Objective: To determine the impact of nutritional status in a multicausal approach of socio-economic, socio-cultural, family, intellectual, educational and demographic variables at the onset of elementary school in 1987 on the educational situation of these children in 1998, when they should have graduated from high school. Setting: Chile's Metropolitan Region. Design: Prospective, observational and 12-year follow-up study. Methods: A representative sample of 813 elementary first grade school-age children was randomly chosen in 1987. The sample was assessed in two cross-sectional studies. The first cross-sectional study was carried out in at the onset of elementary school in 1987 and the second was carried out in 1998, 12-years later, when they should be graduating from high school. In 1998, 632 adolescent students were located and their educational situation was registered (dropout, delayed, graduated and not located). At the onset of elementary school were determined the nutritional status, socio-economic status (SES), family characteristics, intellectual ability (IA), scholastic achievement (SA) and demographic variables. Statistical analysis included variance tests and Scheffe's test was used for comparison of means. Pearson correlation coefficients and logistic regression were used to establish the most important independent variables at the onset of elementary school in 1987 that affect the educational situation 1998. Data were analysed using the statistical analysis system (SAS). Results: Logistic regression revealed that SES, IA, SA and head circumference-for-age Z score at the onset of elementary school in 1987 were the independent variables with the greatest explanatory power in the educational situation of school-age children in 1998. Conclusions: These parameters at an early school age are good predictors of the educational situation later and these results can be useful for nutrition and educational planning in early childhood.