Just Say No to Drugs

Editor’s note: This article first appeared in Midwifery Today, Issue 76, Winter 2005.
Join Midwifery Today Online Membership

Are mothers more likely to give birth by c-section if they receive an epidural?

With rates for both c-sections and epidurals at all-time highs, many obstetricians have put two and two together and come to the conclusion that the latter contributes to the former—especially in first-time moms.

Rather than coming straight out and saying so (and maybe discouraging mothers from having an epidural), most hospitals simply refuse to give one until a mother is well into active labor—anywhere from 4 to 7 cm dilation. The rationale is that the later a mother gets an epidural, the less time is left for side effects to occur and complicate the normal progression of birth. The American College of Obstetricians and Gynecologists (ACOG) said as much in 2002 when they issued Practice Bulletin #36, advising their colleagues to “delay the administration of epidural in nulliparous women.”

However, if obstetricians suspect an early epidural is a harbinger of the kind of slow-paced labor that often leads to an emergency c-section, anesthesiologists beg to differ. The way one team of anesthesiologists tells it in a recent article in the New England Journal of Medicine (NEJM)(17 Feb 2005), rather than the early epidural per se, the need for pain relief early in labor often signals another risk factor for a cesarean, such as dysfunctional labor; which hurts more, lasts longer and often ends in a c-section—with or without an early epidural.

To prove their point and stop what they see as the inhumane policy of depriving women of their right to an epidural whenever they want one, the Department of Anesthesiology at the Feinberg School of Medicine of Northwestern University in Chicago, Illinois, set up a test trial of their own, dividing 750 carefully screened and matched mothers into two groups: a “test” group allowed to have the first half of a combined spinal/epidural as soon as they asked for pain relief and a “control” group, made to wait for their epidurals until reaching 4–5 cm of cervical dilation or by their third request for pain relief, whichever came first.

In the final analysis, a strange thing happened: The early epidural “test” moms had a slightly lower rate of c-sections than the “control” moms in the late epidural group. Their labor-to-delivery times were also approximately 90 minutes shorter. Perhaps best of all, their babies scored slightly better on their Apgar scores.

These were surprisingly good test results for epidural advocates everywhere, which were proudly presented in the prestigious NEJM and quickly leaked to the public media, which spread the word all over the world under such eye-catching headlines as, “No Gain, No Pain, in Delaying Epidural” (New York Times, 18 Feb 2005), “Numbing News: Early Epidural Not Too Risky” (Ivanhoe News Wire, 17 Feb 2005) and the more down to earth: “Gimme That Epidural” (www.usnews.com/usnews/health/briefs/pain/hb050217a.htm. Accessed 7 Nov 2005).

But was it too good to be true?

Surely breaking news reports don’t lie, but neither do they always tell the truth, the whole truth and nothing but…. Omitted information can be every bit as misleading as misquoted data. So, too, can word choice.

Mothers-in-waiting who read all the nice things being written in the news about receiving an early epidural would be best advised to also read the full report as it appeared in the NEJM, under the more cautious title: “Risks of Cesarean Delivery with Neuraxial Analgesia Given Early versus Late in Labor.”

Those who do so may come to the conclusion (as I did) that if this test trial proved anything at all, it was that all epidural combinations lead to more c-sections, whenever they’re given—early or late in labor. The high rates of c-sections in both the “early” and “late” epidural groups, indicated as much.

Only carefully screened and selected first-time mothers were allowed to participate in this study. No breech babies, multiples or mothers with diabetes mellitus or other conditions were studied. Nevertheless, 17.8% of the women in the early epidural group and 20.7% in the control group delivered their babies by c-section. Although the difference was deemed statistically insignificant, the c-section rate for each group was unusually high for healthy, first-time moms.

This three-year study began in November 2000, at a time when the primary c-section rate for all American mothers—with and without complications—was 16.9%. Why would such “cream of the crop” nulliparae end up having so many c-sections? What common risk factor did they share, other than epidural anesthesia?

None, although a good look at the data in the full report shows that many differences existed between the two groups and even between the types of epidurals they received. Could some of these differences explain why mothers in the “early” epidural group fared better than mothers in the “late” group?

To begin with, mothers given a late epidural began labor in a sorrier state than moms in the early epidural group. More of the mothers in the late group checked in at the hospital with unforeseen complications such as pre-labor ruptured membranes, which eventually led to induction—by itself a recognized risk factor for a c-section. More of them received oxytocin and at higher infusion rate than mothers in the early epidural group.

If (as the authors suggest) the greater pain of a dysfunctional labor is truly why most mothers take an early epidural, this study certainly didn’t prove it. On the contrary, data listed in the full report show it worked the other way around: Mothers who did not take an early epidural were quicker to ask for pain relief than those who did.

Data presented in this article show that a full 42% of the mothers in the “late” epidural group requested analgesia before reaching even 1.5 cm dilation—significantly more than the 30.9% who did likewise in the early epidural group. Only 27.3% of “late epidural moms” waited for pain relief before reaching 3 cm or more. This is significantly less than the 33.6% of the early epidural moms who managed to get that far without asking for pain relief.

Could the harder, quicker onset of pain in the “late epidural” moms have been an early warning sign of dysfunctional labor? If so, was dysfunctional labor a reason for their higher rate of c-sections? Would that also explain their surprisingly longer labors? More painful, dysfunctional labor tends to drag on and on.

Taking the authors’ own “early pain = early epidural = more cesareans” hypothesis a step further, might the shorter, less painful labors in the early epidural group (prior to taking an epidural) also explain their lower rate of c-sections?

Assuming that it did, this would still not explain why more mothers in the early epidural group needed mechanical assistance to give birth than mothers in the late epidural group—the ones with the longer, more painful labors. Usually the opposite is true: Longer labors are more likely to end with a mechanical delivery.

Nevertheless, data published in NEJM clearly showed that as many as 13% of the 364 mothers in the late epidural group required mechanical assistance versus 16% of the 336 mothers in the early epidural group. Statistically, these figures are not bad for first-time moms on epidural anesthesia (who tend to need outside help during delivery four times as often than non-medicated moms), but odd nevertheless, considering that mothers in the early group spent an average of 90 minutes less time in labor than mothers in the control (late) group.

Ninety minutes is a lot less time to a woman in the throes of childbirth; yet apparently for many of the mothers in the early group, their reportedly “shorter” labors were not short enough to avoid the trauma of a mechanical delivery.

So what was the rush? Were some of those shorter labors (in the “early” epidural group), labors suddenly cut short by fallen fetal heartbeats that dipped too low and lasted too long to be ignored?

If so, the authors didn’t say. They didn’t have to. Only “indications for cesarean delivery” were a prespecified secondary outcome—not indications for assisted deliveries. Simply stated, what mattered was not why or how urgently a baby needed to be born but only by which mode of delivery: spontaneous, assisted or by c-section.

The “primary outcome” of this study was “method of delivery.” In the final analyses, when counted and matched in both groups, all c-sections were considered equal, regardless of fetal condition or outcome, giving a slight edge to the early epidural group with their higher rates of instrumental deliveries and fetal heart rate decelerations—but fewer c-sections!

The authors did write that babies in the early epidural group developed “persistent variable and late fetal heart-rate decelerations” more often than their peers in the late group, yet surprisingly, they also stated that bradycardia never led to an emergency c-section. Did it (bradycardia) lead to a quick birth by vacuum extractors and forceps? More importantly, who got to decide which baby would be delivered which way?

For all the publicity generated by this “landmark” study, it was not a blind study. Simply put, those physicians who decided which babies to deliver quickly either by instrumental assistance or a c-section, and which mothers could be given a bit more time to labor, knew exactly which mothers belonged to which group. Could some, unconsciously or consciously, have been rooting for the early epidural group to come out ahead? The authors of this study think not and make a point to let us know it by writing: “it is unlikely that knowledge of the type of analgesia biased obstetricians’ decisions regarding the method of delivery.”

Is it?

In a medical trial supported by their hospital’s own department of anesthesiology, might not some participating obstetricians have been just a little bit slower to pick up a scalpel, or a bit faster to reach for forceps and vacuum extractors, in the early epidural test group?

No one will ever know for sure, but one thing is certain, the “early epidural” boldly printed in many newspaper captions was not an epidural at all, but a faster-paced, opioid-based spinal, where a single intrathecal injection of a lipid-soluble opioid is first injected into the mother’s spinal fluid to be followed up with a longer lasting low-dosed epidural. This is a kind of “double hitter” technique for squelching pain faster and longer with higher doses of narcotics and lower doses of anesthetics, yet with all the risks of both. Of course, to most mothers (and reporters) an epidural by any other name: “combined,” “walking,” intrathecal, “mobile,” “low-dose” or “neuraxial” is still only an epidural; but a fetus can tell the difference.

So, too, can those few researchers who have taken the trouble to compare the newer generation of low-dosed epidurals (such as the combined epidural used in this study) with older, standard epidurals, for fetal side-effects.

Data compiled in 2002, from 24 trials involving 3513 women, (1) found that all the newer, lower-dosed combinations of epidurals led to more cases of fetal bradycardia (decreased heartbeats). These were not just the usual dips and dives that often show up on a fetal monitor during labor, but deeper, longer decelerations with a flatter line, that fell a bit too low and lasted a bit too long to leave any doubts that a fetus must be born—at times right away!

The good news rumored about these new and improved “designer epidurals” is that they don’t slow down dilation the way older, more traditional epidurals do, which is a big plus for an anesthesiologist defending an early epidural. Especially because the current edition of Williams Obstetrics (22nd) lists many studies whichclearly show that epidural anesthesia slows first stage labor by as much as one hour when compared with Pethedine, another drug with its own slowing-down propensities.(2)

The bad news about the new “low-dose” epidurals is that narcotics and babies don’t mix well, regardless of where they are injected in a mother. Given alone as a simple low-dose infusion, or in combination with a short, one-shot spinal first (as in the early epidural group of this study), the basic formula for all “low-dosed” epidurals drug “mixes” is the same: Lower the anesthetic and add more narcotics. Every anesthesiologist is free to design his or her own mixture (hence the term “designer epidurals”), but the tricky part is always finding just the right amount of narcotic (opioid) to keep the mother pain-free without sending her baby’s heart beat into a free fall.

Newborns and narcotics also don’t go well together. In the well-known COMET study (involving 1054 moms who had taken some form of epidural for labor), low-dosed babies fared worse than babies whose mothers were given a more traditional epidural.(3) Newborns whose moms had taken the simpler low-dose infusion epidural were the most likely of all those tested to need high-level resuscitation—probably the worst thing that could happen to a baby at birth. On average, all “low-dosed” epidural babies scored lower than the others on their APGAR scores.

Why then, did they do better now, in this study?

Apgar scores weren’t especially high in either group, early or late—not surprising for babies born under the influence of drugs—but why would those exposed to the more risky low-dose combined spinal/epidural (with more narcotics) have scored slightly higher than those given a more old-fashioned epidural? They never did in past studies.

Could it be that this time low-dosed epidural babies looked better only because babies in the control group looked worse? Then again, could some of them have, in fact, been early epidural babies?

According to the rules of this study, late epidural moms were also eligible for an early epidural if they asked for pain relief a third time before reaching the designated 4–5 cm of dilation. How many of them did just that?

Quite a few, judging from the figures provided with these study results. Apparently, as many as 12.0% of mothers in the late group complained of pain a third time and received an early low-dose epidural before reaching as little as 2 cm of dilation. So too, did another 17.3%, when just past 2 cm but still below 3 cm. No figures are given for those who were given an early epidural at between 3–4 cm but simple math shows that at the very least a full 29.3%—almost a third of mothers in the late epidural group—were given an early low-dose systemic epidural. Add to that the 11 mothers mistakenly given a spinal/epidural in this group (and surely a few of the 70.7% listed as having received some form of epidural at 3 cm dilation and above), and it becomes apparent that many of the late epidural babies with the higher rates of c-sections, were in fact “early epidural” babies.

Of course, all mothers who asked for analgesia in the late epidural group a first and second time before dilating to 4–5 centimeters were given hydromorphone to tide them over until they qualified for a more traditional late epidural.

Hydromorphone is a powerful drug for a baby. The popular Physician’s Desk Reference states that it is contraindicated for labor and birth. In animal and human tests hydromorphone, a synthetic derivative of morphine (more potent than the real thing), had a slowing effect on contractions, making it a poor choice (of drug) for a study designed to test whether early epidural placement does or does not slow labor and lead to more c-sections.

The authors of this study are especially proud of those surprisingly shorter labors in the early epidural group, pointing to them as being, “clinically the most important finding of this study.”

In retrospect, one might question whether labors ended up being approximately 90 minutes shorter in the early epidural group only because they were prolonged by 90 minutes or so in the late epidural group? Not simply because of their ostensibly higher rates of dysfunctional labor but because they were medically slowed, not once, but twice: first by the slowing effects of hydromorphone and then again by the slowing down effects of epidural anesthesia, given late in labor.

If so, what did this study prove—or disprove?

End points for any medical study are clearly defined well in advance. Researchers know exactly what they are attempting to prove. This particular study focused on determining whether women who received a regional anesthetic early in labor were more likely to have cesareans than those who received an epidural later in labor (most assumedly due to stalled labor progress). The study was not designed to determine whether epidurals, per se, lead to more c-sections or if when given early in labor, a “low-dosed” epidural (less likely to slow down dilation) is safe for a fetus.

To find the answers to those all important questions, researchers will have to methodically compare mothers who receive an epidural, at various times during labor, with those who take no medication at all. That is the way the Food and Drug Administration (FDA) does it. New drugs are tested against a placebo—not against other drugs.

Until a proper study is done, mothers in labor would be wise to hold off from taking any kind of epidural—at any time.


  1. Mardirosoff, C. 2002. Fetal bradycardia due to intrathecal opioids for labour analgesia: a systematic review. BJOG 109(3): 274–81.
  2. Cunningham, G., et al. 2005. “Selected Labor Events in 2703 Nulliparous Women Randomized to Epidural Analgesia or Intravenous Meperidine Analgesia.” In Williams Obstetrics, 22d ed. p. 485. New York: McGraw-Hill.
  3. Comparative Obstetric Mobile Epidural Trial (COMET) Study Group UK. 2001. Effect of low-dose mobile versus traditional epidural techniques on mode of delivery: a randomized controlled trial. Lancet 358(9275): 19–23.

About Author: Esther Marilus

Esther Marilus was trained as a pre/postnatal therapist and childbirth educator by the Swiss Society of Professional Maternity Gymnasts. Her experience with emergency births led to training as an EMT and becoming director of a private ambulance corps in Zurich. Esther is currently working on a book on exercising to correct many childbirth complications, from misplaced umbilical cords to malpresentations. She also lectures, trains and councils labor coaches.

View all posts by

Skip to content