OBJECTIVE: To estimate the effect of income at birth on adulthood tooth loss due to dental caries in 539 adults from the 1982 Pelotas birth cohort. METHODS: Family income was collected at birth. Tooth loss was clinically assessed when individuals were aged 31. Dental visit and oral hygiene at age 25 were considered mediators. Confounders included maternal skin color, and individual's skin color, sex, and income in adulthood. Marginal structural modeling was used to estimate the controlled direct effect of income at birth on tooth loss due to dental caries that was neither mediated by the use of dental service nor oral hygiene. RESULTS: Forty-three percent of the individuals of low income at birth lost one/two teeth, and 23% lost three or more; among those non-poor, the prevalence was 30% and 14%, respectively. Poor individuals at birth had a 70% higher risk for missing teeth in adulthood than those non-poor. The risk of losing one/two (risk ratio 1.68) and three or more teeth (risk ratio 3.84) was also higher among those of low income at birth. CONCLUSIONS: Economic disadvantage at birth had an effect on tooth loss due to dental caries at age 31 not mediated by individual risk factors.