Thursday 14 June 2012

C-reactive protein, inflammation and autism

More than a few people have argued that the balance between inflammation and er, not-inflammation, is the single most important battle being waged between health and disease. Such an assertion is made on the back of inflammation being implicated in all manner of conditions including obesitydementia and potentially even epilepsy.

The many faces of inflammation have also been quite a consistent theme across quite a few investigations on autism and related neurodevelopmental conditions. Although not by any means an expert, I have always found this to be a fascinating 'association' given that many people would probably view something like inflammation as being associated with more somatic signs and symptoms as for example, in response to an insect bite or such like. If one however assumes that the core embryonic material used to make skin for example is the same for other organs such as gut or brain, it makes it slightly easier to understand that inflammation is not just an overt process and just because you don't see it, doesn't mean that it's not happening.

Inflammation covers quite a lot of ground with autism in mind; ranging from neuroinflammation as per studies like this one and this one, through to comorbid inflammation of other tissues associated with various bowel findings for example (here), which still fuels some heated discussions in autism research and lay circles. I note that even the recent findings on maternal fever and offspring autism risk have been tied back to inflammation, albeit the suggestion of effects from ".. an acute inflammatory response". Inflammation, rightly or wrongly it seems, has had its card well and truly marked.

In this quite long post I want to discuss some of the very limited findings related to a particular protein called C-reactive protein (CRP) which is intimately linked to inflammation. The recent paper by Khakzad and colleagues* will also figure quite strongly and their reporting of significantly elevated levels of CRP in cases of autism.

A short description of CRP is in order. Apparently a pattern recognition molecule according to this quite detailed review by Black and colleagues** (full-text); levels of CRP increase dramatically following tissue injury or infection as part of a cascade of bodily responses. Interestingly there is a link between CRP and cytokines; in particular one familiar cytokine, IL-6, which has already received quite a bit of interest in autism research circles.

CRP has been mentioned previously on this blog in a couple of post discussing (a) some of the ins-and-outs of HBOT and how some quality time under hyperbaric conditions seemed to have some quite impressive effects on reducing levels of CRP in cases of autism, and (b) elevations in CRP detected in a group of parent carers with children with autism. Inflammation was the name of these games and higher, not lower levels of CRP as a marker of such inflammation seemed to be the important findings although I should stress CRP levels have not been consistently detected as being elevated in every study of autism.

Outside of these studies, there is the odd mention of CRP but more often than not, it is the cytokines, the balance between pro- and anti-inflammatory cytokines, which seemed to have stolen the show in autism research. It was therefore of interest to see the paper by Khakzad and colleagues appear, and to summarise:

  • Based on the use of a high-sensitivity measure for CRP (hs-CRP), serum levels were measured in 39 participants with autism compared with 30 age-matched asymptomatic controls in a snapshot study. Participants with autism were further divided on the basis of CARS scores into moderate vs. severe presentation - most fell into the former category.
  • Group levels of CRP were significantly (very significantly) elevated in the autism group compared with controls (P<0.0001); mean CRP values in the autism group were 540.1 ng/ml compared with 1.3 ng/ml in the control group. Severity of symptoms based on CARS scores were also related to CRP levels in that those with the most severe presentation of autism tended to show greater serum levels of CRP than those presenting with more moderate symptoms. There was a significant correlation between symptom severity and CRP levels although given the sample size, strength of association and less than impressive scatterplot of the data, I would perhaps chance that this finding requires some further replicative work.
  • To quote the authors: "These findings affirm the role of inflammation in autism". 

Appreciating the small scale nature of the Khakzad paper and other factors relating to the participants they looked at (I assume they were all Iranian although the paper does not actually make this clear) and the snapshot view of CRP levels, there are some interesting comments to make from this research.

First and foremost is that whether you overtly see it or not, inflammation is a facet of at least some cases of autism. There are of course the arguments about which came first - autism or inflammation - and indeed whether the two are linked or whether inflammation is just purely epiphenomenal or a manifestation of comorbidity for example. I can't provide any an opinion either way aside from saying that the authors did collect some questionnaire data on among other things past medical history of participants. Aside from 'family history' being significantly different from controls (?), the presence of comorbid epilepsy was the only statistically significant difference between the groups.

An association between certain types of epilepsy and elevated CRP has been suggested and so one cannot perhaps discount this as being an interfering variable bearing in mind the lack of data on what types of epilepsy were present in the Khakzad study. Indeed given the link between IL-6 and CRP levels, elevations in IL-6 linked to seizure disorders is further evidence for a link and an interference when it comes to assessing association and causality.

The possibility of other comorbidity accounting for elevations in levels of CRP is another source of bias. I say this on the back of quite a lot of research suggesting elevated CRP levels in relation to inflammatory bowel diseases (here), asthma (here) and type-1 diabetes (here) for example; all of which have been mentioned to some extent in relation to autism as a comorbidity. Don't even get me started on the links suggested between depression and CRP (here).

The suggestion of a positive correlation between increasing CRP levels and increasing severity of autism symptoms is an interesting assertion despite the need for further study. I note for example that similar suggestions have been reported in cases of schizophrenia for example as per this article by Fan and colleagues***. Similar suggestions in relation to mania and CRP have also been reported by Faith Dickerson and colleagues**** (a name mentioned before on this blog). Exactly how this might play out from a mechanistic point of view is slightly uncertain but nonetheless an important next research step alongside ascertaining whether specific endophenotypes / subgroups on the autism spectrum may be more prone to elevated CRP levels as a function of things like symptom onset (regression vs. no regression) and other parameters.

As per the opening sentences to this post, inflammation is readily becoming the villain under lots of different scenarios. The data emerging in cases of autism warrant further research attention with regards to inflammation and CRP levels to understand if autism spectrum conditions should be added to the list of possible connections and importantly, if and how inflammation links to presented behavioural symptoms and what can be done about it (no advice intended). From a strictly health point of view, the growing implications of chronic elevations of CRP for cardiovascular health are for example, a worry for any group/person with potential long-term, low grade inflammation established by CRP measurement.

To finish, y'know we can dance if we want to. Who said Morris dancing was boring?


* Khakzad MR. et al. The complementary role of high sensitivity C-reactive protein in the diagnosis and severity assessment of autism. Research in Autism Spectrum Disorders. 2012; 6: 1032-1037
DOI: 10.1016/j.rasd.2011.10.002

** Black S. et al. C-reactive protein. Journal of Biological Chemistry. 2004; 279: 48487-48490

*** Fan X. et al. Elevated serum levels of C-reactive protein are associated with more severe psychopathology in a subgroup of patients with schizophrenia. Psychiatry Research. 2007; 149: 267-271.

**** Dickerson F. et al. Elevated serum levels of C-reactive protein are associated with mania symptoms in outpatients with bipolar disorder. Progress in Neuropsychopharmacology & Biological Psychiatry. 2007; 31: 952-955.


  1. Nice article on C-reactive protein. Though inflammation is involved in most disorders, it is also worth remembering that inflammation is also a part of healing or repair mechanism.. Complete control of inflammation in a disorder may also be not so effective mode of therapy..Thanks...

  2. Thanks for the comment.

    I agree that inflammation does play an important role in human health as part of healing and repair. As alluded to in the post, where however acute inflammation turns into more chronic inflammation seems to be the issue in relation to its more 'negative' role. Further, longitudinal research is therefore required specifically in cases of autism, to establish whether CRP levels reflect short- or long-term inflammatory pathology.

  3. Hi Paul Whiteley -

    Oh snap you've up and kicked the hornets nest with this one! I've thought about the participation of inflammation in neurological disorders including autism a lot; here are a few things that you might find of interest.

    Up and above associative findings of increased states of inflammation in schizophrenia, depression, bi-polar, etc, we also have some data from the clinic that argues against these associations being artifacts; i.e.,

    Two RDBC studies of anti-inflammatories (aspirin and a COX-2 inhibitor) as add on to standard drug treatment for schizophrenia that showed an effect.

    Or an analysis on depression, the relationship between SSRIs, immune markers, and clinical improvement.

    Suppression of proinflammatory cytokines does not occur in depressed patients who fail to respond to SSRIs and is necessary for clinical recovery.

    There's more, but I won't blast away on your blog.

    Regarding autism, this is a touchy place to go; anytime you start to mention that there just might, maybe be immune mediated component, you usually run into a wall of opposition so people who know better don't have to acknowledge the frailty of our vaccination analysis.

    Exactly how this might play out from a mechanistic point of view is slightly uncertain but nonetheless an important next research step

    Indeed. Are you interested in some wild speculations? It just so happens, the immune regulators of the CNS, the microglia, are also being found to be highly active participants in lots of critical and time dependent operations that we suspect are impaired in the autism population; i.e., optimization of the neural network/synaptic pruning (see: for a nice review.)

    But that's a big problem; we know that if you invoke the immune system outside the CNS, the microglia are also activated, they may not necessarily be causing enough problems to be degenerative, but they are still acting differently. So, take a fetus whose mother has a state of increased inflammation; say, obesity, diabetes, asthma, an "acute state of inflammation" associated with an infection, or one of the other ways we have introduced inflammation into our lives. We can't pretend that the microglia will be participating in brain formation the same way in a mother with a different immuno-milieu.

    Finally, anyone who wants to get really depressed/terrified should take a look at:

    Microglia in the developing brain: a potential target with lifetime effects (Harry et all) []

    Which is a great review of the literature that tells us that microglia are malleable during critical developmental timeframes such that an insult can persistently modify their number, morphology, and actions within the brain into adulthood.

    That is a big, big problem for the notion of a static rate of autism.

    - pD

  4. Thanks for the comment pD.

    Very interesting papers on schizophrenia and the COX-2 inhibitors. I have to say that I've always thought about the role of aspirin on inhibiting prostaglandin production as being involved but the Th-1 / Th-2 "rebalance" sounds worthy of more investigation.

    Insofar as the immune system being involved in autism, at least a proportion of cases of autism, I don't really see how anyone can dispute this. Indeed I am drawn back to the post on autism and SPAD based on the work of Harumi Jyonouchi and colleagues:
    as one example where immune function is aberrant; there are lots of other studies showing similar (or indeed polar opposite issues).

    I am interested in wild speculation as it happens(!) and your point about microglia and autism is something which has been cropping up quite a lot in some of my reading. Indeed I don't know if you read Paul Patterson's blog 'Infectious Behaviour' but he had rather an interesting post recently titled: Microglia eating synapses in autism?

    (I also read your post on the topic:
    with some interest).


Note: only a member of this blog may post a comment.