Head circumference and brain size in autism meta-analysed

I read with interest the paper by Roberto Sacco and colleagues [1] providing some much needed clarity on the topic of head circumference and brain size in relation to autism spectrum disorder (ASD).

Detailing the results of a systematic review and meta-analysis based on "27 studies defining percentages of macrocephalic patients and 44 structural brain imaging studies providing total brain volumes for patients and controls", researchers provided "conclusive effect sizes and prevalence rates for macrocephaly and brain overgrowth in autism."

So: "Head circumference was significantly larger in autistic compared to control individuals, with 822/5225 (15.7%) autistic individuals displaying macrocephaly." Further: "Brain overgrowth was recorded in 142/1558 (9.1%) autistic patients." Something of an interaction was also reported between age and total brain volume "resulting in larger head circumference and brain size during early childhood."

As per previous discussions on this topic, there was always a degree of confusion about the the research reporting abnormal head size to be linked to autism (see here). Previous, quite sweeping generalisations, about 'big heads' being linked to the presentation of autism turned out to be a little too sweeping despite some investigations even talking about head size and screening opportunities (see here). The recognition of specific 'types' of autism perhaps being linked to head size and brain enlargement (see here) perhaps offered a more 'real-world' perspective to this issue in-line with quite a lot of research direction in these days of more 'plural' autism (see here). The evidence for this endophenotype concept related to head size comes in a large part from the various studies of specific genetic issues being associated with autism and head size as per the example from Nebel and colleagues [2].

Being careful not to over-generalise the issue of age related to head size [3] and understanding that where one draws comparison population norm data from might be important [4], I'd like to think that there are still some research benefits to continued study of head and brain size when it comes to autism. Not least are the various studies talking about connectivity with reference to brain regions in relation to autism, and whether brain size might be an important factor. That accelerated head circumference might be part and parcel of other growth parameters [5] is also an interesting observation and begs some interesting questions about [some] autism in relation to whole body physiology...

'Everything I do' (but perhaps not the way remember).

----------

[1] Sacco R. et al. Head circumference and brain size in autism spectrum disorder: A systematic review and meta-analysis. Psychiatry Res. 2015 Sep 28. pii: S0925-4927(15)30057-3.

[2] Nebel RA. et al. Reciprocal Relationship between Head Size, an Autism Endophenotype, and Gene Dosage at 19p13.12 Points to AKAP8 and AKAP8L. PLoS One. 2015 Jun 15;10(6):e0129270.

[3] Cederlund M. et al. Pre-schoolchildren with autism spectrum disorders are rarely macrocephalic: a population study. Res Dev Disabil. 2014 May;35(5):992-8.

[4] Raznahan A. et al. Compared to what? Early brain overgrowth in autism and the perils of population norms. Biol Psychiatry. 2013 Oct 15;74(8):563-75.

[5] Chawarska K. et al. Early generalized overgrowth in boys with autism. Arch Gen Psychiatry. 2011 Oct;68(10):1021-31.

----------

Sacco R, Gabriele S, & Persico AM (2015). Head circumference and brain size in autism spectrum disorder: A systematic review and meta-analysis. Psychiatry research PMID: 26456415

MIND the maternal autoantibodies in autism

It's been a few weeks since I posted on research coming out of the MIND Institute. Regular readers will probably already know that I'm quite interesting in the various investigations on autism from UC Davis, and in particular how they are very usefully starting to think about subgroups and endophenotypes when it comes to autism as per the quite significant heterogeneity (and comorbidity) present across the spectrum.

By your command @ Wikipedia  

Today's offering is focused on the paper by Nordahl and colleagues* looking at how the presence of specific maternal IgG autoantibodies to foetal brain protein might be (a) linked to cases of autism and (b) form a specific endophenotype of autism with a focus on brain enlargement.

Before wandering through the latest offering, I should point out that both the concept of maternal autoantibodies and brain enlargement have been talked about before on this blog with the MIND link in mind(!) Indeed the last time I posted about the work of Christine Wu Nordahl was on how regression (yes it does happen) might link into brain overgrowth in cases of autism (see here) bearing in mind the complexity and heterogeneity within the field of looking at head size (see here).

As for maternal autoantibodies, well I've talked before about what happens to the offspring of pregnant mice (yes, mice) when they receive a transfusion of IgG brain reactive antibodies derived from mums with a child diagnosed with autism (see here) but I should also say that I'm very glad that I MET you... (see here). Other related findings on this topic are also worth mentioning (see here).

• The latest Nordahl paper reports that in an independent cohort of mums of 181 young children (most of whom had a child diagnosed with an autism spectrum disorder, ASD; n=131), there were some interesting findings related to the presence of specific maternal IgG autoantibodies (37 + 73-kDa) overlapping with offspring ASD. 
• A reported 7% of children (n=10) with ASD were "born to mothers with the 37/73kDa IgG autoantibodies" compared with none in the control, typically developing group. 
• More than that, when the ASD-IgG (those born to IgG positive mothers) were compared with the non ASD-IgG-ers, there were some differences between the rate of abnormal brain enlargement (12.1% vs. 4.4%) compared to controls. 
• Without professing any knowledge or wisdom on the specific structure of the pink stuff floating in our skull (yes, pink), I note the authors reported that "the frontal lobe is selectively enlarged in the ASD-IgG group and that both gray and white matter are similarly affected". 

I'm interested in these findings. Interested because not only was there a subgroup of mums with children with autism who were positive for these autoantibodies but also because "all TD [typically developing] controls were negative for these paired autoantibodies".

Of course such interest needs to be balanced; and in particular the temptation to make too much out of these findings based on the numbers reported on this particular occasion. So for example, 10/131 (7%) of the ASD group were reported to be born to IgG positive mothers. Applying this rate to the control group (n=50) would, by my crude calculations, mean that 3 or 4 of those 50 typically-developing children would have been expected to show the same rate had there been no differences between the groups. You can perhaps see that the difference in the numbers of cases is not exactly overwhelming from this point of view. Indeed, similar principles go for the reported rate of brain enlargement between the IgG positive and negative ASD groups. More shades of grey over a stark black-and-white difference. Even the MRI findings on brain enlargement need to be brought into context with other reviews on structural issues being related to autism as per my post on the Vasa paper (see here).

All that being said and following the whole sub-group / endophenotypes mantra, I'm taken back to some interesting commentary by another of the study authors, David Amaral elsewhere, and his notion of "autism type A, or type B, or type C" as the endpoint of much of this autism subtype research.

I suppose the next set of questions should be something along the lines of how such antibodies come about, how exactly maternal autoantibodies to foetal brain might lead to brain enlargement, and what if anything can be done to mitigate any effect on the developing child. Without wishing to speculate too much, the family of IgG antibodies are normally part and parcel of the on-going struggle against the various bacterial and viral infections which we're all faced with day-to-day. They also have the quite exceptional ability to cross the placenta from mum to baby which the Braunschweig study** seemed to reiterate. Certainly knowing all that, there seems enough to go on with as to where research should be looking with regards to a potential source. The peer-reviewed literature also suggests a couple of other potential angles which may or may not be important (see Zhang and colleagues*** for example) but like everything linked to autism research, it's probably going to be complicated.

Again with my non-expert hat on, and armed with that pinch/dollop of salt, one has to wonder whether for example, there is a role for maternal autoimmune disease in the formation of such autoantibodies as per the research in other autoimmune conditions such as systemic lupus erythematosus (SLE). I was particularly taken by the paper from Lee and colleagues**** (open-access) who reported a really interesting connection between maternal SLE and N-methyl-D-aspartate receptor (NMDAR)-specific autoantibodies (in a mouse model). Indeed Palmeria and colleagues***** (open-access) present quite a good overview of IgG placental transfer with some discussion on maternal autoimmunity and its possible effects. Likewise whether maternal immune activation during critical periods of pregnancy might have some role in this process of IgG brain autoantibodies remains to be seen, following the recent maternal C-reactive protein findings published by Prof. Alan Brown and colleagues (see here).

Speculations aside, the Nordahl results add to the growing body of literature on immune function showing some connection to cases of autism and how the earliest days might well be important for at least some of the autisms and beyond (stress on some of the autisms) as per the news recently on folic acid. The precise mechanics of this biological relationship currently still remain, in the most part, hidden from view, and complicated by all that heterogeneity and comorbidity present in cases. The continued focus however on subgroups and endophenotypes represents the way forward in autism research, combining also with the notion that the presentation of autism may be much greater than the sum of its behavioural dyad.

Final note: the pretty picture of a cylon included in this post, which bears no relation to the content of the post, is simply there because I've recently been revisiting some of the golden days of TV Sci-Fi and wanted to share some of them with you. The original Battlestar Galactica series with Lorne Green was a favourite, but then again so was another series about a chap returning to earth some 500 years later, give or take a year or two, and making some new friends.

----------

* Nordahl CW. et al. Maternal autoantibodies are associated with abnormal brain enlargement in a subgroup of children with autism spectrum disorder. Brain Behav Immun. February 2013.

** Braunschweig D. et al. Maternal autism-associated IgG antibodies delay development and produce anxiety in a mouse gestational transfer model. J Neuroimmunol. 2012; 252: 56-65.

*** Zhang Y. et al. Induction of autoimmunity to brain antigens by developmental mercury exposure. Toxicol Sci. 2011; 119: 270–280.

**** Lee JY. et al. Maternal lupus and congenital cortical impairment. Nat Med. 2009; 15: 91–96.

***** Palmeria P. et al. IgG Placental transfer in healthy and pathological pregnancies. Clinical and Developmental Immunology. 2012: 985646.

----------

Nordahl, C., Braunschweig, D., Iosif, A., Lee, A., Rogers, S., Ashwood, P., Amaral, D., & Van de Water, J. (2013). Maternal autoantibodies are associated with abnormal brain enlargement in a subgroup of children with autism spectrum disorder Brain, Behavior, and Immunity DOI: 10.1016/j.bbi.2013.01.084

Brain enlargement and regression in autism

I will warn you that I have blogged about this research before. By 'this research' I mean the latest paper from Christine Wu Nordahl and colleagues* from the MIND Institute and Harvard University. The paper suggested that brain enlargement, a feature mentioned more than once in autism research, might be tied into a specific type of autism characterised by symptom onset with regression of previously acquired skills. The blog post in question was on the presentation of possible biological phenotypes in autism; that is sub-groups of people with autism with shared, homogeneous biological findings. Homogeneous is not normally a word used in autism and even now I use it very sparingly.

OK lets take a step back. Regression associated with the onset of autistic symptoms is a topic previously covered on this blog. The summary of that post was that for a long time, the general idea was that regression (a loss in previously acquired skills) seemed not to be possible in cases of autism. This then changed to 'of course its possible' and the detailing of quite a few different kinds of regression including the concept of delays + regression. There is still some debate on how widespread regression is in autism and the best criteria to score regression (language, behaviour, adaptive skills, etc). Having said that with the rise of the digital age and the fact that most parents these days have a photographic and video archive of their child's growth and development from very early ages, the process is made slightly easier to examine.

Head size and growth have also been discussed in relation to autism. The general consensus being that larger heads and early accelerated brain growth shows more than a passing connection to cases of autism but with some interesting ethnic variables and certainly, by no means, an exclusive phenomenon to just autism.

So putting head size and regression together, Nordahl and colleagues present the results of a small, but quite important study which in my humble opinion charts the future course for autism research, looking at smaller sub-groups with shared commonalities. The paper itself is open-access but here is a brief summary:

• Participants were enrolled as part of the Autism Phenome Project and included 114 children with autism. Controls (n=66) were matched for age and gender (well perhaps not so much for gender).
• Children with autism were divided into two groups: regression (n=61) and non-regression (n=53) with regards to their symptom onset history using the ADI-R. Before you say ADI-R...mmm, there was some other follow-up to determine the presence of regression or not. 
• Imaging scans were conducted to ascertain total cerebral (brain) volume which was accompanied by inspection of infant records charting head circumference.
• Results suggested a few things: brain enlargement is not consistent across all cases of autism but perhaps showing slightly more connection to those cases where regression in onset was cited. Having said that brain enlargement seemed to be more of a boy than girl thing and there was a large amount of overlap between the autism and control groups. Head circumference measurements showed that head enlargement in cases of autism appeared to become more pronounced after about 4 months of age and between 6-13 months of age, there was some clear water in measurements comparing the autism-regression vs. the autism-non-regression and control groups. The regression group showing larger measurements.

Allowing for the relatively small participant numbers and the reliance on stock items from the ADI-R to denote regression or not, this is an interesting paper. If we add these findings to that of the recent neuron counting in the prefrontal cortex data allowing for difference in things like age, a very fuzzy picture starts to emerge of when and possibly where brain overgrowth might occur in some cases of autism.

More of this phenotype research please.

* Nordahl et al. Brain enlargement is associated with regression in preschool-age boys with autism spectrum disorders. PNAS. November 2011.

Head size in autism is complicated

Size and growth are some pretty important concepts related to lots of different features of the human condition. Although universal connections related to health rarely (never?) exist, there are some interesting data linking small birth weight and later childhood intellectual development for example, as well as brain size and the intelligence quotient (at least to a degree). Size and growth tend to be determined by a variable combination of both genes and environment. This meta-analysis for example, showed how the physical environment affects various growth measures, height and weight, but perhaps not as much parameters such as head circumference. This leads me to the subject of this post - head size and autism - and some of the published research and its potential implications.

Macrocephaly is the technical term where head circumference is larger than about 2 standard deviations from the average (mean); the opposite condition being microcephaly, where head circumference is smaller. There are various reasons why a larger head may occur, such as an enlargement of the brain and cases of hydrocephalus (water on the brain). In most cases in infants and children, larger head size is thought to be governed by more genetic factors such as heredity or the presence of one or more various genetic disorders.

Macrocephaly has, down the years, been associated with some cases of autism spectrum conditions. Kanner first noted that some of his original cohort had larger heads (among other things). An observation which has subsequently been reported again and again and again with prevalence estimates of macrocephaly in autism ranging from approximately 10-30%. The data does suggest some degree of heritability linked to macrocephaly, although the relationship is not entirely straight forward and the heritability implications not immediately clear. In more recent years there has been some debate about macrocephaly and autism and how factors such as ethnicity might also show an effect. This study for example first published as an abstract at IMFAR 2009, based on an Israeli cohort found little evidence of elevated rates of macrocephaly in autism in comparison to asymptomatic controls.

Why might macrocephaly be important to autism? Well I mentioned that outside of heritability, generally macrocephaly is tied into one or more genetic conditions. The theory in autism is that such findings might indicate some genetic component and hence help point towards candidate areas of interest. To see this theory in action, readers may wish to have a look at these studies (here and here) on how macrocephaly from both a case study and group study point of view might help untangle a few research strands in such a heterogeneous condition. The 'endophenotypes' way of study is something which really is the future of autism research.

There are perhaps more immediate effects potentially tied into macrocephaly presenting in autism also. This study suggested specific autistic behaviours tend to be more severe in those with larger heads, including delays  in acquiring language; the language side of things however seems to be the source of some speculation. Uta Frith joins a number of other researchers in saying that macrocephaly may imply abnormal neural connectivity. One might also expect that a larger head may indicate a larger brain and hence the maturational issue of brain growth could be implicated. Certainly studies of total brain volume in autism seem to suggest some discrepancy with control populations, driven by things like an increased cortical surface area. Such studies also seem to point to a 'brain overgrowth' as potentially being present, at least during early infancy with lots of different areas of the brain involved. Thinking about this, I was reminded of some work I have heard discussed quite a few years ago about synaptic pruning in relation to autism and the various instruments of such pruning including the MHC and endogenous opioid peptides. Indeed the work of Zagon and McLaughlin is I think absolutely fascinating, particularly their characterisation of the zeta opioid receptor and their various studies on the effects of opioid antagonists such as naltrexone on rat postnatal development. I don't know if and how it might tie into these latest findings on synaptic pruning.

I think it is important also to note that larger head size is not exclusive to autism spectrum conditions. This paper reported similar trends in head size and various other growth parameters from non-autism controls with a psychiatric disorder. Importantly however they indicated that there may be subtle differences in the timing of measurements between autism and controls which might provide some clues for the underlying processes involved.

What we can ascertain from this collected body of work is that larger heads show more than a passing association to autism spectrum conditions, and perhaps tie into some elements of the presented condition including some interesting findings on serotonin. Having said that, the recent meta-analysis of neonatal factors in connection to autism included as part of this post suggesting that the connection to head circumference might not be all it has been cracked up to be, perhaps needs to be taken on board. What I would like to see is more information on is any link to important autism co-morbidities such as epilepsy and if such 'overgrowth' is confined to early infancy, and whether it is controlled by genes or environment (or both)?

Regarding size matters...