Autism Spectrum Disorders: Exploring Pregnancy and Birth Factors
Are there factors associated with pregnancy or birth that increase a child's risk of developing an autism spectrum disorder?
Researchers are striving to understand what causes or contributes to the development of autism spectrum disorders (ASDs). Some of the factors they continue to explore are those related to pregnancy or birth.
Thus far, research on a range of such factors has not yielded a single one that seems to be associated with a large number of ASD cases. Rather, having a less than optimal pregnancy and/or birth in general has been linked with higher risk of ASD. Here, "less than optimal" refers to a variety of circumstances such as being born feet first ("breech presentation"), Cesarean delivery, "fetal distress," or prematurity.1,2,3,4,5,6,7 It is clear from reviewing many studies, all done in different ways and each with its own shortcomings, that we still have few concrete answers. Even studies that have shown a certain factor may be associated with risk for ASD have generally not been able to pinpoint the underlying mechanism – that is, how the factor actually, biologically leads to the development of autism.8
At the same time, some have argued that it is more likely a child's ASD contributes to a less-than-perfect pregnancy or birth rather than being caused by it. They suggest that both mother and baby contribute to the unfolding of a normal pregnancy, and if the baby has a developmental issue like ASD, it is more likely some aspect of the pregnancy or birth will be less than ideal. 9,10,11
Part of what makes untangling these issues so difficult is the perplexing variability of ASDs. That is, a person with ASD may face challenges in communication, social understanding, repetitive behaviors, sensory sensitivities, or any number of other areas, and it is rare for two people with ASD to have the exact same combination of symptoms. Why is this? One proposed explanation is that ASD is caused in different ways in different people who then end up with a slightly different version of ASD. We know there are all kinds of genetic factors that may contribute to ASD,12 and it is possible a variety of environmental factors play a part as well. In fact, the "triple hit hypothesis" proposes that ASD is caused by 1) an environmental stressor interacting with 2) an underlying genetic vulnerability 3) at a critical moment of brain development.13
An "environmental stressor" is some external factor a person encounters, such as a pesticide, a virus, or a medication. A person with a "genetic vulnerability" has one or more genes which may turn "on" or "off" or otherwise be changed due to the action of an environmental stressor. Clearly, there are a great many possible variations, with different genes interacting with different environmental factors, and having a different impact depending on timing. Because an especially critical period of brain development occurs when a baby is still in the womb or newly born, researchers have been particularly interested in potential environmental exposures and other stressors encountered during pregnancy or birth. Below, we highlight a number of topics researchers are investigating.
Prematurity and Low Birth Weight
As more and more very small babies are born and survive, researchers have worked to assess what risks they may face for a variety of problems, including ASD. Is being born preterm associated with ASD? If so, what explains this?
A number of studies have assessed risk of general psychiatric and behavioral problems in children born with a low or very low birth weight. (A birth weight of less than 2000 grams, or about 4 pounds 6 ounces, is considered "low"; a birth weight of less than 1500 grams, or about 3 pounds 5 ounces, is considered "very low.") One consistent finding is that such children often suffer from higher rates of anxiety and depression, as well as attentional and social difficulties. 14,15,16 Research has shown that "preterm children tend to be socially isolated, to play by themselves, are less likely to initiate social behaviors and have poorly developed adaptive skills." 17 But do they actually have ASDs more often than children who were born full term?
Researchers who explored this question did find elevated rates of autistic symptoms or traits among low birth weight children compared to normal birth weight children. 17,18,19,20
Others went on to investigate rates of actual diagnosed ASD among children who had been born at a very low birth weight. For example, a study at the U.S. Centers for Disease Control and Prevention (CDC) compared children who had been born at a normal birth weight with those who had been born weighing less than 2500 grams (or about 5 pounds 8 ounces). These researchers found that the risk of autism was approximately doubled for the low birth weight children overall, but that the picture was actually more complicated than that. If they looked at ASD, co-occurring intellectual disability, and gender all together, they found that low birth weight girls with intellectual disability were four times more likely to have ASD than normal birth weight children with no such disability. On the other hand, low birth weight boys with no intellectual disability didn't have a statistically significant increased risk of ASD. 21
Likewise, researchers in the U.K. compared 219 children born at less than 26 weeks gestation with 153 children born full term. (A baby is considered "full term" if it has spent 37 or more weeks in the womb; it is considered "premature" or "preterm" if it is born earlier.) They found that the preterm children had significantly more autistic traits as measured by the Social Communication Questionnaire (SCQ), a screening instrument focused on autism symptoms. 22 They also found that extremely preterm children had a much higher risk of having a diagnosed ASD compared to the full term group. Although they did not separate out gender, like the CDC researchers, they found that ASD in preterm children appeared to be associated with intellectual disability much more often than ASD is in the general population. One researcher speculated that ASD in children born very early, who are so fragile they spend many weeks in transparent incubator units called Isolettes, may be akin to the ASD-like symptoms witnessed in Romanian orphans who experienced pervasive deprivation, including minimal contact with caregivers. 23
For some time now, there has been clear evidence that specific drugs taken during pregnancy are associated with an increased risk of ASD. Thalidomide, which was once given to treat morning sickness, but was banned because it led to all kinds of birth defects, has been associated with autism, 24 as have valproic acid and other anti-seizure medications taken during pregnancy. 25,26,27
More recently, one topic of discussion has been a category of medication called selective serotonin reuptake inhibitors (SSRIs). These medications, which include drugs like Zoloft and Prozac, are used to treat depression. Is there any possibility that SSRIs might contribute to the development of ASD in a child whose mother took them while pregnant? This idea, based on animal research, suggests that excess levels of serotonin in an infant's brain during pregnancy may affect regulation of the pro-social hormone, oxytocin, and the overall development of a crucial area of the brain called the amygdala. In turn, poorly regulated oxytocin and problems with the amygdale may be associated with symptoms of autism. 28,29 Keep in mind, however, that this notion is based on research on rats; no one has yet investigated this possible link in human mothers and their babies. (For more on oxytocin and ASD, see "Pitocin (Induced Labor)," below.)
It is also important to note that a number of studies have shown that depression and other mood disorders are more common in relatives of people with ASD, and in people with ASD themselves, than in the general population. Based on this, it has been suggested that there may be some kind of genetic link between mood disorders and ASD. 30,31,32,33 If that is the case, more moms of children with ASD might have been taking antidepressants during pregnancy because more of them suffered from depression, but those antidepressants might have nothing to do with causing ASD. 34
As far back as the 1970s, researchers have wondered if illnesses suffered by mothers while pregnant might be associated with a child's ASD. One early study looked at measles, rubella, and mumps infections during pregnancy. The researchers found children with ASD were more likely to have been in the womb during such a maternal infection compared with unaffected siblings. However, there were so few cases of infection during pregnancy in the whole group of mothers of children with ASD that the researchers thought it unlikely this could explain most cases of autism. 35
There have been reports of ASD associated with a variety of other maternal infections during pregnancy, including toxoplasmosis, cytomegalovirus, and herpes. 36,37 Two recent studies, one in California and one in Denmark, have looked at maternal infection during pregnancy in general. Did it result in a greater risk of ASD in children?
In the California study, 407 children with ASD were compared to 2,075 children who did not have ASD. The team looked at trimester of exposure as well as type of infections (bacterial, viral, etc.). They found that maternal infection in the second half of pregnancy, particularly bacterial infection, is associated with a modest increase in risk of having a child with an autism spectrum disorder. 38 They also found that women who had been hospitalized in the second trimester for an infection were at significantly increased risk of having a child who would later be diagnosed with ASD.
The Danish study, which looked at all children born in the country from 1980 through 2005, resulted in comparable findings. If the researchers looked at any infection over the entire pregnancy, they found no increased risk of ASD, which is what an earlier Danish study had also found. 34 When they narrowed their focus by timing and type of infection, however, they found that admission to the hospital for a viral infection during the first trimester, or for a bacterial infection during the second trimester, were both associated with an increased risk of ASD in the child. 39
It has been proposed that it may be a mother's immune response rather than a specific infection that is linked with vulnerability to autism. In this case, it would be the reaction of a mother's body as it tried to fight an infection, rather than the infection itself, that caused a problem. Research on immune response during pregnancy is still at the "animal model" stage, that is, they are using mice to test these hypotheses. 40,41
We know that extreme levels of stress experienced by a mother during pregnancy can have an impact on her child, resulting in eventual cognitive, behavioral, and emotional problems. It is believed this is because the stress hormone, cortisol, can cross the placenta and disturb developmental processes. 42,43
Could any such process be associated with increased risk of ASD? A study looking at more than a million children born in Denmark from 1978 to 2003 explored whether stress during pregnancy was linked to ASD by comparing mothers who had lost a close relative during pregnancy, or up to one year before pregnancy, with those who had not. At first, it appeared there was greater risk of having a child with ASD if a mother had experienced bereavement, but once they took into account other factors (like child's gender, birth year, and socioeconomic status) they found no association between a mother's having lost a close relative during or shortly before her pregnancy and increased risk of ASD in her child. This was the case whether the bereavement occurred in the year before or during pregnancy. 44
The difficulty here may be how to measure stress during pregnancy. What events are defined as stressful, and how can researchers assess the way different women experience different stressors during different windows of vulnerability?
Assisted Reproductive Technologies
We do know that children born as a result of some kind of assisted conception are more likely to experience a variety of issues, including cerebral palsy, behavior problems, and more childhood hospitalizations. 45,46,47 Because of this, researchers have wondered whether any of the methods used to help infertile couples get pregnant could be associated with increased risk of ASD in the children that result. The types of technologies or techniques coming under scrutiny include fertility drugs and in vitro fertilization. 48
Fertility drugs, which are also called "ovulation inducing medications," are a treatment for women who have ovulation disorders. (In other words, they are not releasing eggs with regularity and so have a hard time getting pregnant.) These drugs work by releasing hormones that trigger or regulate ovulation. 49
In vitro fertilization (IVF) is a process where a woman's eggs (or ova) are fertilized outside of her body in a liquid medium. If a fertilized egg results, it is inserted into the woman's uterus in the hopes it will then develop just as it would have via natural conception. 50 An additional procedure, used in conjunction with IVF, is intracytoplasmic sperm injection (ICSI), which involves injecting a single sperm into an egg. 51
Thus far, research focused specifically on any association between use of these reproductive technologies and ASD in offspring has been limited. One review of eight studies including ASD as a possible topic of interest, and thirty on developmental delay, found no consistent results, and there were no studies targeting just ASD. 48
A Danish study that later did look specifically at assisted conception and ASD found children conceived by such methods at lower risk of ASD. 52 In contrast, a small Dutch study (n = 264) comparing children conceived by ICSI combined with IVF, IVF alone, or with no assistive technologies found higher ASD risk for offspring conceived by ICSI combined with IVF. 53 Yet another study, this one involving 564 children and taking place in Israel, found a higher risk of ASD in children who were conceived by IVF. 54 Meanwhile, a U.S. study focusing on 111 mothers with children with ASD and 3,874 control mothers found use of fertility drugs was more common in the ASD mothers than the control mothers (34.2% vs. 23.8%). 55
One challenge is teasing apart all the factors that may add to a mother's risk of having a child with ASD, but which tend to occur together, like infertility, use of assisted reproductive technologies, multiple births, prematurity, and advanced maternal or paternal age. 8
Advanced Maternal or Paternal Age
It has been suggested that older parents may be at higher risk of having a child with an ASD. Although some studies have shown no association between either a mother's or father's age and risk of ASD in their child, 5 many others have shown a link between one or both and such risk. 34,56,57,58,59,60,61
Two separate studies, one based on ASD cases from the Israeli draft board medical registry 57 and one based on cases in Denmark, 56 found that increasing paternal, but not maternal age, was associated with higher risk of ASD in children. These studies did include data on both mothers and fathers. Maternal age appeared important until they controlled for paternal age. "Controlling for" is a statistical way of accounting for the contribution of factors that influence each other to see which one is the driving force behind an association. It turned out that only paternal age mattered. In the study based on the Israeli data, offspring of men age 40 or older were more than 5 times more likely to have a child with ASD compared to men under 30. The Danish study also found increased risk for older fathers.
Four U.S. studies, three in California and one multi-state study, found that increasing maternal and paternal age were both independently associated with higher risk of ASD in children. (In other words, both mattered.) These were all large scale studies. The multi-state study looked at more than 250,000 children, 59 while the other three looked at more than 132,000, 58 nearly 5 million, 61 and more than 7 million, 60 respectively.
Wrote one group of researchers, "Regardless of the relative importance of paternal and maternal age, it is noteworthy that so many credible epidemiologic studies...have now confirmed that paternal and maternal age at birth are related to autism risk. The evidence is substantial enough to justify a search for the underlying mechanisms." 62
One "underlying mechanism" that has been discussed is the possibility of permanent changes to a gene's DNA within a specific sperm or egg cell, which is called a de novo mutation. 56,62 Because a father's sperm cells are continuously manufactured, they are especially susceptible to such changes. These mutations are more likely to occur as men age, whether due to the process of aging itself or due to a lifetime of exposure to everything from pesticides to food additives to cosmic rays. A woman's eggs, in contrast, are with her when she is born. They may age, but they are not still being created when she is older.
It is interesting to note that one group of researchers studying twin sets in which at least one twin had an ASD found that both fathers over 50 and fathers under 25 had a higher risk of having a child with ASD compared with fathers in the middle. The group hypothesized that older fathers might have more de novo mutations due to aging and a lifetime of exposure to environmental toxins, while younger fathers might be at higher risk for such mutations because of life style factors like drug abuse. 63
Pitocin (Induced Labor)
Oxytocin is a natural hormone which plays a major role not only in labor and lactation, but also in a variety of social behaviors including mother-infant bonding and social memory. 64 It has been suggested that a problem with the system that regulates oxytocin may play a role in the development of ASD. What's more, several small studies have shown oxytocin can be used to treat the core social dysfunction that characterizes autism. In one study, adults with ASD taking an oxytocin infusion significantly reduced repetitive behaviors. 65 In two other studies, oxytocin improved the social cognition 66 and emotion recognition 67 of adults on the spectrum.
At the same time, some researchers have hypothesized that Pitocin, which is a synthetic form of oxytocin used to induce (or "jump start") labor, may be somehow associated with ASDs in children. Researcher Eric Hollander commented, "In some individuals whose oxytocin system could be genetically vulnerable, a strong environmental early hit while the brain is still developing could down-regulate the oxytocin system, leading to developmental problems." 68 (What he means is that it's possible something could alter the brain's settings for producing and controlling oxytocin, and a person would forever after have to function on levels of oxytocin that were too low.)
A study involving 41 boys with autistic disorder and 25 age- and IQ-matched controls failed to show any such link. 69 Yet another study, this one involving 15 children with ASD, 42 with attention deficit hyperactivity disorder (ADHD), 21 with anxiety or depression, and 15 with learning disabilities, did seem to show an association between Pitocin administered during labor and ASD, specifically. Those in the ASD group were far more likely to have received Pitocin than the children in the other groups. 70
Studies so far have been very small, and contradictory. On the one hand, this may be a case where it is a child's ASD that contributes to problems with labor. In that case, Pitocin would be needed more often during the birth of children with ASD although it had nothing to do with causing the ASD. On the other hand, there is at least some theoretical explanation for why Pitocin might lead to ASD, and more research on a larger scale is certainly called for.
Bilirubin, a yellow pigment created in the body during the normal recycling of old red blood cells, is processed by the liver. This task is performed by a mother's body through the placenta before a baby is born. After birth, an infant's body has to take over this task, and many babies will develop a yellowish color in their skin and the whites of their eyes as they are making this transition. Usually this yellow color, or jaundice, gets better on its own, but sometimes bilirubin levels can become dangerously high and treatment is required. 71
Studies focusing on whether infants who had neonatal jaundice are at increased risk for ASD have had mixed results. One U.S. study focused on 338 children with ASD and nearly 2000 typical children born in northern California. About 28% of both groups had been tested for high bilirubin levels in the first 30 days of life. The children who did have a high bilirubin level were at no greater risk of developing ASD than those who did not. Therefore, this study did not provide evidence that newborn jaundice, or hyperbilirubinemia, is a risk factor for ASD. 72
In contrast, a Danish study involving 473 children with autism and an equal number without found that children who had been diagnosed with high bilirubin levels after birth had nearly four times the risk of developing autism. This applied only to full term infants. 73 The authors speculated that they may have gotten a different result than the researchers in California because they included children with autism only, and not those with other ASDs like Asperger's syndrome. Another possible reason for the difference was statistical. "The confidence intervals for high levels of serum bilirubin in our study and the study by Croen et al. are very broad and the two study estimates may differ by chance," wrote the Danish researchers. In other words, when you're using statistics and probability to estimate risk, as researchers do, there is some chance that you'll get an extreme or false finding by chance, just as it is possible for a person tossing a coin 100 times to get "heads" 99 times and not the 50 times you'd expect.
The solution? More research. In fact, this is how science works: by incremental steps, and a weight of evidence building that eventually makes clear what was fluke and what is a valid association.
The IAN Pregnancy & Birth Questionnaire
As this review demonstrates, we know that a less than ideal pregnancy or birth is associated with a higher risk of ASD. What we do not know is whether a baby's ASD leads to various pregnancy and birth issues, or is caused by them. In any case, researchers are hoping to discover which specific factors are associated with this overall effect, and if there are ways to intervene to minimize any ASD risk for future children.
Towards that end, the Interactive Autism Network (IAN) Research project is currently developing a detailed Pregnancy and Birth Questionnaire. We will ask mothers a series of questions about their pregnancy and birth experiences with regard to each of their children, both those with ASD and those without. (It's important to have information about the children who do not have ASD for purposes of comparison.) There will be questions on all the topics discussed above, and more.
Current IAN Research participants will receive a notice by e-mail when the questionnaire is ready for them to answer. If you are not yet registered with IAN Research, but would like to participate in the upcoming Pregnancy & Birth Questionnaire, visit IAN Research to register today.