A Longitudinal Study of Head Circumference Trajectories in Autism and Autistic Traits

We examined HC trajectories between birth and 15-years in young people with ASD and those with elevated autistic traits compared to controls. Children with ASD had significantly larger HC across this period. Differences in HC were attributed to generalized dysregulation of growth in those with ASD. Conversely, children with elevated autistic traits had reduced HC compared to controls. These preliminary findings go against the idea of a pathophysiological continuum of autistic traits and ASD, highlighting potentially distinct aetiological differences that may separate clinical versus non-clinical cases.

Previous research on HC in ASD populations is typically restricted to the first 2–3 years of life (Constantino et al., 2010), and few studies prospectively chart HC beyond this period. Our study provides evidence that HC is consistently larger in ASD, as demonstrated by the lack of an age-by-group interaction, whilst also being robust against adjustment for multiple confounders. The timing of this increase has remained elusive in the literature, with some studies reporting the onset between 1 and 3 months-of-age consistent with the findings from our study (Fukumoto et al., 2008; McKeague et al., 2015), whereas others report enlarged HC from 6 months onwards (Courchesne et al., 2003; Hazlett et al., 2005; Webb et al., 2007). Our study supports the view that these changes in HC are continuous and not time-limited to early development.

Previous literature has suggested that enlarged HC may only affect a subgroup of individuals with ASD. For example, one meta-analysis reported macrocephaly (HC > 97th percentile) in 15.7% of individuals with ASD (versus 3% controls) and was more common in individuals with IQ scores < 70 (Sacco et al., 2015). In our subgroup analyses, the largest effect-size differences between ASD and controls were observed when participants with ASD had CLN, suggesting that shared mechanisms might lead to larger HC and compromised intellectual functioning, or alternatively that cognitive impairments may be an important mediator in larger HC within ASD. Importantly, enlarged HC was observed in the ASD group regardless of additional CLN, however the coefficient size was reduced when the CLN group were removed. In the general population, head growth during infancy is a significant predictor of later IQ (Gale et al., 2006). A birth cohort study found greater increases in HC and height during the first 5 years of life were associated with higher IQ in childhood, whereas greater weight gain after 1 year was not (Kirkegaard et al., 2020). Our study suggests there may be an opposing effect of IQ on head growth in ASD samples, however confirmation will require larger samples.

Whether large HC in ASD is related to a more general marker of dysregulated growth is debated. When controlling for height, there were no longer group differences between ASD and controls in HC, suggesting that general growth, irather than brain growth specifically, may be dysregulated in ASD. This is consistent with a review (Green et al., 2015), citing increased physical growth - including HC, height, weight and BMI - in children with ASD compared to typically developing individuals. The review also highlights that due to the paucity of longitudinal studies in adolescent and adulthood, it is unknown whether height and weight follow relative increased acceleration in childhood followed by deceleration in those with ASD as seen for HC and brain growth. The timing of dysregulation may differ for height compared to HC, as increases to HC according to unadjusted post-hoc t-tests occurred earlier (at 2 months) than height increases (at 9 months) in those with ASD. These results were no longer significant when controlling for multiple comparisons, however it is also possible that small sample sizes at individual timepoints lack sufficient statistical power. This contrasts to one study which reported overgrowth in length/height earlier at 4-months, and increased HC at 8-months in ASD (Campbell et al., 2014). Some studies report that HC in ASD varies independently of height (Dawson et al., 2007; Miles et al., 2000), while others suggest that they covary, reflecting more generalised growth processes (Chaste et al., 2013; Chawarska et al., 2011; Dissanayake et al., 2006). Our findings suggest that general growth, including HC trajectories, are altered in ASD.

There are several possible reasons for the association between autistic traits and reduced HC. First, the aetiology of ASD may not be best explained through a continuum but instead by distinct symptom profiles. Whist research has shown that the SCDC has strong specificity and sensitivity to detect ASD, only 65% of the autistic traits group were diagnosed with ASD in our sample. The SCDC predominately captures deficits in social communication/reciprocity and does not include items on repetitive behaviours or restricted interests. There is evidence linking enlarged brain size during childhood in ASD and poor performance on repetitive behaviour scales, as well as with delayed language and skill regression (Courchesne et al., 2003; Lainhart et al., 2006; Nordahl et al., 2011). It is therefore possible that neurobiological mechanisms underlying enlarged HC show a greater affinity for certain clinical outcomes over others. Furthermore, it is possible that the SCDC captures general psychopathology, as social difficulties are common across many disorders. A more detailed clinical assessment of the SCDC group is needed to ascertain whether the reductions in HC can be attributed to autistic traits or other disorders.

HC and brain size are strongly correlated (Bartholomeusz et al., 2002; Hshieh et al., 2016; Piven et al., 1996), hence HC can be used as a proxy measure of neurodevelopment.³⁷ Atypical cell proliferation and reductions in neuronal pruning are thought to underlie the growth differences in HC observed in ASD and cause an overabundance of cortical neurons, particularly in frontal and temporal brain networks which underlie social and language processes (Courchesne et al., 2018). The are several potential genes associated with ASD which affect brain size and growth in general (e.g., mutations to chromodomain helicase DNA binding protein 8 (CHD8) (Sugathan et al., 2014) (Adam et al., 2022) However, while larger HC may be a viable biomarker for ASD, it is also a biomarker for several other genetic conditions (McCaffery & Deutsch, 2005).

Several limitations should be considered when interpreting this study. As with most longitudinal cohorts, missing data reduced sample size and representativeness. The majority of participants (94%) were of white ethnicity, our findings therefore have limited generalizability to other ethnic groups. The use of parent confirmation of a diagnosis of ASD may be less valid than formal diagnostic tools. However, the number of ASD cases was similar to that of previous ALSPAC studies which used multidisciplinary assessments from health and educational records (Williams et al., 2008). The subgroup of ASD with CLN was small and may lead to the model overfitting the data when regression models are used, and so these findings require replication in a larger sample. Greater clinical profiling of the ASD and autistic traits group beyond CLN subgroups would be useful. We did not observe any sex-by-group interactions with HC, although we were likely underpowered to do so due to females representing only 21% (n = 15) of the ASD sample. Strengths of the study include the use of a representative birth cohort with multiple (> 10) timepoints for HC measurements. Previous longitudinal studies typically examine 3–4 timepoints, often limited to early development (Dissanayake et al., 2006; Fukumoto et al., 2008; Hazlett et al., 2005).

Findings from this study are consistent with previous literature that large HC may be considered an accessible biomarker for ASD, at least for a subgroup of individuals with ASD, and potentially complement behaviour assessments for diagnosis throughout childhood (Denier et al., 2022; Muratori et al., 2012; Sacco et al., 2015). Greater profiling of ASD growth and clinical trajectories is required to gain a better picture of this subgroup, particularly in adulthood. Our findings suggest that HC enlargement is specific to those with clinical diagnoses as we do not find evidence of enlargement in the subclinical trait sample. It will be important for future studies to continue to monitor HC across a greater time-period to see the full extent of the developmental trajectory.