Early Labor Assessment

The goal of early labor assessment is to delay hospital admission until a woman is in active labor because early admission is associated with increased rates of obstetric intervention, including cesarean.^110-111 Two Canadian trials of fair quality assessed the effect of early labor assessment strategies (see Table 6 below).^{68, 102} A trial of 1,459 nulliparous women compared home-based triage (n=728) with telephone triage (n=731).¹⁰² Women in both groups were evaluated for labor progress and abnormalities, fetal movement, and maternal coping. In addition, those women who had home visits also had maternal vital signs checked, abdominal palpation, fetal heart rate auscultation, assessment of contractions, and cervical examination. Women in the telephone group were given suggestions for coping with contractions while women and their partners in the home visit group received education about comfort measures. Criteria for advising women to proceed to the hospital were the same for both groups except cervical dilatation, which was used as an additional criterion for the home visit group. The percentage of women who had a cesarean birth was higher in the home-based triage group (28.6%) than the telephone triage group (25.4%), but this difference was not statistically significant (RR=1.12, 95% CI: 0.94, 1.32). Five-minute Apgar scores (RR=1.52, 95% CI 0.54-4.23), admission to a Level II nursery (RR=0.93, 95% CI: 0.63, 1.37), and admission to a Level III nursery (RR=2.35, 95% CI: 0.90, 6.08) were comparable (Tables 10 and 11).

Table 6

Summary of effectiveness of labor management cesarean reduction strategies: early labor assessment.

Table 10

Maternal outcomes for labor management strategies to reduce cesarean births.

Table 11

Neonatal outcomes for labor management strategies to reduce cesarean births.

In a trial that enrolled 209 women without prior births, participants were randomized to early labor assessment or direct admission when they presented to the hospital in labor.⁶⁸ Women in the early labor assessment group were evaluated and only admitted to the labor and delivery unit if they were in active labor, defined as the presence of regular, painful contractions and cervical dilation greater than 3 cm. Women who were not in active labor were advised to walk outside or return home until active labor began. Women in the control group were not assessed prior to admission to the labor and delivery unit. The percentage of women who had a cesarean birth was lower in the early labor assessment group (7.6%) than the direct admission group (10.6%), but this difference was not statistically significant (OR=0.70, 95% CI: 0.27, 1.81).⁶⁸ As shown in Table 11, the percentage of infants with Apgar scores lower than seven at five minutes was comparable (p=0.318).⁶⁸

Midwifery-Led Care

According to the U.K.'s National Institute for Health and Clinical Excellence, low-risk women who give birth in a midwife-led unit are more likely to have a normal birth with less intervention.¹¹² A poor quality trial from Norway compared outcomes for nulliparous and multiparous women who gave birth in 1) a midwife-led unit for low-risk women who wanted as little intervention as possible where epidural and augmentation outside of the second phase of second stage of labor were not available (n=412); 2) a normal unit for women with expected normal births where extended surveillance, epidural, operative vaginal birth, cesareans, and induction of labor were offered (n=417); and 3) a special unit for women requiring extended surveillance in the antepartum period, during labor, or after birth (n=282).⁷² Women expecting normal births can give birth at any of the three units; however, the midwife-led unit only accepts low-risk women. Eligible women who desired to participate were randomized to one of the three units at the onset of spontaneous labor. The percentage of women who had a cesarean birth was lower in the midwife-led unit (16.0%) than the normal unit (18.0%) or special unit (18.8%), but these differences were not statistically significant (midwife-led unit vs. normal unit RR=0.90, 95% CI: 0.67, 1.22; midwife-led unit vs. special unit RR=0.87, 95% CI: 0.62, 1.20).⁷² The percentage of women with postpartum hemorrhage, neonates with Apgar score less than seven at five minutes, and NICU admissions did not differ significantly across groups (Tables 10 and 11).⁷²

Measurement of Labor Progress

The purpose of measuring labor progress is to remain vigilant for and intervene early in abnormal progress, also known as labor dystocia, which is the most common indication for primary cesareans in the United States.¹¹³ When dystocia is identified, strategies to improve labor progress, such as augmentation with oxytocin, can be used. Various methods of tracking labor progress are available including the partogram, which is a paper form used to record labor examination findings such as cervical dilation, fetal descent, and contraction frequency. The partogram provides a graphical representation of labor progress and alerts clinicians to abnormal progress.¹¹⁴ Four trials investigated strategies to measure labor progress. Three involved partograms,^{51, 82, 91} and the fourth used a computerized reference range (Table 7).⁹⁴

Table 7

Summary of effectiveness of labor management cesarean reduction strategies: measurement of labor progress.

Use of a partogram in conjunction with standard written labor progress notes was compared to documentation of labor progress solely with sequential notes in a Canadian study with 1,962 participants.⁸² The proportion of women who had a cesarean birth was lower in the group whose labors were assessed with the partogram compared with those who only had standard notes (24.7 percent vs. 25.4%), but this difference was not statistically significant (p=NS reported with no specific value given).⁸² The differences in rates of maternal fever, five-minute Apgar scores less than seven, and NICU admission were not statistically significant (test results not reported, see Tables 10 and 11).

Partograms typically include pre-printed alert and action lines. The alert line represents the labor progress of the slowest (less than or equal to the 10^th percentile) of nulliparous women. The action line, which prompts clinicians that intervention may be warranted for slow labor progress, is placed a number of hours (usually two to four) after the alert line.¹¹⁴ In two trials in the United Kingdom, women whose labor progress crossed the action line had management for prolonged labor (oxytocin augmentation with amniotomy if membranes were intact).^{51, 91} In the first trial,⁵¹ 315 women had a partogram with a two-hour action line, 302 women had a partogram with a three-hour action line, and 311 women had a partogram with a four-hour action line. The rate of cesarean birth was lowest in the women in the 4-hour group (8.4%) followed by the two-hour group (11.1%) and the three-hour group (14.2%).⁵¹ When the intervals were compared (2-hour vs. 3-hour, 3-hour vs. 4-hour, and 2-hour vs. 4-hour), only the results for the 4-hour versus the 3-hour were significant (OR=1.8, 95% CI: 1.1, 3.2).⁵¹ In the second trial,⁹¹ 1,490 women had a partogram with a 2-hour action line, and 1,485 women had a partogram with a 4-hour action line. The rate of cesarean birth in both groups was identical (9.1%).⁹¹ Both trials had rates of postpartum hemorrhage, 5-minute Apgar scores less than seven, and NICU admission that did not differ significantly across groups (see Tables 10 and 11).^{51, 91}

One trial in 4,988 women evaluated the addition of a computerized reference range to standard measurement of labor progress by plotting cervical dilatation against time in nulliparous women.⁹⁴ The software combined the results of clinical examinations with contraction frequency from uterine monitoring to produce a percentile comparison to the reference range. Cesareans were performed in 17.6 percent of the experimental group and 16.9 percent of the control group. This difference was not statistically significant (RR=1.04, 95% CI: 0.92, 1.18).⁹⁴ The only maternal or neonatal outcome of interest that was reported was the percentage of newborns in each group who had five-minute Apgar scores lower than seven, which was nearly identical (see Table 11).

Active Management of Labor

Active management of labor is a general term for a multifaceted approach to labor care that includes some or all of the following: “patient education, strict criteria for the diagnosis of labor, strict criteria for the determination of abnormal progress of labor, high-dose oxytocin infusion, one-to-one nursing support in labor, strict criteria for interpretation of fetal compromise, and peer review of operative deliveries.”¹¹³ The purpose of active management of labor is to decrease the incidence of dystocia, which should in turn decrease the cesarean rate.⁶⁶ A Cochrane review of seven active management trials initially did not find a significant reduction in the cesarean rate, but the difference was significant (RR=0.77, 95% CI: 0.63, 0.94) when a trial in which more than one-third of participants become ineligible after randomization was removed from the analysis.¹¹⁵ The Cochrane review differed from this report in that it specifically sought studies examining active management regardless of low-risk status or primary aim of reducing cesarean.¹¹⁵

Six trials examined use of active management of labor for cesarean reduction; one demonstrated effectiveness (Table 8).^{50, 64, 66, 71, 83, 87} This trial of good quality in South Africa combined use of a partogram with aspects of active management.⁵⁰ Nulliparous women in active labor, defined as regular and painful contractions with cervical dilatation of four or more centimeters, were assigned to aggressive (n=344) or expectant (n=350) management.⁵⁰ A partogram with a single alert line was used for the aggressive management group who had a repeat vaginal examination two hours after the first examination. If cervical dilatation had progressed on or above the alert line, the cervix was reexamined in two to four hours depending on when complete cervical dilatation was expected. If cervical dilatation crossed the alert line, fetal heart rate was normal, and gross cephalopelvic disproportion was not present, oxytocin was started. The expectant management group had a partogram with an alert line and a four-hour action line. Their cervical examination was repeated four hours after initial examination. If cervical dilatation had progressed on or above the alert line, the cervix was reexamined in two to four hours depending on current dilatation and expected time of complete dilatation. If cervical dilatation had moved to the right of the alert line, the cervix was reexamined at the time it was anticipated the action line would be crossed. If cervical dilatation reached or crossed the action line, fetal heart rate was normal, and gross cephalopelvic disproportion was not present, oxytocin was started. Amniotomy was not performed due to the high prevalence of HIV and thus need to prevent vertical transmission. The cesarean rate was 16.0 percent in the aggressive management group and 23.4 percent in the expectant management group (RR=0.68, 95% CI: 0.50, 0.93).⁵⁰ As shown in Table 11, three newborns (one of which was a known intrauterine fetal death [IUFD] prior to enrollment) in the aggressive management group and none in the expectant management group had Apgar scores less than eight at 10 minutes.⁵⁰ There were three perinatal deaths in the aggressive management group (including the known IUFD) and none in the expectant management group, but this difference was not statistically significant (RR=7.12, 95% CI: 0.37, 137.37).⁵⁰

Table 8

Summary of effectiveness of labor management cesarean reduction strategies: active management of labor.

A New Zealand trial of poor quality randomized 551 nulliparous women to active (n=320) or routine (n=331) management when labor was diagnosed, which was defined as regular, painful contractions every five minutes or more lasting at least 40 seconds accompanied by either spontaneous rupture of membranes or complete cervical effacement with cervical dilatation of at least two centimeters.⁸³ In the active management group, amniotomy was encouraged at the time of labor diagnosis, and the cervix was assessed every two hours. Oxytocin augmentation was initiated if cervical dilation was less than one centimeter per hour, descent of the fetal head had not occurred after 30 minutes of pushing, or contractions were more than five minutes apart without imminent birth during the second stage of labor. Oxytocin was started at 6 mU per minute and increased by 6 mU every 15 minutes to a maximum dose of 36 mU per minute. In the routine management group, the caregiver determined use of amniotomy, frequency of cervical examination, and use of oyxtocin.⁸³ The cesarean rate was 9.4 percent in the active management group and 9.7 percent in the routine management group (p=0.5).⁸³ The groups did not differ significantly in maternal infection (RR = 1.12; 95% CI: 0.72, 1.74) or postpartum hemorrhage (RR=1.04, 95% CI: 0.72, 1.51; see Table 10).⁸³

In a U.S. trial of poor quality, nulliparous women were randomized to active management (n=1,009) or usual care (n=906) before 30 weeks' gestation.⁸⁷ Women in the active management group attended special childbirth classes to explain the protocol, while women in the usual care group were given payment to attend the classes of their choice. Active management was provided in a separate unit by nurse-midwives and nurses who worked exclusively for the study. Active management included one-to-one nursing care and standardized criteria for labor diagnosis (painful contractions with effacement of at least 80 percent, bloody show, or spontaneous rupture of membranes). Women in the active management group had amniotomy within one hour of labor diagnosis and cervical examinations at least every two hours. Women who had inefficient uterine action (cervical dilation of less than one centimeter per hour during first stage or greater than one hour between full dilatation and the fetal head reaching the pelvic floor during second stage) or more than 30 minutes elapsed between the fetal head reaching the pelvic floor and the birth received oxytocin (started at 4 mU per minute and increased by 4 mU every 15 minutes to a maximum dose of 40 mU per minute).⁸⁷ Care of women with failure to progress (cervical dilation less than one centimeter per hour after efficient uterine action was established with oxytocin or prolonged second stage) was assumed by the woman's regular provider. No constraints were placed on management of the usual care group. Use of amniotomy, initiation of oxytocin, and cervical examination were at the provider's discretion. When oxytocin was used, it was typically started at a dose of one to two mU per minute and increased by 1-2 mU per minute periodically. The rate of cesarean was 19.5 percent in the active management group and 19.4 percent in the usual care group (RR=1.0, 95% CI: 0.8, 1.2).⁸⁷ Among the protocol-eligible subgroup (n=678 active management, n=585 usual care), incidence of maternal fever was lower in the active management group than the usual care group (n not provided; RR=0.6, 95% CI: 0.4, 0.9).⁸⁷ There was no significant difference between groups in admission to the NICU (n and statistical test result not provided).⁸⁷

A U.S. trial of poor quality with 705 nulliparous women enrolled participants in spontaneous labor, which was defined as regular, painful contractions at least every five minutes plus complete cervical effacement or spontaneous rupture of membranes.⁶⁶ Active management included amniotomy within one hour of labor diagnosis, hourly cervical examinations for the first three hours then examinations every two hours, and high-dose oxytocin augmentation (started at 6 mU per minute and increased by 6 mU per minute every 15 minutes) for cervical dilation of less than one centimeter per hour in the first stage of labor or fetal descent of less than one centimeter per hour in the second stage of labor. In the traditional management group, the physician decided when to perform amniotomy, how often to examine the cervix, and what criteria were used to diagnose inadequate labor progress. When oxytocin augmentation was used, it was typically started at 1 mU per minute and increased by 1-2 mU per minute every 15 minutes until there were eight contractions per 20 minutes. The cesarean rate was 10.5 percent in the active management group, and 14.1 percent in the traditional management group (p=0.18).⁶⁶ The active management group had a significantly lower rate of chorioamnionitis than the traditional management group (4.6% vs. 9.9% percent, p<0.01).⁶⁶ Differences in five-minute Apgar scores less than seven and NICU admission rates were not significant (test result not reported, see Table 11).⁶⁶ No neonatal deaths occurred in either group.⁶⁶

Another U.S. trial of poor quality enrolled 405 nulliparous women.⁶⁴ Participants in the active management group were diagnosed with labor when they had regular, painful contractions every two to five minutes with at least 80 percent cervical effacement, regardless of dilatation. They had amniotomy within two hours of admission, cervical examination every two hours, and high-dose oxytocin augmentation (started at 6 mU per minute and increased every 15 minutes) for cervical dilation of less than one centimeter per hour in the first stage of labor or fetal descent of less than one centimeter per hour in the second stage of labor. Participants in the control group were admitted when they had regular, painful contractions every two to five minutes and three to four centimeters of cervical dilatation, regardless of effacement.⁶⁴ Amniotomy was performed at the physician's discretion. If the cervix did not change 1.25 centimeters per hour once the active phase of labor began, low-dose oxytocin was started (1 mU per minute and increased by 1 mU/min every 30 to 40 minutes). The cesarean rate was 7.5 percent in the active management group and 11.7 percent in the expectant management group, which was not a significant difference.⁶⁴ Maternal and neonatal outcomes were similar in both groups (not statistically significant, no test results reported, see Tables 10 and 11).⁶⁴

In a poor quality trial of 960 nulliparous women in Thailand,⁷¹ labor was diagnosed when by regular, painful contractions lasting at least 40 seconds and occurring at least once per five minutes plus spontaneous rupture of membranes or bloody show with cervical dilatation and complete effacement. The active management group had amniotomy within one hour of admission, cervical examination every two hours, and high doses of oxytocin (started at 6 mU per minute and increased by 2 mU per minute every 30 minutes until there were five contractions per 10 minutes or the rate was 40 mU per minute) if cervical dilatation was less than one centimeter per hour in the first stage of labor.⁷¹ The conventional management group did not have a protocol for amniotomy, cervical examination, or oxytocin initiation. The use of cesarean was lower in the active management group (11.9%) than the conventional management group (14.7%), but this difference was not significant (p-value not reported).⁷¹ There were no significant differences (test results not reported) between groups in rates of maternal fever, chorioamnionitis, and one-minute Apgar scores less than seven (Tables 10 and 11).⁷¹

Management of Abnormal Labor

Five trials assessed strategies for managing abnormal labor (see Table 9).^{41, 54, 79, 90, 104} One resulted in proven reduction of cesarean. A poor quality trial of early versus delayed oxytocin included 412 nulliparous women with primary dysfunctional labor, diagnosed by cervical dilatation of two centimeters or less over four hours from initial dilatation of three to six centimeters.⁹⁰ Women with intact membranes had amniotomy prior to randomization. The early oxytocin group started oxytocin within 20 minutes of randomization, while the delayed oxytocin group did not receive oxytocin for eight hours unless intervention was warranted.⁹⁰ Cesarean risk was identical in the two groups (14%), and there were no significant differences in postpartum hemorrhage (OR=0.87, 95% CI: 0.5, 1.4), five-minute Apgar scores (OR=1.6, 95% CI: 0.4, 7.0), NICU admission (OR=1.2, 95% CI: 0.4, 3.9), or neonatal death (OR=0.98, 95% CI: 0.06, 16) (Tables 10 and 11).⁹⁰

Table 9

Summary of effectiveness of labor management cesarean reduction strategies: management of abnormal labor.

Two trials, one of good quality in the U.S.⁴¹ and one of fair quality in Finland,¹⁰⁴ compared oxytocin alone and with propranolol, a beta receptor blocking agent thought to have the potential to enhance uterine activity, for treatment of abnormal labor progress. These were the only trials related to management of labor that included both nulliparous and multiparous women. The first trial defined dysfunctional labor as no cervical dilatation for at least two hours in the active phase of labor or a deceleration phase of at least three hours in nulliparas and one hour in multiparas.⁴¹ All women continued to receive oxytocin. Propranolol (n=49) or placebo (n=47) were administered intravenously and repeated after one hour if cervical dilatation did not change. Cesarean was performed if there was no response within an hour after the second dose.⁴¹ The cesarean rate was 26.5 percent in the propranolol group and 51.1 percent in the placebo group (RR=0.58, 95% CI: 0.35, 0.93; p=0.02).⁴¹ Differences in Apgar scores and NICU admissions were not significant (Table 11). The second trial defined arrest of labor as hypocontractility with other causes of dystocia, such as disproportion, excluded.¹⁰⁴ At the time of diagnosis of arrested labor, oxytocin was initiated along with a dose of propranolol (n=55) or placebo (n=52), which was repeated an hour later if the cervical status was unchanged.¹⁰⁴ The timing of cesarean was not specified as it was in the first trial. The cesarean rate was 11 percent in the propranolol group and 4 percent in the placebo group (p=0.154). The difference in NICU admissions was not significant (Table 11).¹⁰⁴

One trial of fair quality, done in the United Kingdom, examined three oxytocin protocols in 60 nulliparous women in labor (diagnosed by complete cervical effacement, dilatation greater than or equal to three centimeters, regular contractions with at least one per five minutes, and cervical progress on a partogram) whose cervical dilatation was less than 0.5 centimeters per hour.⁵⁴ Amniotomy was performed prior to randomization for women with intact membranes. Participants were randomized into three arms: delayed oxytocin for eight hours (n=20); automatic infusion system (AIS) oxytocin (2 mU per minute increased by 2 mU every 15 minutes) for women whose uterine activity was less than 700 kPas per 15 minutes (n=21, 13 received oxytocin); and high-dose oxytocin (7 mU per minute increased by 7 mU every 15 minutes, n=19).⁵⁴ The cesarean rate was 45 percent in the delayed oxytocin group, 33.3 percent in the AIS oxytocin group, and 26.3 percent in the high-dose oxytocin use. Differences across groups were not statistically significant.⁵⁴ Five-minute Apgar scores lower than seven (Table 11) did not differ significantly between groups (test result not reported).⁵⁴

A poor quality U.S. trial of early intervention included 150 primigravid women who had been admitted in labor (diagnosed by contractions plus cervical dilatation of three centimeters or ruptured membranes) and had an inadequate contraction pattern (less than 3 contractions lasting 40 seconds each in a 10-minute time period).⁷⁹ The early intervention group had amniotomy if membranes were intact, insertion of a fetal electrode and an intrauterine pressure cannula, and initiation of oxytocin infusion, all of which were performed within 30 minutes of admission. The control group had external fetal monitoring and oxytocin infusion if the cervical dilatation did not change for more than two hours or if there was no change in station for one hour during the second stage of labor.⁷⁹ The cesarean incidence was 13 percent in the early intervention group and 15 percent in the control group, a difference that was not statistically significant.⁷⁹

Amniotomy

Amniotomy, the surgical rupture of fetal membranes, has been purported to shorten the duration of the first stage of labor although a Cochrane review of 15 studies found no statistical difference in first-stage labor length or the rate of cesarean when amniotomy was performed.¹¹⁶ A fair quality trial in Nigeria, randomly assigned 128 women to amniotomy or no amniotomy upon presentation in labor.⁸⁸ Both groups had identical risk of cesarean (1.6 percent; p>0.05). The only neonatal outcome reported was five-minute Apgar scores, which were comparable between groups (test result not reported, see Table 11).⁸⁸

Increased Intravenous Fluids

One U.S. trial of good quality evaluated the effect of increased intravenous hydration in labor based on the rationale that adequate fluid replacement might improve labor progress, and subsequently reduce the cesarean rate, similar to the effects of adequate hydration on the exercise performance of athletes.⁵⁶ Women received increased (250 ml/hour, n=101) or standard (125 ml/hour, n=94) rates of lactated Ringer's solution or isotonic sodium chloride solution.⁵⁶ The percentage of women who had a cesarean birth was lower in the 250-ml group (9.9%) than the 125-ml group (17.0%), but this difference was not statistically significant (p=0.22). Maternal and infant outcomes were similar (no test of statistical significance reported, see Tables 10 and 11).⁵⁶

Oral Carbohydrate Solution

One trial of fair quality from the Netherlands evaluated the effect of an oral carbohydrate solution during labor based on the rationale that carbohydrate intake might reduce the cesarean rate and improve labor progress, similar to the effects of carbohydrate intake on the exercise capacity and fatigue among athletes.¹⁰⁶ Women received a carbohydrate (n=102) or placebo (n=99) solution to drink at will. They were not offered other food or drinks but could have small amounts of either upon request.¹⁰⁶ The percentage of women who had a cesarean birth was higher in the carbohydrate group (20.6%) than the placebo group (7.1 percent; RR=2.91, 95% CI: 1.29, 6.54).¹⁰⁶ Maternal and infant outcomes we extracted were not reported (Tables 10 and 11).¹⁰⁶

Doulas as Providers of Labor Support

Three doula support studies were included (Table 12).^{42, 46, 48} All studies of doula support were conducted with participants in labor who had uncomplicated pregnancies at term and were having their first birth. One RCT in Mexico⁴² compared 50 women who received labor support from childbirth educators who had doula training to 50 women who received usual labor care. The study was conducted at a public hospital with an overall cesarean rate of 40 percent. Doulas were Lamaze-trained childbirth educators, who received doula training as part of the Lamaze Childbirth Education curriculum. During labor and childbirth, doulas provided advice and information, physical assistance and emotional support to the study participants and worked actively to promote natural childbirth.⁴²

Table 12

Summary of effectiveness of cesarean reduction strategies using doula support.

The proportion of births by cesarean for the doula supported and usual care groups were 2 and 24 percent respectively (p<0.003).⁴² Labor duration did not differ significantly for the two groups (p-value not reported). Incidence of oxytocin administration for the doula supported and usual care groups were 42 percent and 96 percent respectively (p<0.001).⁴²

In a U.S. trial,⁴⁶ 224 women were assigned to doula support and 196 to usual care. The study was conducted at a university hospital with an overall cesarean rate of 24 percent. All doulas completed training requirements that were equivalent to the Doulas of North America International doula certification. The doula met couples shortly after random assignment and stayed with them throughout labor and birth providing verbal encouragement, reassurance, teaching, and touch, to support the laboring woman and her partner.⁴⁶ Women in the doula supported group had significantly fewer cesarean births (13.4% vs. 25.0%, p=0.002) and fewer epidurals (64.7% vs. 76.0%, p=0.008). Five-minute Apgar scores did not differ (p=0.30) (Table 14).⁴⁶

Table 14

Neonatal outcomes for psychosocial/labor support strategies to reduce cesarean births.

Another U.S. trial⁴⁸ randomized 412 women to doula support or monitoring by an inconspicuous observer. The authors also selected an additional 204 women for an additional “control” comparison group.⁴⁸ The trial was conducted at a public hospital where companions were not routinely permitted to be with a woman during labor and birth. For study participants, brief visits by family members were allowed if the labor area was not too busy. All doulas completed a three-week training period.⁴⁸ They stayed at the patient bedside from admission through birth providing touch, encouragement and information about the labor and childbirth process. The observer stayed in the labor room, but at a distance from the mother, and did not interact with the laboring woman.

The proportions of cesarean births for the doula support, observed, and control groups were 8, 13 and 18 percent respectively (p<0.0001 for all three groups).⁴⁸ When pair wise comparisons were made, significant differences remained (doula vs. observed, p=0.009 and doula vs. control, p=0.004). This study reported forceps births for 8.2 percent of women in the doula support group and 21.3 percent for women in the observed group (p=0.0006).⁴⁸ Forceps-assisted births occurred in 26.3 percent of women in the control group (p=0.006, doula support group vs. control group). Among participants who had spontaneous vaginal births, epidural use varied significantly for the doula, observed, and control groups (p<0.0001).⁴⁸ The mean duration of labor was significantly shorter for the doula group compared to the observed (p<0.02) and control (p<0.02) groups.⁴⁸ Labor augmentation occurred less frequently in the support group compared to the control group (p<0.0001).⁴⁸ The percentages for oxytocin use for labor augmentation were 17, 23 and 43.6 percent, across groups (p<0.0001).⁴⁸ Maternal fever was more common in the observed and control groups than in the supported group, but there was no statistical analysis provided (Table 13).⁴⁸ The authors noted that the proportion of newborns who remained in the hospital more than 48 hours because of medical problems was lower in the supported group (Table 14).⁴⁸

Table 13

Maternal outcomes for psychosocial/labor support strategies to reduce cesarean births.

Trained Female Friend or Family Member as Provider of Labor Support

In the RCT that used family or friend supports,⁹⁶ 291 women were assigned to the supported group and 295 women were assigned to usual care. The participant selected a female friend or family member who participated in two 2-hour learning sessions.⁹⁶ A research assistant who was a doula certified by Doulas of North America conducted training that included: anatomy and physical changes during labor and childbirth, assessing labor progression, coping strategies, how to provide anticipatory guidance, comfort measures and reassurance to laboring women.⁹⁶ There were no restrictions on visitors or other support for laboring women at the hospital. The primary cesarean rate for the study facility during the enrollment period was 17.9 percent.

In this study, support from a trained friend or family member did not reduce cesarean births. The proportions of cesarean births for the intervention and usual care groups were 18.9 and 17.9 percent respectively (p=0.7).⁹⁶ Women in the supported group had significantly shorter lengths of labor (p=0.004), greater cervical dilation at the time of epidural (p=0.007), and a higher proportion of five-minute Apgar scores above six (p=0.006).⁹⁶

Nurses and Midwifery Students as Providers of Labor Support

Two studies in the U.S. and Canada investigated the effects of continuous labor support by nurses,^{73, 86} and one in Finland examined the effects of labor support by midwifery students.⁶⁷ The effectiveness of these strategies are presented in Table 15.

Table 15

Summary of effectiveness of cesarean reduction strategies: labor support by nurses and midwifery students.

One randomized controlled Canadian trial⁷³ of poor quality, assigned 209 women to one-to-one intrapartum support from a nurse and 204 women to usual care. Usual care consisted of two to three patients per nurse with variable labor support techniques provided. The strategy was almost continuous one-to-one nursing care from the time of randomization until one hour after the birth. In addition to usual intrapartum care, the support nurse provided physical comfort, emotional support, instruction on relaxation and coping with pain, and support for the father. The support nurses completed an initial 30-hour training period and quarterly refresher workshops.⁷³

Cesarean risk was not significantly different.⁷³ The proportion of cesareans in the supported and usual care groups was 13.9 and 16.2 percent, respectively (RR=0.86, 95% CI: 0.54, 1.36). Use of oxytocin stimulation for women in the nurse supported group was 17 percent lower (RR=0.83; 95% CI: 0.67, 1.04).⁷³ There were no significant differences in epidural analgesia (RR=0.96, 95% CI: 0.84, 1.09), instrumented vaginal births (RR=1.06, 95% CI: 0.74, 1.53), five-minute Apgar scores (Mean difference=-0.1, 95% CI: 0.1, 1.05), and NICU admissions (RR=1.46, 95% CI: 0.64, 3.18).⁷³

A RCT⁸⁶ randomized 6,290 participants to continuous labor support or usual care. Study sites included nine tertiary care and four community hospitals. The strategy was continuous labor support by a specially trained nurse from the time of randomization to birth. Nurses volunteered to participate and completed a two-day training program conducted by an expert labor nurse doula trainer.⁸⁶ Usual care varied depending on a patient's stage of labor, condition, and nurse workload but did not include care by a nurse with special labor support training.

The proportion of cesarean births in the intervention and usual care groups was 12.5 and 12.6 percent, respectively (p=0.44).⁸⁶ Labor augmentation for the continuous labor support and usual care groups was 30.1 and 27.2 percent, respectively (p=0.008).⁸⁶ Assisted vaginal births, duration of labor, and use of epidural did not differ. There were no significant differences in neonatal outcomes including neonatal deaths (p-value not provided), need for higher level of nursery care (p=0.70), five-minute Apgar scores (p=0.50) and length of hospital stay (p-value not provided).⁸⁶

A study conducted in Finland⁶⁷ included three trials: one small pilot using laywomen as labor support persons and two trials using midwifery students as labor support persons, one conducted in 1987 and one in 1988. The pilot study with laywoman support was stopped for economic and other reasons.

These studies were conducted at a university hospital.⁶⁷ At this hospital, normal births are attended by midwives who do not stay with the patient constantly, and usually take care of more than one laboring woman at a time. Fathers are present for 60 to 70 percent of births. The hospital's cesarean birth rate was 9.8 percent. In 1987, 11 midwifery students volunteered to provide support in labor for study participants, and in 1988 all 16 midwifery students were required to participate.⁶⁷

The 1987 trial randomized 79 women to midwifery student labor support or usual care.⁶⁷ The 1988 midwifery student trial randomized 161 women to either a midwifery student for labor support or usual care. The authors reported outcomes for each trial year and for both combined.⁶⁷ Cesarean births were equally common: none among supported and eight percent in the usual care in 1987, three and four percent in 1988, and five and five percent for both years combined. The supported group had significantly shorter labors from admission to birth (p<0.05). Among women giving birth for the first time, a smaller percentage of women in the supported group had labor durations of 11 hours or more (p<0.01).⁶⁷ The percentage of women whose contractions stopped was significantly lower in the supported group (5 vs. 15%; p<0.01). Postpartum complications (infections, discharge diagnoses and proportion of mothers not nursing at discharge) were rare and similar in both groups. Mean Apgar scores were higher for neonates in the supported group (p<0.05).⁶⁷

Fetal Pulse Oximetry

Three studies^{43, 74, 101} evaluated whether the addition of fetal pulse oximetry to FHR monitoring, fetal scalp sampling or both, would improve fetal assessment and reduce operative birth rates without increasing adverse outcomes for the women, the fetus or the newborn. These trials enrolled women who were at least 36 weeks gestation, with a singleton fetus in vertex presentation, in labor with ruptured membranes (or if not ruptured consented for artificial rupture), and nonreassuring FHR. The effectiveness of these strategies is presented in Table 19.

Table 19

Summary of effectiveness of cesarean reduction strategies: fetal pulse oximetry.

The first study was a U.S. multicenter trial at nine centers.¹⁰¹ Patients gave consent for possible study inclusion but were only randomized once one of the pre-determined concerning FHR patterns developed while in labor. Included women also needed to be at least two centimeters dilated and at minus two station or lower in the pelvis. All randomized participants (n=1,010) underwent FHR monitoring with either Doppler or scalp electrode, or both.¹⁰¹ For patients in the intervention group a fetal oxygen sensor was placed against the fetal cheek and connected to a monitor. For women in the control group, only an electronic FHR monitor was used. For both groups the FHR was defined as reassuring, nonreassuring, or ominous. An ominous FHR pattern, defined as FHR persistently less than 70 beats per minute for at least seven minutes, required immediate birth in either group. The difference in management between the two groups occurred among the patients with nonreassuring FHR patterns. In the intervention group a fetus with a nonreassuring FHR tracing was considered to be normally oxygenated if the fetal oxygen saturation was greater than 30 percent at any time between two contractions. However, if the fetal oxygen saturation remained less than 30 percent for the entire interval between two contractions, the clinician reverted back to the FHR and used the same criteria as that used for the standard control group: if the FHR was persistently nonreassuring the clinician could rule out acidosis by examining spontaneous or induced FHR accelerations or scalp pH to rule out fetal acidosis. If reassurance could not be established, cesarean or operative vaginal birth was undertaken.¹⁰¹

Despite randomization, the intervention arm included more women with induced labors and use of prostaglandins for cervical ripening.¹⁰¹ There was a 50 percent relative reduction, in the number of cesareans performed for nonreassuring fetal status (intervention group 4.5% vs. 10.2% for controls) with no significant difference in overall cesarean use between the two groups (29% in the intervention group and 26% in the control group). An independent reviewer evaluated all electronic FHR tracings. The reviewer identified three cesarean births in each group done for nonreassuring fetal status in which there were protocol violations. There were no differences between the two groups in adverse maternal outcomes (including placental abruption, postpartum hemorrhage, wound infection, intrapartum fever, and endometritis).¹⁰¹ Neonatal outcomes were similar with five neonatal deaths, three in the intervention group and two in the control group.¹⁰¹ Four of the five deaths were caused by complex congenital heart anomalies. The fifth occurred in an infant from the intervention group whose five-minute Apgar score and umbilical cord pH were normal, yet postnatally developed bilateral tension pneumothoraces (Table 21).¹⁰¹

Table 21

Neonatal outcomes for fetal assessment strategies to reduce cesarean births.

In a German trial,⁴³ 146 patients were recruited with nonreassuring FHR patterns (defined by International Federation of Gynaecology and Obstetrics [FIGO] score less than 8).¹²⁹ Women in the study also needed to be at least two centimeters dilated and at minus two station or lower in the pelvis. They were randomized to two groups: triple fetal surveillance with a FHR monitor, fetal scalp sampling, and fetal pulse oximetry (n=73) or a control group in which women received only FHR monitoring and fetal scalp sampling (n=73).⁴³ After randomization, management of labor varied by group. In the intervention group, if cardiotocography (CTG) was nonreassuring, fetal blood sample (FBS) pH was greater than 7.25, and fetal pulse oximetry was greater than 30 percent, they continued CTG and fetal pulse oximetry and attempted vaginal birth. If the CTG was nonreassuring and the fetal blood sample pH was greater than 7.25, but the fetal pulse oximetry was less than 30 percent, they still continued CTG and fetal pulse oximetry; however, a followup fetal scalp sampling was performed. If that repeat scalp sampling pH was less than or equal to 7.25, clinical intervention was necessary (either tocolysis for intrauterine resuscitation, cesarean, or assisted vaginal birth). In the control group, if CTG was non-reassuring and fetal scalp blood sample pH was greater than 7.25, they continued CTG and attempted vaginal birth. If the fetal blood sample pH was less than or equal to 7.25 clinical intervention was necessary. In either group if CTG was ever found to be pathologic, patients were delivered immediately.⁴³

There was no difference between the two groups in incidence of nonreassuring FHR patterns in the first and second stages of labor.⁴³ The first scalp pH sampling for baseline assessment in the two groups was also similar. The proportion of cesareans was significantly lower in the intervention group than in the control group (16.4% vs. 37%), and the proportion of operative vaginal births for nonreassuring fetal status was also significantly lower in the intervention versus control group (17.8% vs. 30.1%) demonstrating almost a 50 percent reduction in the risk of operative births for nonreassuring fetal status. There were no adverse maternal or neonatal events in either group (Tables 20 and 21).⁴³

Table 20

Maternal outcomes for fetal assessment strategies to reduce cesarean births.

The third trial, called the FOREMOST trial, was conducted in four Australian academic hospitals.⁷⁴ Six-hundred and one women with nonreassuring fetal monitoring were randomly assigned to a group with fetal pulse oximetry with CTG or a control group monitored using conventional CTG alone.⁷⁴ Monitoring continued from the time a sensor was placed until as close to birth as possible. In either group, if the CTG became reassuring, labor was continued unless otherwise indicated. In both groups, ominous FHR patterns prompted birth. In the control group, a nonreassuring CTG prompted evaluation and management of the FHR pattern. In the intervention group, if oxygen saturation levels were less than 30 percent for 10 minutes or not recording, evaluation and management of FHR pattern was recommended. Evaluation and management could include maternal position change, supplemental oxygen, hydration, correction of hypotension, discontinuation of oxytocin infusion, or birth. Fetal blood scalp sampling was available to all without restriction and the study did not regulate management based on fetal scalp pH or lactate values.⁷⁴

The primary outcome measured was operative birth (cesarean, forceps, vacuum) for nonreassuring fetal status.⁷⁴ This study reported a significant reduction in operative births for nonreassuring fetal status in the intervention group compared to those in the control group (24.9% vs. 32.2%; RR=0.77, 95% CI: 0.599, 0.999; p=0.048). Cesareans for nonreassuring fetal status represented 13.8 percent of intervention group and 20 percent of control group births (RR=0.69, 95% CI: 0.48, 0.99; p=0.042).⁷⁴ However, the overall rate of operative births between the two groups did not differ (intervention group 73% vs. control group 71%; RR=0.77, 95% CI: 0.599, 0.999; p=0.48). Women in the intervention group were more likely to have an operative birth secondary to dystocia than those in the control group. This difference in indication was significant for assisted vaginal birth, not for cesareans. Fetal scalp blood sampling was performed more often in the control group. There were four cases of endometritis in the intervention group and one in the control group (p=0.192). Postpartum maternal fever was similar between the groups (p=0.792) (Table 20). There was also no difference in neonatal outcomes (Apgar score, cord pH/fetal acidosis, NICU admission), or serious adverse events (Table 21).⁷⁴

The fourth study was conducted by the National Institute of Child Health and Human Development Maternal-Fetal Medicine Units Network.⁸⁰ This RCT enrolled women at 14 academic centers to test whether fetal oximetry in addition to conventional EFM would reduce the overall use of cesarean. The study was launched after a rigorous training phase with the oximetry equipment. During the study, all women recruited underwent placement of fetal pulse oximeter after placement of a standard internal electronic FHR monitor and intrauterine pressure catheter.⁸⁰ If the oximeter was not placed within three attempts or if a signal registration was not accomplished by 15 minutes the attempt was considered unsuccessful, and the patient was not randomization. One hundred seventy women had failed attempts at sensor placement and 42 other attempts were abandoned secondary to prolonged FHR decelerations during placement. Ultimately, 5,341 women were randomly assigned to open fetal oximetry in addition to conventional electronic fetal monitoring or to masked fetal oximetry with conventional electronic fetal monitoring.⁸⁰

Nonreassuring FHR patterns were defined according to the criteria used by Garite.¹⁰¹ Intrapartum management in both groups was otherwise left to the discretion of the attending physician.⁸⁰ Fetal pulse oximeter sensors were removed before study completion in 238 women in the intervention group and 267 women in the masked group (reasons: patient request, n=244; physician request, n=196; technical problems, n=65). Discomfort accounted for 92 percent of patient requests for sensor removal and interference with cervical examination or management of labor for 67 percent of physician requests.⁸⁰

Cesarean births did not differ between the oximetry and masked groups (26.3 and 27.5%; p=0.31).⁸⁰ Cesarean births for nonreassuring FHR (7.1 and 7.9%; p=0.30) and dystocia (18.6 and 19.2%; p=0.76) were also similar. Results were similar in the subgroup (n=2,168) of women in whom a nonreassuring FHR was detected prior to randomization as well as those with normal baseline FHR. Maternal and neonatal outcomes did not differ significantly between groups (Tables 20 and 21). One neonatal death occurred due to sepsis in the masked group.⁸⁰

Fetal ST-Segment Analysis Studies

Two RCTs assessed ST-segment analysis (STAN) to provide additional information.^{70, 99} The effectiveness of these strategies is presented in Table 22 below.

Table 22

Summary of effectiveness of fetal assessment cesarean reduction strategies: fetal STAN.

A study at an academic institution in Finland examined whether intrapartum monitoring with STAN could reduce the rate of neonatal acidemia and operative intervention during labor.⁹⁹ A total of 1,483 women were randomly assigned to the intervention group, with monitoring by STAN through a scalp electrode, or the control group with monitoring by a conventional FHR monitor either via an internal intrauterine scalp electrode or an external ultrasound signal sensor.⁹⁹ Fetal blood sampling was optional in both groups, based on clinician judgment. If scalp pH was less than 7.20, immediate birth was recommended. Primary outcome measures were neonatal acidemia (defined as umbilical artery pH less than 7.10), operative interventions (cesarean and vacuum birth), and need for fetal blood sampling.⁹⁹

An umbilical artery blood gas was available for 714 neonates in the intervention group and 722 in the control group.⁹⁹ There were 83 cases of inadequate monitoring, five in the control group (secondary to technical difficulties) and 78 in the intervention group. Failure in the intervention group was attributed to: monitoring stopped more than 20 minutes before birth (n=25), poor signal quality (n=21), technical difficulties with scalp electrode (n=19), total recording time less than 20 minutes (n=7) and one case of unsuccessful recording. In the control group, 83 percent used an internal scalp electrode for monitoring, and 17 percent used an external ultrasound sensor. Overall cesarean rate did not differ between the two groups (6.4% intervention vs. 4.7% control; p=0.24).⁹⁹ Fetal blood sampling was used less in the intervention group compared to the control group (7% vs. 15.6%; p<0.001). When evaluating fetal pH using values less than 7.10 as a criterion of neonatal acidemia, there were no differences between the groups. There were no maternal complications related to the STAN, electronic fetal monitor or fetal blood sampling (Table 20). Neonatal outcomes did not differ between groups with no difference in neonatal acidemia, Apgar scores, need for intubation, or admission to the NICU (Table 21). There were no neonatal deaths.⁹⁹

In the second study,⁷⁰ a multicenter trial in France, the authors sought to assess whether STAN reduced operative births for nonreassuring fetal status or reduced need for at least one scalp pH during labor. The study population included participants who either had an abnormal FHR pattern (by FIGO classification) or thick meconium stained amniotic fluid during labor. After rupture of membranes, 799 women were randomized to intervention with both STAN and an electronic fetal monitor for CTG to monitor fetal status in labor or to the control group with only an electronic fetal monitor for CTG to monitor fetal status.⁷⁰ In the STAN group, fetuses were monitored continuously through a scalp electrode and recommendations were based on STAN guidelines. Scalp pH testing was optional in both groups. If scalp blood pH was less than 7.20, immediate birth was recommended. As soon as possible after birth, umbilical cord artery and vein gases were obtained and analyzed.

The proportion of operative births for nonreassuring fetal status did not differ between the two groups (33.6% for the study group vs. 37% for the control group; RR=0.91, 95% CI: 0.75, 1.10).⁷⁰ Use of operative interventions for dystocia also did not differ between groups. The percentage of women whose fetus had at least one blood scalp pH measurement during labor was substantially lower in the intervention group (27% compared with 62% in the control group; RR=0.44, 95% CI: 0.36, 0.52). Neonatal outcomes did not differ between groups (acidosis, Apgar score, and NICU admission), with one fetal death in the CTG only group (Table 21).⁷⁰

Acupuncture

Two RCTs from the same institution evaluated the use of traditional Chinese medicine acupuncture to initiate labor, with secondary objectives including reducing the rate of cesarean.^{53, 105} In the first trial,⁵³ 30 women were randomized to receive acupuncture on three of four consecutive days for initiation of labor, with 26 women randomized to usual care. The incidence of cesarean was 17 percent in the intervention group and 39 percent among control patients (p=0.07). In the second trial,¹⁰⁵ participants were randomized to up to five acupuncture treatments over two weeks (n=20), sham acupuncture (n=29), or usual care (n=30). The cesarean rate was 20 percent in the acupuncture group as compared with 57 percent in the usual care group; however, the sham treatment had the lowest incidence of cesarean, at 7 percent (p=0.37 for comparison across the three groups). Both studies found similar maternal and neonatal outcomes for intervention as compared with control participants (Tables 28 and 29).

Table 28

Maternal outcomes for unique strategies to reduce cesarean births.

Table 29

Neonatal outcomes for unique strategies to reduce cesarean births.

Devices

Two trials evaluated the utility of devices for reducing use of cesarean. The larger of these studies¹⁰⁰ randomized women to use of an inflatable obstetric belt to provide fundal pressure during contractions (n=260) or to usual care (n=240), finding a similar incidence of cesarean between the groups (5.8 and 3.8% respectively; p=0.29). There were fewer malpositions at birth in the belt group as compared with usual care (15% vs. 20.8%) but the difference was not statistically significant. Other neonatal and maternal outcomes were similar between the two groups (Tables 28 and 29).

Matsuo and colleagues⁵⁵ assessed whether the use of a molded dental device during active pushing (n=32) had an effect on cesarean as compared with usual care (n=32). The device was designed to optimize dental occlusion, based on evidence indicating this may improve isometric muscle strength.¹³³ The observed incidence of cesarean was 12.5 percent in the intervention group as compared with 25.0 percent in patients treated per usual care (no test of statistical significance reported).

Medical Interventions

One small RCT ⁶⁰ found a modest, insignificant effect of a single intravenous 2 mg dose of propranolol at admission for beta-adrenergic blockage and prevention of dysfunctional labor (n=34), with 11 percent of intervention participants having a cesarean compared with 17.6 percent those receiving usual care (p=0.367). Duration of the first stage of labor was similar between the two groups, while second stage duration was significantly longer in the medication group as compared with usual care (median 31 vs. 19 minutes; p<0.001). Neonatal and other outcomes were similar between the two groups (Tables 28 and 29).

Activities

Two RCTs assessed whether simple changes in activities during labor may influence risk of cesarean. A large RCT of walking during the first stage of labor (n=536) found similar use of cesarean as compared with patients who were restricted to bed (n=531), 4 percent and 6 percent respectively (p=0.25).⁷⁵ Investigators noted that 22 percent of women randomized to walk did not elect to walk. Duration of labor and other maternal and infant outcomes were similar between the walking and usual care groups (Tables 28 and 29).

Another large RCT assessed the incidence of cesarean among women encouraged to eat a light diet during labor (n=1,227) as compared with those limited to water and ice consumption (n=1,216). This trial also found similar incidence of cesarean in both groups (30% in each).⁹² In terms of adherence to the intervention, 29 percent of those randomized to the eating group chose not to eat, while 20 percent of those randomized to usual care with restricted intake elected to eat during labor. The overall incidence of vomiting was similar between those randomized to eating as compared with those limited to water consumption; no cases of pulmonary aspiration were observed in either group (Tables 28 and 29).

Overview

The outcomes of systems-levels strategies are summarized in Table 30. Both randomized trials and pre-post study types are included. The indication that cesarean risk was the “same” in a study is based on small effect size with lack of statistical significance. Indication of higher risk means the risk was statistically higher in the intervention portion of trials or at the end of the intervention period than at baseline.

Randomized Clinical Trials

Each of the four trials grouped hospitals in pairs matched for key characteristics. The researchers randomly assigned a member of the pair to implement the strategy while the other member of the pair continued usual practice.^{78, 85, 89, 94} Each trial evaluated a different type of strategy and none demonstrated effectiveness for reducing use of cesarean.

• Nationwide trial of annual audit and feedback in the Netherlands. The intervention consisted of annual reports summarizing each department's cesarean profile in the context of anonymized data from other departments. Several analyses were provided that included graphs and figures to make clear the status of a particular department. The departmental leadership was contacted after receipt of the report to followup and to answer questions. This ensured the reports were reviewed. Over the course of three years during which the reports evolved from only departmental data to include individual provider data in context, the variation in cesarean use within hospital in the intervention decreased but there was no overall difference at the end of the trial between those randomized to receive or not receive reports and followup.⁸⁹
• Implementation of the World Health Organization partogram with four hour action line to guide active management of labor. The trial was conducted in Indonesia, Thailand, and Malaysia. The authors do not report an intention-to-treat analysis of the hospitals as randomized but do reveal the overall change from baseline at intervention hospitals was a 1.7 percent reduction that was not statistically meaningful.⁷⁸
• Requirement for a second opinion of a higher or equal rank physician with application of evidence-based guidelines for each category of indications for cesarean (e.g., elective, breech, failure to progress, emergent, etc.). In the intervention group cesarean was the route for 24.9 percent of births, compared with 24.7 in the hospitals that did not implement the requirements.⁸⁵
• Use of a novel computer system for evaluation of labor progress in a consortium of Canadian and U.S. hospitals. The computerized system featured visual display of labor curves with addition of references ranges (5^th, 50^th, and 95^th percentile norms); 16.9 percent of women who gave birth in hospitals that continued their usual care patterns had cesareans compared to 17.6 in the hospitals using the computerized system.⁹⁴

Management of Abnormal Labor

Two trials of strategies used to manage abnormal labor reported various adverse effects of the intervention.^{54, 90} Bidgood and colleagues reported seven women with hyperstimulation within 15 minutes after infusion, all in the high-dose oxytocin group.⁵⁴ In a trial of early and delayed oxytocin there were no significant differences in maternal fever, postpartum hemorrhage, need for blood transfusion, or major depression.⁹⁰

Pain Management

Two U.S. studies comparing combined spinal-epidural anesthesia and IV meperidine analgesia⁹⁵ and epidural⁹⁷ reported several adverse effects. Maternal fever greater than 38°C was more common in the combined spinal-epidural group (22% vs. 3%; p<0.005).⁹⁵ In addition this study reported that eight infants were delivered by emergency cesarean due to profound fetal bradycardia within one hour of analgesia. Norris and colleagues reported higher rates of accidental dural puncture, intravascular catheterization, failed epidural and positional headache among women who received the combined spinal-epidural.⁹⁷ However, proportion of participants experiencing a blood patch was higher in the epidural group.⁹⁷ Maternal fever did not differ significantly among participants in the Croatian trial of continuous versus intermittent epidural.⁴⁹ Hypotension was more common among women who received the intermittent epidural.⁴⁹

Fetal Assessments

Only one study of fetal pulse oximetry reported adverse effects.⁸⁰ There were no significant differences in the proportion of chorioamnionitis (10.7% in each group), endometritis, or wound complications. The sensor used in the trial resulted in facial marks on 5.8 percent of infants in the open group versus 3.4 percent in the masked group.⁸⁰

Amnioinfusion

Three of the four studies of amnioinfusion reported higher numbers of maternal fever in the amnioinfusion group compared to standard care without amnioinfusion.^{44, 59, 98} However, this was not seen in the study by Choudhary and colleagues.³⁹

Unique Strategies

Four studies of unique strategies to reduce cesarean report adverse effects.^{60, 92, 100, 105} Chorioamnionitis was more common among women who received acupuncture in the study of acupuncture compared to sham acupuncture and no acupuncture (23% vs. 21% and 7%).¹⁰⁵ There were no significant differences in postpartum hemorrhage and uterine atony across the three groups. Adamsons and colleagues compared propranolol during labor to usual care and reported no anesthesia-related morbidity among study participants.⁶⁰ In the study examining the effectiveness of an inflatable obstetrical belt,¹⁰⁰ significantly more women in the control group experienced a 3rd degree perineal tear (6.5% vs. 0.4%). Six women in the obstetrical belt group needed catheter insertion for urinary retention compared to two women in the standard care group.¹⁰⁰ In the study by O'Sullivan and colleagues there was no difference in the percent of women who vomited among those who were allowed to eat in labor versus those only allowed to have ice chips and water.⁹²

Overall the included trials were small and unlikely to detect rate, but potential important events. Those strategies that were directly associates with adverse effects were primarily procedural such as risk of dural puncture with epidural. While this is a known risk of epidural it is not the case that it would be uniquely associated with the particular dosing strategy to be used, for instance continuous versus intermittent dosing. We have considered these sorts of complications as not specific to the intention of the use, and because this is not a review of all the uses of these categories of strategy (social support in labor, fetal monitoring devices, etc), we do not provide estimates per se of these sorts of adverse effects. Overall no adverse effects were identified that were unique or notably exacerbated by use of the intervention for the purpose of attempting to decrease use of cesarean.