the most important factors for predicting the likelihood of successfully inducing labor. For this reason, a cervical examination should be performed before initiating attempts at induction. There are several cervical scoring systems available for this purpose [38], although the modified Bishop score is the system most commonly used in clinical practice in the United States [39]. This system tabulates a score based upon the station of the presenting part and four characteristics of the cervix: dilatation, effacement, consistency, and position (Table 50.2). If the Bishop score is high (variously defined as ≥5 or ≥8), the likelihood of vaginal delivery is similar whether labor is spontaneous or induced [40]. In contrast, a low Bishop score is predictive that induction will fail and result in cesarean delivery. These relationships are particularly strong in nulliparous women who undergo induction [11, 41], although Bishop scoring was originally described in multiparous women. Of note, the relationship between a low Bishop score and failed induction, prolonged labor, and a high cesarean birth rate was first described prior to widespread use of cervical ripening agents [42].

In observational studies, other characteristics associated with successful induction include multiparity, tall stature (over 5 ft 5 in.), increasing gestational age, non‐obese maternal weight or body mass index, and infant birth weight less than 3.5 kg [43, 44]. However, these characteristics are predictive of success even in spontaneous labors, which suggests they are more predictive of the route of delivery than the likelihood that the patient will reach the active phase of labor.

Because of the risk of cesarean delivery and the rising costs of health care associated with labor induction, some researchers have tried to identify, with varying success, biochemical and biophysical assays to predict the probability of vaginal delivery following labor induction [45–48]. These measures include digital evaluation of the cervix, ultrasonographic cervical length measurements, and use of fetal fibronectin (fFN) before labor induction.

Cervical length is predictive of the likelihood of spontaneous onset of labor post‐term [49]. Sonographic assessment of cervical length for predicting the outcome of labor induction has been evaluated in numerous studies. A systematic review of 20 prospective studies found that cervical length was predictive of successful induction (likelihood ratio of a positive test, 1.66; 95% CI 1.20–2.31) and failed induction (likelihood ratio of a negative test, 0.51; 95% CI, 0.39–0.67) [50]. However, sonographic cervical length performed poorly for predicting vaginal delivery within 24 hours (sensitivity 59%, specificity 65%), vaginal delivery (sensitivity 67%, specificity 58%), achieving active labor (sensitivity 57%, specificity 60%), and delivery within 24 hours (sensitivity 56%, specificity 47%), and did not perform significantly better than the Bishop score for predicting a successful induction. These data are limited by substantial heterogeneity among the studies. The role of ultrasound examination as a tool for selecting women likely to have a successful induction is uncertain at this time.

In a study from Verhoeven et al., they performed a systematic review and meta‐analysis to assess the predictive capacity of transvaginal sonographic assessment of the cervix for outcome of induction. This study included 31 studies reporting on both cervical length and outcome of delivery. Sensitivity of cervical length in the prediction of cesarean delivery ranged from 0.14 to 0.92 and specificity ranged from 0.35 to 1.00. For cervical wedging in the prediction of failed induction of labor summary point estimates of sensitivity/specificity were 0.37/0.80. They concluded that cervical length measured sonographically at or near term have moderate capacity to predict the outcome of delivery after induction [51] (Level of evidence B). Overall, the role of ultrasound examination as a tool for selecting women likely to have a successful induction is uncertain at this time.

The presence of an elevated fFN concentration in cervicovaginal secretions has also been used to predict uterine readiness for induction. fFN is thought to represent a disruption or inflammation of the chorionic‐decidual interface. In several studies, women with a positive fFN result had a significantly shorter interval until delivery than those with a negative fFN result [52] and there was reduction in the frequency of cesarean delivery [53]. Positive fFN results were predictive of a shorter interval to delivery, even in nulliparas with low (<5) Bishop scores [54]. However, there are other investigations which did not confirm these findings [48, 55].

The Bishop score appears to be the best available tool for predicting the likelihood that induction will result in vaginal delivery. This conclusion is based on systematic reviews of controlled studies that found the Bishop score was as, or more, predictive of the outcome of labor induction than fFN [45] or sonographic measurement of cervical length [45, 52, 56], and that dilatation was the most important element of the Bishop score [45] (Level of evidence B).

1. 3. In a pregnant patient with an unfavorable cervix (population), how do pharmacologic ripening methods (intervention) compare to mechanical methods (comparison) in terms of achieving vaginal deliveries (outcome)?

Cervical ripening is a complex process that results in physical softening and distensibility of the cervix, ultimately leading to partial cervical effacement and dilatation [57]. The methodology falls into two main categories: (i) mechanical (physical), such as disruption of the fetal membranes or insertion of dilators or a balloon catheter, and (ii) application of cervical ripening agents, such as prostaglandin compounds or oxytocin (Table 50.3).

Mechanical methods are among the oldest approaches used to promote cervical ripening. Advantages of these techniques compared to pharmacologic methods include their low cost, low risk of tachysystole, few systemic side effects, and convenient storage requirements (no refrigeration or expiration) [58]. Comparing mechanical methods with placebo or no treatment [58], tachysystole with fetal heart rate changes was not reported. The risk of cesarean birth was similar between groups (34%; RR 1.00; 95% CI: 0.76–1.30, n = 416, 6 studies). There were no reported cases of severe neonatal and maternal morbidity among them. The risk of tachysystole was reduced when compared with all prostaglandins. Compared with oxytocin in women with unfavorable cervix, mechanical methods reduce the risk of cesarean delivery. Disadvantages of mechanical methods include a small increase in the risk of maternal and neonatal infection from introduction of a foreign body [59], the potential for disruption of a low‐lying placenta, and some maternal discomfort upon manipulation of the cervix. The most common mechanical methods are stripping (or sweeping) of the fetal membranes, placement of hygroscopic dilators within the endocervical canal, and insertion of a balloon catheter above the internal cervical os (with or without infusion of extra‐amniotic saline).

The efficacy of membrane sweeping was demonstrated in a meta‐analysis of 22 trials in which 20 compared sweeping of membranes to no treatment, three compared sweeping to prostaglandin administration, and one compared sweeping to oxytocin administration before formal induction of labor [60]. Compared to no intervention, membrane sweeping was associated with reduced frequency of pregnancy continuing beyond 41 weeks (RR 0.59, 95% CI 0.46–0.74) and 42 weeks of gestation (RR 0.28, 95% CI 0.15–0.50), and reduced frequency of formal induction (RR 0.72, 95% CI 0.52–1.00). The cesarean delivery rate was not altered; the change in Bishop score was not assessed. Overall, eight women would need to undergo membrane sweeping to avoid one formal induction of labor. There was no increased risk of maternal or neonatal infection, but minor maternal discomforts were common. Compared to no intervention, weekly membrane stripping at term shortens the interval of time to onset of spontaneous labor and reduces the need for formal induction. Current recommendations are to begin stripping membranes at more than 37 weeks gestation in patients who wish to hasten the onset of labor [6] (Level of evidence A).

Amniotomy appears to be an effective method of labor induction, but can only be performed in women with partially dilated and effaced cervices. A Cochrane review of randomized trials found the combination of amniotomy plus intravenous oxytocin administration was more effective than amniotomy alone for induction of labor [61] (Level of evidence A). With the combined regimen, fewer women were undelivered at 24 hours than with amniotomy alone (RR 0.13, 95% CI 0.04–0.41). To achieve the greatest impact on duration of labor, amniotomy should be performed as early as possible and oxytocin should be initiated immediately thereafter [62]. There are inadequate data for assessing the efficacy of the combination of amniotomy plus intravenous oxytocin administration compared to intravenous oxytocin alone [63]. There are limited data suggesting the efficacy of amniotomy plus oxytocin is similar to that of prostaglandins alone [61].

Hygroscopic dilators are safe and effective for dilating the cervix, although they are used primarily during pregnancy termination rather than for pre‐induction cervical ripening of term pregnancies. A meta‐analysis of randomized trials comparing hygroscopic dilators to placebo or no treatment found that pregnant women in both groups had similar rates of not achieving a vaginal delivery by 24 hours (RR 0.90; 95% CI 0.64–1.26), cesarean deliveries (RR 0.98; 95% CI 0.74–1.30), and infection [58]. These data suggest that although hygroscopic dilators can dilate the cervix, they are inadequate for improving the outcome of induction. However, no large trials have been performed and there are no high‐quality comparative studies evaluating the optimal use of hygroscopic dilators with other modalities, such as amniotomy, to improve the rate of successful induction. A significant disadvantage of the use of laminaria for cervical ripening is patient discomfort both at the time of insertion and with progressive cervical dilatation. With other equally effective agents available, there is no obvious benefit to support their routine use for labor induction at term (Level of evidence A).

Transcervical balloon catheters appear to be as effective for preinduction cervical ripening as prostaglandin E2 (PGE2) gel and intravaginal misoprostol in most studies [64–70] (Level of evidence A). A meta‐analysis on intravaginal misoprostol vs. transcervical Foley catheter (nine studies included, n = 1603) revealed no significant difference in mean time to delivery (mean difference 1.08 ± 2.19 hours shorter for misoprostol, p = 0.25), rate of cesarean delivery (RR 1.0; 95% CI 0.77–1.28), or rate of chorioamnionitis (RR 1.13; 95% CI 0.61–2.09) [70]. As anticipated, transcervical balloon catheters were associated with a lower incidence of tachysystole. The combination of a balloon catheter plus administration of a prostaglandin does not appear to be more effective than prostaglandins alone [69]. While the risk of infection may theoretically be associated with the insertion of a foreign object in the cervix, existing meta‐analysis data did not show evidence of an increased risk of infectious morbidity. This technique is a superior method of preinduction cervical ripening when compared with intravenous oxytocin and has been associated with a lower rate of cesarean delivery in one investigation [58]. Some studies show more rapid cervical ripening, a shortened induction to delivery interval, and reduced frequency of patients undelivered in 24 hours when combining a transcervical balloon catheter with a pharmacologic method of cervical ripening such as a prostaglandin [66], whereas others do not [64]. Sciscione et al. [69], in a study examining 126 women, found no increased risk of preterm delivery in subsequent pregnancies following the placement of balloon catheters in the lower uterine segment (Level of evidence A).

The use of the Atad double‐balloon device has also been described in a limited group of studies [70–72]. One investigation included 95 women with Bishop scores <4 and randomly assigned them to vaginally administered PGE2, Atad balloon dilator technique, or continuous oxytocin for labor induction. They found a significant mean change in Bishop score after 12 hours in the PGE2 group and Atad balloon dilator group of 5 compared with 2.5 in the oxytocin group. In addition they found a higher rate of failed induction in the oxytocin group (58%) compared with 20% in the PGE2 and 5.7% in the Atad balloon dilator groups. Vaginal delivery rates in the oxytocin group were 26.7% compared with 77% and 70% in the Atad balloon dilator and PGE2 groups, respectively. There are no comparative studies of single to double‐balloon catheters.

In comparison to mechanical methods, randomized trials have established that prostaglandins (PG E2, F2‐alpha, and E1) are also effective for both cervical ripening and labor induction [73–76]. The efficacy of locally applied prostaglandins (vaginal or intracervical) for cervical ripening and labor induction as compared with oxytocin (alone or in combination with amniotomy) has been demonstrated in a Cochrane review involving more than 10 000 women. Vaginal PGE2 compared with placebo reduced the likelihood of vaginal delivery not being achieved within 24 hours, the risk of the cervix remaining unfavorable or unchanged, and the need for oxytocin. There was no difference between cesarean delivery rates, although PGE2 use increased the risk of uterine tachysystole with fetal heart rate changes. The various administration vehicles (tablet, gel, and timed‐release pessary) appear to be as efficacious as each other [76]. The optimal route, frequency, and dose of prostaglandins of all types and formulations for cervical ripening and labor induction have not been determined (Level of evidence A).

Although oxytocin is an effective means of labor induction in women with a favorable cervix, it is less effective as a cervical ripening agent. Many randomized controlled trials comparing oxytocin with various prostaglandin formulations and other methods of cervical ripening confirm this observation. Lyndrup et al. [77] compared the efficacy of labor induction with vaginal PGE2 with continuous oxytocin infusion in 91 women with an unfavorable cervix (Bishop score < 6). They found PGE2 more efficacious for labor induction in 12–24 hours, with fewer women undelivered at 24 hours. However, by allowing the inductions to proceed for 48 hours, they found no difference in vaginal delivery rates after 48 hours between the two groups. In a larger study involving 200 women with an unfavorable cervix undergoing labor induction, vaginally applied PGE2 was compared with continuous oxytocin infusion [78]. These investigators found a shorter time interval to active labor, a significantly greater change in Bishop score, fewer failed inductions, and fewer multiple‐day inductions with PGE2 compared with oxytocin. No difference in the rate of cesarean delivery was found between the groups overall. In a Cochrane review of 110 trials including more than 11 000 women comparing oxytocin with any vaginal prostaglandin formulation for labor induction, oxytocin alone was associated with an increase in unsuccessful vaginal delivery within 24 hours (52% vs. 28%, RR 1.85, 95% CI 1.41–2.43). There was no difference in the rate of cesarean delivery between groups. When intracervical prostaglandins were compared with oxytocin alone for labor induction, oxytocin alone was associated with an increase in unsuccessful vaginal delivery within 24 hours (51% vs. 35%, RR 1.49, 95% CI 1.12–1.99) and an increase in cesarean delivery (19% vs. 13%, RR 1.42, 95% CI 1.11–1.82) [79] (Level of evidence A).

1. 4. In a pregnant patient undergoing labor induction with oxytocin (population), do low dose protocols (intervention) lead to more cesareans (outcome) than high dose protocols (comparison)?

Oxytocin is a polypeptide hormone produced in the hypothalamus and secreted from the posterior lobe of the pituitary gland in a pulsatile fashion. It is identical to its synthetic analog (pitocin), which is among the most potent uterotonic agents known. Synthetic oxytocin is an effective means of labor induction [79]. Oxytocin is most often given intravenously because when given orally the polypeptide is degraded to small, inactive forms by gastrointestinal enzymes. The plasma half life is short, estimated at three to six minutes [80], and steady‐state concentrations are reached within 30–40 minutes of initiation or dose change [81].

The optimal regimen for oxytocin administration is debatable, although success rates for varying protocols are similar. Protocols differ as to the initial dose (0.5–6 mU min⁻¹), incremental time period (10–60 minutes), and maximum dose (16–64 mU min⁻¹) [3]. Success rates for the varying protocols are strikingly similar. Several experts have suggested that implementation of a standardized protocol is desirable to minimize errors in oxytocin administration [82–84]. A literature review of randomized clinical trials of high vs. low dose oxytocin regimens for augmentation or induction of labor concluded high‐dose oxytocin decreased the time from admission to vaginal delivery, but did not decrease the incidence of cesarean delivery compared with low‐dose therapy [85]. Only one double‐blinded randomized trial has been published, and had the same findings [86]. High dose regimens are associated with a higher rate of tachysystole then low dose regimens, and in some studies this has resulted in a higher rate of cesarean for fetal distress [87], but no significant differences in neonatal outcomes have been noted [88]. A large observational study produced by the Consortium on Safe Labor evaluated 7775 nulliparous and 7280 multiparous patients, with similar results to the randomized trials [89]: no differences in rate of cesarean delivery or other perinatal outcomes. The Safe Labor Project is a retrospective observational study conducted by the Eunice Kennedy Shriver National Institute of Child Health and Human Development, National Institutes of Health, in collaboration with several institutions across the United States. In this particular evaluation of oxytocin regimens, six hospitals provided data on 15 054 women who were grouped based on their starting oxytocin doses (1, 2, or 4 mU min⁻¹). Interestingly, the high dose regimen (starting dose of 4 mU min⁻¹ with increases of 4 mU min⁻¹) was associated with reduced risks of meconium staining, chorioamnionitis, and newborn fever in multiparous patients.

The oxytocin dose is typically increased until there is normal progression of labor, or strong contractions occurring at two to three minutes intervals, or uterine activity reaches 150–350 Montevideo units (i.e. the peak strength of contractions in mmHg measured by an internal monitor multiplied by their frequency per 10 minutes). There is no benefit to increasing the dose after one of these endpoints has been achieved. In addition, two randomized trials found there was no significant benefit in continuing oxytocin infusion after the onset of active labor [90, 91].

Low‐dose protocols attempt to mimic endogenous maternal physiology [3]. Oxytocin is initiated at 0.5–1 mU min⁻¹ and increased by 1 mU min⁻¹ at 40–60‐minutes intervals. Slightly higher doses beginning at 1–2 mU min⁻¹ increased by 1–2 mU min⁻¹, with shorter incremental time intervals of 15–30 minutes have also been recommended [92]. Pulsatile oxytocin administration at 8–10‐minutes intervals is considered a variant of low‐dose oxytocin administration and may better simulate normal labor. It has the advantage of reducing total oxytocin requirements by 20–50% compared to nonpulsatile regimens [93–95].

High‐dose oxytocin regimens are often employed in active management of labor protocols. Examples of these protocols include an initial oxytocin dose of 6 mU min⁻¹ increased by 6 mU min⁻¹ at 20 minutes intervals [93] or an initial dose at 4 mU min⁻¹ with 4 mU min⁻¹ incremental increases [41]. It is important to note that active management of labor protocols do not consist merely of the dosing for oxytocin, but are actually multi‐faceted strategies which include one to one nursing care and early amniotomy within an hour of active labor diagnosis [96] (Level of evidence B).

In a systematic review comparing high dose and low dose infusions for induction of labor, nine trials involving 2391 women and their babies were included in the review. Their findings did not provide evidence that high‐dose oxytocin increases vaginal delivery within 24 hours or the cesarean delivery rate. Additionally, they did not find a decrease in induction to delivery time, although results may be confounded by poor quality trials [97] (Level of evidence B).

1. 5. In pregnant patients undergoing induction of labor (population), what constitutes a failed induction (outcome)?

Vaginal delivery is the goal of the induction process; however, this occurs less often than when women labor spontaneously. It is important to allow adequate time for cervical ripening and development of an active labor pattern before determining that an induction has failed. One group proposed that failed induction be defined as the inability to achieve cervical dilatation of 4 cm and 80% effacement or 5 cm (regardless of effacement) after a minimum of 12–18 hours of both oxytocin administration and membrane rupture [98]. They also specified that uterine contractile activity should reach 5 contractions per 10 minutes or 250 Montevideo units, which is the minimum level achieved by most women whose labor is progressing normally. The goal is to minimize the number of cesarean deliveries performed for failed induction in patients who are progressing slowly because they are still in the latent phase of labor [41, 99, 100]. Once induced women enter active labor, progression should be comparable to progression in women with spontaneous active labor, or faster [101].

The utility of administering oxytocin for at least 10–12 hours after membrane rupture is illustrated by the following examples:

• An Australian researcher evaluated a group of 978 nulliparous women after either artificial or spontaneous rupture of membranes to determine factors that could predict failed induction [102]. There was a direct correlation between increasing duration of the latent phase and the probability of cesarean birth. After 10 hours of oxytocin administration, the 8% of women not in the active phase of labor had an approximately 75% chance of being delivered by cesarean for failed induction; after 12 hours of oxytocin administration, the chance of cesarean was almost 90%.
  
• Two large studies that required a minimum of 12 hours of oxytocin administration after membrane rupture before diagnosing failed labor induction reported that vaginal delivery occurred in 75% of all nulliparas and that failed labor induction was eliminated as an indication for cesarean [100]. Also, for nulliparous women with an unfavorable cervix, the overall rate of vaginal delivery was 63%, with most of the 95% of women who completed latent phase delivering vaginally and approximately 40% of the remaining 5% of women achieving vaginal deliveries as well [103].
  
• Membrane rupture and oxytocin administration should in most cases be a prerequisite before diagnosing a failed induction of labor. Additionally, experts have proposed waiting at least 24 hours in the setting of both oxytocin and ruptured membranes before making the diagnosis [104] (Level of evidence C).