[ Editorials | Letters | Selected Articles | Signature Series | Patient Handouts | Index ]

SCHOLARLY DEBATE >>

Tilting At Windmills: Selected Term Vaginal Breech Delivery Should Not Be Abandoned

Richard K. Silver, MD

It should be stated from the outset that this article unequivocally and unapologetically supports selected vaginal breech delivery at term. This author has trained countless residents and colleagues over the past 15 years in operative vaginal delivery techniques, particularly vaginal breech delivery, and has participated in vaginal triplet delivery—the ultimate expression of confidence in vaginal breech birth.^1,2 Indeed, the author’s enthusiasm for teaching proper patient selection and key operative techniques could potentially cloud any assessment of the facts in this debate. However, all the facts have not yet been clarified, despite the performance and publication of a major randomized clinical trial on the subject.³ Furthermore, those in charge of specialty training programs in obstetrics and gynecology must take specific steps to reverse the self-fulfilling prophecy of dwindling interest in vaginal breech delivery due to inadequate training. Adopting these steps would eventually eliminate the presumption of inexperience among future practitioners. Toward this end, the specialty’s leadership must be prepared to tilt at a very popular windmill, ie, the randomized clinical trial (RCT). While this form of scientific inquiry provides outcome data so vital to academic obstetrics, altering clinical practice patterns in response to a single RCT is only appropriate to the extent that the findings can be generalized to most clinical settings.

back to top

The Research Dilemma

Clinical medicine has always struggled to find the best approach to situations characterized by variable patient presentation and management alternatives that rely substantially on clinical judgment and operator skill. Clinical questions, such as comparison of antibiotic regimens for puerperal infection or of oral hypoglycemic treatment versus insulin therapy for gestational diabetes, are both amenable to randomized study and largely insensitive to physician variability. By contrast, studies designed to determine the preferable surgical technique (eg, vaginal hysterectomy versus laparoscopically assisted vaginal hysterectomy) are far more problematic because the surgeons must be equally experienced in their performance of both alternatives to evaluate the merits of each procedure fairly.

The preferred route of delivery for fetuses with breech presentation at term exemplifies a research dilemma that seems to be beyond rigorous scientific resolution. To their credit, Hannah and colleagues³ pursued and completed a multicenter RCT comparing planned vaginal delivery to planned cesarean delivery for breech presentation, appearing to put this issue to rest. In addition, endorsement and dissemination of their findings was both swift and definitive, giving clinicians little time to weigh their own training and experience.^4-6 Indeed, Charles J. Lockwood, MD, who provides the counterpoint for this debate, was one of those who sanctioned the findings of the multicenter trial.⁵

However, objective reflection suggests that shortcomings in the design of this RCT may significantly limit its general applicability and perhaps its interpretation (Table 1). Ironically, these findings may actually support the contention that selected vaginal breech delivery within stringent guidelines is as safe (and therefore preferable) as routine cesarean delivery.

 

View this table

Table 1. Significant Limitations of the Term Vaginal Breech RCT3

back to top

Patient Criteria.—The 121 centers in 26 countries that accrued and randomized the subject cohort handled patient inclusion as well as possible. Consequently, the randomized subjects reflected the universe of breech cases. However, this superficial methodologic strength may have ultimately confounded comparison of the delivery route because the inclusion criteria were too broad and inadequately enforced. Thus, patient accrual disproportionately influenced perinatal outcomes among planned labors compared with scheduled cesarean deliveries.

First, a significant percentage of the estimated fetal weights were not obtained by ultrasonography (845/2,083, or 40%), yielding a frequency of unanticipated macrosomia (91 cases, or 4%) that suggests the need for a more rigorous approach to fetal weight estimation when contemplating vaginal breech delivery. Second, pelvic adequacy was not documented beyond clinical assessment in most patients (1,890/2,083, or 91%), and the attitude of the fetal head was only determined clinically in a minority of cases (645/2,083, or 31%). It is possible that many of the “difficult” vaginal deliveries (48/1,042, or 4.6%) among the randomized cases intending to deliver vaginally, could have been mitigated by quantitative pelvimetry or imaging studies to rule out deflection.³ Of greatest potential to confound the comparison of delivery methods are those pregnancies that did not have continuous fetal heart rate (FHR) monitoring, especially during the second stage of labor. The distribution of intermittent versus continuous monitoring is not described in the original paper, but in a subsequent letter the authors reported that only 34% of the women who went into labor were monitored continuously.⁷ The virtues of continuous FHR monitoring versus intermittent auscultation in general are debatable,^8,9 especially as FHR skills and interpretation are not uniform across centers or between caregivers.¹⁰ However, the need for more complete FHR data accompanying a trial of labor with breech presentation would seem self-evident. That continuous FHR monitoring was not the standard of care in this trial was underscored by the occurrence of cases in which the FHR “disappeared,” resulting in unanticipated fetal death prior to vaginal delivery.³

In addition, fetuses thought to have intrauterine growth restriction should also be excluded, to the extent that sparing of head growth coupled with a smaller abdominal circumference is a risky combination for breech delivery. A similar argument could be made to exclude patients with prolonged pregnancy (longer than 41 weeks’ gestation), where the risk of macrosomia and cephalopelvic disproportion is highest. Hannah and coworkers may contend that these characteristics were evenly distributed between groups so that their individual influence on outcome should not bias the study. Nonetheless, all of these factors would have a disproportionately negative effect on labor and vaginally delivering patients than on those undergoing a planned cesarean section prior to the onset of labor.

Variability in the Delivery Environment.—Another criticism of the Hannah study involves the inclusion criteria for participating OB/GYNs and institutions, which allowed for considerable disparity in key characteristics of the delivery environment. This criticism should not be interpreted as scientifically ethnocentric; having observed high-quality obstetric care in many settings in the United States and abroad, this author has no a priori bias toward Western operative obstetrics. However, the difficulty in controlling for key provider and institutional variables cannot be ignored. It is insufficient to simply propose an assortment of post-hoc justifications for the differences in perinatal mortality rates or operator experience among centers. Another potential environmental variable not well documented in the study is the rate of enrollment per center as a percentage of all eligible cases. At least one correspondence asserted that selected participating centers enrolled less than 1% of eligible cases, suggesting the possibility of ascertainment bias.¹¹

What Constitutes the “Ideal” Breech Delivery?—Due to the excess perinatal morbidity in breech infants assigned to be delivered vaginally (13 of the 16 perinatal deaths in the study were allocated to vaginal delivery), the “term vaginal breech trial” paradoxically helps to identify the ideal antepartum and intrapartum prerequisites for vaginal breech delivery. These characteristics contradict those in the Canadian consensus publication on this subject,¹² which established the methodology for the subsequent RCT. For example, had a normal (nonacidemic) intrapartum continuous FHR tracing been required, up to four of the cases with FHR abnormalities contributing to the 16 perinatal deaths might have been mitigated. The same could be said for formal documentation of pelvic capacity, which was only performed in one of 10 women assigned to planned vaginal birth. It is not surprising that six additional cases of difficult vaginal delivery ensued, resulting in stillbirth or neonatal death.

The choice of 0.5 cm per hour as the cutoff defining the adequacy of active-phase cervical dilation is also problematic, especially for the multiparous subjects. This passive approach to the active phase of labor could have contributed to the occurrence of difficult delivery, to the extent that desultory labor may be a proxy for cephalopelvic disproportion. Likewise, allowing up to 3 hours from complete dilation until expulsion only serves to increase the risk of a difficult breech delivery.

What is the Ideal Setting for Vaginal Breech Delivery?—The characteristics of caregivers and available services across centers must also be considered in the context of providing the optimal environment for vaginal breech delivery. Although Hannah and colleagues³ worked hard to characterize the OB/GYNs as similar between randomized cohorts, equivalent years of experience with vaginal breech delivery or the percentage of licensed practitioners in each study arm does not necessarily ensure sound clinical judgment and/or refined operative skills. The marked variability in anesthesia and neonatology services among centers and tolerance of up to a 30-minute delay in responding to a depressed infant are examples of nonideal environments for vaginal delivery, regardless of breech or cephalic presentation.

Type III Error.—While most researchers are aware of how type I and II statistical errors can interfere with validity, the type III error (as described by Friedman¹³) characterizes the Hannah trial.³ Specifically, a study can suffer from a type III error (implementation failure) to the extent that the intervention under evaluation is inadequate on its own merits, independent of any planned study comparisons. Implementation failure can be overlooked as a reason for either null or misleading results. With the best of intentions, the vaginal breech delivery trial set out to make a valid comparison between routes of delivery. The intervention arms were appropriately chosen, but the interventions themselves were flawed as outlined above. The case could be made that a similarly powered study with equivalent design could have been as likely to identify cesarean section as preferred over vaginal delivery in the cephalic-presenting fetus that is compromised before labor, inadequately monitored during labor, or suboptimally resuscitated after delivery.

back to top

Piper Forceps

It would be unfair to critique the Hannah trial³ without offering an alternative. Table 2 describes the essential patient characteristics, labor management protocol, and delivery methodology for selected term breech pregnancies. Compared with other contemporary protocols in the literature,^14,15 the cornerstone of this paradigm is to move from use of forceps for the aftercoming head without adequate planning or as a last resort to prophylactic application in all cases. Second only to the anxiety generated by the occasional nuchal arm is that associated with the uncertainty of successfully and safely delivering the aftercoming head. This anxiety is not restricted to vaginal breech birth, but also occurs during cesarean delivery. The irony is that a low-segment uterine incision provides a similar dimension for the fetal head to pass, while simultaneously reducing the placental perfusion pressure prior to eventual delivery.

 

View this table

Table 2. Patient Characteristics and Management Protocol for Trial of Labor in Term Vaginal Breech Delivery

Piper forceps are perhaps the most elegant instrument devised for operative obstetrics, and are the key to teaching safe term vaginal breech delivery. The classic text by Dennen¹⁶ contains a concise narrative extolling the virtues of Piper forceps. It provides clear instructions on their appropriate use. The design of this instrument, with the handles below the level of the blades, ensures proper flexion of the fetal head and protection of the cervical spine (Figure 1). The length of the shanks affords a spring-like quality to the blades that reduces compression of the fetal head. In addition, proper use of the Piper forceps eliminates the need for any traction on the fetal body, which is by contrast unavoidable even with careful use of the Mauricceau-Smellie-Velt maneuver during either vaginal delivery or cesarean delivery. Moreover, pelvic application of the Piper forceps (a one-step procedure) is actually easier than standard forceps that require cephalic application (two distinct steps relying on specific fetal landmarks). Once Piper forceps are applied, the breech fetus is delivered as a unit, head and torso simultaneously, a mechanism that further protects the cervical spine.

 

Click to enlarge

Figure 1. Piper forceps are uniquely designed to provide an atraumatic application to the aftercoming fetal head, maintain flexion of the head onto the fetal chest, and limit traction on the cervical spine. The handles (A) articulate easily, immediately after pelvic application. The excess length of the shanks (B) provides a spring-like quality that minimizes head compression, and the blades (C) are below the handles to effect persistent flexion of the head.

Contrary to the notion that vaginal breech delivery skills cannot be taught in the current obstetric environment, there are many opportunities to teach the same techniques at the time of cesarean section for breech presentation and vaginal delivery for noncephalic second twins. With the exception of the initial breech extraction, delivery of the torso, arms, and aftercoming head requires identical methodology during both vaginal and abdominal delivery of breech fetuses. A shortened-shank, cesarean-delivery forceps (Simpson type) could be employed much like a vectis to facilitate delivery of the head through the uterine incision in the cephalic-presenting infant.

back to top

CONCLUSION

Despite the best efforts of Hannah and colleagues,³ the conclusions drawn from the RCT cannot be generalized. Their study lacks external validity as described by Grimes in a recent biostatistical primer for clinicians.¹⁷ Instead, while condemning unselected breech delivery because of the higher risk of adverse perinatal outcomes, their observations leave selected breech delivery as an open question. Furthermore, some centers in their study lacked the key elements of the medical environment essential not only to successful vaginal breech delivery, but also to the practice of obstetrics in general.

Obstetric trainees can still learn vaginal breech delivery techniques if their teachers exploit the numerous opportunities to instruct them. Ultimately, the cesarean delivery rate for term breech fetuses with a contraindication to external cephalic version or a failed attempted version will remain higher than that for cephalic-presenting infants, but it need not approach the 100% rate that the Hannah trial³ authors and their supporters advocate.

back to top

REFERENCES

1. Adams DM, Sholl JS, Haney EI, et al. Perinatal outcome associated with outpatient management of triplet pregnancy. Am J Obstet Gynecol. 1998;178:843-847.
2. Grobman WA, Peaceman AM, Haney EI, et al. Neonatal outcomes in triplet gestations after a trial of labor. Am J Obstet Gynecol. 1998;179:942-945.
3. Hannah ME, Hannah WJ, Hewson SA, et al. Planned caesarean section versus planned vaginal birth for breech presentation at term: a randomised multicentre trial. Term Breech Trial Collaborative Group. Lancet. 2000;356:1375-1383.
4. Lumley J. Any room left for disagreement about assisting breech births at term? Lancet. 2000;356:1369-1370.
5. Lockwood CJ. The end of term breech delivery. Contemp Ob/Gyn. 2001;9:10,13.
6. American College of Obstetricians and Gynecologists. ACOG Committee Opinion No. 265. Mode of term singleton breech delivery. Intl J Gynaecol Obstet. 2002;77(1):65-66.
7. Hannah ME, Hannah WJ, Willan A. Author’s reply. Lancet. 2001;357:227-228.
8. Vintzeleos AM, Nochimson DJ, Guzman ER. Intrapartum electronic fetal heart rate monitoring versus intermittent auscultation: a meta-analysis. Obstet Gynecol. 1995;85:149-155.
9. hacker SB, Stroup D, Chang M. Continuous electronic heart rate monitoring for fetal assessment during labor. Cochrane Database Syst Rev. 2001;(2):CD000063.
10. Parer JT, King T. Fetal heart rate monitoring: is it salvageable? Am J Obstet Gynecol. 2000;182:982-987.
11. Stuart IP. Term breech trial [Letter]. Lancet. 2001;357:228.
12. Society of Obstetricians and Gynaecologists of Canada. The Canadian consensus on breech management at term. SOGC Policy Statement No. 31. J SOGC. 1994;16:1839–1848.
13. Friedman LM, Furberg C, Demets DL. Fundamentals of Clinical Trials, ed 3. New York City: Springer-Verlag; 1998.
14. Albrechtsen S, Rasmussen S, Reigstad H, et al. Evaluation of a protocol for selecting fetuses in breech presentation for vaginal delivery or cesarean section. Am J Obstet Gynecol. 1997;177:586-592.
15. Diro M, Puangsricharem A, Royer L, et al. Singleton term breech deliveries in nulliparous and multiparous women: a 5-year experience at the University of Miami/Jackson Memorial Hospital. Am J Obstet Gynecol. 1999;181:247-250.
16. Dennen EH, Dennen PC, (Hale RW, ed) Dennen’s Forceps Delivery, ed 4. Washington, DC: American College of Obstetricians and Gynecologists; 2001.
17. Grimes DA, Schultz KF. An overview of clinical research: The lay of the land. Lancet 2002;359(9300):881-884.

back to top