Induction and Circular Logic
Editor’s note: This article first appeared in Midwifery Today, Issue 63, Autumn 2002.
Join Midwifery Today Online Membership
It’s been said that induction has reached epidemic proportions in this country, but I dispute that statement. Induction is so common that it is no longer of “epidemic” proportions; it is now “pandemic.”
Recently collected data on interventions in labor shows a “dramatic rise” in inductions during a 10-year period. Induction of labor doubled between 1987 and 1997, with the rate jumping from 9% of births to 18.4%! Meanwhile, the rate of “chemical stimulation” increased from 11% to 17.4%. Over one-third of women had their labors induced or augmented, but the researchers say this number is low and that induction of labor is frequently underreported. The data also shows, surprisingly, that midwives are inducing as often as doctors.
I’m a midwife with an innate faith in childbirth as a normal, natural function. When I hear that one-third of American women are given chemicals to start labor, I have to either conclude that women have somehow lost the ability to give birth or that we are witnessing a societal change. If women are being induced for the legitimate reasons of health and safety, then mortality and morbidity statistics should be improving. Yet the statistics are quite flat. We see little change in US statistics, except for in the category of tiny, preterm babies. An induction and augmentation rate of over 35% has not seemed to improve the health of mothers or babies. It has, however, strongly impacted the cesarean rate and—subsequently—the rate of VBAC and uterine rupture after prior cesarean section.
Induction is so common that many people are unaware of the risks. Even a “simple,” uncomplicated induction can begin an avalanche of interventions. It often starts with a cervical stretch and sweep to “ripen” the cervix, IV Pitocin, electronic fetal monitoring (EFM) and amniotomy; then, perhaps, it’s on to an intrauterine pressure catheter, amnioinfusion for unusual fetal heart tones, an epidural for the pain of Pitocin-induced contractions and malrotation or poor descent because of the epidural; then maybe a vacuum extraction or cesarean is performed for “failure to progress.” It goes on and on. The mother ends up with lifelong injury to her uterus. Her baby may be stressed and separated from the family. A normal birth may turn into a nightmare. And that’s if all goes well! If there are complications or a surgical emergency, then the nightmare really begins.
Induction is a minefield, a setup for complications. An induced labor forces the baby out before the body is ready before the complex hormone interaction has primed the cervix and often before the baby has reached his full intrauterine maturity. We have drugs now that can produce contractions and soften the cervix, but this is only a small part of the complicated process of labor. We can make a woman have contractions but we don’t always succeed in forcing her body to release the baby and give birth. If we start a labor with chemicals, we may very well have to finish the labor with the surgeon’s scalpel.
In some studies, induction raises the risk of cesarean by 800%. EFM must be used in all chemical induction methods because of the risk of hypertonic contractions and fetal distress. Electronic fetal monitoring alone increases the risk of cesarean and of vacuum extraction or forceps. Amniotomy increases the risk again. Cesarean for fetal distress is even more common—whether the distress is real or a result of EFM artifact—since non-reassuring fetal heart tones are frequently observed. Meconium staining, meconium aspiration syndrome and even shoulder dystocia are directly associated with inductions. The rise in induction closely mirrors the rise in cesarean delivery, as does the rising incidence of post-cesarean rupture. A woman with a prior cesarean is unlikely to suffer a uterine rupture (odds are usually given under 1 percent). But if she is induced, her risk may rise to 2–4 percent.
If the data shows that induction is a risky procedure, and we see little statistical benefit, then why are we inducing so often? Doctors and midwives will express many reasons for induction, but many of those reasons are colored by a misunderstanding of the risks involved. The risks created by induction are sometimes ignored—induction seems simple and easy. Any complications or problems are seen as simple chance—the “normal risk” of birth—caused by the situation that prompts the induction.
There are several reasons for induction. Some of them are valid and some of them have little data to support their validity. Intrauterine growth restriction (IUGR), placental malfunction, oligohydramnios, fetal distress, infections, deteriorating preeclampsia and diabetes are often good reasons for inducing labor, but these conditions are rare. Other reasons for induction are more common and the arguments vague, based on controversial data or a misunderstanding of the data.
Labors may be induced in an attempt to avoid “macrosomia”—a baby who is too big to be born safely. Yet study after study shows this not effective. Inducing to avoid macrosomia merely substitutes an induced labor for a spontaneous labor and raises the risk of cesarean section. Labors induced for “macrosomia” show a cesarean rate twice as high as those allowed to begin spontaneous labor, even if the babies are bigger. Allowing an extra week or two of growth will not allow the baby to increase enough to be obstetrically significant. And waiting for spontaneous labor allows the greatest likelihood of a vaginal birth (1).
Postdates and prelabor rupture of membranes are the most common reasons for homebirth midwives to induce labor. And yes, even homebirth midwives are inducing labor. In fact, national statistics show little difference between the induction rates of midwives and those of doctors. Sometimes we are forced by our state protocols to initiate labor even though good evidence shows that women with prelabor rupture of membranes can be managed conservatively and monitored for at least several days without raising the risk of infection. Waiting for natural labor to begin is more likely to result in a normal vaginal birth than is stimulating labor. But some state boards don’t allow the luxury of waiting, instead mandating transfer of care unless labor can be initiated with home methods. (If we disagree with the state policies the issue should be addressed to the state agencies that oversee licenses or certification.) However, these “protocol-forced inductions” are a minority of cases.
Postterm is the big issue, the great fear factor that controls obstetrics and midwifery today. Is there reason to be concerned if labor hasn’t yet begun weeks after the due date? Yes, of course. Babies do have a timing of maturity. The placenta does seem to slow function after a certain gestational week. But that timing is not necessarily related to the calendar. The placenta can begin to fail at any point in pregnancy. Placental insufficiency can result in poor fetal growth, damage to the baby’s organ systems and even eventually in death. This is rare but we watch for it at every stage of pregnancy. In a perfect world we pick up on these cases and act before the baby’s reserves are drained. We induce labor when needed—even facilitate and advise a cesarean without labor—if the baby is in trouble. But how common is this trouble? And how likely is it to occur as the due date passes? The data might just be surprising.
The U-Shaped Graph
I imagine most of us have seen a very famous, deeply curved graph demonstrating the rate of perinatal mortality by week of gestation. It starts high on the left side of the chart at 37 weeks and begins to drop, hits the nadir at 40 weeks, and rises again, showing a swift rise on the right side, to almost double, at 42 weeks. It is a frightening chart and everyone who sees it is impressed with the danger of any pregnancy going past “term.” That U-shaped curve is a strong visual teacher!
Yet, when I researched that famous chart, I was very surprised at what I found. This chart comes from a study published by McClure-Brown in 1963. Yes—1963—from data gathered in 1958. This is the study that originated the phrase “Stillbirth doubles at 42 weeks.” But did you know that modern studies don’t show this at all? Modern studies completely contradict that famous graph. (See the included chart and table.)
In the statistics published in 1963, the “line” is about a 35° rise. In the 1982 statistics, it is so flat the rise is almost imperceptible to 42 weeks, then there’s about a 15° (very slight) rise from 42 to 43 weeks, and a steep close to 45° from 43 to 44 weeks. That famous U-shaped graph only works if you use the McClure-Brown statistics. All other modern data contradicts it. Perinatal mortality does not double at 42 weeks; it rises almost imperceptibly and doesn’t show a steep rise until after 43 weeks, not 40 weeks.
I can’t think of any other instance where 60-year-old data is used to support modern methods, especially when newer, carefully collected data contradicts the older data. If we accept this U-shaped graph, we would also be accepting a 20 in 1000 mortality rate! Not so. I think the McClure-Brown data must have reflected something else in older obstetric management. During the 1950s almost all births were conducted under general anesthetics and the majority of babies were delivered by forceps. Perhaps McClure-Brown data showed that older or larger babies were more likely to be damaged by that type of obstetric management? I don’t know for certain what made the McClure-Brown data so different from today’s statistics. I just know it does not apply. There is no “U-shaped” graph of fetal loss at term. The truth is that perinatal mortality is an almost flat line from week 38 to week 43.
There is a sort of creeping overreaction in dealing with postdate pregnancies. It is true that the stillbirth and fetal distress rates rise sharply after 43 weeks, even though it is also true that the vast majority of babies at 43 weeks are still fine and healthy. We should “react” to this rise by monitoring postdate pregnancies carefully. However, it is an overreaction to see the rise in problems at 43 weeks and apply that risk to 42 weeks and then to 41 weeks. Postmaturity syndrome is a continuum (one that affects less than 10 percent of babies past 43 weeks). It becomes more likely as weeks progress past the due date but does not start on the due date. A baby is not at risk for being postterm until she is postterm.
If a person accepts the false data from McClure-Brown, then they will believe that perinatal mortality “doubles by 42 weeks,” and they will be anxious to conclude pregnancies well before that date. Others will use a new study comparing induction at 42 weeks with induction at 41 weeks (showing a lower cesarean rate) as a jumping-off point to reason that inducing at 41 weeks (or earlier) is better than spontaneous labor at a later date. This is not what the study shows. It merely shows a slightly lower cesarean rate when comparing induced labor to induced labor. It does not compare induced labors to spontaneous labors. Large studies have been done on that question and show rather clearly that waiting for spontaneous postdate labor results in fewer cesareans without any rise in the stillbirth rate (2, 3).
One study stands out in particular. This was a retrospective study of almost 1800 postterm pregnancies with reliable dates—compared with a matched group that delivered “on time” (between 37 and 41 weeks) (4). The perinatal morality was similar in both groups (0.56/1000 in the postterm group and 0.75/1000 in the on-time group). Perhaps contrary to our expectations, the rate of meconium, shoulder dystocia and cesarean were almost identical. The rate of fetal distress, instrumental delivery and low Apgar was actually lower in the postdate group than in the on-time group. This is only one of several studies showing postdate pregnancies can be safely monitored until delivery. The key to this safety is, of course, “monitoring.”
This brings us to another issue: monitoring the postdate pregnancy. Some doctors and midwives would rather induce than incur the expense, worry and hassle of monitoring a postdate pregnancy. But how are they monitoring? Some start at 40 weeks plus 4 days and many start at 41 weeks. When should they begin? They should begin biweekly monitoring of a normal pregnancy at 42 weeks, according to the American College of Obstetrics and Gynecologists (ACOG) practice guideline (5). Biweekly monitoring before 42 weeks is a needless exercise, as the risk of problems before 42 weeks is very low.
Postdates is an almost overwhelming issue. There is a world of data and a world of fear. If we are going to carefully study the available data to help us navigate this uncertain landscape, then one of the first things we should do is agree on the definition of postdates, the likelihood of postdates and the risk of postdates.
We should also agree on the importance of getting good dates—and of adjusting the due date according to the mother’s individual menstrual cycle. Since data shows that risk doesn’t rise appreciably until after 42 weeks, I would suggest using the “classic” definition and consider pregnancies of 37 to 42 weeks to be “term,” and use “postterm” to describe pregnancies that continue past 42 full weeks. By that definition only about 3 percent of women will go “postterm,” meaning they will actually reach 42 weeks.
Forty-two weeks is a sort of magic number, the point at which ACOG suggests we begin fetal surveillance testing. Some practitioners would like to induce at 42 weeks and perhaps they can find some evidence to support the intervention. However, very few women are actually induced at 42 weeks. They are induced before 42 weeks in order to “prevent” them from becoming postdate. In other words, they are subjected to a risky medical intervention in order to prevent them from reaching a point in pregnancy—a point that is still low-risk, and one that they are unlikely to reach anyway.
Some women are not actively induced, but are given serial cervical sweeps, or instructed to use herbs, evening primrose oil or other folk medicines to “prevent” postmaturity, even though 97 percent of them will deliver before 42 weeks without any gadgetry or interventions.
This is circular reasoning. It’s as if I were to push you down the stairs in order to prevent you from falling down them. Then again, circular reasoning is common in obstetrics—such as the idea that we should do an episiotomy on all women because some women will tear and need stitches. Or when a newborn baby is taken away from his mother’s arms and put under a radiant heater to keep him warm when he wouldn’t need to be kept warm by the heater if he had been allowed to stay in his mother’s arms. Or a baby who refuses to nurse because he has already been given sugar water in the nursery and is then given more sugar water since he won’t nurse. Circles!
When we worry unnecessarily about the woman who has passed her due date but is not yet in the realm where worry is reasonable, we are likely to let our fear push us into interventions that are not proven to be effective and that may in fact be harmful. Inducing labor without reason is risky. The most frequent problem we see with home inductions is slow labor, “failure to progress,” maternal exhaustion, possibly fetal distress, and postpartum hemorrhage. These risks that accompany induction must be lower than the risk of continuing the pregnancy for induction to be a reasonable option. The only way we can tell the true risk vs. benefit is by examining each individual pregnancy, each mom and baby, as a unique condition with their own timetable, their own growth charts, their own time of maturity. Each baby must be watched for all the reassuring signs that he is thriving or for any of the subtle signs that he may soon get into trouble. However, we should be doing this at every point in gestation, not just because the due date has come and gone.
If we could only begin to see women as unique, rather than as a set of “statistical risks,” we would be on the path to understanding pregnancy and childbirth as the complex miracle it is. And we would learn to respect its cycles and its powers; we would learn to work with natural mechanisms rather than trying to intervene and control them. We have a birth industry in this country that is unable to help nearly one-third of its women achieve birth. These women are induced or have their labors stimulated with chemicals or surgeons who cut their babies out of them. Yet babies in this country are no healthier than they used to be a few years ago, when most of their mothers were able to give birth without drugs and instruments. Pitocin is now one of the most widely used drugs in the US and induction and augmentation of labor are common medical procedures. Yet it isn’t working, is it? It isn’t an accident that induction and cesarean mirror the same rising line on a graph. The induction pandemic is linked to the cesarean epidemic. Induction of labor needs to be exposed as an intervention that is overused, dangerous and not even effective at achieving its stated goal—successful childbirth. In fact, if a woman wishes to achieve a successful birth, then a chemical induction of labor is the one thing she should strive to avoid!
Maybe we just have trouble seeing the forest for the trees. Induction is easy. It is seductive. It seems to take all the worry out of pregnancy. We create such fear in modern women that most of them are relieved when they finally go into labor. We teach them through subtle means that their healthy babies could suddenly die, that between one heartbeat and the next, the heart could stop. We teach them that the womb is a very dangerous place. Isn’t this the true, albeit unconscious, message of electronic fetal monitoring? Of all our tests and scans? We say we do a test to “make certain everything is OK.” But the unconscious mind supplies its own interpretation. It says we do the tests and the scans to discover the terrible thing that is wrong. And when a woman nears her due date, we bury her under “informed consent,” listing every possible disaster that could occur, no matter how unlikely they actually are. And when we start to discuss the risks of postmaturity—risks that are lower than we perceive—it’s no wonder that women jump at a chance to be induced. Women worry about their babies, and we are responsible for much of that worry. Worry and fear become the primary motives for inductions and other interventions.
I always say that worry is not healthy for pregnant women. But it isn’t healthy for care providers either. We need to learn to face our fears, to put “risks” in their correct perspective, to learn the true risks rather than the rumored risks and to rightly decide when an intervention is truly warranted.
- Combs C.A., N.B. Singh, and C. Khoury. 1993. “Elective induction versus spontaneous labor after sonographic diagnosis of fetal macrosomia.” J Obstet Gynecol 81 (4): 492–6.
- Matijevic, R. 1998. “Outcome of post-term pregnancy: a matched-pair case-control study.” Croat Med J 39 (4): 430–4.
- Yeast, J.D., A. Jones, and M. Poskin. 1999. “Induction of labor and the relationship to cesarean delivery: A review of 7001 consecutive inductions.” Am J Obstet Gynecol 180 (3, pt 1): 628–33.
- Weinstein D., et al. 1996. “Expectant management of post-term patients: observations and outcome.” J Matern Fetal Med 5 (5): 293–7.
- American College of Obstetricians and Gynecologists. October 1997. Management of postterm pregnancy. 6:6.