by Henci Goer

Commentary on: Landon, M. B., Hauth, J. C., Leveno, K. J., Spong, C. Y., Leindecker, S., Varner, M. W., et al. (2004). Maternal and perinatal outcomes associated with a trial of labor after prior cesarean delivery. N Engl J Med, 351 (25), 2581-2589. [Abstract]

Study design and results: prospective multicenter study of 17,898 women defined as planning VBAC and 15,801 women planning elective repeat cesarean (ERC). Study included all women with a prior cesarean and a singleton pregnancy of 20 weeks or more or whose newborn weighed at least 500 g. Investigators defined ERC as any woman having a cesarean without labor and without any medical reason for cesarean delivery. Investigators defined planned VBAC as any woman presenting in labor with cervical dilation of at least 4 cm or who received oxytocin. Women having cesareans in early labor were excluded from analysis because their planning status could not be determined.

VBAC rate: 73.4%

Maternal complications: VBAC vs. ERC
The following complication rates were similar between groups:

• maternal death: 3 (0.2 per 1,000) vs. 7 (0.4 per 1,000)
• hysterectomy: 41 (2 per 1,000) vs. 47 (3 per 1,000)
• embolism, including deep leg vein and pulmonary: 7 (0.4 per 1,000) vs. 10 (1 per 1,000)
• other adverse events, including broad-ligament hematoma, bladder puncture, bowel injury, and ureter injury: 64 (4 per 1,000) vs. 52 (3 per 1,000)

The following complication rates differed significantly (meaning the difference is unlikely to be due to chance):

• uterine rupture: 124 (7 per 1,000) vs. 0, odds ratio cannot be calculated
• transfusion: 304 (17 per 1,000) vs. 158 (10 per 1,000), odds ratio: 1.71
• endometritis: 517 (29 per 1,000) vs. 285 (18 per 1,000), odds ratio: 1.62
  one or more of the above: 978 (55 per 1,000) vs. 563 (36 per 1,000), odds ratio: 1.56

The VBAC group had excess rates (absolute risk increase) of the following compared with the ERC group:

• uterine rupture: 7 per 1,000
• transfusion: 7 per 1,000
• endometritis: 11 per 1,000
• any adverse event: 19 per 1,000

Maternal complications according to whether labor ended in vaginal birth or repeat cesarean: 309 / 13,139 (2.4%) women having vaginal births had one or more of the complications above vs. 669 / 4,759 (14.1%) having unplanned repeat cesareans. odds ratio: 6.81; 95% confidence interval 5.93 to 7.83

Uterine rupture rate in VBAC labors: Women having induced or augmented labors were significantly more likely to experience uterine rupture compared with women having normal labor:

• spontaneous labor: 24 / 6,685 (3.6 per 1,000)
• induced labor (27.1%): 48 / 4,708 (10.2 per 1,000) p < 0.001
• augmented labor in women beginning labor spontaneously (34.5%): 52 / 6,009 (8.7 per 1,000) p < 0.001
• not classified: 0 of 496

The excess rate (absolute risk increase) of uterine rupture was 6 per 1,000 with induction and 5 per 1,000 with augmentation compared with women having normal labor.

Perinatal outcomes for term infants (37 wks or more, including deaths associated with congenital malformations): VBAC vs. ERC

The following complication rates were similar between groups:

• intrapartum death: 2 / 15,338 (0.1 per 1,000) vs. 0 of 15,014
• neonatal death: 13 / 15,338 (0.8 per 1,000) vs. 7 of 15,014 (0.5 per 1,000)

The following complication rates differed significantly:

• antepartum death: 34 / 15,338 (2.2 per 1,000) vs. 13 / 15,014 (0.9 per 1,000)
• hypoxic-ischemic encephalopathy (symptoms of brain injury such as seizure that may or may not result in permanent damage): 12 / 15,338 (0.8 per 1,000) vs. 0 of 15,014, p < 0.001
  one or more of the above: 59 / 15,338 (3.8 per 1,000) vs. 20 / 15,014 (1.3 per 1,000), p < 0.001; odds ratio: 2.90; 95% confidence interval 1.74 to 4.81

The VBAC group had excess rates (absolute risk increase) of the following compared with the ERC group:

• antepartum death: 13 per 10,000
• hypoxic-ischemic encephalopathy: 8 per 10,000
• any adverse perinatal event: 25 per 10,000
• any adverse perinatal event excluding antepartum death: 11 per 10,000 Note: This difference is probably no longer significant.

Perinatal outcomes subsequent to uterine rupture in term pregnancies: 2 of 13 neonatal deaths and 7 of 12 cases of hypoxic-ischemic encephalopathy occurred in association with uterine rupture.

Corrected perinatal death rate at term after excluding deaths associated with congenital malformations: 4.0 per 10,000 with VBAC vs. 1.4 per 10,000 in women undergoing ERC.

Problems include but are not limited to the following:

• Despite the description as “prospective,” the study suffers from the fundamental weakness of retrospective VBAC studies: the inability to determine planning status and eligibility of participants: The major advantage of conducting a prospective VBAC study should have been the ability to enter participants based on pre-specified criteria and to track and exclude them should they develop an indication for repeat cesarean. Clearly this did not happen since the investigators defined planned VBAC as any woman receiving oxytocin or achieving 4 cm or more dilation and excluded women admitted in early labor because their planning status was unknown. They also acknowledged that the excess of antepartum stillbirths in the VBAC group was likely due to those women being more likely to be encouraged to plan VBACs. In addition, two of the three maternal deaths in the VBAC group were in seriously ill women, and five women having hysterectomies in the ERC population had them because they had cancer. This raises the question of how many other women had medical conditions that could have affected outcomes regardless of birth route. These problems hopelessly compromise the study's ability to determine the risks of planned VBAC versus ERC, making it useless for guidance on this matter.
  
  
• The ways in which the investigators chose to report outcomes biases against VBAC:
  • The investigators reported comparisons as odds ratios: Odds ratios in the absence of other calculations such as “absolute excess risk” or “number needed to harm” give an exaggerated impression of risk when absolute differences between groups are small.
  • The investigators included antepartum deaths in their calculation of adverse perinatal outcomes at term: The authors justify this on grounds that elective cesarean at 39 weeks may have prevented some of them, but, in fact, as they themselves acknowledge, a more likely reason for excess numbers in the VBAC group is that women whose babies die before labor are more likely to be encouraged to plan VBAC. In other words, choosing VBAC did not lead to antepartum stillbirth; antepartum stillbirth led to choosing VBAC.
  • The investigators chose not to report differences in rates of permanent harm: In trying to determine which is the better option, the key issue is the odds of experiencing an adverse event that cannot be fixed. If one limits the calculation of maternal adverse events to death or hysterectomy, the only two on the “adverse outcomes” table from which there is no recovery, the absolute difference between VBAC and ERC falls from 2 more women per 100 in the VBAC group experiencing an adverse event to 1 fewer woman per 1,000 in the VBAC group dying or having a hysterectomy. On the infant side, hypoxic-ischemic encephalopathy doesn't always result in brain injury. With only 12 cases, some follow up could have determined how many babies suffered permanent damage. Finally, it is important to consider that accumulating another cesarean scar is also an irremediable injury. As the number of cesareans goes up, so do the odds of serious complications for women and the babies of any future pregnancies. (3)
    
    
• The two groups of women studied were dissimilar, making interpretation of the results problematic. The authors emphasized that the women planning VBAC were significantly more likely to have had a prior vaginal birth, which is known to be associated with a higher VBAC success rate and better outcomes. However, they report but do not emphasize many other differences between the VBAC and ERC groups. For instance, women planning VBAC were more likely to smoke, have public insurance (Medicaid), have a preterm birth, and deliver a low birth-weight baby. It is unknown whether or how these variables might affect outcomes. The researchers took care to exclude the preterm births from the analysis of neonatal outcomes, but their analysis of maternal outcomes included these births. This might account for this study's findings related to postpartum endometritis (infection of the womb). Unlike previous studies comparing VBAC and ERC, this study found a significantly higher rate of endometritis in women electing VBACs. Some common risk factors for endometritis are also known to cause preterm birth. Since 14 percent of the women planning VBAC in this study were preterm, it is possible that much of the excess risk for postpartum endometritis reported in this study could have been prevented by excluding preterm births in the analysis of maternal outcomes.
  
• Obstetric management, not labor, caused many of the adverse outcomes in the VBAC group: The uterine rupture rate in women laboring spontaneously was 3.6 per 1,000. Inducing and augmenting labor increased the uterine rupture rate to 10.2 and 8.7 per 1,000 respectively. Based on these rates, if every woman had labored without stimulation, 63 women would have had uterine ruptures instead of 124. In addition, 5 of the 12 cases of hypoxic-ischemic encephalopathy and 11 of 13 neonatal deaths in term pregnancies occurred in the absence of uterine rupture. Here, too, obstetric management may have played a role. The increased rates of miscarriage, preterm birth, low-birth weight, abnormalities of placental attachment, and late fetal demise found in subsequent pregnancies tell us that a scarred uterus can be an unhealthy environment for growing a baby. (3) This suggests that babies of women with prior cesareans may be less able to tolerate stress. Oxytocin, especially when used to induce labor or in high-dose, “active management” augmentation protocols, is notorious for causing overly strong, overly long contractions (uterine hyperstimulation). Other common obstetric procedures and practices, including epidural analgesia (which can reduce maternal blood pressure), rupturing membranes, and having women push long and forcefully while lying on their backs may also introduce more stress than an already compromised baby can tolerate.
  
  
• It seems highly unlikely that no woman having ERC had a uterine rupture: Two other large studies looking at this issue reported rates of uterine rupture in the ERC population of 1.6 per 1,000 and 1.9 per 1,000. (2; 4) If you multiply these rates by 15,801, the number of women having ERC in this study, you see that there should have been 25 to 30 cases of the scar giving way in the ERC population. One potential explanation for the discrepancy could lie in the exclusion of women arriving at the hospital before 4 cm dilation who had cesareans. Another could lie in how obstetricians distinguished ruptures from dehiscences (harmless “windows” in the scar) discovered during elective repeat surgery.
  
  
• Only one of three maternal deaths in the VBAC group was possibly related to the scar, whereas the investigators attribute two of seven deaths in the ERC group to the cesarean surgery: In the ERC group, the surgically related deaths were due to hemorrhage in one case and complications of anesthesia in the other. By contrast, two of the three women who died in the VBAC group were critically ill. One had severe preeclampsia with liver failure and another had heart failure secondary to a sickle cell crisis. The third woman died of a postpartum hemorrhage, but we are not told whether the hemorrhage followed uterine rupture. Therefore, it is possible that none of the maternal deaths in the VBAC group were attributable to scar complications.

Comment: The conclusions that can be drawn from this study are the same as those that can be drawn from every VBAC study for the past 20 years, namely:

• Planned VBAC is safe for the mother: Even without considering the ways in which this study stacks the deck against VBAC, a mere 2 more women in every 100 planning VBAC had some adverse outcome compared with ERC. And when one looks at permanent harm, one more women in every 1,000 having ERC died or had a hysterectomy compared with women planning VBAC.
  
  
• Planned VBAC is a reasonable choice for the baby: The absolute increased risk of perinatal death in a normally formed baby with VBAC versus ERC was 2.6 per 10,000. Compare this with an excess pregnancy loss rate of 60 per 10,000 resulting from amniocentesis, a procedure that no one disallows on grounds of safety. (5) Keep in mind as well that these numbers include deaths before labor. As discussed above, women whose babies died before labor onset—deaths that would be unrelated to choice of birth route—were much more likely to have VBACs.
  
  
• Conventional obstetric management is a major contributor to the risks of VBAC: Half the uterine ruptures and the life- and health-threatening dangers that accompany them could have been avoided if women had simply been left to labor without artificial stimulation. It is possible that some of the cases of hypoxic-ischemic encephalopathy and neonatal death unrelated to uterine rupture might also have been avoided.

Nonetheless, debating the safety or lack thereof of VBAC misses the real point: the right of every woman to make an informed decision about her care, based on a thorough understanding of the benefits and risks. Michael Greene ends his New England Journal of Medicine editorial on this study by saying, “[R]isk, like beauty, is in the eye of the beholder,” by which he means the obstetrician. (1) The “beholder' should, however, be the pregnant woman, not her doctor, not a hospital administrator, and certainly not her doctor's malpractice insurance company!

References:

1. Greene, M. F. (2004). Vaginal birth after cesarean revisited. N Engl J Med, 351 (25), 2647-2649.
2. Lydon-Rochelle, M., Holt, V. L., Easterling, T. R., & Martin, D. P. (2001). Risk of uterine rupture during labor among women with a prior cesarean delivery. N Engl J Med, 345 (1), 3-8.
3. Maternity Center Association. (2004). Effects of cesarean and vaginal birth on mothers, babies, and future reproductive capacity: A rapid systematic review. from http://www.childbirthconnection.org/article.asp?ck=10271&ClickedLink=200&area=2
4. Rageth, J. C., Juzi, C., & Grossenbacher, H. (1999). Delivery after previous cesarean: A risk evaluation. Swiss working group of obstetric and gynecologic institutions. Obstet Gynecol, 93 (3), 332-337.
5. Seeds, J. W. (2004). Diagnostic mid trimester amniocentesis: How safe? Am J Obstet Gynecol, 191 (2), 607-615.