Cesarean Delivery Rate in Nulliparous Women in the Second Stage of Labor When Using Zhang Compared With Friedman Labor Curves

A Systematic Review and Meta-analysis

Abstract

OBJECTIVE: 

To estimate the cesarean delivery rate in the second stage of labor in nulliparous women when using the Zhang compared with the Friedman labor curve by performing a systematic review and meta-analysis.

DATA SOURCES: 

We explored MEDLINE, EMBASE, Web of Science, CINAHL (Cumulative Index to Nursing and Allied Health Literature), ClinicalTrials.gov, and Cochrane Central Register of Controlled Trials for studies comparing the cesarean delivery rate in the second stage of labor in nulliparous women when using the Zhang compared with the Friedman labor curve.

METHODS OF STUDY SELECTION: 

Two reviewers separately ascertained studies, obtained data, and gauged study quality. The cesarean delivery rate was compared, and odds ratios (ORs) with 95% CIs were estimated.

TABULATION, INTEGRATION, AND RESULTS: 

Five retrospective cohort studies and two randomized controlled trials were identified with 20,165 nulliparous women (10,861 with the Zhang labor curve vs 9,304 with the Freidman labor curve). Studies evaluated women with delivery from June 2010 to May 2017. Two studies were from the United States; two were from Israel; and one each was from China, France, and Norway. The median point prevalence of cesarean delivery in the second stage was 4.0% (95% CI 2.1–14.5%) in cases managed with the Zhang labor curve compared with 5.6% (95% CI 1.8–28.6%) in cases managed with the Friedman labor curves. The overall cesarean delivery rate in the second stage of labor was similar when the Zhang labor curve compared with the Friedman labor curve was used (pooled OR 0.86, 95% CI 0.47–1.57, I²=93%). When the two randomized controlled trials were excluded, the results of the cohort studies continued to demonstrate that the overall cesarean delivery rate in the second stage of labor was similar (pooled OR 0.84, 95% CI 0.42–1.69, I²=92%).

CONCLUSION: 

Nulliparous women have similar cesarean delivery rates in the second stage of labor when either the Zhang or Friedman labor curve is used.

SYSTEMATIC REVIEW REGISTRATION: 

PROSPERO, CRD42022346425.

In Brief

Nulliparous women have similar cesarean delivery rates in the second stage of labor when using the Zhang compared with the Friedman labor curve.

In 2014, guidance that defined the normal progression of labor in the United States was updated to recommend that the second stage of labor in nulliparous women be extended to 3 hours for individuals without an epidural and 4 hours for those with an epidural to facilitate achievement of vaginal delivery.¹ This update was based on a new analysis of labor curves from a large contemporary cohort (the Consortium of Safe Labor) of pregnant individuals in the United States by Zhang et al² suggesting that more time in labor be allowed before labor dystocia is diagnosed, which may reduce the rate of intrapartum cesarean deliveries. For 60 years before this guidance change, expert opinion based on historical data by Emanuel Friedman suggested that most nulliparous women entering the second stage of labor should deliver within 2 hours for those without epidural analgesia and within 3 hours for those with an epidural.^3,4

Despite this recommended guidance change, the influence on the second stage cesarean delivery rate in the nulliparous individual has been mixed. Some trials have shown decreased rates^5,6; other trials have shown no effect⁷; and some trials have shown an increased rate of cesarean delivery in nulliparous women.⁸ Similarly, the proportions of maternal and neonatal adverse outcomes varied when the Zhang labor curve was implemented, with some studies observing higher rates and others noting no differences.^7,8

With each hour of increase in the second-stage duration of labor, the proportion of women achieving a spontaneous vaginal birth without serious maternal or neonatal morbidity steadily decreases.⁹ Therefore, the optimal labor curve for the second stage of labor in the nulliparous woman that balances the greatest chance of vaginal delivery with the lowest maternal and neonatal morbidity is unclear. Therefore, our primary objective is to estimate the rate of cesarean delivery in the second stage of labor in nulliparous women when using the Zhang compared with the Friedman labor curve by performing a systematic review and meta-analysis. Our secondary objective is to compare maternal and neonatal adverse outcomes when using the Zhang compared with the Friedman labor curve.

SOURCES

This review follows MOOSE (Meta-analysis of Observation Studies in Epidemiology) guidelines.¹⁰ The protocol was recorded with PROSPERO (registration number CRD42022346425) on July 24, 2022, before introductory searches were started. The study was judged exempt from Baylor College of Medicine Institutional Review Board endorsement because it was deemed non–human subjects research. An electronic inquiry with no language restrictions using the databases of MEDLINE through the PubMed interface, Web of Science, EMBASE (Excerpta Medica Database), CINAHL (Cumulative Index to Nursing and Allied Health Literature), ClinicalTrials.gov, and Cochrane Central Register of Controlled Trials was conducted to identify articles that reported the cesarean delivery rate in nulliparous pregnant women in the second stage of labor when using the Friedman compared with the Zhang labor curve. The primary investigation was performed from database origin through December 2022 by one author (M.A.T.) with assistance from a medical librarian in systematic reviews. Unpublished composed articles (ie, gray literature) were assessed by exploring Google Scholar and the ProQuest Dissertations & Theses database. For Google Scholar searches, the initial 100 results were checked on the supposition that the most germane results would emerge first.¹¹ Articles were detected with a combination of the following MeSH terms: “obstetric care consensus” OR “prolonged second stage of labor” AND “cesarean delivery” (see Appendix 1, available online at https://links.lww.com/AOG/D127). References from recovered articles were also reviewed for supplementary articles. The final form of the MeSH terms was used by a second author (M.M.L.) separately, and no added articles were ascertained.

STUDY SELECTION

All identified articles were fully recovered and evaluated separately by two reviewers (M.A.T. and M.M.L.) for inclusion. Studies qualified for inclusion if they assessed nulliparous pregnant women in the second stage of labor who were evaluated with the Freidman labor curve and were compared with a group of nulliparous pregnant women in the second stage of labor who were evaluated with the Zhang labor curve. Data were accepted from randomized controlled trials, cohort, case-control, and cross-sectional studies. Abstracts presented at meetings were excluded because of incomplete data being available for assessment.

Two reviewers (M.M.L. and S.C.S.) independently rated the quality of the included articles. Discrepancies were resolved through discussion with the third author (M.A.T.). An intraclass correlation coefficient was calculated to estimate interrater reliability.¹² For randomized controlled trials, the risk of bias for each study was assessed with the Cochrane Risk of Bias tool.¹³ The assessment focused on seven different aspects of each study: 1) the randomization sequence generation, 2) allocation concealment, 3) reporting bias, 4) other bias, 5) performance bias, 6) detection bias, and 7) attrition bias. Each aspect was categorized as having a high, low, or unclear risk of bias. The overall qualities of nonrandomized studies were rated with a modified Newcastle–Ottawa Scale (NOS) for cohort studies.¹⁴ These were rated as follows: low risk of bias (NOS score 7–9), high risk of bias (NOS score 4–6), and a very high risk of bias (NOS score 0–3).¹⁵

The primary outcome is the primary cesarean delivery rate in nulliparous women in the second stage of labor, as reported by the authors of each study in the published article. Descriptive variables to be compared (if available) between groups were age in years, body mass index (BMI, calculated as weight in kilograms divided by height in meters squared), gestational age at delivery in weeks, and rate of epidural use. Planned secondary outcomes (dependent on outcomes available in identified studies) were the duration of the second stage of labor, rate of operative vaginal delivery, rate of third- or fourth-degree vaginal lacerations, rate of postpartum hemorrhage, proportion receiving maternal blood transfusions, rate of chorioamnionitis, rate of neonatal shoulder dystocia, 5-minute Apgar score less than 7, and neonatal admission to the intensive care unit. The means and SDs were estimated from this information for studies that reported any continuous variable as the median and range.¹⁶ Although subgroup analysis of stratifying institutions by the rate of primary cesarean delivery in the second stage of labor was not planned in the originally registered protocol, analysis of studies was noted to have different pooled effects if an institution's primary cesarean delivery rate for the second stage of labor using the Friedman labor curve was in the upper quartile. Therefore, for a potential explanation of the degree of heterogeneity noted, studies with a primary cesarean delivery rate in the second stage of labor greater than the upper quartile were compared with studies that were lower than the upper quartile.

Two-by-two tables were generated for ordinal data, and odds ratios (ORs) with their 95% CIs were computed. Pooling of results was performed with the Der Simonian–Laird random-effects model.¹⁷ A random-effects model was applied because of the probable heterogeneity of included studies. Mean differences or ORs for nulliparous pregnant women in the second stage of labor when using the Friedman labor curve as the reference group were pooled separately from nulliparous pregnant women in the second stage of labor when using the Zhang labor curve, and a forest plot was constructed.

Sensitivity analyses (omitting trials to be judged a high risk of bias) and influence analysis (removing each study sequentially from the pooled estimates) were executed. For ORs, an α level of 0.05 was considered significant. Statistical heterogeneity was calculated with the Cochrane Q test and Higgins I² statistic.¹⁸ Statistically significant heterogeneity was considered if the Cochran Q was P<.1 or Higgins I²>30% to address the low power of heterogeneity tests.¹⁸ Publication bias was assessed graphically by visual inspection of funnel plots and by the Peters or Eggers test as indicated.^19,20 The meta-analysis was performed with RevMan 5.4.

RESULTS

A flow diagram of the study collection method is shown in Figure 1. A total of 779 related studies were initially identified. After studies not relevant to the subject of interest and meeting abstracts were eliminated, nine were retrieved for complete analysis.^{5,6,8,21–26} Two studies were excluded: one for including only pregnant women who had a cesarean delivery²¹ and one because of the inability to determine from published information the number of cesarean deliveries performed in the second stage of labor.²² The corresponding author for the latter study was contacted and indicated that they could not provide the additional information requested.²² If not available in the original publication, additional patient-level data were requested from included studies. Five corresponding authors from included studies were contacted, with all providing additional supplementary data for inclusion in either the primary or secondary analyses.^8,23–26 In total, seven studies with 20,165 nulliparous pregnant women in the second stage of labor were included in the review: 10,861 when using the Zhang labor curve and 9,304 when using the Friedman labor curve.^{5,6,8,23–26}

Fig. 1.: Flow diagram of studies identified in the systematic review. CINAHL, Cumulative Index to Nursing and Allied Health Literature; CENTRAL, Cochrane Central Register of Controlled Trials.

Study characteristics and results were summarized and verified for accuracy independently by two authors. Features of the included studies are shown in Table 1. Of the seven studies included, six were from single institutions, and one was a cluster randomized controlled trial with 14 separate obstetric units. Five were retrospective cohort studies,^6,8,23,24,26 and two were randomized controlled trials.^5,25 Studies evaluated pregnant women who underwent delivery from June 2010 to May 2017. All studies included singleton pregnancies at 36 weeks or more of gestation. Two studies were from the United States; two were from Israel; and one each was from China, France, and Norway. The median (range) of the number of pregnant women in each group was 1,368 (43–3,627) when using the Zhang labor curve compared with 1,267 (39–3,141) when using the Friedman labor curve. Three studies reported maternal age to allow a pooled analysis.^5,6,24 Four studies provided the BMI and estimated gestational age at delivery of nulliparous pregnant women between groups.^5,6,24,25 Age was similar between individuals when using the Zhang labor curve compared with the Friedman labor curve (29.8±2.0 years vs 29.8±2.0 years, respectively), a mean difference of 0.05 years (95% CI −0.31 to 0.40, I²=0%). Body mass index was similar between individuals when using the Zhang labor curve compared with the Friedman labor curve (24.1±3.8 vs 24.1±2.6, respectively), with a mean difference of 0.39 (95% CI −0.53 to 1.31, I²=91%). Gestational age at delivery was similar between individuals when using the Zhang labor curve compared with the Friedman labor curve (40.1±0.4 weeks vs 40.1±0.3 weeks, respectively), a mean difference of −0.03 weeks (95% CI −0.13 to 0.07, I²=44%). All seven studies reported the rate of epidural use in women.^{5,6,8,23–26} The median point prevalence rates of epidural use were 81.5% (interquartile range 55.8–100%) in women when using the Zhang labor curve and 78.2% (interquartile range 55.3–100%) in women when using the Friedman labor curve.

Table 1.: Study Characteristics of Nulliparous Pregnant Individuals When Using the Zhang Compared With the Friedman Labor Curve in the Second Stage of Labor

^{The five cohort studies were rated as having a low risk of bias (mean NOS score: Jing and Xian23 8, Wilson-Leedy et al6 9, Thuillier et al24 9, Zipori et al26 9, and Kadour-Peero et al8 9). The intraclass correlation coefficient comparing total NOS scores between the two reviewers for the quality of included studies noted excellent agreement (r=0.75). The two reviewers identically rated the overall risk of bias for the randomized controlled trials.5,25 The study by Bernitz et al25 rated all domains of bias as low. For the trial by Gimovsky and Berghella,5 all bias domains were rated as low apart from the two. Both reviewers gave a high risk of bias for the domains of performance and detection bias. This bias was attributable to the study design in that participants (both patient and obstetric health care professional) could not be blinded to the intervention group to which women were randomized.The five cohort studies were rated as having a low risk of bias (mean NOS score: Jing and Xian23 8, Wilson-Leedy et al6 9, Thuillier et al24 9, Zipori et al26 9, and Kadour-Peero et al8 9). The intraclass correlation coefficient comparing total NOS scores between the two reviewers for the quality of included studies noted excellent agreement (r=0.75). The two reviewers identically rated the overall risk of bias for the randomized controlled trials.5,25 The study by Bernitz et al25 rated all domains of bias as low. For the trial by Gimovsky and Berghella,5 all bias domains were rated as low apart from the two. Both reviewers gave a high risk of bias for the domains of performance and detection bias. This bias was attributable to the study design in that participants (both patient and obstetric health care professional) could not be blinded to the intervention group to which women were randomized.}

All seven studies reported the cesarean delivery rate in the second stage of labor.^{5,6,8,23–26} The median point prevalence of cesarean delivery was 4.0% (95% CI 2.1–14.5%) in women when using the Zhang labor curve compared with 5.6% (95% CI 1.8–28.6%) with the Friedman labor curve. The overall cesarean delivery rate in the second stage of labor in women when using the Zhang labor curve was similar to that of women when using the Friedman labor curve (pooled OR 0.86, 95% CI 0.47–1.57, I²=93%) (Fig. 2). Visual inspection of the funnel plot (Fig. 3) suggested no evidence of publication bias for studies on the overall cesarean delivery rate in the second stage of labor in women when using the Zhang compared with the Friedman labor curve (Peter test P=.74). When the two randomized controlled trials were excluded, the results continued to demonstrate that the cesarean delivery rate in the second stage of labor in women when using the Zhang labor curve was similar to that for women managed with the Friedman labor curve (pooled OR 0.84, 95% CI 0.42–1.69, I²=92%). An influence analysis of the seven included trials (removing each study in turn from the pooled estimates) was performed. The overall pooled results were unchanged. In addition, a subgroup analysis of studies was undertaken to compare institutional primary cesarean delivery rates in the second stage of labor using the Friedman labor curve in the upper quartile (9.9% or greater) compared with those lower than the upper quartile. Two institutions reported rates of 9.9% or more (13.7–43.2%).^5,26 The overall cesarean delivery rate in the second stage of labor in women when using the Zhang labor curve was reduced to that of women using the Friedman labor curve (pooled OR 0.48, 95% CI 0.41–0.56, I²=0%, 6,650 participants) in this sensitivity analysis. Five institutions had rates less than 9.9% (range 1.5–6.1%).^6,8,23–25 The overall cesarean delivery rate in the second stage of labor in women when using the Zhang labor curve was similar to that of women when using the Friedman labor curve (pooled OR 1.32, 95% CI 0.94–1.86, I²=51%, 13,515 participants).

Fig. 2.: A. Forest plot of the overall cesarean delivery rate in the second stage of labor in nulliparous women's cases managed with the Zhang vs the Friedman labor curve. B. Forest plot of the cesarean delivery rate in the second stage of labor in nulliparous women's cases managed with the Zhang vs the Friedman labor curve in the randomized controlled trials. C. Forest plot of the cesarean delivery rate in the second stage of labor in nulliparous women's cases managed with the Zhang vs the Friedman labor curve in the retrospective cohort studies. M-H, Mantel Haenszel; df, degrees of freedom.

Fig. 3.: Funnel plot of the overall cesarean delivery rate in the second stage of labor in nulliparous women's cases managed with the Zhang vs the Friedman labor curve.

Two studies reported the second-stage labor duration.^24,26 The duration of the second stage of labor was similar between individuals when using the Zhang labor curve and individuals using the Friedman labor curve (3.0±1.7 hours vs 2.4±0.6 hours, respectively), with a mean difference of 0.57 hours (95% CI −0.09 to 1.23, I²=99%). Table 2 lists the pooled results of the preplanned analysis of secondary maternal and neonatal adverse outcomes in women when using the Zhang compared with the Friedman labor curve. Only one study reported neonatal intensive care unit admission, and pooled analysis could not be performed for this outcome.⁵ Although most adverse outcomes noted increased absolute rates associated with women when using the Zhang compared with the Friedman labor curve, overall pooled secondary outcome results were similar between groups. Removing the randomized controlled trials from the pooled analyses did not change the comparison of the outcomes between the two groups for any of the secondary maternal and neonatal adverse outcomes (data not shown).

Table 2.: Secondary Outcomes for the Zhang Compared With the Friedman Labor Curve in Nulliparous Women in the Second Stage of Labor

DISCUSSION

This meta-analysis demonstrates a similar rate of cesarean delivery in nulliparous pregnant women when using a contemporary (Zhang) labor curve in the second stage of labor compared with a historical approach (Friedman labor curve). Furthermore, the overall rates of adverse maternal and neonatal outcomes were comparable in pregnant women with either labor curve.

Low-risk births are defined by the Centers for Disease Control and Prevention as those occurring in nulliparous women at 37 or more weeks of completed gestation with a singleton pregnancy and a vertex presentation. The total low-risk cesarean delivery rate in the United States has decreased by nearly 8% from 28.1% in 2009 (before Zhang labor curve publication) to the most current reported rate of 25.9% in 2020.^27,28 However, outcomes from birth certificate data do not allow the determination of the rate of cesarean delivery performed in the second stage of labor. Institutions from various parts of the world have shown mixed results on the effect on the total cesarean delivery rate in nulliparous women after implementing the Zhang labor curve. Trials have shown either a reduced total cesarean delivery rate^6,22–24,26 or no difference.^7,25

Two randomized controlled trials have evaluated the effect of implementing contemporary compared with historical labor curves on the cesarean delivery rate in nulliparous women.^5,25 Bernitz et al²⁵ performed a multicenter cluster randomized trial comparing total cesarean delivery rates in nulliparous women when using the World Health Organization partograph (based on Friedman’s historical data) or the Zhang labor curve. The total cesarean delivery rates did not differ between those for whom the Zhang labor curve was used (6.8%) compared with those for whom the World Health Organization partograph was used (5.9%, adjusted relative risk 1.17, 95% CI 0.98–1.40). The authors noted that the intrapartum cesarean delivery rate was reduced in both groups, which they speculated could have been caused by a Hawthorne effect. The lower rate of cesarean delivery noted in the control group than that used for sample size calculations (9.2%) may have contributed to a potential type II error in the primary outcome. Gimovsky and Berghella⁵ carried out a single-center randomized controlled trial comparing nulliparous women who reached the second stage of labor and whose labor course was managed by extending the time in the second stage (ie, Zhang labor curve) compared with a usual labor curve (ie, Friedman labor curve). The total cesarean delivery rate was reduced in women when using the Zhang labor curve (19.5%) compared with the Friedman labor curve (43.2%, relative risk 0.45, 95% CI 0.22–0.93). However, because of the inability to blind obstetric care professionals and patients to the randomized intervention, performance or detection bias could have unintentionally affected outcomes. The current meta-analysis noted similar cesarean delivery rates in nulliparous women in the second stage of labor when using the Zhang compared with the Friedman labor curve. This similar rate was observed in the pooled analysis even when the randomized controlled trials were considered separately from the cohort studies. However, the pooled rates between groups in the cesarean delivery rate contained a high heterogeneity level, posing interpretive challenges. The high rate of heterogeneity noted in Figures 2 and 3 suggests that in the presence of large sample sizes, individual studies may be providing a highly accurate estimate of their population effect size, which results in statistically significant differences in both directions. Thus, most of the observed variability in the pooled analysis may reflect differences in these population effect sizes and possible true heterogeneity. Suggestions of heterogeneity may be founded on data or study design, including differences in target populations, obstetric practices in different regions of the world, the timing of interventions, or outcome measurements.

We focused our analysis on the second stage of labor for two reasons. First, compliance with study protocols of labor management can be challenging.²⁹ The attention and focus on assessing labor progression may influence the cesarean delivery rate.²⁵ In comparison, once a pregnant woman reaches 10 cm of cervical dilation, entry into the second stage of labor can be easily identified. Second, studies have shown increased rates of adverse maternal and neonatal outcomes the longer the duration of the second stage of labor.^9,30 Grobman et al³⁰ evaluated more than 26,000 nulliparous women who reached the second stage of labor as part of the Eunice Kennedy Shriver National Institute of Child Health and Human Development Maternal-Fetal Medicine Units Network Assessment of Perinatal Excellence study. The risk of postpartum hemorrhage, third- or fourth-degree lacerations, and the composite adverse neonatal outcome increased in frequency with a longer duration of active pushing. In a retrospective analysis of the Consortium of Safe Labor data (ie, Zhang labor curve) on the duration of the second stage of labor in more than 43,000 nulliparous women,⁹ vaginal delivery rates diminished with increasing second-stage duration. In the Consortium for Safe Labor analysis, each of the individual morbidity rates tended to increase for nulliparous women without an epidural. In the current meta-analysis, although increased absolute maternal and neonatal adverse outcomes rates were noted in women when using the Zhang compared with the Friedman labor curve, studies did not provide rates of adverse events stratified by the duration of the second stage of labor. From the trial data available, it is unknown what proportion of pregnant individuals with either labor management strategy exceeded the 95th percentile of the duration of the second stage of labor. Despite having more than 20,000 nulliparous pregnant women in the current analysis, because of the rarity of adverse events, the sample sizes for the various secondary outcomes may not have been large enough to detect meaningful clinical differences. Furthermore, secondary outcomes should be evaluated in the context of hypothesis generation, not clinical management.

This study has several strengths, including a prospectively registered protocol, a comprehensive search stratagem with no language restrictions that included populations from different parts of the world, individual patient-level data obtained from study authors, and studies of contemporary labor practices for management of the second stage in nulliparous women. However, the limitations of this study regarding the interpretation for clinical practice should be noted. Most studies were from single institutions, and external validity may be narrow. Because most of the studies were retrospectively based on administrative data sets, determining the actual duration of the second stage of labor (eg, delayed compared with active pushing or the decision to incision time for cesarean delivery) can be imprecise. As with all meta-analyses, the utility of the conclusion is conditional on the quality of the primary studies included. Unidentified or unmeasured effects not systematically reported could have altered the identified associations.

In summary, this meta-analysis notes similar rates of cesarean delivery with comparable rates of adverse maternal and neonatal outcomes in nulliparous women when using contemporary compared with historical labor curves. Although the considerable degree of unexplained heterogeneity may limit these meta-analysis results, this demonstrates the difficulty of comparing clinical diversity in labor management in geographic regions around the globe. Further research will need to be conducted to determine what rate of cesarean delivery in nulliparous patients in the second stage of labor is clinically meaningful (eg, optimizes maternal and neonatal outcomes) to determine the best labor management approach.

For any pregnant patient, weighing risks and benefits of different labor curves will have to balance an assortment of clinical factors such as estimated fetal weight, maternal or fetal status evaluation, and the degree of continued fetal descent in the pelvis with pushing. In addition, a specific fixed quantity of time spent in the second stage of labor outside of which all women should undergo operative delivery has not been established.¹ Obstetric care professionals and pregnant individuals will need to engage in shared decision making when weighing the prospect of vaginal delivery against the risk of serious maternal and neonatal morbidity when considering continuing the second stage of labor.