Eating less frequently is associated with increased obesity risk in older children but data are potentially confounded by reverse causation, where bigger children eat less often in an effort to control their weight. Longitudinal data, particularly in younger children, are scarce. We aimed to determine whether eating frequency (meals and snacks) at 2 years of age is associated with past, current or subsequent BMI.

Cohort analysis of a randomised controlled trial. Eating frequency at 2 years of age was estimated using 48 h diaries that recorded when each child ate meals and snacks (parent-defined) in five-minute blocks. Body length/height and weight were measured at 1, 2 and 3·5 years of age. Linear regression assessed associations between the number of eating occasions and BMI Z-score, before and after adjustment for potential confounding variables.

Prevention of Overweight in Infancy (POI) study, Dunedin, New Zealand.

Children (n 371) aged 1–3·5 years.

On average, children ate 5·5 (sd 1·2) times/d at 2 years of age, with most children (88–89 %) eating 4–7 times/d. Eating frequency at 2 years was not associated with current (difference in BMI Z-score per additional eating occasion; 95 % CI: −0·02; −0·10, 0·05) or subsequent change (0·02; −0·03, 0·06) in BMI. Similarly, BMI at age 1 year did not predict eating frequency at 2 years of age (difference in eating frequency per additional BMI Z-score unit; 95 % CI: −0·03; −0·19, 0·13).

Number of eating occasions per day was not associated with BMI in young children in the present study.