Background: Refined grains, a major source of dietary carbohydrates, have been related to impaired glucose homeostasis and obesity. Emerging animal data suggest that in utero exposure to dietary refined carbohydrates may predispose offspring to an obese phenotype, indicating a potential role for nutritional programming in the early origins of obesity, but intergenerational human data are lacking.Objective:We prospectively investigated refined-grain intake during pregnancy in association with offspring growth through age 7 y among high-risk children born to women with gestational diabetes mellitus (GDM).Design:The analysis included 918 mother-singleton child dyads from the Danish National Birth Cohort. Offspring body mass indexzscores (BMIZs) were calculated by using weight and length or height measured at birth, 5 and 12 mo, and 7 y. Overweight or obesity was defined by WHO cutoffs. Linear and Poisson regressions were used, with adjustment for maternal demographic, lifestyle, and dietary factors.Results:Refined-grain intake during pregnancy was positively associated with offspring BMIZ (adjusted β per serving increase per day: 0.09; 95% CI: 0.02, 0.15) and risk of overweight or obesity at age 7 y [adjusted RR (aRR) comparing the highest with the lowest quartile: 1.80; 95% CI: 1.09, 2.98;P-trend = 0.032]. The association appeared to be more pronounced among children who were breastfed <6 mo. The substitution of 1 serving refined grains/d with an equal serving of whole grains during pregnancy was related to a 10% reduced risk of offspring overweight or obesity at 7 y of age (aRR: 0.90; 95% CI: 0.82, 0.98). No associations were observed between refined-grain intake and infant growth.Conclusions:Higher maternal refined-grain intake during pregnancy was significantly related to a greater BMIZ and a higher risk of overweight or obesity at age 7 y among children born after pregnancies complicated by GDM. The findings highlight pregnancy as a potential window of susceptibility associated with offspring growth and obesity risk among this high-risk population. Data with longer follow-up are warranted.