Low birth weight and autism: rise of the population attributable risk

RIP Cheggers.

"LBW [low birth weight] accounted for 6.0% of all ASD [autism spectrum disorder] cases, 2.4% of BCD [behaviour and conduct disorder], and 6.8% of LD [learning disability] among the study population."

Those were the observations made by Sandie Ha and colleagues [1] and with it, another example of the use of the population attributable risk/fraction in the context of autism (see here for another occasion). Published in 2014 but only recently appearing on PubMed, Ha et al report results based on data from the 2011 (US) National Survey of Children’s Health, (NSCH) - a "random-digit-dial phone survey conducted between February 2011 and June 2012" - where data on birth weight and receipt or not of a diagnosis of "attention deficit/hyperactivity disorder (ADHD), autism spectrum disorder (ASD), behavior and conduct disorder (BCD) and learning disability (LD)" were available. As an aside, I've talked about the other studies arising from the NSCH program before on this blog (see here and see here for examples).

Including data pertinent to around 81,000 children aged between 2 and 17 years of age, researchers reported that around 9% of the cohort were "born with a LBW as reported by their parent" in response to the question: "What was [sampling child’s] birth weight?" There were some interesting correlates alongside those responses regarding LBW status: "children who were female, non-Hispanic black, had single mothers, had less educated mothers, were poorer, lacked insurance, were exposed to in-home smoking, or born prematurely were more likely to have LBW compared to those with normal BW."

Insofar as 'neurobehavioural disorders' (ND) also asked about: "The weighted prevalence of parent-reported ND among children ages 2 to 17 was approximately 9.9% for ADHD, 2.3% for ASD, 4.1% for BCD, and 10.6% for LD." Yes, this was a telephone-based survey where "both exposure and outcome are based on parental reporting, and thus the information may not represent actual diagnoses" but with the size of the participant numbers included, these prevalence/frequency figures still make for important reading.

Then to the main event - the population-attributable risk percentage (PAR%) and the finding headlining this post: "LBW [low birth weight] accounted for 6.0% of all ASD [autism spectrum disorder] cases, 2.4% of BCD [behaviour and conduct disorder], and 6.8% of LD [learning disability] among the study population." The authors caution that "maternal age at delivery, gestational age, and pregnancy complications could be important confounders" and were not taken into account in their analyses and could be "potential reasons for LBW" alongside undetected "congenital anomalies or genetic disorders." Caution is required.

It's not new news that birth weight might impact on something like autism risk (see here and see here). One also has to bear in mind that something like LBW may not necessarily appear in isolation to other pregnancy or birth events (see here) so a wider research agenda perhaps needs to be followed. But the size of the PAR% talked about by Ha and colleagues is not easily ignored. Taking into account that LBW for some may very well have some 'genetic' influences, one is left asking whether those more 'social' variables linked to LBW might be to some degree 'influenced' with a corresponding effect on neurodevelopmental 'consequences' reported. I say this in the context that poverty as a variable, has already been linked to some diagnoses included in the Ha study (see here)...

----------

[1] Ha SU. et al. Population attributable risks of neurobehavioral disorders due to low birth weight in US children. Adv Pediatr Res. 2014;1. pii: 2.

----------

1 in 5 children "met criteria for low language at 7 years"

Although primarily looking at the potential predictors of language outcome, the study results published by Cristina McKean and colleagues [1] revealed the rather important title heading this blog entry: "Almost 19% of children (22/1204;18.9%) met criteria for low language at 7 years."

The source of the finding was a cohort of some 1900 infants "recruited at age 8 to 10 months" who were followed until aged 7 years old and subject to quite a bit of research inspection looking at "early life factors", maternal factors and "child language ability" at various points through the childhood years. I believe this was part of the The Early Language in Victoria Study (ELVS) initiative; something that has previously created a bit of stir in speech and language circles. The authors reported that alongside the quite high percentage of children who met 'low language' criteria (based on standardized receptive or expressive language scores "≥1.25 SD below the mean"): "Child language ability at 4 years more accurately predicted low language at 7 than a range of early child, family, and environmental factors." Said low language abilities at 7 years old were also "associated with a higher prevalence of co-occurring difficulties."

Aside from pointing out that language ability at 4 years old might be quite important to language ability at 7 years old, the question that should be in most people's minds is 'why?' Why are nearly 1 in 5 children presenting with low language ability at 7 years old (and presumably at 4 years old too)? Yes, there are variables such as adverse early life factors (prematurity, birth weight, coming from a non-English speaking background, etc) that will no doubt influence various aspects of language ability, but the authors note that such factors only account for roughly 15% at most of the variation in language scores seen in their cohort (not including "child language scores at ages 2 and 4"). Ergo, there are other factors involved with regards to these findings.

In light of the McKean findings, I'm also going to draw your attention back to another occasion when language ability has been discussed on this blog (see here) and specifically: "At school entry, approximately two children in every class of 30 pupils will experience language disorder severe enough to hinder academic progress." [2] Low language (ability) is not necessarily the same as a diagnosed language disorder, and probably accounts for the variation between the studies (1 in 5 vs. 1 in 15). But in amongst a spectrum of language ability (disorder?) the questions about 'why?' still very much remain (and please, no sweeping generalisations about us 'just being better at diagnosing').

----------

[1] McKean C. et al. Language Outcomes at 7 Years: Early Predictors and Co-Occurring Difficulties. Pediatrics. 2017 Feb 8. pii: e20161684.

[2] Norbury CF. et al. The impact of nonverbal ability on prevalence and clinical presentation of language disorder: evidence from a population study. Journal of Child Psychology and Psychiatry. 2016. May 16.

----------

McKean C, Reilly S, Bavin EL, Bretherton L, Cini E, Conway L, Cook F, Eadie P, Prior M, Wake M, & Mensah F (2017). Language Outcomes at 7 Years: Early Predictors and Co-Occurring Difficulties. Pediatrics PMID: 28179482

Low gestational age is associated with risk for autism

The results published by Robert Joseph and colleagues [1] provide some blogging fodder today, observing that as part of the ELGAN (Extremely Low Gestational Age Newborns) Research Study, gestational age might matter when it comes to offspring risk of autism spectrum disorder (ASD).

Gestational age is a measure of how far along a pregnancy is but in the context of the Joseph study refers to premature birth or those "born at least 3 months early." Researchers included data for nearly 1000 children born extremely premature using a prospective (rather than retrospective) methodology who, at age 10, were "evaluated for ASD and ID [intellectual disability]." They also took into account various other 'pregnancy information' derived from both medical records and interviews with mums. This included instances of cervical-vaginal ‘infection’ among other things.

The results: over 90% of their cohort were assessed for autism/ASD and ID. Rates of ASD alone (without ID) were 3.2%. Some 3.8% of participants screened positive for autism and ID and 8.5% of participants presented with ID but not autism. Whilst these autism (with or without ID) rates might seem high, I'm not actually convinced that they are 'significantly higher' than that suggested in modern times (see here). The authors also noted that: "The lowest gestational age category (23-24 weeks) was associated with increased risk of ASD+/ID+... and ASD+/ID-."

Also: "Maternal report of presumed cervical-vaginal ‘infection’ during pregnancy was associated with increased risk of ASD+/ID+." The sorts of things included under the heading of cervical-vaginal infection were "bacterial infection (n = 4), bacterial vaginosis (n = 30), yeast infection (n = 62), mixed infection (n = 4) or other/unspecified infection (n=43; e.g., chlamydia, trichomonas or herpes, etc.)."

I agree with the authors that "low gestational age is associated with increased risk for ASD" and this research does seem to tally with other studies on this topic (see here for example). The one caveat I do want to make however is that the absolute numbers of children diagnosed with autism (with or without ID) were quite low (27 and 32 children respectively out of a total of 840) and somewhat dwarfed in comparison to those presenting with ID without autism (71 children out of 840). Still, preferential screening for autism and ID might be implied from these results; assuming that is, that the screening instrument in question is up to the job [2].

As to potential mechanisms of effect... well, it's rather difficult to pin any 'excess autism risk' to any one mechanism in light of the various factors that can accompany prematurity. Aside from the immaturity of various biological system associated with premature birth (including the brain), the obvious effect of a reduced birth weight is something to consider (see here). I am also interested in the idea that infection (using the term quite broadly) during pregnancy might also impact on offspring autism risk in light of other data (see here). Specific pathogens during pregnancy affecting offspring developmental risks have already made a mark on the peer-reviewed research literature (see here for example) and could provide a template for the processes pertinent to infection plus prematurity too.

----------

[1] Joseph RM. et al. Extremely low gestational age and very low birth weight for gestational age are risk factors for ASD in a large cohort study of 10-year-old children born at 23-27 weeks gestation.  American Journal of Obstetrics and Gynecology. 2016. Aug 13.

[2] Kim SH. et al. Predictive Validity of the Modified Checklist for Autism in Toddlers (M-CHAT) Born Very Preterm. J Pediatr. 2016 Nov;178:101-107.e2.

----------

Joseph, R., Korzeniewski, S., Allred, E., O’Shea, T., Heeren, T., Frazier, J., Ware, J., Hirtz, D., Leviton, A., & Kuban, K. (2016). Extremely low gestational age and very low birth weight for gestational age are risk factors for ASD in a large cohort study of 10-year-old children born at 23-27 weeks gestation American Journal of Obstetrics and Gynecology DOI: 10.1016/j.ajog.2016.11.1009

Risk of type 2 diabetes in autism

"Adolescents and young adults with ASD [autism spectrum disorder] were more likely to develop type 2 DM [type 2 diabetes mellitus] during the follow-up. In addition, those with ASD using atypical antipsychotics exhibited a high risk. Therefore, further research is necessary to investigate the common pathophysiology of ASD and type 2 DM."

So said the findings reported by Mu-Hong Chen and colleagues [1] as, yet again, Taiwan and their very useful National Health Insurance Research Database (NHIRD) continues to give to autism and related research.

Type 2 diabetes - the one where the pancreas don't produce enough insulin or the body's cells don't react to insulin - was the focus of NHIRD interrogation this time around, following in the scientific footsteps of previous research looking at both type 1 and type 2 diabetes in relation to autism (see here). Enrolling over 6,000 adolescents and young adults diagnosed with an autism spectrum disorder (ASD) and nearly 25,000 age and sex-matched controls "between 2002 and 2009", researchers followed participants until the end of 2011 watching for who and how many would be diagnosed with type 2 DM.

Dividing groups into adolescents and young adults, researchers reported that those diagnosed with autism were at "higher risk of developing type 2 DM than those without ASD" after adjusting for various potential forms of bias including "atypical antipsychotics use, and medical comorbidities." Further: "Short-term... and long-term... use of atypical antipsychotics were associated with a higher likelihood of subsequent type 2 DM." Ergo, yet another important, growing and potentially life-changing comorbidity appears to be 'over-represented' when it comes to the label of autism.

Accepting that it is not necessarily new news that certain pharmaceutical formulations can affect risk of type 2 diabetes [2], these are interesting and potentially important results. The focus on how maternal diabetes 'exposure' might modify risk of autism in offspring (see here) coupled to ideas about how autoimmunity inferred by type 1 diabetes history might link into some autism (see here) have tended to predominate in this area of the autism research landscape. The Chen results tap into a pretty under-appreciated idea that for one reason or another, a diagnosis of autism could potentially raise the risk of something like type 2 diabetes.

There is additional research to do on this topic. So, in these days of greater appreciation that 'autism genes' might not necessarily just be 'genes for autism' (pleiotropy), one could reasonably ask the question of whether there are subtle genetic (or epigenetic) issues influencing risk of type 2 diabetes. As per my use of the word 'epigenetic' in that last sentence, I'm also minded to bring in the work from people like the late David Barker, and the idea that birth weight might also 'program' for a greater risk for type 2 diabetes and related health issues (see here) given what's known about this area with autism in mind (see here).

That type 2 diabetes is not a wholly genetic issue is something else to bear in mind as issues such as weight and eating patterns come into the frame. Minus any sweeping generalisations, weight issues and particularly obesity are not uncommon research topics when it comes to autism (see here) and as for eating patterns and habits, well, let's just say there is some science there too (see here). Whether singularly or combined (and with potential added mention of exercise) one might already see how screening for type 2 diabetes should be added to the list of monitoring required when a diagnosis of autism is received. I might also mention some recent research chatter about leptin as being important to type 2 diabetes in kids and where that might go with regards to other autism research (see here). I'm sure there other factors too.

And, if and when type 2 diabetes in diagnosed, there are a number of positive changes that can be made to lifestyle including the idea that 'reversal' is not something totally unheard of...

So:  The Force Awakens and Dark Side of the Moon... synch or no synch?

----------

[1] Chen MH. et al. Risk of Developing Type 2 Diabetes in Adolescents and Young Adults With Autism Spectrum Disorder: A Nationwide Longitudinal Study. Diabetes Care. 2016 Mar 22. pii: dc151807.

[2] Galling B. et al. Type 2 Diabetes Mellitus in Youth Exposed to Antipsychotics: A Systematic Review and Meta-analysis. JAMA Psychiatry. 2016 Mar 1;73(3):247-59.

----------

Chen, M., Lan, W., Hsu, J., Huang, K., Su, T., Li, C., Lin, W., Tsai, C., Tsai, S., Lee, Y., Chen, Y., Pan, T., Chang, W., Chen, T., & Bai, Y. (2016). Risk of Developing Type 2 Diabetes in Adolescents and Young Adults With Autism Spectrum Disorder: A Nationwide Longitudinal Study Diabetes Care DOI: 10.2337/dc15-1807

Late/moderately preterm kids at risk for a positive autism screen

If you're sick of the sight of tinsel and/or turkey (delete as appropriate), I promise no more mention of them in this post. Just a brief introduction to the the paper by Alexa Guy and colleagues [1] (open-access) who concluded that: "LMPT [late and moderately preterm] infants are at significantly increased risk for positive autistic screen."

Megamind, incredibly handsome criminal
genius and master of all villainy!

Based on the [final] analysis of data from some 600 LMPT infant and 760 term-born infants taking part in a study initiative here in the UK, researchers looked at M-CHAT (Modified Checklist for Autism in Toddlers) scores derived from parental questioning when the child was 2 years old. They found: "LMPT infants were at significantly increased risk for a positive M-CHAT questionnaire screen compared with term-born infants, and this remained significant after application of the follow-up interview and exclusion of infants with neurosensory impairments." Further that their findings: "provide empirical evidence that screening for ASD is especially confounded in preterm populations" as a function of the the high false-positive rate found compared against the term-born controls.

M-CHAT has cropped up a few times on this blog (see here for example). In recent times, the schedule has undergone a bit of (further) revision, evolving into the M-CHAT-R/F [2] (see here for further details) to further improve on its sensitivity statistics when it comes to assessing risk of an autism spectrum disorder (ASD). Indeed, as per the Guy quote: "An M-CHAT follow-up interview is essential as screening for autism spectrum disorders is especially confounded in preterm populations" one can perhaps see where M-CHAT R/F might be particularly handy for further study of this important group.

These results add to an already sizeable volume of peer-reviewed research literature hinting that the timing of our very earliest introduction into the world might influence risk of neurodevelopmental issues such as autism and/or a positive screen for potential autism. That being said, the various factors linked to time in-utero such as birth weight as potentially impacting on development illustrate how complex an issue this might be. How we screen for autism in among the various complexities of preterm births (and lower birth weight) remains a question on many people's minds, including the possibility of using more than one instrument [3].

Music to close: Norwegian Wood.

----------

[1] Guy A. et al. Infants Born Late/Moderately Preterm Are at Increased Risk for a Positive
Autism Screen at 2 Years of Age. J Pediatrics. 2014. 5 December.

[2] Robins DL. et al. Validation of the modified checklist for Autism in toddlers, revised with follow-up (M-CHAT-R/F). Pediatrics. 2014 Jan;133(1):37-45.

[3] Dudova I. et al. Comparison of three screening tests for autism in preterm children with birth weights less than 1,500 grams. Neuropsychiatr Dis Treat. 2014 Nov 17;10:2201-2208.

----------

Guy, A., Seaton, S., Boyle, E., Draper, E., Field, D., Manktelow, B., Marlow, N., Smith, L., & Johnson, S. (2014). Infants Born Late/Moderately Preterm Are at Increased Risk for a Positive Autism Screen at 2 Years of Age The Journal of Pediatrics DOI: 10.1016/j.jpeds.2014.10.053

Neonatal jaundice and increased risk of ADHD

The findings from Chang-Ching Wei and colleagues [1] suggesting an over-representation of the diagnosis of attention-deficit hyperactivity disorder (ADHD) following a history of neonatal jaundice provides fodder for today's brief post. Based in Taiwan, one of the most impressive countries when it comes to the use and analysis of 'big data' (see here), researchers were able to identify some 25,000 participants diagnosed with neonatal jaundice and compare them with almost 70,000 non-jaundiced controls to calculate: "the incidence rate and hazard ratios (HRs) of physician-diagnosed ADHD".

"Imperial troops have entered the base"

They observed that the "incidence of ADHD was 2.48-fold greater in the jaundice cohort than in the nonjaundice cohort (3.84 vs. 1.51 per 100,000 person-years) in the study period" between 2000 and 2008. Several other variables also seemed to affect the HR of ADHD including being male, being born preterm and being a low birth weight infant. To my mind, finding such variables already potentially connected to a heightened risk for a diagnosis of ADHD [2] strengthens the Wei results when it comes to jaundice potentially also being a risk factor. The authors conclude: "A risk alert regarding neurologic consequences is urgently required after a neonatal jaundice diagnosis" bearing in mind the need for further research on the potential mechanisms at work in this proposed relationship.

I'm becoming pretty interested in the cognitive and developmental outcomes associated with cases of neonatal jaundice. As per previous posts on this blog on jaundice and autism (see here and more recently here), there are definitely grounds for quite a bit more investigation in this area. The overlap between something like ADHD with autism (at least some of the autisms) also begs the question whether there may be some tie up between the various diagnoses and something like neonatal jaundice. It certainly wouldn't be the first time that ADHD and autism have been mentioned with the same risk factors in mind as for example, with the increasingly interesting area of asthma and childhood neurodevelopmental issues (see here and see here)...

Music to finish. I assume most people have heard the very sad news about actor and comedian Robin Williams this morning. Just the other day I was introducing my brood to Mork and Mindy... RIP.

----------

[1] Wei CC. et al. Neonatal jaundice and increased risk of attention-deficit hyperactivity disorder: a population-based cohort study. J Child Psychol Psychiatry. 2014 Jul 24.

[2] Chu SM. et al. The relationship between attention deficit hyperactivity disorder and premature infants in Taiwanese: a case control study. BMC Psychiatry. 2012 Jul 23;12:85.

----------

Wei CC, Chang CH, Lin CL, Chang SN, Li TC, & Kao CH (2014). Neonatal jaundice and increased risk of attention-deficit hyperactivity disorder: a population-based cohort study. Journal of child psychology and psychiatry, and allied disciplines PMID: 25056274

A bottom-up approach to autism research?

If anything else, the recent discussions on the principles of RDoC - Research Domain Criteria (see here) - being applied to psychiatry have prompted many to question whether our current diagnostic labels are actually fit for purpose when it comes to answering the big research questions of how and why.

Bottoms up? @ Wikipedia 

RDoC, for those of you who might not know, is described by the US National Institute of Mental Health (NIMH) as an attempt to: "define basic dimensions of functioning (such as fear circuitry or working memory) to be studied across multiple units of analysis, from genes to neural circuits to behaviors, cutting across disorders as traditionally defined".

In other words, by all means use the DSM or ICD as your diagnostic key, but don't necessarily expect the genes (and epigenetics), biochemistry and brain / body physiology to also fit neatly into those categories that you diagnose with. The proof for this disparity is laid out in front of us all as labels such as autism or schizophrenia or ADHD remain in the most part a mystery when it comes to aetiology, common biological signatures (including replicative work) and consequently generalised intervention options to improve quality of life and functioning.

Even before RDoC I, like quite a few others, often thought that the reliance on something like the label of autism or autism spectrum as a research starting point offered little when it came to autism science. In a previous post, I did talk about the possibility of overcoming this issue as per the discussions on focusing on phenotypes - smaller subgroups on the autism spectrum - characterised by something like response to intervention for example (see here). Indeed, my recent droning on the area of dietary intervention for autism (see here) exemplified how such a shift in focus might yet yield some testable results.

Enter then another modified example of this approach with the paper by Lisa Unwin and colleagues* (open-access version here) and their discussions on a bottom-up approach to autism research. I admit that I was always going to be interested in this paper because of (a) the presence of Andrew Whitehouse on the authorship list (see previous posts here and here) and (b) mention of the 'autisms' over the more 'unitary' label of autism (see here).

As Unwin et al describe, the bottom-up approach used in the paper refers to "known aetiological risk factors, and whether individuals exposed to these risk factors have a more homogenous phenotype". In this case, low birth weight and independently maternal use of SSRIs during pregnancy (see here) were the starting points and then "examining the homogeneity within the groups based on medical complaints such as sleep problems and gastrointestinal complaints in addition to core features of ASD such as social behavior, language characteristics, and severity". Actually on the point of anti-depressant use during pregnancy and autism risk, I'm minded to bring in other, more recent evidence that suggests risk may have been over-inflated?

The first part of the Unwin paper looked at those children with an ASD where maternal SSRI use was reported during pregnancy and suggested that "children with ASD whose mothers took SSRIs during pregnancy were significantly more likely to experience gastrointestinal complaints during childhood" compared with no maternal SSRI history. The second study included in the paper examined children with ASD with a low birth weight compared against those with a normal range birth weight. They suggested "greater sleep disturbances" to be present in the former group.

Whilst the participant numbers included in this paper were low and the results are crying out for replication, I find this approach to be a breath of fresh air when it comes to autism research. Not only because of the realisation that there may be different phenotypes within the autism spectrum as a function of suggested risk factors - of which there may several (see here and here for example) - but also because the authors expanded their horizons outside of just looking at the core behavioural symptoms associated with autism. Indeed the focus on bowel issues (see here) and sleep issues (see here) reflect how frequently these variables crop up and how, for some people, they can so significantly affect quality of life.

There are some other nuggets of research gold to come out of this paper which are worthy of comment. So, "it seems reasonable that environmental factors may be related to the expression of non-core ASD symptoms among these children rather than to any variance in core symptomatology". In other words, autism might actually be greater than the sum of its triadic (sorry dyadic) parts. As per my chatter about the 'autisms' previously, the authors also note: "A key question facing the field is whether the long-held view that autism is a unitary disorder with a single causal pathway is correct, or whether autism may best be conceptualized as an umbrella term for a collection of behavioral disorders resulting from a range of causal pathways, analogous to cerebral palsy". Not surprisingly the authors conclude that the plural autism might be "a more accurate representation".

Tom Insel of the NIMH is quoted along similar lines with his phrase: "Clearly autism is many disorders". I'd have to say that I agree and that 'range' of conditions falling under the autism umbrella seems to be growing all the time.

Music maestro.... Common People by Pulp?

----------

* Unwin LM. et al. A "bottom-up" approach to aetiological research in autism spectrum disorders. Front Hum Neurosci. 2013 Sep 19;7:606. doi: 10.3389/fnhum.2013.00606.

----------

Unwin LM, Maybery MT, Wray JA, Whitehouse AJ. (2013). A "bottom-up" approach to aetiological research in autism spectrum disorders. Front Hum Neurosci. DOI: 10.3389/fnhum.2013.00606

Autism risk and interpregnancy interval

Pregnancy issues and adverse birth factors 'associated' with the development of autism in offspring is a topic which has cropped up more than once on this blog, as autism research strives to identify as many possible 'risk' factors potentially linked to symptom onset. Lots of different birth-related variables have been analysed and put forward as potential candidates related to risk (see this post) ranging from birth weight (see here) to birth order (see here) and even season of conception/birth (see here).

Let me out... @ Wikipedia  

That's not to say however that anything concrete in terms of generalised offspring risk of autism has emerged from these various lines of research inquiry. Indeed, as per the study by Schieve and colleagues* (which I've already blogged about) hinted, many of these factors may be contributory but not necessarily 'causative' of autism when looked at on a population scale. This bearing in mind that I've not introduced the various 'exposure' events during pregnancy which have also been linked to offspring autism risk (e.g. the emerging valproate story) and the idea that population risk does not necessarily always translate into personal circumstances and risk.

Another factor which has seen some research action is the idea that having children in close temporal succession to one and another - a short interpregnancy interval - might also elevate the risk of the second child presenting with autism. The paper by Cheslack-Postava and colleagues** (full-text) hinted at this effect as per their conclusion: "children born after shorter intervals between pregnancies are at increased risk of developing autism". I note that Dr Emily Deans over at Evolutionary Psychiatry carried some discussion on this paper too (see here).

The more recent paper by Nina Gunnes and colleagues*** adds to the literature on this topic; indeed coming to pretty much the same conclusion: "interpregnancy intervals shorter than 1 year were associated with increased risk of autistic disorder in the second-born child". Based on yet more analysis out of Norway (although I am unsure whether this was a MoBa study or not), researchers looked at the records of several thousand sibling pairs in order to identify the length of the interpregnancy interval (IPI) and whether autism was mentioned in the records of second-born children. Their conclusion about short IPI and an elevated risk of autism in second-born children seemed to be particularly pertinent to those children born 9 months after their sibling compared with those born 3 years or later after their sibling.

A few points are worthy of mention. The very discerning readers out there might have already spotted a couple of familiar names attached to the Gunnes paper authorship in the form of Mady Hornig and Ian Lipkin (see this quite recent post).

I note also the authors suggest that a "depletion of micronutrients" might have something to do with the explanation for the short IPI-autism association, which carries hints of the late David Barker's hypothesis (see here) and is pretty much in line with what Dr Deans previously mentioned. Indeed, to reiterate her discussion about baby 'sucking out' whatever nutrients it needs from its host (i.e. mum) I can remember similar words being told when my/our brood were due for an appearance. With all the current fascination on things like folic acid and autism (see here and here), one might very easily say that there is a possible link to be had there, bearing in mind Gunnes and colleagues did not assay for or report on maternal or offspring folic acid levels at any point during their study.

If also I had to play devil's advocate on such 'association' research I might point out that looking at the IPI alone and knowing relatively little about the family or offspring in terms of their lives is still methodologically problematic. We don't for example know about any medical or psychiatric familial history which might also be an important modifier of offspring risk. I assume the authors already controlled for whether sibling number one had a formal diagnosis of autism (as per the autism recurrence data previously discussed), but did they for example, ask about the potential presence of sub-clinical signs and symptoms associated with something like the broader autism phenotype for example? Were they also able to comment on any additional siblings after child number two and the elevated risk or not for them presenting on the autism spectrum either alone or as a function of IPI?

Then there's the volume of research suggesting that a short IPI might also increase the risk of reduced birth weight**** or the risk of preterm birth***** which I assume have been controlled for, but still one wonders about their impact on the presentation of offspring autism and any wider links (see here). I might also draw readers' attention to an interesting correspondence from Downs & Jonas****** (full-text) with regards to research suggesting a link between short IPI and risk of offspring schizophrenia. In short(!), one has to be careful of making too much of such association data at the current time.

That being said, I don't want to take anything away from the Gunnes study and results. It was a well-powered study and they got what they got. Their data also add to the various other information suggesting that when it comes to having children, mums (and dads) are advised to give themselves a bit of breather between kids.

Some music to finish. How about Robbie & Kylie?

----------

* Schieve LA. et al. Have secular changes in perinatal risk factors contributed to the recent autism prevalence increase? Development and application of a mathematical assessment model. Ann Epidemiol. 2011 Dec;21(12):930-45.

** Cheslack-Postava K. et al. Closely spaced pregnancies are associated with increased odds of autism in California sibling births. Pediatrics. 2011 Feb;127(2):246-53.

*** Gunnes N. et al. Interpregnancy Interval and Risk of Autistic Disorder. Epidemiology. 2013 Sep 16.

**** Smits LJ. et al. The association between interpregnancy interval and birth weight: what is the role of maternal polyunsaturated fatty acid status? BMC Pregnancy Childbirth. 2013 Jan 25;13:23.

***** De Franco EA. et al. A short interpregnancy interval is a risk factor for preterm birth and its recurrence. Am J Obstet Gynecol. 2007 Sep;197(3):264.e1-6.

****** Downs JM. & Jonas S. Short inter-pregnancy interval and schizophrenia: overestimating the risk. Br J Psychiatry. 2012; 200: 160.

----------

Gunnes N, Surén P, Bresnahan M, Hornig M, Lie KK, Lipkin WI, Magnus P, Nilsen RM, Reichborn-Kjennerud T, Schjølberg S, Susser ES, Oyen AS, & Stoltenberg C (2013). Interpregnancy Interval and Risk of Autistic Disorder. Epidemiology (Cambridge, Mass.) PMID: 24045716

Autism and ID: Born again?

The recent papers published by Amanda Langridge and colleagues* (open-access) and Venla Lehti and colleagues** set some cogs running in my grey/pink matter recently with their research focus on our very earliest days and what (if any) risk for the autism spectrum disorders (ASDs) there may be when things don't go as smoothly as expected during that magical nine months and just before/after.

A new child is Björn @ Wikipedia  

I've previously talked about factors during pregnancy and parturition (child birth) in relation to autism as per posts like this one from quite a while back and all that newer stuff from initiatives like CHARGE. Combined with investigations examining everything from season of birth to birth weight to birth order, there is quite a bit of peer-reviewed literature on this area with autism in mind.

Having said all that, I don't want to give any false impression that there are any hard and fast rules about in-utero or birth factors being linked to autism, because there aren't. Indeed, it's all even further complicated by lots of other conditions potentially being linked also to pregnancy and birth and a flurry of speculation.

Nevertheless, there are some interesting points raised by each of the articles presented today which are worthy of discussion. So in turn:

The Langridge paper:

• Open-access so please do have a look at it for yourself. A huge data mining study examining the "total population data sets of children diagnosed with ID and ASD in Western Australia (WA)" between 1984 and 1999 (included among N=383,153) to see if there were any links between "maternal conditions and perinatal factors for all WA children subsequently diagnosed with ASD, with or without ID, and children with varying severity of ID, and compare findings to the rest of the birth cohort of unaffected children". Lots of description of the population and how autism and intellectual disability (ID) were ascertained. Also how this huge dataset were analysed.
• Results: yes, quite a few depending on how the data were modelled and what background characteristics were controlled for. Notably suggesting that those diagnosed with an ID were more likely to have experienced some episode or event in their earliest days when compared with those diagnosed with an ASD. I'll leave you to pick out the associations but there were some interesting findings related to gestational diabetes, threatened abortion before 20 weeks and pregnancy hypertension.
• Soundbite: "small head circumference was associated with reduced risk of ASD" bearing in mind head size and autism is a complicated area.
• Another soundbite: "These findings support the concept that ID and ASD may lie on a continuum, as opposed to being different clinical entities, and may explain why there are various ID subtypes of ASD (i.e. ASD with and without ID)". I can't fault their logic. Perhaps another addition to the spectral model of behaviour and psychiatry since that is the direction things seem to be heading these days.

The Lehti paper examined IVF (In vitro fertilisation) and...

• A very organised paper which bluntly asked: "Does IVF increase the risk of autism spectrum disorders (ASDs)?"
• Another very big participant group (autism: n=4164, matched controls: n=16,582); all born in Finland.
• Result: "This study showed no increased risk of ASDs in children born after IVF" as per other studies in this general area.
• But...  and it is a but, the connection in particular between IVF and boys with Asperger syndrome (AS) might need to be looked at with greater assiduity. Having said that, any study had better be done sooner rather than later since AS as an independent diagnostic entity is due to be phased into ASD in the latest DSM revision in May (2013). Please note: this does not mean that people will lose their AS diagnosis though.

Combined, these papers add quite a bit to the literature on birth factors in relation to autism bearing in mind risk is risk and not proof. Although events in-utero and during childbirth have always been on the autism research radar, I am noting a renewed interest in all things early development with autism in mind in recent years as per areas like maternal immune activation and autism risk (think Paul Patterson and mice) and some continued interest in the environment (think pollution and correlation for a start). The suggestion that the placenta for example may play a "potent role in autism risk" represents quite an exciting research opportunity, mirroring research on other health-related conditions. Thin-fat body anyone?  

----------

* Langridge A. et al. Maternal conditions and perinatal characteristics associated with autism spectrum disorder and intellectual disability. PLoS ONE. 2013; 8: e50963.

** Lehti V. et al. Autism spectrum disorders in IVF children: a national case-control study in Finland. Hum Reprod. 2013 Jan 4.

----------

Amanda T. Langridge, Emma J. Glasson, Natasha Nassar, Peter Jacoby, Craig Pennell, Ronald Hagan, Jenny Bourke, Helen Leonard, & Fiona J. Stanley (2013). Maternal Conditions and Perinatal Characteristics Associated with Autism Spectrum Disorder and Intellectual Disability PLoS ONE : 10.1371/journal.pone.0050963

Birth weight and autism again (and again)

Following on from the news that low birth weight might put a person at greater risk of autism, there has been a flurry of activity on this topic. Lots of headlines about the 'increased risk' and lots of discussion about it on various social media and other platforms.

Lo and behold pretty much 24 hours after the 'hype' about the paper, details of another paper drop into my inbox, this time from Schieve and colleagues* published in the Annals of Epidemiology. After some more mathematical and statistical wizardry, Schieve and co. come to the conclusion that various birth factors might well be linked to some cases of autism but on the whole, ".. the contribution of many of these factors to the recently observed ASD increase is likely minimal".

I can't pretend to be able to offer a definitive overview of the mathematical model used in this recent study but they applied it to various pregnancy- and birth-related factors including: low birth weight (and very low birth weight), preterm births, multiple births, caeasarean section deliveries, breech presentation, and use of assisted reproductive technologies, all based on existing datasets. Their main message, that less than 1% of the observed increase in cases of autism is likely due to these factors.

Understanding that this paper does not rule out an increased risk of autism from various pregnancy and birth factors, it does suggest that the current data on such factors does not yet provide a satisfactory answer to questions of why so many more children (and adults) are being diagnosed with an autism spectrum condition. I may be looking at apples and oranges here (risk vs. reasons for an increase in numbers of cases) but ultimately increased risk is going to potentially tie into the wider argument on the numbers of cases.

Whilst not advocating that anyone place their universal trust in maths and statistics, particularly when applied to heterogeneous conditions like autism, I do wonder whether we are seeing another piece to the puzzle when it comes to investigating autism on the basis of phenotypes and clustering of symptoms and characteristics. My mind wanders back to the recent news from Prof. David Amaral on biological phenotypes and whether this is just one more factor to add to the array of others proposed. Such pregnancy and birth factors may be a major component of one person's 'route' to an autism spectrum condition; for others, an insignificant drop in an ocean of other risks.

* Schieve LA. et al. Have secular changes in perinatal risk factors contributed to the recent autism prevalence increase? Development and application of a mathematical assessment model. Annals of Epidemiology. 2011.

Birth weight and autism again

This is not the first time that the issue of birth weight in connection to autism has been discussed on this blog. Earlier this year a meta-analysis of various birth issues potentially linked to autism was published and blogged about which included birth weight, low birth weight, as a potential connector. The strength of that paper lay in the fact that it was a meta-analysis; so kinda the top of the tree level of inquiry.

The journal Pediatrics again publishes a study discussing the birth weight - autism connection. Jennifer Pinto-Martin and colleagues* report on a prospective population-based study that suggests that children born weighing less that 2kg (4.4lbs) may be five times more likely to present with an autism spectrum condition than those born heavier.

The details:

• A regional cohort of children (n=1,105) followed as part of the Neonatal Brain Hemorrhage Study during 1984 and 1989 were included for analysis. Infants weighing between 501 and 2000g were prospectively enrolled and followed through childhood into adulthood. 
• At 16 years old, just over half of the original group (n=623) were screened for autism based on either a previous diagnosis of autism or 'liberal' cutoff scores on the SCQ or other measure. Positive screens were detected in 117 of the 623 (18.8%). 
• Author diagnostic testing of a proportion of positive-screeners and negative-screeners was carried out when participants were 21 years old using those gold-standards, ADOS and ADI revealing that of 70 positive screeners at 16 years old, 11 (14.3%) were confirmed as presenting with an autism spectrum condition and 3 out of 119 negative-screeners (2.5%) met criteria for an autism spectrum condition.
• Balancing out positive and negative-screener results suggested an estimated prevalence rate of 5% for the entire cohort. This is compared with a 1% figure produced by the CDC back in 2006.

There are a few other details from this study which, whilst not significant at 5%, were nevertheless interesting. So the lower the birth weight, the higher the risk of autism and low birth weight was more often linked to autism in boys than girls. There are also some limitations set by the authors on the current dataset in view of the lack of normal birth weight control groups and the quite high loss at follow-up statistics.

I note that this study has again 'hit' the media headlines hard with places like the BBC reporting on it. With all due credit they also cite people like Prof. Dorothy Bishop who provides a good balance in that many children with autism are not born low birth weight and many low birth weight children do not have autism. Others have also discussed the important issue of low birth weight and learning disability as being perhaps a more primary relationship.

You can tell that this research, as with any longitudinal investigation, is a real labour of love for Pinto-Martin and colleagues. The results are interesting and due credit should be given to the possible 'risk' that comes with low birth weight without however getting too excited about the 'directness' of any relationship between autism and birth weight.

* Pinto-Martin JA. et al. Prevalence of autism spectrum disorder in adolescents born weighing <2000 grams. Pediatrics. October 2011.