Autism Head Growth Characterized by Growth Surges Throughout Childhood

Background: Delineation of ASD heterogeneity requires characterization of biological phenotypes which occur in a significant proportion of ASD. Macrocephaly, HC ≥97%, is an endophenotype occurring in ~30% of ASD individuals and 37-47% of their parents (Miles, 2002). Longitudinal data, however, are conflicting; some report normal or low birth HC with accelerated growth in the first months or between 2 and 3 years; others gradual growth or decline. 

Objectives: Better characterize the "macrocephaly" phenotype.   

Methods:  Examined HC curves in a homogeneous group of 63 children (58M, 5F) with classic AD, essential phenotype & birth HC plus 3+ later measurements. Followup 8.8yr ±6.

Results: Birth HC was normal in 94% (59/63) (Z=-1 to 0.5); 48% of these became macrocephalic & 52% remained normocephalic. In both groups head growth occurred in surges of >1SD throughout childhood. In those who became macrocephalic, 47% had 1 surge of > 1SD, 44% had 2 surges and 9% had 3 surges. Growth surges of >1SD also occurred in 61% of children whose HC remained in the normal range; 52% had one growth surge and 9% had two. For both groups timing of the first surge was similar: <3yr (43%, 53%), 3-8yr (38%, 42%) & >8yr (19%, 5%). Both groups also had similar parental HC profiles; 68% of macrocephalic probands had macrocephalic fathers, compared to 60% of normocephalic; 16% of macrocephalic probands had macrocephalic mothers compared to 11% of normocephalic probands. At birth few ASD probands were macrocephalic (1.6%) or microcephalic (4.7%); all 4 individuals had surges into the macrocephalic range. Ultimate head size correlated with the number of growth surges.                                                                                                                         

Conclusions: Surges in brain growth is a more informative autism endophenotype than macrocephaly and continue throughout childhood. Ultimate head size appears to be a continuous variable, which may explain why macrocephaly has not been predictive of behavioral phenotypes or outcomes.