When Research is Flawed: Does the Timing of an Epidural Influence Risk for C-Section?

by Henci Goer

Commentary on: Wong, C. A., Scavone, B. M., Peaceman, A. M., McCarthy, R. J., Sullivan, J. T., Diaz, N. T., et al. (2005). The risk of cesarean delivery with neuraxial analgesia given early versus late in labor. N Engl J Med, 352 (7), 655-665. [Abstract]

Study design and results:randomized controlled trial of 728 nulliparous (no prior births) women with cervical dilation less than 4 cm. All participants were healthy women with full-term singleton pregnancies who were either in spontaneous labor or who had spontaneous rupture of membranes. The study had the statistical ability to detect a 50% difference in cesarean rate 80% of the time provided there were 350 women in each group. The study concluded that women “who request pain relief early in labor can receive neuraxial analgesia at that time without adverse consequences” and that starting “early neuraxial analgesia” did not increase the cesarean rate.

At first request for pain medication, 366 women assigned to what investigators called the “intrathecal group” were given an intrathecal (spinal) injection of a narcotic (fentanyl). At second request, a continuous epidural infusion of fentanyl-bupivacaine was begun, a technique called a “combined spinal-epidural.” The 362 women assigned to the “systemic opioid” group were given intramuscular and intravenous injections of the narcotic hydromorphone (Dilaudid) at their first requests for pain medication, and the dose was repeated if the woman made a second request and was still less than 4 cm dilated. If cervical dilation was 4 cm or more at the time of the second request or if a third request was made (regardless of dilation), the woman received a continuous epidural infusion of fentanyl-bupivacaine.

• Cervical dilation at initiation of epidural:
• intrathecal group: 138 women less than 4 cm dilated; 226 women dilated to 4 cm or more
• systemic group: 100 women dilated to 3 cm or less; 241 women more than 3 cm dilated [how many women were more than 3 cm dilated but still less than 4 not reported]
• Adverse effects:
  • abnormal fetal heart rate (FHR): Compared with the systemic group, an excess 3 babies per 100 in each category in the intrathecal group experienced either prolonged FHR decelerations (< 100 bpm for > 60 sec) or a non-reassuring change (persistent variable decelerations, late decelerations) concerning enough that it “would lead to an obstetrical intervention if observed in real time by the perinatologist.” Differences between the intrathecal and systemic groups were statistically significant (meaning unlikely to have occurred by chance) in all three cases (p-values ranged from 0.003 to 0.02).
  • fever: No difference between groups, but percentages experiencing fever not reported. [The percentages of women developing fever is important because it is associated with increased use of cesarean section, neonatal septic workups, and administration of IV antibiotics to mother, newborn, or both. (3) ]
  • vomiting: 2% in the intrathecal group vs. 17% in the systemic group, p < 0.001
• Mode of delivery: No significant differences between groups.
  • cesarean surgery: 18% in the intrathecal group vs. 21% in the systemic group
  • vaginal instrumental delivery: 20% in the intrathecal group vs. 16% in the systemic group

Problems include but are not limited to the following:

• The trial doesn't compare early versus late epidurals: The introduction to the study states that the study will address early versus late epidurals, and the Associated Press story quotes the lead author of the study as saying: “There is no reason why women who want an epidural should not get it when they first request it.” (1) In point of fact, neither group of women were given an epidural initially. An “epidural” refers to the route of administration of pain-relief medication (i.e., the medication is injected into the epidural space, a gap between two membrane layers that wrap the spinal cord). An “epidural,” as it is commonly understood, is the injection of local anesthetic, or more usually these days, a mixture of anesthetic with narcotic, into this space. In this study, both groups of women were given pure narcotic first, only by different routes. The “intrathecal” group had fentanyl injected beneath the epidural space while the “systemic” group had intramuscular and intravenous injections of hydromorphone (Dilaudid). This sequence resulted in large numbers of women in both groups having a true epidural initiated around the same time. Only 138 women of the 366 in the intrathecal group, the supposed “early epidural” group, received epidural anesthetic prior to 4 cm dilation. Meanwhile, in the systemic group, the supposed “late epidural” group, we know that 100 of the 362 women assigned to this group received an epidural at 3 cm dilation or less, but we don't know how many additional women had their epidural initiated before 4 cm because the investigators don't tell us this.
  
  The distinction between intrathecally injected narcotic and epidurally injected anesthetic is important. The problem with early epidurals is that they are associated with higher cesarean rates, and maternal fever is associated with epidural duration, which would naturally be longer with earlier placement. But, unlike epidural anesthetic, intrathecal (spinal) narcotic has not been shown to slow labor nor to cause fevers. The fact that this trial didn't evaluate early versus late epidurals makes any statements about the effects on cesarean or fever rates invalid.

• The trial isn't capable of determining if there was a clinically important increase in cesarean rate: Leaving aside that it didn't compare what it claimed to compare, the trial wasn't big enough to show meaningful differences in c-section rates. To begin with, it wasn't designed to test for a difference in rates that is smaller than 50%. Differences much less than this would still be important differences. Also, the trial needed 350 women in each group to have a reasonable chance of showing a difference of this magnitude. As stated in the bullet above, far fewer women than 350 actually received early epidurals in the early epidural group and vice versa in the late group. This means that even if there had been a true 50% difference in cesarean rates, this trial was unlikely to detect it.

• Babies were significantly more likely to experience alarming fetal heart rate changes in the intrathecal group: Investigators stated that the spinal epidural technique was “without adverse consequences,” yet an additional 3 babies per 100 experienced fetal distress. The trial had too few women to find out if these extra cases of fetal distress were associated with poor health outcomes for the babies or greater cesarean section risk for the mother. Moreover, these were all healthy, full-term babies and thus best able to tolerate stress. Incidence and severity of abnormal fetal heart rate might be much higher in a less optimal population.

• Women were given a pain relief medication they should not have been given: As the Maternity Center Association critique of this trial points out, the FDA package insert specifically states that hydromorphone (Dilaudid) should not be used in labor. (5; 6) Considering the number of acceptable opioid medications for labor pain, this would indicate at best indifference to observing practice standards and at worst a disregard for patient well-being.

• The absence of an undrugged comparison group masks the true risks of epidurals: The investigators report that more babies experienced fetal distress in the combined spinal-epidural group than the group having systemic narcotic followed by an epidural. We do not know what the excess would be compared to babies whose mothers had no pain medication, but it would almost certainly be greater because both groups of babies here were exposed to the same problematic drugs. While we are told only that maternal fever rates did not differ between groups, research shows that rates are much higher in women having an epidural than in women who don't. (3) Cesarean rates can also be much lower in women who don't have epidurals, a point known by the authors of this study. The 50% expected difference in cesarean rates investigators used to calculate the required size of this trial came from a study of induction carrie d out previously at the same hospital. (7) That study reported cesarean rates of 19% with epidural placement before 4 cm dilation and 9% with placement at or greater than 4 cm. But the cesarean rate was only 3% in women having no epidural.
  
  Both the study and the media coverage make the implicit assumption that all women will be having epidurals; the only question is timing. But women may well wish to choose non-drug options for coping with labor pain. (4) Most women find comfort measures such as immersion in warm water, maintaining mobility, massage and cognitive strategies that enhance relaxation, relieve anxiety, and increase confidence to be helpful. Research shows that continuous support from a doula (a trained labor companion) is especially effective. None of these alternatives incur the risks or drawbacks of pain relief medications. In fact, they tend to promote good labor progress. These advantages make doulas, comfort measures, and other strategies the first choice for laboring women, and many women will need nothing more for the entire labor.

Comment: The misrepresentations and omissions in the reporting of this trial seriously misinform both the public and professionals about the risks of early versus late epidurals as well as the safety of epidurals in general. Even more disturbing, this trial is part of a recurring pattern of heavily publicized, gravely flawed obstetric studies for which exaggerated and unwarranted claims are made in press releases or media interviews, a pattern in which the New England Journal of Medicine has played a prominent role. (2) Recent decades have seen researchers and clinicians strive to build modern medical practice on the foundation of the scientific evidence, an effort jeopardized in obstetrics if this pattern continues unchecked. When the obstetric research is completely riddled with poor quality, misleading studies such as this one, we will lose an objective standard on which to base care. If you can't trust researchers at a university-based medical school and the New England Journal of Medicine to provide sound scientific data, who can you trust?

References:

1. Donn, J. (2005, Feb 18, 2005). Early epidural won't raise c-section risk. Los Angeles Times.
2. Goer, H. (2003). "Spin doctoring" the research. Birth, 30 (2), 124-129.
3. Lieberman, E., & O'Donoghue, C. (2002). Unintended effects of epidural analgesia during labor: A systematic review. Am J Obstet Gynecol, 186 (5 Suppl Nature), S31-68.
4. Maternity Center Association. How will I cope with labor pain? from http://maternitywise.org/mw/topics/pain/
5. Maternity Center Association. (2005). Effect of labor pain medication timing on cesarean section: New England Journal of Medicine study, February 2005. from www.maternitywise.org/nejm_epidural_response.html
6. PDR drug information for Dilaudid. Retrieved Mar 3, 2005, from www.drugs.com/PDR/Dilaudid_Injection.html
7. Seyb, S. T., Berka, R. J., Socol, M. L., & Dooley, S. L. (1999). Risk of cesarean delivery with elective induction of labor at term in nulliparous women. Obstet Gynecol, 94 (4), 600-607.