Autism and Induced Labor

A recently published epidemiological study in JAMA Pediatrics looked at the association between induction and enhancement of labor and the risk of autism. The researchers found a positive association, especially with males. The study has been variously reported in the popular press with causal interpretations not justified by the data.

The study itself is very robust – the authors looked at 625,042 live births, including 5,500 children with a diagnosis of autism. They found:

Compared with children born to mothers who received neither labor induction nor augmentation, children born to mothers who were induced and augmented, induced only, or augmented only experienced increased odds of autism after controlling for potential confounders related to socioeconomic status, maternal health, pregnancy-related events and conditions, and birth year. The observed associations between labor induction/augmentation were particularly pronounced in male children.

Although this is a large study, it is one study, and so the correlation needs to be independently confirmed. But if we assume the correlation is accurate, the next question is – what is the arrow of causation? Observational studies can only indicate an association. By themselves they cannot prove causation, although multiple observational studies may be able to triangulate to the most likely causal interpretation.

There are always three generic interpretations when A is associated with B: A causes B, B causes A, or C causes both A and B.

In the case of this study, either inducing/augmenting labor triggers autism in some children, children with autism are more likely to require induced labor, or some other factor(s) is a risk factor for both developing autism and needing to induce or augment labor. This current study does not contain data that can differentiate among these possibilities.

We can, however, explore plausibility. Inducing and augmenting labor includes a variety of generally safe procedures and medications. An unknown medication side effect, however, is always possible. Anything that adds stress to a newborn can also plausibly be a trigger for autism.

Twin studies and other genetic studies suggest that autism has a large genetic predisposition, but that genes are not 100% of the causal factors for autism. It may be that we simply do not yet have the full genetic picture. But it is also likely that the environment in the womb interacts with the genetic predisposition in determining the development of autism.

The environment of the womb could include delayed delivery, which results in induction, or some other factor that delays delivery. It is therefore plausible that either the delay in delivery itself, some other factor that delays delivery, or the procedures used to induce or augment labor when delivery is delayed can (separately or together) be an environmental trigger in genetically susceptible children.

It is also plausible that children who will become autistic may also be slow to deliver – that the delayed delivery is a symptom of, rather than a cause of, autism.

Evidence is mounting that autism begins in the womb, and not just from genetic evidence. Several studies, although individually small, have all shown that children with autism have significantly more brain cells, especially in their frontal cortex.

Other studies indicate that people with autism have abnormal networks in the brain, and specifically there is a deficit of connectivity between different parts of the brain, especially interhemispheric connectivity (the two sides of the brain do not talk to each other as robustly as a typical brain).

Clinically, the features of autism appear to be present starting at about 6 months. Such early presentation of clinical features is consistent with autism being a congenital disorder (i.e. present at birth).

Further, the specific changes in the autistic brain that research so far is demonstrating – changes in neuron number and connectivity – are the kinds of changes that occur during fetal development.


While there is still room for much further research, all the information we have at this time suggests that autism is something that occurs as the fetus is developing the womb, is primarily genetically determined, but may also be determined by factors present in the environment of the womb. The latest study identifies delivery as another possible factor, but it is not clear what role it is playing, and delayed delivery may simply be another symptom or marker of the underlying cause(s) of autism.

All of this is incompatible, of course, with theories about environmental triggers that cause autism in children older than 6 months. The MMR and other vaccines come to mind.

Posted in: Neuroscience/Mental Health, Obstetrics & gynecology, Vaccines

Leave a Comment (42) ↓

42 thoughts on “Autism and Induced Labor

  1. Brett McCoy says:

    Very interesting. My son is autistic, and was an induced labor (my wife has a blood-clotting issue and the entire process had to be very controlled).

  2. WilliamLawrenceUtridge says:

    I don’t know nearly enough about pregnancy and labour, but I do know that autistic children, on average, tend to have larger heads than nonautistic peers. Can that delay labour? For that matter, what sorts of conditions can delay labour? I’m guessing it’s quite a list.

    Plus, Dr. Gorski has discussed diagnostic substitution quite frequently, where “autistic” is preferred over “developmentally delayed” due to greater access to resources. I wonder if the higher rates of autism are seen in the less severe end of the spectrum, where the symptoms of autism shade towards more general cognitive issues, which I could see being caused by problems during birth.

    Or, perhaps the drugs used in induction and augmentation are one of the environmental triggers for autism. That would suck, parents will have to play chicken with “do I want an autistic baby” or “do I want a dead baby”, and feel guilty no matter the outcome.

    1. lilady says:

      IIRC, I read a study or two ~ 7-8 years ago, about the head circumference growing unusually fast during the first year of life i.e. 25th percentile for an autistic child at birth, then 75th (or more) percentile at one year old. I was “impressed” that it could be an excellent early diagnostic tool for babies where there were strong suspicions that the child was showing ASD signs. (No guesswork there, because head circumference, length and weight are plotted on graph paper forms kept in the infant’s chart).

      It turns out that those studies were incorrect (except for identified genetic disorders), according to this recent SFARI article.

      1. Chris says:

        I blame the issues I had with my first birth was marrying someone with a huge Dutch head. Fortunately the last two took more after me. ;-)

        Though in reality, there was not indication of labor after 42 weeks, so induction was part of the protocol, and his big head had to be guided out with forceps. The second would have been inducted at 41 weeks, but he came before his predicted birthday. The third went to a beat of her own drummer, and I am just glad I was not at home.*

        * I don’t care want home birthers say, I know that I would be expected to clean up afterwards. Number three was almost born at home, and, yeah, I know I would have had to clean it up!

  3. Emily B. says:

    I look forward to seeing more on these studies. My son is also autistic, and I was also induced, then given more of the drug to speed the process when labor was slow to progress.

  4. calliarcale says:

    I seem to recall reading that autism was more prevalent in postdates babies — and we already know there is a remarkable correlation to head size as well. Both of these things tend to encourage hastening of labor, because you don’t want the baby to get any bigger, and part of me wonders whether that is at all relevant. It certainly seems plausible.

    The decision to induce should never be taken lightly. Some do in order to schedule the birth for a convenient or auspicious time, but that’s not sensible, IMHO. I’d like solid numbers on that practice (truly elective scheduling), but I suspect it’s not that much of a problem. I know when my first labor was induced, I’d been having false labor pains for two weeks, basically since my due date, and my baby was getting rather quiescent. Also, I’d started to dialate, so they decided it was time to encourage things. She ended up being an emergency c-section when her heart stopped. She’s had no heart problems since then, interestingly enough. But she is autistic. And she has a very big head; she can wear adult-sized hats already, at age 10. (I have a big head too, as does my father; hat-shopping is frustrating for us.)

    So I don’t think parents should let this factor into the decision of whether or not to induce. Not yet, anyway. It’s already a complicated decision, and like most labor interventions, should be based on the specific medical situation. And you need to consider too that the general alternatives to induction are not at all without risk themselves. If the baby absolutely must come out now and we can’t wait any longer for the labor to get going on its own, the alternative is c-section, and the numbers suggest that is inherently riskier. Faster, though, so in a dire emergency it’s preferable to everything, but before saying no to induction, you have to consider that you will next have to decide whether to risk waiting longer, or risk a c-section. (I’ve had two c-sections. They’re actually not bad at all — when everything goes right. There are horror stories of when they go wrong, which can happen. It’s major surgery; it’s sort of a big deal.)

    1. Young CC Prof says:

      Looking at those truly elective inductions is an interesting idea, but they ARE very rare, and the families who make that choice may be sociologically unusual in some way. I really don’t think one could find a large enough sample to effectively study the autism rate, since autism is only about 1% of the population under current conditions.

  5. My son is also Autistic and was an induced labor. I understand the correlation, but the causal link is unknown. The drugs used during induction are an unlikely cause for the abnormalities as they are synthesized naturally-occurring hormones and because there exists thousands of other traits which could cause abnormalities that led to the induction being required.

    The causal link will be more than likely linked to another trait of the mother/baby which required her to be induced – such as preeclampsia or simply high blood pressure. (Please understand that I made up those two examples as links to Autism – I have no data or understanding of the two).

    One last note the “synthesized naturally-occurring hormones” which I mentioned:
    This means that the hormones are naturally occurring but were not extracted from a living host. synthesized “naturally-occurring anything” is chemically similar or identical to “the real thing”.

  6. windriven says:


    “That would suck, parents will have to play chicken with “do I want an autistic baby” or “do I want a dead baby”, and feel guilty no matter the outcome.”

    Caesarian section would appear to be a better choice for at least some of these parents.

    1. Meghan says:

      Sections aren’t without risk, to the woman and the infant. We’re also very quick to deliver for a “big baby” that ends up being of perfectly average size (ultrasound isn’t particularly accurate at weight estimation), or for other reasons. ACOG considers expectant management to be acceptable post-dates management, and they’re hardly proponents of low-intervention birth.

    2. When I had my second child I actually elected to do a c-section and refused induction after my due date had passed and I did not have a single contraction or sign of labor. She is doing amazingly well in contrast to my son who is diagnosed with autism. I was induced with him, spent 2 days in labor, ended up in emergency c-section and he came out not breathing, limp, white and a low pulse. I made that choice with my daughter because I always regretted the stress I placed my son under attempting to have a ‘natural’ birth.

  7. rork says:

    I’ll note that earlier term babies had more autism, not less (in itself, this doesn’t disprove some of Novella’s conjectures though). Gestation <34 weeks had about the same effect size as induction.
    I also worry that the lay press will fail to notice that being non-hispanic black, older, college educated, or having diabetes gave odds ratios bigger than induction. Year child was born was also very significant (Before 1996 worse than after). Some of those things could (easily) be confounded with induction, but they were in the model, so you'd hope it was accounted for properly, but there could still be something funny under the hood (like these things having interactions that we don't model).
    Oh, mother smoking gave reduced odds.
    My point is just that induction is not going to explain that many of the cases – I don't think they did the math to try to make that clear.

  8. Thanks for the very timely article, my daughter was induced nearly 2 years ago, and I feel better prepared to answer the questions that will undoubtedly come as this story gets more and more attention.

  9. inconscious says:

    I’m really glad SBM wrote something in response to this study so quickly.

    You make some very good points that the media are not always quick to follow up on when reporting something like this – though at least NPR did showcase a physician that expressed some of the same concerns with interpreting these data in terms of causal effects.

    Speaking of which: Most inductions for labor are done with pitocin (oxytocin). When I first heard about this my first thoughts were along the lines of “how could oxytocin possibly cause autism, especially if given just before birth?” This would seem to follow the whole SBM mantra of trying to find a scientifically plausible mechanism. I’m no OB/GYN, but presumably infants are exposed to quite a degree of oxytocin just before birth ANYWAY as it is released in vivo during the birthing process (thus the use of it exogenously to induce labor). I don’t think anyone seems to be picking up on this line of logic.

    Combined with the necessity to not infer causality based on one association study (no matter how good it is) the above line of logic would seem to refute a plausible mechanism for pitocin/oxytocin CAUSING autism in any way.

    1. rork says:

      They discuss oxytocin. It’s sexually dimorphic, so might help explain higher male incidence. Here are PUBMED IDs of the references:
      15288368, 16884725, 9414463, 17000015, 19845972. I’ve only looked at their abstracts. Last one might be worth reading – it’s about the receptor, and it’s from folks at Duke so maybe the intro or discussion might review the literature.
      (Thankyou moderators…)

    2. @inconscious –

      From the women I’ve known that have had induced and non induced births, the difference between getting pitocin injected and what they were felt during non induced births was quite significant.

      In any case, from a *possible* mechanism of action point of view, we could look (again) to the realm of developmental programming and the evidence that events that occur very early in life are capable of programming malleable systems with lifelong effects.

      In this case, the interconnected nature of oxytocin and the HPA axis may be of interest; both have been observed as altered in the autism population.

      Food for thought.

      1. Calli Arcale says:

        I’ve had labor induced, and had labor start on its own. The main difference was that the induced one got organized much more quickly. Then again, it’s very hard to generalize from any one delivery; the experience can vary tremendously even when it’s the same mother and there are no interventions at all. It really is the exact same hormone as that released during labor, which is why it works, but it’s probably a higher dose, hitting faster. Now, oxytocin has been implicated in autism, but in very unclear ways. (Hasn’t stopped people attempting to treat autism by giving oxytocin to children, though, which I doubt is effective.) There is a suggestion that autistic people may have problems with their oxytocin receptors; this suggests a plausible reason for induction to be more common in the births of autistic children — maybe their mothers have a similar issue, and so they need more oxytocin to get the contractions going.

  10. WilliamLawrenceUtridge says:

    calliarcale, I hope you don’t mind but I chuckled at the image of your family trying on comically-large hats. I felt bad at the same time.

    If it makes you feel any better, I’d read years ago that people with disproportionately large heads tended to be more successful. Oprah? Big head. Tom Cruise? Big head. Disproportionate number of corporate CEOs? Big heads.

    I’ve got a tiny little pinhead, I have head envy. But my anatomical snuffbox is big enough to house a family of four.

    1. windriven says:

      “Disproportionate number of corporate CEOs? Big heads.”


      Disproportionate number of corporate CEOs? Fat heads.

    2. Calli Arcale says:

      We love to tease one another about our huge heads, so no offense taken. ;-) My mom has a very small head, so she has your problem. My 10 year old daughter’s head is already larger than hers. She has a terrible time buying eyeglasses, because in addition to having a child-sized head, she has a severe nickel sensitivity. She requires titanium frames, and those are hard to find in the children’s section. She misses the eighties, when plastic eyeglass frames were all the rage. ;-) Too big, too small; it’s a pity hats don’t have crown size printed inside of them. It would save a lot of time.

      I also have large hands, so I have the same problem glove shopping. (Yet I’m a small woman, only 5’2″, so people don’t expect it.)

  11. Harriet Hall says:

    Thanks for writing about this. I had similar thoughts when I read about the study.
    I hope it won’t unduly frighten pregnant women. There are legitimate medical indications for induction and augmentation, and we need to consider whether NOT using them might lead to worse outcomes in some cases. Common sense (even before this study) tells us to avoid elective inductions and to keep use of medications to a minimum.

    One question: how did the study define augmentation? Did they include manually stripping membranes, artificially rupturing membranes, forceps deliveries, vacuum devices, etc.?

    1. rork says:

      Maybe not clear.
      “Using NCDBR variables for labor induction and augmentation, we constructed a 4-category variable to indicate that during delivery a mother was either not induced or augmented (reference), induced and augmented, induced only, or augmented only.”
      Yes that “(reference)” part is actually what appears. I’m looking at the version:
      Abstract helps with “To examine whether induced (stimulating uterine contractions prior to the onset of spontaneous labor) and/or augmented (increasing the strength, duration, or frequency of uterine contractions with spontaneous onset of labor) births are associated with increased odds of autism.”
      I’ll criticize that as describing intentions rather than acts. I hate that.
      They give examples, saying we need more work about “the specific treatments and dosing used to induce/augment labor (eg, exogenous oxytocin and prostaglandins).”

      Maybe a doc is needed to interpret the writing, or maybe it’s hiding in “reference”.
      (Thankyou moderators, sorry to be a constant source of work.)

      1. Meghan says:

        My biggest problem with the methodology of the study is that the authors did not stratify for induction reason or for the method used for induction. From a theoretical perspective, both of those are very important if one is to make any causative suggestions about induction and autism.

  12. Paul de Boer says:

    Steve says “children with autism have significantly more brain cells, especially in their frontal cortex”. Not to say this directly correlates to a larger head size…


    Many people here have been saying that a larger head would necessitate induced labour in some cases.

    However, this study shows a smaller head size at birth and larger as the child with autism developes.

    And many others fail to replicate this larger head phenomenon.

  13. Thomas E. Myers MD says:

    As a practicing OBGYN, thank you to the author and all commenters for a timely response to the autism/induction paper. Your critical evaluation and interpretation of the data is appreciated.

  14. Shelley says:

    So, since I can’t access the article directly, how large was the relationship and how did they control for other variables?

    1. rork says:

      Not easy to summarize as they fit several models. They are all multivariable logistic regressions, and they fit simple ones with few (or no) covariates, and then add more covariates to fit ever larger models. Chris’s link shows some of the betas and their confidence intervals, and remarks (as I did above) that they aren’t very large (about 1.25, but depends on model, and they also fit models where the betas for male and female are separate). I did like that they gave summaries of results for several of the models – to show the beta estimates were pretty stable (that rather surprised me). That link and my comment above tell just a few of the covariates (mostly, the significant ones).
      Really, I think interested people need to get ahold of the whole thing – too many models and covariates to summarize. Figure out if one of your friends has access, or make a friend of someone who does have access. How else are you keeping up on wolf (or steelhead) genetics and ecology without being able to get at the papers? PS: Emailing the corresponding author for a copy would seem kinda rude to me.

    2. rork says:

      Crap, 1.25 isn’t the estimate of beta, that’s the anti-log of the estimate of beta (euler’s number, e, to the power beta), which is estimating the odds ratio. Sorry.

  15. Iolaire says:

    Well, just for a contrary anecdote, I am on the autistic spectrum, have a small head, was born two weeks before my due date and my mother wasn’t induced. However, she might have wished for her labour to have been augmented as it lasted 48 hours. When i finally emerged my head had been squashed into an unusual pointed shape (the midwife assured my mother that it would soon go back to normal and it did, at least on the outside). I’d always assumed my autism was inherited as there are lots of people with autistic traits on both sides of the family and i just happened to have been born with more than most. Have there been any studies on the relationships between autism and difficult births in general?

  16. daedalus2u says:

    The increase for induced and augmented is 1.23

    What that means is that the average incidence is 5500/625,042 = 0.0088 = 1 per 114 births. 1.23 times that is 1.23 per 114 births or 1 per 93 births.

    I highly recommend the article by Emily Willingham. She points out that the “risk” for autism also correlates with maternal education, with an odds ratio of 1.33 for the mother going to college.

    If you want to reduce autism, don’t let women go to college.

  17. Andrey Pavlov says:

    PZ Myers happened to write about this as well. And made a succinct and accurate point.

    Don’t miss the forest for the tree. Squint hard enough at enough data and you can find whatever pattern you want to.

  18. C says:

    I find this information very interesting. My son is waiting for an evaluation for autism. He was not induced, but delivery was a big deal. He was head-down but facing up during delivery, which made the whole thing long and terrible. This was after 3 days of labor.. When he finally came out after a 2-1/2 hour delivery, his head was badly misshapen and remains slightly so even almost 5 years later. His head was large as an infant, but he has finally grown into it. Still, no medical professional reported any concerns about his delivery or possible damage to his head. It wasn’t until his preschool teachers called us in for a meeting earlier this year that we realized he had so many telling signs of autism.

    1. Chris says:

      Best of luck to you. My son had a big head that got stuck, and then had seizures two days after birth.

      Though one promising sign is that it took a while for others to realize your son had issues. My oldest was evaluated when he was about 27 months for lack of speech. And that was early about twenty two years ago, though his milestones were being looked at closely due his history of seizures. Other parents usually got the “wait and see” approach.

      Now his younger brother also turned out to have a language delay. But it was not noticed until he was three years old by his older brother’s speech therapist when she had them to cooperative play therapy together. My younger son was diagnosed with a language delay, not a disorder like his older brother. It took a year of therapy with student therapists at the local university (an hour or two per week), plus a couple of years with the school district (a half hour per week), and he was within one standard deviation (almost normal) for his age before kindergarten.

      While my oldest still has some issues (even after a decade of speech/language therapy), and gets disability services at a local community college, my younger son is okay. Younger son lives on his own (okay in a house with a bunch of other college students, soon to move into a house with nine other friends*), and is entering his last year of college as a math major on the teaching track, all while working half time as a lifeguard and swim instructor.

      * Yes, you too may experience in a bit over a dozen years the annual summer migration of the college student. This is where you get a call that starts with “Can we borrow the van?” to “Oh no! I need to work today, can you come and pack for me!?” Did you know that male college students actually collect beverage cans and stack them as a wall next to their kitchen? Your mileage may vary, and good luck.

  19. Colleen says:

    Makes me nervous. I had to be induced as I was stuck at 5 cms. They only gave me a lick of pitocin and that did it for me. But I really had no choice. My daughter is 12 months and shows no signs of autism as of now. Plus, yes, what about all the studies on how premature birth causes autism? Maybe both do? Maybe it’s induction PLUS vaccines? In last 20 years there have been many changes that happened that could cause autism: plastics, pollution, pesticides, more induction, more vaccines, lest trees to purify the air, more Perscriptions, more Pharma in the water. Maybe autism is on the rise simply because all of these things are on the rise? Just pondering not making debate. Also I’m curious as to what type of autism the article refers to. Was it regressive autism? Or Aspergers? Or…?

    1. Chris says:

      Or perhaps it is the rise of the internet! That has more plausibility.

      Actually, it is more likely the introduction of the DSM IV in 1994.

  20. M. says:

    Interesting. Though my own anecdotal experience is quite the opposite. My one child who is on the spectrum was not early, nor induced. He was a precipitous labor at 42 w 3 days. No pitocin. He was briefly stuck. His next younger brother was born at the exact same gestation, also precipitous labor, also no pitocin.

    The child with processing disorders has significant differences in physical features from his siblings. He revieved the same care in the home as his siblings (extended breastfeeding, etc) and the same routine medical care from the same doctor, including vaccines. His youngest sibling was born after an induction at 38 weeks, involving pitocin. At 18 months she shows none of the early signs of asd. the child who has it did.

    Obviously our family does not constitute a proper study. But the combination of existing information and my own experience makes me lean heavily towards this being something that was part of him from conception onward. Maybe the brief oxygen deprivation triggered something? I dont know. But I’m fairly confident that he did not get it from anything I did. And obviously not induction, in his case.

  21. Kasey Chang says:

    Personal anecdote…

    One of my uncles is probably autistic by modern standards. According to grandma (RIP) they were about to take a long stormy boat ride so they induce labor as to have the baby before the boat ride. I doubt there were many meds available back in WW2 days, it’s probably just some oxygen or breathables? Not that I know anything about it.

    My uncle was a quiet child and lived otherwise unremarkable childhood, at least to grandma, don’t cry much, was hazed a lot in compulsory military service (probably due to somewhat anti-social appearance) that required a superior’s intervention, and is now completely withdrawn into himself.

    Of course, a few cases does not make a trend, but this is an interesting angle to look at. Autism is such a fascinating field.

  22. Blues Girl Babe says:

    Suggestion to researchers – look at all the estrogen mimickers we have now – and in particular the use of synthetic hormones that women use for years to prevent birth – and then 10-15-20 years later decide they want a baby. That baby is in a bath of excess estrogen from years of the mother taking the pill. Does anyone really know the effects of messing with nature? Birth control pills were brought to the mass market 50 years ago, in the 60’s and use has steadily increased since then – so much to the point that birth control pills are now being given like candy to young adolescents to supposedly prevent acne. What will happen when they try to get pregnant at age 30 when they have been on synthetic hormone pills virtually since they hit child-bearing years? It’s a Pandora’s box.

    1. Harriet Hall says:

      You are under a couple of misconceptions. When a woman stops taking the birth control pill, the pill hormones are quickly eliminated from her body. The baby is not in “a bath of excess estrogen.” Quite the contrary: a woman’s levels of estrogen and progesterone must be normal for her to become pregnant. In fact, the pill has been used to IMPROVE fertility: women with irregular menstrual cycles were put on the pill, and when the pill was stopped, they were more likely to become pregnant. And after decades of use, we know that the babies of women who took the pill are just as healthy as babies of those who didn’t.

      “What will happen when they try to get pregnant at age 30 when they have been on synthetic hormone pills virtually since they hit child-bearing years?”

      They will get pregnant easily and have normal babies. We know that because we have studied it. I’m a case in point: I took the pill for 14 years, stopped it, and got pregnant 2 months later.

      In this case, we really DO know “the effects of messing with nature.” The effects are to give women control over their fertility, improve their health (compared to those who regularly get pregnant), and yes, relieve acne, menstrual pain, and other problems. All without any significant adverse effects.

Comments are closed.