“…The quality criteria in which articles presented the greatest limitation were, in particular, proof of representativeness of the sample studied in relation to the study population (in observational and experimental studies); definition of main confounding factors (in observational studies) and description of loss of participant (in observational and experimental studies). Considering that all the articles had an average or high quality and that the main risks of bias were associated with the analysis of loss of participants and confounding factors, without negative effects on the objectives of the review, we decided to systematize all the studies The analysis of sample representativeness indicates that 15 studies 16,17,19,20,29,31,34,36,[38][39][40]42,[44][45][46] used a representative sample and random selection of participants, five 15,18,21,22,37 studied a large population percentage and ten 23,24,26,27,30,32,33,35,41,43 studied populations with no characteristic of representativeness. With the exception of two studies, 25,30 representativeness was guaranteed in all others that focused on preschool children (Tables 1 and 2).…”