Optimizing Management of the 2nd Stage of Labor: Multicenter Randomized Trial

A. SPECIFIC AIMS The central hypothesis is that immediate pushing in the second stage of labor increases spontaneous vaginal delivery, shortens duration of the second stage, and reduces adverse neonatal and maternal outcomes in nulliparous women. They will pursue the following specific aims: 1\) Primary Aim: Assess the effectiveness of immediate pushing at complete cervical dilation on the rate of spontaneous vaginal delivery in nulliparous women. Hypothesis: Nulliparous women will have an increase in spontaneous vaginal delivery rates with immediate, as compared to delayed, pushing. 2\) Secondary Aims: i.Determine the effect of immediate pushing on the rate of neonatal composite morbidity. Hypothesis: Immediate pushing will reduce the rate of neonatal composite morbidity, defined as one or more of: neonatal death, major neonatal injury, umbilical cord arterial acidosis, suspected neonatal sepsis, respiratory distress, transient tachypnea, meconium aspiration with pulmonary hypertension, hypoxic-ischemic encephalopathy, hypoglycemia, and need for hypothermia treatment. ii.Determine the impact of immediate versus delayed pushing on objective and subjective measures of maternal pelvic floor morbidity. Hypothesis: Immediate pushing will be associated with a reduction in the rate of acute levator ani and anal sphincter muscle injury, and lower rates of pelvic floor disorders identified on subjective and objective clinical assessments. B. STUDY DESIGN B1. Design Overview This is a multicenter randomized trial of nulliparous women to compare the effectiveness of immediate and delayed pushing in the second stage of labor. The investigators aim to test the central hypothesis that immediate pushing in the second stage of labor increases spontaneous vaginal delivery, shortens duration of the second stage, and reduces adverse neonatal and maternal outcomes. The investigators chose a randomized controlled trial, the 'gold standard' of clinical research design, with the goal of obtaining the highest quality evidence to inform clinical practice. Randomly allocating subjects to different interventions minimizes selection bias and results in groups that are comparable with regards to important confounding variables, both measured and unmeasured. Additionally, the broad inclusion criteria, simplicity in relationship of the interventions to current practice, and the multicenter design with regional representation and practice diversity increase generalizability and direct application of the findings. The investigators will follow the Consolidated Standards of Reporting Trials (CONSORT) guidelines wherever appropriate in the conduct and reporting of this trial. They will use computer-generated random sequences, stratified by study site, to assign participants to the two interventions. Labor management will be similar in the two groups except for the timing of maternal pushing effort. Analysis will follow the intention-to-treat principle. The use of broad inclusion criteria and intention-to-treat analysis will allow a more conservative estimate of differences in outcomes between the two strategies and allow a better estimate of effectiveness and public health implications of practice change than would pure estimate of efficacy alone. B2. Study Sites and Populations 1. Washington University Medical Center - Coordinating center (PI/PD-Cahill/Tuuli): The obstetrical service at WUMC delivers babies for approximately 4000 women annually. WUMC is a tertiary care hospital and draws all ethnic groups typical of a large urban referral hospital. Sub site: Missouri Baptist Medical Center is a private hospital affiliated with WUSM. The obstetric service delivers 4000 women annually, predominantly by private physicians. 2. University of Alabama at Birmingham (Site PI-Tita): UAB Hospital is the major perinatal referral facility for Alabama. Approximately 4200 women (53% African American, 25% Caucasian, 20% Hispanic, and 2% other ethnic groups) are delivered annually in the Women and Infants' Center. 3. Hospital of the University of Pennsylvania (Site PI-Srinivas): The obstetrical service at Penn delivers approximately 4100 women per year. Penn is an urban tertiary care hospital with a diverse population. Patients are managed by faculty of the Penn School of Medicine. Sub site: Pennsylvania Hospital is a close affiliate of Penn. The obstetric unit is staffed by both private obstetricians and midwives. Of the 4600 - 5000 deliveries performed annually, 4000 - 4500 are performed by private physicians and 300 - 600 by midwives. 4. Oregon Health \& Science University (Site PI-Caughey): The Labor and Delivery Unit at OHSU delivers approximately 2700 women per year. OHSU is a tertiary care hospital and the only academic medical center in Oregon. B3. Inclusion and exclusion criteria The investigators will use broad inclusion criteria to ensure generalizability of our results. Exclusion criteria will be limited to conditions for which vaginal delivery is contraindicated or expedited delivery is required. B4. Outcome Measures B4-1. Primary aim Primary outcome measure: The primary outcome is spontaneous vaginal delivery, defined as delivery that occurs without the use of forceps, vacuum, or cesarean, as the primary outcome measure. This is because spontaneous vaginal delivery is the most desirable outcome for laboring women in the second stage of labor. Compared to cesarean delivery, spontaneous vaginal delivery is associated with lower maternal and neonatal morbidity, including hemorrhage, postoperative wound infection, endometritis, and prolonged hospitalization. Further, once a cesarean is performed, the risk of delivery by cesarean for future pregnancies is dramatically elevated, further increasing maternal morbidity. Finally, although operative vaginal delivery is associated with overall lower maternal morbidity than cesarean delivery, it carries higher risks of neonatal and maternal pelvic floor injury than spontaneous vaginal delivery. Secondary outcome measures: Operative vaginal delivery (forceps or vacuum), cesarean delivery, total duration of second stage, duration of active pushing, postpartum hemorrhage, endometritis. B4-2. Secondary aim#1 Outcome measure: The outcome measure for this aim will be neonatal composite morbidity, defined as occurrence of any of the morbidities: * Neonatal death * Major neonatal injury * Acidemia * Respiratory distress * Transient tachypnea * Meconium aspiration with pulmonary hypertension * Hypoxic-ischemic encephalopathy * Hypoglycemia * Suspected sepsis B4-3. Secondary aim#2 Outcome measures 1. Rate and extent of acute levator ani muscle (LAM) injury in nulliparous women. i.Overt LAM injury, defined as clinically diagnosed 2nd, 3rd, and 4th degree perineal lacerations. ii.Occult LAM injury, defined as interruption in LAM on 3D transperineal ultrasound) injury in nulliparous women. 2. Rates and extent of signs of POP on objective clinical examination at the 4 - 8 weeks, and 5 - 7 months postpartum follow-up visits using the validated Pelvic Organ Prolapse Quantification (POP-Q) system. 3. Rates of patient-reported symptoms of urinary incontinence (UI), fecal incontinence (FI), and pelvic organ prolapse (POP) on validated quality-of-life questionnaires at 1 - 5 days, 4 - 8 weeks, and 5 - 7 months postpartum. C. STUDY PROCEDURES C1. Recruitment The investigators will employ familiar efficient recruitment techniques we have used in recent randomized trials in the same settings. All women admitted to the labor and delivery units of the participating medical centers will be screened against inclusion and exclusion criteria. Eligible subjects will be approached for written consent to participate in the study once they are committed to a vaginal delivery attempt. Although consent to participate will be obtained irrespective of cervical dilation, randomization will be performed only when complete cervical dilation is confirmed. This is necessary to avoid situations in which subjects are randomized and subsequently undergo cesarean delivery in the first stage of labor for indications such as failure to progress or non-reassuring fetal heart tracing. C2. Randomization Enrolled subjects will be randomly assigned in a 1:1 ratio to immediate or delayed pushing. A web-based randomization sequence will be prepared using blocks of variable sizes, stratified by study site, and maintained centrally by the study statistician. The advantage of this method is that it provides a good probability of balance, and future assignments are unpredictable. In addition, it allows an explicit randomization analysis to be conducted with relative ease. Stratification by center assures balance between the two treatment groups within each center to account for possible differences in patient management. A subject's group assignment will be obtained only after the subject is confirmed to continue to meet inclusion criteria, and a study number and 10-cm cervical dilation are entered and locked in a secure web site upon complete cervical dilation. C3. Blinding Although blinding of both subjects and physicians would be ideal, blinding is clearly not possible in this trial. We will minimize systematic bias by applying the same standard procedures for managing labor and delivery in all patients. Further, the group assignment of subjects will not be taken into account by research personnel collecting neonatal and maternal outcomes and maternal pelvic floor morbidity. Importantly, the primary outcome of spontaneous vaginal delivery is an objective measure. C4. Interventions Interventions to be compared in this trial are the two most common approaches to initiating maternal pushing in the second stage of labor: 1. Immediate pushing: Women in this group will be instructed to initiate pushing as soon as complete cervical dilation is documented. Maternal position, technique (closed or opened glottis), and duration and frequency of maternal pushing effort will be at the discretion of each subject's nurse or physician. 2. Delayed pushing: Women assigned to the delayed pushing group will be instructed to wait for 60 minutes from complete cervical dilation before pushing. Women will be allowed to push earlier if they feel an irresistible urge to push or there is a clinical indication to initiate pushing. We chose 60 minutes as the delay time because our meta-analysis suggested that the duration of delay (ranging from 60 to 180 minutes) did not modify the effect of delayed pushing on spontaneous vaginal delivery rates. Additionally, a 60-minute delay will enable us to assess the effect of delayed pushing without unduly increasing total duration of the second stage. This is important because, although there is no established threshold, a number of large cohort studies suggest incremental increase in the risks of operative delivery, maternal infection, hemorrhage, and laceration, as well as neonatal Apgar score depression, worsening cord pH, and neonatal morbidity, with each additional hour in the second stage, regardless of whether patients pushed immediately or delayed pushing. Finally, patients indicated in our survey that they would prefer to wait no longer than 60 minutes after complete cervical dilation before pushing if they were randomized to the delayed pushing group. C5. Pelvic Floor Assessment These assessments will be overseen by Co-I Dr. Richter, a Urogynecologist and site the UAB PI of the NIH-funded Pelvic Floor Disorders Research Network. Patients willing to participate in the pelvic floor assessment part of the study will be consented. C5-1. Postpartum three-dimensional transperineal ultrasound imaging Traditionally, MRI has been used to detect LAM injury, but investigators have recently validated the use of 3D ultrasound to assess acute LAM injury. We will use ultrasound because it is a much more readily available and inexpensive tool than MRI. Subjects enrolled at WUMC, UAB, OHSU and PENN will be asked to consent for pelvic floor assessment including postpartum 3D transperineal ultrasound examination. Examinations will follow methods developed and validated by Dietz and Simpson. Dr. Shobeiri will undertake all image processing and interpretation. Drs. Shobeiri and Lockhart, as well as other trained personnel, will be blinded to the group assignments of the subjects. C5-2. Assessment of pelvic floor morbidity The investigators will conduct a comprehensive assessment of the pelvic floor to include all components outlined in the proceedings of the NIH Terminology Workshop for Research in Pelvic Floor Disorders. These assessments will be overseen by Co-I Dr. Richter, a Urogynecologist and site PI of the NIH-funded Pelvic Floor Disorders Research Network with extensive experience in the use of these instruments for pelvic floor research. 1. POP-Q staging of pelvic organ prolapse (4 - 8 weeks, and 5 - 7 months postpartum): Objective assessment of pelvic organ prolapse will be based on the POP-Q system, an objective, site-specific system for describing, quantifying, and staging pelvic support in women. We will follow standard guidelines established by the International Continence Society. Pelvic examinations will be performed in the dorsal lithotomy position with the subject straining maximally. 2. Symptom and Quality-of-Life Questionnaires (1-5 days, 4 - 8 weeks, and 5 - 7 months postpartum): We will use four validated Symptom and Quality-of-Life Questionnaires to assess subjective symptoms of pelvic floor morbidity: Fecal Incontinence Severity Index (FISI), Modified Manchester Questionnaire, Pelvic Floor Distress Inventory-20, Pelvic Floor Impact Questionnaire-7 (PFIQ-7). D. DATA MANAGEMENT D1. Data Collection The investigators will collect extensive antepartum, intrapartum, and postpartum information from participants and their infants. Data will be collected on standardized forms on which nearly all responses will be precoded. D2. Data Management The study statistician (Dr. Liu) and the data manager (Mr. Colvin) at WUSM will be responsible for coordinating the overall data management. Data will be collected and managed with REDcap (Research Electronic Data Capture), an established, secure, web-based data capture and management tool developed at Vanderbilt University and supported by the Division of Biostatistics at WUSM (http://www.biostat.wustl.edu/redcap/). The database is backed up periodically throughout each day and is backed up offsite nightly. assessment. Data cleaning will be performed prior to quarterly reports. E. STATISTICAL CONSIDERATIONS E1. Sample size and Power E1-1. Total sample size for the trial: The sample size for the trial is estimated based on the primary outcome. The investigators then estimate, on the basis of the sample size for the primary aim, the power they will have to detect clinically significant differences in the secondary outcomes. All sample size and power estimates are based on two-tailed tests. This is important because we will be powered to detect both increases and decreases in outcomes with immediate compared to delayed pushing. The baseline spontaneous vaginal delivery rate with delayed pushing used for the sample size estimation is based on our observational cohort. We estimate that a total of 3184 (1592 delayed and 1592 immediate pushing) will be sufficient to detect a 5% absolute difference (estimated 72% versus 77%) in spontaneous vaginal delivery with 90% power (alpha of 0.05). E1-2. Estimated power for secondary aim #1: The sample size of 3184 for the primary outcome will be sufficient to detect a 2% absolute difference in neonatal composite morbidity between the two groups with \>80% power and significance level of 5%. This represents the difference between a 6% neonatal composite morbidity rate with delayed (based on our observational cohort study) and 4% with immediate pushing. E1-3. Estimated power for secondary aim #2: We anticipate that at least 630 will consent for and complete the pelvic floor aspects of the study. This will provide adequate power for the key outcomes under secondary aim #2. The rate of POP (defined as any POP at or below the hymen) or FI (based on FISI) in nulliparous women at 6 - 12 months postpartum from prior studies were 20.2% \[38\] and 12.6%\[38, 39\], respectively. Using these as baselines, the 630 women will provide \>80% power to detect a 40% difference in rates pelvic organ prolapse or fecal incontinence between the two groups. For LAM injury on postpartum 3D transperineal ultrasound examination, a prior study found an average rate of 38% \[40\]. Using this baseline rate, 630 women will provide \>90% power to detect at least a 30% difference in occult LAM injury. For overt anal sphincter injury, the sample size will be sufficient to detect a 40% difference in 2nd, 3rd, or 4th perineal lacerations (anticipated 27% to 16.2% with \>80% power). E2. Interim Analyses The investigators anticipate two interim analyses after 50% and 75% of the sample size are recruited, but the exact timing will be at the discretion of the DSMB. Analyses will be performed by the study statistician and presented to the DSMB, which will make recommendations regarding further conduct of the trial. At their first meeting, the DSMB will establish thresholds and rules for trial stoppage based on safety and efficacy limits. Although early stopping decisions cannot be based purely on a mathematical stopping rule, the Haybittle-Peto stopping rule will be used as a guide. Under this rule, the interim analyses of the primary outcome would have to demonstrate an extreme difference between groups (p \<0.001) to justify premature disclosure. This rule has the advantages that the exact number and timing of interim analyses need not be specified in advance and the overall type I error is preserved at 0.05. Therefore, samples size adjustments are not needed. E3. Data Analysis Plan E3-1. Overview Data analyses will adhere closely to the CONSORT guidelines. Analyses will follow the intention-to-treat principle in which subjects will be analyzed in the group to which they were randomized, regardless of whether or not they received the assigned intervention. E3-2. Primary Analysis Descriptive statistics will characterize the group of individuals recruited and investigate comparability of the two groups at baseline. Formal statistical testing will be limited to selected baseline characteristics considered to be prognostic factors for the primary outcome, such as fetal station at complete cervical dilation, birth weight, and duration of the first stage of labor. The categorical prognostic factors will be compared between trial groups by using the Chi-squared or Fisher's exact tests as appropriate. Distributions of continuous prognostic factors will be assessed by visual inspection of histograms and the Kolmogorov-Smirnov test. The two-group independent t-test will be used to compare normally distributed variables. If variables are not normally distributed, the Mann-Whitney U test will be used to make comparisons between the trial groups. The analysis of primary and secondary outcomes will be controlled for study site, considering that the randomization is stratified by study site. The primary outcome (spontaneous vaginal delivery) and other categorical secondary outcomes will be compared between trial groups by using the Cochran-Mantel-Haenszel test. The estimates of the common relative risk and confidence intervals associated with the primary and secondary outcomes will be calculated. The Breslow-Day test for homogeneity of the odds ratios will be reported as well. Distributions of continuous secondary outcome measures such as duration of the second stage of labor within each site will be assessed by visual inspection of histograms and the Kolmogorov-Smirnov test. The mixed model in which study site is treated as a fixed effect will be used to compare normally distributed variables. If variables are not normally distributed, the van Elteren test will be used to make comparisons between the trial groups. E3-3. Secondary Analyses The investigators will perform other analyses aimed at obtaining adjusted assessments of treatment effectiveness, adjusting for baseline patient characteristics (covariates). The objectives of these analyses are to estimate the influence of covariates on the outcome and to use covariates to improve the estimated difference between treatment groups. The stepwise logistic regression model stratified by study site will be used to identify and estimate the effect of multiple prognostic factors on the probability of spontaneous vaginal delivery and other categorical outcomes. For continuous secondary outcomes such as duration of the second stage, the mixed model in which study site is treated as a fixed effect will be considered to adjust for prognostic factors. Interaction tests will be used to determine whether the effectiveness of the pushing strategy significantly differs across these subgroups. These analyses will be considered exploratory in nature and will not be viewed as providing confirmatory tests of hypotheses. The following prespecified subgroup analyses will be conducted: 1. Study site 2. Fetal station at complete cervical dilation (high versus low) 3. Fetal position at complete cervical dilation (occiput-anterior versus occiput-posterior) 4. Duration of delay prior to pushing (\<30 min. versus 30 - 60 min. versus \>60 min.) 5. Maternal age (\<35 years versus ≥35 years) 6. Maternal Race (Black versus White versus Other) 7. Obesity (Obese versus non-obese) 8. Birthweight (\<average weight versus ≥average weight) 9. Fetal sex (male versus female) 10. Pushing technique (spontaneous versus directed) The following planned secondary analyses will be conducted: 1. Effect of labor onset (spontaneous versus induced) and type of induction method 2. Effect of duration of the first stage of labor 3. Effect of Oxytocin use in the second stage Yes versus No 4. Effect of contraction pattern 5. Effect of magnesium sulfate use 6. Effect of support person in labor (Sindhu) 7. Effect of duration of pushing 8. Effect of chorioamnionitis (Tita) 9. Effect of BMI (Tita) 10. Effect of gestational age 11. Fetal position at complete cervical dilation (Caughey) 12. Effect of prior LEEP (Sindhu) 13. Effect of Amnioinfusion (Sindhu) 14. Length of hospital stay 15. Differences in indications for cesarean and operative vaginal delivery 16. Effect on cord blood lactate 17. Cost-effectiveness of delayed versus immediate pushing 18. Pelvic floor related outcomes i. Within and across intervention groups, correlate UI, FI and POP symptoms and impact and POPQ exams in women with/without LAM injury (control for LAM injury when comparing pelvic floor symptoms between immediate and delayed pushing groups) ii. Characterize rates of regression/persistence of pelvic floor symptoms/impact and objective (POP-Q and LAM) measures from 4-8 weeks to 5-7 months in the delayed versus immediate pushing groups iii. Characterize risk factors associated with LAM injury independent of intervention group iv. Characterize risk factors associate with persistence of pelvic floor symptoms v. Correlate overt and occult LAM injury to objective and subjective measures of pelvic floor morbidity vi. Rate of LAM defect on follow-up imaging after repair of overt laceration vii. Predictors of overt and occult LAM injury viii. Characterize rates of 3rd and 4th degree tears by randomized group ix. Characterize predictors of anal sphincter injury x. Predictors of anal sphincter defect following primary repair E3-4. Patient Satisfaction The investigators will assess patient satisfaction with their birthing experienced using the modified Mackey Childbirth Satisfaction Rating Scale. This is a simple validated childbirth satisfaction scale which has been modified to focus on the second stage experience. Face and construct validity as well as reliability have been established. E3-5. Economic Analysis The investigators will conduct economic analysis to estimate medical costs associated with management of the second stage and associated outcomes quantify potential cost savings attributable to immediate or delayed pushing.