For follow-up at 4 years of age, we estimated a priori that 300 children (intervention=120;control=180) were required to detect a difference of 35% in the occurrence of ECC (RR £ 0.65),considering a study power of 80%, statistical
significance of 5%, unexposed ⁄ exposed ratio of 3:2 and a caries prevalence of 48%, the frequency documented in a recent cross-sectional study (18). Figure 1 shows the diagram of the randomized trial, from recruitment of mother–child pairs to the
assessment of outcomes at the age of 4.