Objective To identify dietary patterns in pregnant women and investigate whether they are associated with sociodemographic factors, lifestyles, gestational diseases and pregestational body mass index.

Design and Methods Longitudinal study was carried out with 1482 pregnant women in two cities in the State of Rio de Janeiro, Brazil. A FFQ was applied retrospectively to assess diet at the third trimester of gestation. Principal components analysis was used to identify dietary patterns. Multiple linear regression model was used to study the associations between diet and covariates.

Results Four factors were identified: “Prudent” pattern (dairies, cracker, fruits and meat); “Traditional” (rice, beans, vegetables, bread, butter and sugar); “Snack” (salty snacks, sandwich cookies, and chocolate); “Western” (fast food, processed meat, eggs, sweet drinks, and other food with high carbohydrate content). The “Prudent” pattern was positively associated with maternal age (B=0.012, p=0.04), per capita family income (B=0.253 p<0.001) and negatively associated with pregestational overweigh (B=-0.202, p=0.008) and gestational anaemia (B=−0.060, p=0.04). The “Traditional” was inversely associated with gestational diabetes (B=−0.551, p=0.002), city (B=−0.189, p=0.002) and positively with smoking (B=0.179, p=0.038). The “Western” was inversely associated with instruction level (B=−0.024, p=0.05), pregestational obesity (B=−0.210, p=0.05) and positively with smoking (B=0.187, p=0.05) and city (B=0.138, p=0.04). The “Snack” pattern was positively associated with alcohol consumption (B=0.274, p=0.01), high social class (B=0.121, p=0.03) and inversely associated with maternal age (B=−0.044, p<0.001).

Conclusion Four distinct dietary patters were identified which may be useful for further research concerning maternal diet and health outcomes among mothers and their offspring.