Science & Sensibility has already reported on this trial, "Effect of Immediate vs Delayed Pushing on Rates of Spontaneous Vaginal Delivery Among Nulliparous Women Receiving Neuraxial Analgesia: A Randomized Clinical Trial", but let's take another pass at it to see if its conclusions hold up to closer scrutiny. Let's begin with a review of the abstract, the study summary that heads up every research paper because it tells us the points trialists are using to make their case.
"Effect of immediate vs. delayed pushing on rates of spontaneous vaginal delivery among nulliparous women receiving neuraxial analgesia: a randomized clinical trial" is a "pragmatic," meaning "real world," trial, as opposed to a trial that attempts to make management uniform except for the intervention under study. Included were 2404 nulliparous people at term who either began labor spontaneously or were induced, had an epidural, and reached full dilation. At that point, they were allocated either to begin pushing "immediately" or to delay pushing for one hour unless instructed to do otherwise, or they developed an irresistible urge to push. Excluded were multiple gestations and scheduled cesareans, which would have included breeches, making this an NTSV (nulliparous, term, singleton, vertex) population.
The abstract reported that results were statistically similar, meaning differences were likely to be due to chance, for incidence of spontaneous vaginal birth (86% vs. 87%), perineal lacerations (46% in both groups), and the occurrence of one or more adverse outcomes in a neonatal morbidity composite (7% vs. 9%). The immediate pushing group had a significantly shorter mean 2nd-stage duration (102 min vs. 134 min, mean difference -32 min) despite longer mean active pushing duration (84 min vs. 75 min mean difference +9 min) and significantly decreased incidence of chorioamnionitis (inflammation/infection of the membranes) (6.7% vs. 9.1%, difference -2.5%) and postpartum hemorrhage (2.3% vs. 4.0%, difference -1.7%).
These outcomes supported the trialists' conclusion that immediate pushing was the better option:
"Among nulliparous women receiving neuraxial anesthesia, the timing of second stage pushing efforts did not affect the rate of spontaneous vaginal delivery. These findings may help inform decisions about the preferred timing of second stage pushing efforts when considered with other maternal and neonatal outcomes."
However, as so often happens in studies conducted by medical-model thinkers in a medical-management environment, if we twist the kaleidoscope, a different picture emerges. Let's see what trialists decided not to include in the abstract.
Is There Any "There" There?
To begin with, trialists chose to tell us the mean (average) difference in length of second stage, not the difference in time from complete dilation to beginning pushing. This is curious because the point of the study is to ascertain the benefits, if any, of delayed pushing. Buried in the text of the full article is that the mean time to pushing in the "immediate pushing" group was 19 minutes, already not exactly immediate, and the standard deviation, a measure of the spread of the values around the mean, was 15 minutes, which is quite wide. In the "delay pushing" group the mean was, indeed, 60 minutes but the standard deviation was 22 minutes, which is, again, quite wide. This means that a considerable proportion of people in the "immediate pushing" group began pushing after 19 minutes, and a considerable proportion of people in the "delay pushing" group began pushing before 60 minutes. In other words, there's a substantial overlap between groups.
Trialists also chose to report perineal laceration rates, 2nd degree or greater perineal injuries, but not anal sphincter laceration rates, 3rd- and 4th-degree anal injuries. While perineal laceration rates were similar, anal sphincter laceration rates were not. Anal injury was significantly more common (6% vs. 5%) in the immediate pushing group. This omission, too, is curious because anal sphincter injury is a serious adverse outcome.
Similarly, trialists reported chorioamnionitis rates, which favored immediate . . . well, "earlier" pushing, to be more accurate, but not endometritis rates, which were similar (0.6% vs. 0.3%). Chorioamnionitis is a catch-all diagnosis for a fairly mild complication while endometritis is both more clearly defined and more serious.
Finally, trialists reported more postpartum "hemorrhages," which they defined as estimated blood loss ≥ 500 mL at vaginal delivery or ≥ 1000 mL at cesarean delivery. This is rather an old-fashioned definition of "hemorrhage" because 500 mL of blood loss or thereabouts has no clinical significance. Five-hundred milliliters is the amount taken at a blood donation, and losses up to 1000 mL do no harm to otherwise healthy people (Begley 2010). Estimated blood loss of 1000 mL or more ("severe" hemorrhage) or need for transfusion is a much more meaningful measure of clinically significant blood loss. The trialists reported this rate in the body of the paper, and, in fact, rates at this level were similar (2% vs. 3%), and the transfusion rate was identical (1%).
Choose different results, and instead of finding that outcomes somewhat favor immediate pushing, we find no clinically meaningful advantages to immediate pushing, and one disadvantage: a small increase in anal sphincter injury. We also find little advantage to delaying pushing, but this shouldn't surprise us because there isn't much difference in time to pushing between groups.
Trialists cherry-picked their data probably because they viewed the trial through the lens of an unconscious bias favoring medical management of labor. That bias also prevents them from seeing what leaps off the page for anyone not inside their bubble: medical management of labor and epidurals plays a far more important role in outcomes than whether people push immediately or delay pushing for some modest amount of time.
The trialists seem satisfied with an 85% spontaneous birth rate, noting in the discussion section that this was higher than the 72% expected in the delayed pushing group based on an earlier study. But the 15% operative delivery rate breaks down into an 8% cesarean rate and a 7% instrumental vaginal delivery rate. This means that 1 person in every 13 who made it to 2nd stage still had a surgical delivery, and an additional 1 in 14 delivering vaginally had an instrumental vaginal delivery. Furthermore, as we just saw, the anal sphincter injury rate was 6%, which means that 1 person in 17 having a vaginal delivery had an anal sphincter laceration (3rd- or 4th-degree perineal injury).
Medical-Model Management: The Elephant in the Room
How do we know that intrapartum management in the trial's participating hospitals falls short of what it should be?
Exhibit A: We already know that the 2nd-stage cesarean rate was 8%. We can calculate the 1st-stage cesarean rate from a figure showing who was excluded and for what reasons: 1005 of the 4963 people who were admitted in labor or for labor induction "never reached 10 cm," which calculates to a 1st-stage cesarean rate of 20%. Combining the two rates,* we get a total cesarean rate of 26%, smack on the U.S. average for NTSV people (Hamilton 2018). Contrast this with a 14% cesarean rate—12 fewer per 100—in NTSV people participating in a program for Medicaid beneficiaries at 45 free-standing birth centers (Jolles 2017).
Exhibit B: A figure in Listening to Mothers in California, a survey of 2539 people giving birth in California in 2016, illustrates how induction (46% of people in the pushing trial), epidurals (100% of people in the pushing trial), and the interaction between them affects the cesarean rate in nulliparous people.
Source: © 2018 National Partnership for Women & Families. Used with permission.
In nulliparous people who labored and were both induced and had an epidural, the cesarean rate was 30%. If they were induced but didn't have an epidural, the rate was 18%. If they had an epidural but began labor spontaneously, the cesarean rate was 19%, but if they were neither induced nor had an epidural, the rate was a mere 1%.
Exhibit C: Trialists argue in the discussion section that immediate pushing is preferable because studies show that maternal and neonatal adverse outcomes increase with 2nd-stage duration. However, as I countered in an earlier Science & Sensibility post, 2nd-stage duration per se isn't the issue; it's when doctors decide “time's up.” To recap my proof of this, a study compared outcomes according to three definitions of prolonged second stage in 42,225 people delivering at a single institution (Cheng 2014). Investigators found that cesarean and instrumental vaginal delivery rates and adverse maternal and neonatal outcomes were similar for second stages deemed not prolonged and increased markedly and to a similar degree in second stages deemed prolonged regardless of the definition of "prolonged." While proportions were similar at the three cut-off points (see the previous S&S post for details), absolute numbers declined with increasingly longer definitions because the number of undelivered people in the "prolonged" category fell from 8005 to 1580. At some point, someone has to make the call that the baby isn't going to come out on its own, but setting that time too early greatly increases the numbers of avoidable cesareans and instrumental deliveries along with their consequent harms.
Finally, this was a pragmatic trial. Other than allocation to early or delayed pushing, no attempt was made to dictate management. We know that common management practices such as frequent use of induction and augmentation, unphysiologic oxytocin regimens, rupture of membranes, frequent vaginal exams in people with ruptured membranes, confinement to bed, unphysiologic pushing positions, Valsalva maneuver pushing, pre-set limits on 2nd-stage duration, and episiotomy contribute to avoidable cesarean and instrumental vaginal deliveries, anal sphincter injuries, perineal injuries, instances of poor neonatal condition at birth, maternal and neonatal infection, and excessive postpartum bleeding.
Should we abandon "laboring down"? The trial doesn't answer this question because, for many people, time to pushing wasn't all that long in the "delay pushing" group and pushing delay overlapped with the "immediate pushing" group. Sadly, though, this hardly matters because medical management will likely overwhelm any small advantage that might be gained by a relatively brief delay before instructing people to push.
*Assume 1000 people. If 20% of 1000 people have a 1st-stage cesarean, then 200 people have a 1st-stage cesarean and 800 people reach 2nd stage. If 8% of people reaching 2nd stage have a cesarean, then 64 people (800 X 8%) have a 2nd-stage cesarean, making the overall cesarean rate 264/1000 or 26.4%, which, rounded down, is 26%.
Begley, C. M., Gyte, G. M., Murphy, D. J., Devane, D., McDonald, S. J., & McGuire, W. (2010). Active versus expectant management for women in the third stage of labour. Cochrane Database Syst Rev(7), CD007412. doi:10.1002/14651858.CD007412.pub2
Hamilton, B. E., Martin, J. A., Osterman, M. J., Driscoll, A. K., & Rossen, L. M. (2018). Births: Provisional data for 2017. NVSS Vital Statistics Rapid Release, Report No. 004.
Jolles, D. R., Langford, R., Stapleton, S., Cesario, S., Koci, A., & Alliman, J. (2017). Outcomes of childbearing Medicaid beneficiaries engaged in care at Strong Start birth center sites between 2012 and 2014. Birth, 44(4), 298-305. doi:10.1111/birt.12302