No Link Between Childhood Infections, Autism

10 May

One of the theories posited about autism causation is that childhood infections can result in autism. A recent study from the Arhus, Denmark, explores this by checking how often children are admitted to the hospital for infectious diseases. Given that maternal infections do appear to be associated with a greater risk of autism, the idea of childhood infections is worth considering. I would add that the attention to mitochondrial disorders and autism that was high in the past couple of years would also suggest this is a valuable area to consider.

Here is the abstract:

OBJECTIVE: To investigate the association between hospitalization for infection in the perinatal/neonatal period or childhood and the diagnosis of autism spectrum disorders (ASDs). DESIGN: A population-based cohort study. SETTING: Denmark. PARTICIPANTS: All children born in Denmark from January 1, 1980, through December 31, 2002, comprising a total of 1 418 152 children. EXPOSURE: Infection requiring hospitalization. MAIN OUTCOME MEASURE: The adjusted hazard ratio (HR) for ASDs among children hospitalized for infection compared with other children. RESULTS: A total of 7379 children were diagnosed as having ASDs. Children admitted to the hospital for any infectious disease displayed an increased rate of ASD diagnoses (HR, 1.38 [95% confidence interval, 1.31-1.45]). This association was found to be similar for infectious diseases of bacterial and viral origin. Furthermore, children admitted to the hospital for noninfectious disease also displayed an increased rate of ASD diagnoses (HR, 1.76 [95% confidence interval, 1.68-1.86]), and admissions for infection increased the rate of mental retardation (2.18 [2.06-2.31]). CONCLUSIONS: The association between hospitalization for infection and ASDs observed in this study does not suggest causality because a general association is observed across different infection groups. Also, the association is not specific for infection or for ASDs. We discuss a number of noncausal explanatory models.

Autistic children *are* admitted to the hospital for infectious diseases more often than the rest of the population. But, in general autistic children are admitted to the hospital more often than the rest of the population.

Bloomberg Businessweek discussed this paper in No Link Between Childhood Infections, Autism. They interviewed Dr. Andrew Zimmerman of the Kennedy Krieger Institute. Dr. Zimmerman concurred with the conclusions of the paper:

“Yes, there is an increased rate of hospitalization preceding the diagnosis of autism, but it doesn’t support a causal relationship between autism and infections,” Zimmerman said.

This is significant, in my view. Dr. Zimmerman is one of the doctors who treated Hannah Poling (the young girl whose case was conceded by the Department of Health and Human Resources in the vaccine court, sparking the public interest in the subject).

It is also worth noting that the study considered specific infections. From the Bloomberg/Newsweek story:

And the researchers could point to no particular infection that upped the risk.

They therefore conclude that childhood infections cannot be considered a cause of autism.

“We find the same relationship between hospitalization due to many different infections and autism,” noted lead study author Dr. Hjordis Osk Atladottir, of the departments of epidemiology and biostatistics at the Institute of Public Health, University of Aarhus in Denmark. “If there were a causal relationship, it should be present for specific infections and not provide such an overall pattern of association.”

13 Responses to “No Link Between Childhood Infections, Autism”

  1. Sullivan May 10, 2010 at 21:01 #

    As an aside–for those who claim that the autism rate in Denmark is dramatically lower than that in the U.S., consider–

    7379 children with ASD’s out of 1,418,152 children

    That’s about 0.52%. For children born between January 1, 1980, through December 31, 2002. Considering that this covers years back to 1980–when the prevalence was much lower–these numbers do not appear much lower than in the US and UK.

  2. Joseph May 10, 2010 at 21:38 #

    As an aside—for those who claim that the autism rate in Denmark is dramatically lower than that in the U.S

    Those who claim that are erroneously comparing the incidence rate reported in some Danish studies with the prevalence rate in the US.

    JB Handley has been known to do that, and you can also find it in Jenny McCarthy’s HuffPo piece in response to Frontline. Dr. Novella went over Handley’s error here, and further explained that one study found the prevalence of autism was 80 in 10,000 in Denmark.

    Wikipedia’s explanation of the difference between incidence and prevalence is pretty good. In general, just remember that incidence can be very low even if prevalence is high.

  3. isles May 10, 2010 at 21:39 #

    This is all consistent with the idea that higher levels of healthcare utilization in general are a factor in the reported incidence of autism. It is logical that children in families that have good access to healthcare would be both more likely to be hospitalized for infections (of the severity where hospital care is prudent but might not happen if the family were uninsured, undocumented immigrants, etc.) and more likely to get an autism diagnosis rather than just be regarded as slow or odd.

  4. Tsu Dho Nimh May 11, 2010 at 04:33 #

    isles … There is no problem with being uninsured or not having access. Denmark has universal health care.

  5. passionlessDrone May 12, 2010 at 03:48 #

    Hi Sullivan –

    What if, a disturbance to the innate immune system could lead to autism, something that happens regardless if the pathogen were bacterial or viral? In that case, what would the pattern of infection by pathogen type look like? The researchers are making the presumption that the insult must be tied to pathogen specific interactions, as opposed to more common pathways. I’m not sure this is a safe assumption to make.

    For example:

    Viral-like brain inflammation during development causes increased seizure susceptibility in adult rats

    Viral infections of the CNS and their accompanying inflammation can cause long-term neurological effects, including increased risk for seizures. To examine the effects of CNS inflammation, we infused polyinosinic:polycytidylic acid, intracerebroventricularly to mimic a viral CNS infection in 14 day-old rats. This caused fever and an increase in the pro-inflammatory cytokine, interleukin (IL)-1beta in the brain. As young adults, these animals were more susceptible to lithium-pilocarpine and pentylenetetrazol-induced seizures and showed memory deficits in fear conditioning. Whereas there was no alteration in adult hippocampal cytokine levels, we found a marked increase in NMDA (NR2A and C) and AMPA (GluR1) glutamate receptor subunit mRNA expression. The increase in seizure susceptibility, glutamate receptor subunits, and hippocampal IL-1beta levels were suppressed by neonatal systemic minocycline. Thus, a novel model of viral CNS inflammation reveals pathophysiological relationships between brain cytokines, glutamate receptors, behaviour and seizures, which can be attenuated by anti-inflammatory agents like minocycline.

    Considered alongside

    Postnatal Inflammation Increases Seizure Susceptibility in Adult Rats

    There are critical postnatal periods during which even subtle interventions can have long-lasting effects on adult physiology. We asked whether an immune challenge during early postnatal development can alter neuronal excitability and seizure susceptibility in adults. Postnatal day 14 (P14) male Sprague Dawley rats were injected with the bacterial endotoxin lipopolysaccharide (LPS), and control animals received sterile saline. Three weeks later, extracellular recordings from hippocampal slices revealed enhanced field EPSP slopes after Schaffer collateral stimulation and increased epileptiform burst-firing activity in CA1 after 4-aminopyridine application. Six to 8 weeks after postnatal LPS injection, seizure susceptibility was assessed in response to lithium–pilocarpine, kainic acid, and pentylenetetrazol. Rats treated with LPS showed significantly greater adult seizure susceptibility to all convulsants, as well as increased cytokine release and enhanced neuronal degeneration within the hippocampus after limbic seizures. These persistent increases in seizure susceptibility occurred only when LPS was given during a critical postnatal period (P7 and P14) and not before (P1) or after (P20). This early effect of LPS on adult seizures was blocked by concurrent intracerebroventricular administration of a tumor necrosis factor (TNF) antibody and mimicked by intracerebroventricular injection of rat recombinant TNF. Postnatal LPS injection did not result in permanent changes in microglial (Iba1) activity or hippocampal cytokine [IL-1? (interleukin-1?) and TNF] levels, but caused a slight increase in astrocyte (GFAP) numbers. These novel results indicate that a single LPS injection during a critical postnatal period causes a long-lasting increase in seizure susceptibility that is strongly dependent on TNF.

    If you pay close attention, you can see that the researchers achieved similar endpoints; a greater succeptibility to seizures, by mimicing either viral or bacterial infection. It didn’t matter what the trigger was, just that there was an innate immune response.

    Furthermore, anti inflammatory agents (minocycline), or pro-inflammatory cytokine antibodies (tnf-antibody) attenuated the effect; implicating the innate immune response as the critical component. Remember, these animals weren’t getting sick, they were being fooled into thinking they were under microbial attack, but we still observed drastic physiological changes. The immediate immune response mounted by your body, the innate immune response, is largely similar if you encounter a viral or bacterial pathogen.

    Given these two results, if you asked the wrong question, one could easily assume that getting either a viral or bacterial infection could not lead to increased seizure succeptibility, but the reality is that having either leads to increased risk. Does this give us any insight towards what we see in hospitilization rates in the Denmark studies?

    Further, we might consider that we have numerous examples of exaggerated innate immune responses in the autism population, specifically Enstrom, Ashwood, and Jynouchi.

    Given that, I’m not sure that detecting the same outcome from bacterial or viral infections is necessarily evidence for a lack of causation, but, indeed, may be the result of asking the wrong question.

    Thanks for posting this.

    – pD

  6. RAJ May 12, 2010 at 04:21 #

    The problem with design of this type of study is that the comparisons are made between a group of children who had poor outcomes associated with a specific event and the larger group unaffected by the type of obstetrical or unfavorable event in the pre, peri or neonatal period.

    Studies that have recruited children who have experienced a POOR outcome associated with a specfic event gives a completly different picture.
    Children born with congenital rubella and poor outcomes reported 7.3% of the poor outcome children were diagnosed with autism (Chess 1973). Thalidomide embryopathy, a condition where poor outcomes were associated with prenatal exposure to thalidomide found 4/100 (4%) of poor outcome children were diagnosed with autism. This Swedish study is based on recruiting every child who had poor outcomes associated with prenatal Thalidomide exposure. A nationwide registy found that only 105 children born in Sweden had poor outcomes compared to the hundreds of thousand of pregnant women who had been exposed to the risk of Thalidomide Embryopathy. 8.9% of children with poor outcomes associated with Valproate Acid Syndrome were diagnosed with autism.

    The phenomena has also been reported in Downs Syndrome with 5-7% diagnosed with autism.

    Congenital rubella, thalidomide embryopathy and valproate acid syndrome are all considered as presenting with the strongest and least unassailable evidence between specific events and autism risk. All three of the studies were based on the Chess methodology which was introduced in 1973 by Chess in her study of poor outcomes associated with congenital rubella.

    • Sullivan May 12, 2010 at 05:42 #

      I’m a bit at a loss here RAJ–

      what children born with congenital rubella syndrome have “good” outcomes?

      Congenital Rubella, Thalidomide, valproate–you are forgetting another one: misoprostal

      What do they all have in common? The window of opportunity is early 2nd trimester. Prenatal.

    • Sullivan May 12, 2010 at 06:03 #


      I’ll admit, I was expecting to see you tomorrow. That said, on Valproate

      6.3% of those exposed during pregnancy had autism. Not 6.3% of those with “poor outcomes”, but 6.3% of the total.


      this is not the strongest study I’ve ever seen. But I do think the focus on environmental causation should focus on prenatal exposures.

  7. RAJ May 12, 2010 at 12:10 #

    This is the study most referenced with respect to Valproate Acid Syndrome:

    The authors looked at children with parent reported behavioral problems. 8.9% met diagnosis for autism.

    Chess’s study was based on the children referred to her clinic in in NYC with developmental problems associated with prenatal rubella infection:

    Unfavorable events can also occur in the posnatal period using the Chess methodology. Children with a history of unprovoked seizure in the first year of life were found to have a significant risk for autism. The children were referred because of concerns about the outcomes identified by high scores on the Social Communication Questionnaire.

    CONCLUSION: These results suggest that the estimated prevalence of ASD is higher in children with history of seizure in the first year of life than it is in the general population. There are indications that support the view that children with ASD and history of seizure in the first year of life have higher prevalence of congenital brain abnormalities and are more often female, than other children with ASD.

  8. Joseph May 12, 2010 at 14:39 #

    These results suggest that the estimated prevalence of ASD is higher in children with history of seizure in the first year of life than it is in the general population.

    That doesn’t mean the seizures caused ASD, though. It only means autistic infants have a higher prevalence of seizures. It doesn’t address the prenatal vs. postnatal issue at all.

  9. passionlessDrone May 12, 2010 at 15:30 #

    Hi Joseph –

    Detangling causation is difficult indeed. However, we might be able to learn something from animal models of seizures and, curiously enough, how the immune response affects the outcome from those seizures.

    For example, <a href=""Glial activation links early-life seizures and long-term neurologic dysfunction: evidence using a small molecule inhibitor of proinflammatory cytokine upregulation

    Early-life seizures increase vulnerability to subsequent neurologic insult. We tested the hypothesis that early-life seizures increase susceptibility to later neurologic injury by causing chronic glial activation. To determine the mechanisms by which glial activation may modulate neurologic injury, we examined both acute changes in proinflammatory cytokines and long-term changes in astrocyte and microglial activation and astrocyte glutamate transporters in a “two-hit” model of kainic acid (KA)-induced seizures. METHODS: Postnatal day (P) 15 male rats were administered KA or phosphate buffered saline (PBS). On P45 animals either received a second treatment of KA or PBS. On P55, control (PBS-PBS), early-life seizure (KA-PBS), adult seizure (PBS-KA), and “two-hit” (KA-KA) groups were examined for astrocyte and microglial activation, alteration in glutamate transporters, and expression of the glial protein, clusterin. RESULTS: P15 seizures resulted in an acute increase in hippocampal levels of IL-1beta and S100B, followed by behavioral impairment and long-term increases in GFAP and S100B. Animals in the “two-hit” group showed greater microglial activation, neurologic injury, and susceptibility to seizures compared to the adult seizure group. Glutamate transporters increased following seizures but did not differ between these two groups. Treatment with Minozac, a small molecule inhibitor of proinflammatory cytokine upregulation, following early-life seizures prevented both the long-term increase in activated glia and the associated behavioral impairment. CONCLUSIONS: These data suggest that glial activation following early-life seizures results in increased susceptibility to seizures in adulthood, in part through priming microglia and enhanced microglial activation. Glial activation may be a novel therapeutic target in pediatric epilepsy.

    The parallels here should be relatively straightforward; having a seizure early in life led to chronically activated microglia, increased GFAP (Vargas / Fatemi, respectively). Behavioral and seizure succeptibility changes were observed in groups that had a seizure. To top it off, as seen in the links I posted above, inhibiting the immune response, once again, has the ability to attenuate the physiological and behavioral effects of the seizure. There are several papers regarding glial priming that may have interest to our population.

    The jump from a rodent to human brain development from a prenatal to postnatal timeframe is one that is fraught with difficulties in reaching certainty on with our current knowledge.

    – pD

  10. passionlessDrone May 12, 2010 at 15:32 #

    Sorry for spamming, folks, I forgot to include the paper that I refernced above.

    Glial activation links early-life seizures and long-term neurologic dysfunction: evidence using a small molecule inhibitor of proinflammatory cytokine upregulation.

    Sorry again.

    (Sullivan / Kev – if you get a preview function up, I promise to use it)

    – pD


  1. The Nobel Prize winner and the unethical autism trial « gimpy’s blog - November 22, 2010

    […] allow various immune mechanisms to act with greater efficiency.  Unfortunately there is no evidence that bacterial or viral infections have a part to play in the causes of autism.  It is not good practice to base a research protocol on assumptions that are inconsistent with […]

Leave a Reply to RAJ Cancel reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.

%d bloggers like this: