Head circumference is an important aspect of childhood growth, and which may be influenced by maternal, in utero and early life exposures. We aim to model head circumference longitudinally within ALSPAC, and to examine relationships between key maternal characteristics (including maternal eating disorders) and childhood head circumference. We will then relate summary measures of change in head circumference to IQ at age 8 - although this has the potential to be extended to child development outcomes at different time points during childhood, including cognitive, motor and behavioural outcomes.

Head circumference was measured at birth and at ages 7 and 15 for the entire ALSPAC cohort, and between ages 4 and 61 months for the children in focus cohort. There are also routine data on head circumference available for all children. We aim to develop growth models for these head circumference data, to describe average patterns of growth in head circumference, and how individuals vary around these patterns. We will relate patterns of growth to prenatal characteristics (e.g. maternal and paternal education, height, and weight) and to measures of socioeconomic status. We will also examine the relationships between maternal characteristics (e.g. parity, age, smoking status, eating disorders) and child head circumference development. In all models we will conduct unadjusted analyses, then adjust for confounders, then adjust for previous head circumference measures and finally also adjust for child's height at approximate time of head circumference measure.

Multilevel models (with measurement occasions nested within children) will be used to relate head circumference to age. Fractional polynomials will be used to examine assumptions of linearity, and linear spline models used to approximate any non-linear relationships. Individual estimates of head circumference at birth, and changes in periods of approximately linear growth, will be derived and added to the ALSPAC resource for use as exposures in future models.

All models will exclude triplets/quads and allow for interactions with twin status. We will examine boys and girls growth separately, and formally test for interactions between sex and growth. We will also examine premature children separately, and formally test for interactions between prematurity and growth. Children with comorbidities known to affect growth will be excluded from all analyses.