OBJECTIVE: To evaluate a universal obesity prevention intervention, which commenced at infant age 4-6 months, using outcome data assessed 6 months after completion of the first of two intervention modules and 9 months from baseline. DESIGN: Randomised controlled trial of a community-based early feeding intervention. SUBJECTS AND METHODS: Six hundred and ninety-eight first-time mothers (mean age 30 ± 5 years) with healthy term infants (51% male) aged 4.3 ± 1.0 months at baseline, Mothers and infants were randomly allocated to self-directed access to usual care or to attend two group education modules, each delivered over 3 months, that provided anticipatory guidance on early feeding practices. Outcome data reported here were assessed at infant age 13.7 ± 1 .3 months. Anthropornetrics were expressed as z-scores (WHO reference), Rapid weight gain was defined as change in weight-for-age z-score (WAZ) of > + 0.67. Maternal feeding practices were assessed via self-administered questionnaire. RESULTS: There were no differences according to group allocation on key maternal and infant characteristics, At follow-up (n = 598 (86%)), the control group infants had higher BMI-for-age z-score (BMIZ) (0.42 ± 0,85 vs 0.23 ± 0.93, P = 0.009) and were more likely to show rapid weight gain from baseline to follow-up (odds ratio (OR) 1.5, confidence interval (Cl) 95% = 1.1-2.1, P=0.014). Mothers in the control group were more likely to report using non-responsive feeding practices that fail to respond to infant satiety cues such as encouraging eating by using food as a reward (15% vs 4%, P=0.0O1) or using games (67% vs 29%, P.<0.001). CONCLUSIONS: These results provide early evidence that anticipatory guidance targeting the 'when, what and how' of solid feeding can be effective in changing maternal feeding practices and, at least in the short term, reducing anthropometric indicators of childhood obesity risk. Analyses of outcomes at later ages are required to determine if these promising effects can be sustained.