BackgroundDevelopment of excess adiposity has been hypothesised to originate in early life. Animal studies have shown that protein intake in pregnancy might influence offspring fat metabolism and adiposity. Macronutrient ratio in human pregnancy seems to be important for child blood pressure and glucose tolerance; less is known about the influence on offspring adiposity. We investigated the relation between maternal dietary protein intake during pregnancy and child anthropometric measures and biomarkers of adiposity and glucose metabolism. MethodsWe used a prospective cohort of 965 Danish pregnant women recruited in 1988–89 with offspring followed up at 20 years. This population was 80% of a consecutive sample of 1212 women carrying singletons and attending prenatal care in the city of Aarhus, Denmark. For the follow-up study, mothers were first informed about the study by mail and then the offspring were asked to fill out a web-based questionnaire and attend a clinical examination. Almost 95% (n=915) of mother–child pairs could be identified in the central registration. 690 of the offspring filled out the questionnaire and 443 participated in the clinical examination. Macronutrient intake was collected in gestational week 30, and we divided protein according to source (animal, vegetable including cereals). Offspring body-mass index (BMI) and waist circumference were recorded at follow-up (n=696), and biomarkers were quantified in a subset (n=443) of participants. We used multivariable linear and logistic regression to calculate effect estimates and 95% CIs for a 1:1 substitution of protein for carbohydrates in percent energy. All models were adjusted for maternal energy intake, maternal age, maternal education, parity, maternal pre-pregnancy BMI, maternal smoking during pregnancy, and sibling overweight. We accounted for missing data by the missing indicator method because the covariates were poorly correlated with the exposure and were therefore weak confounders. The study was approved by the Danish Data Protection Agency and the Danish Council of Ethics. FindingsAt baseline, mean maternal age was 29·1 years (SD 4·1), 373 (54%) of 695 women reported higher academic or university education, and 411 (59%) were nulliparous. Mean pre-pregnancy BMI was 21·3 kg/m2 (SD 2·9) and 242 women (35%) reported smoking during pregnancy. Offspring mean BMI was 22·4 kg/m2 (SD 3·1) and waist circumference was 79·8 cm (9·3) for women and 84·3 cm (9·3) for men; 124 (18%) of 695 participants had a BMI of at least 25 kg/m2. We found that a 1:1 substitution of animal protein for carbohydrates increased the risk of a BMI of at least 25 kg/m2 among female (quartile 4 vs quartile 1 risk ratio 3·36, 95% CI 1·52–7·42) and male (2·24, 0·93–5·40) offspring in a multivariable model that adjusted for covariates including maternal pre-pregnancy BMI and smoking, and sibling overweight. These results seemed to be accounted for by protein from meat sources. Consumption of vegetable protein in the highest quartile was not associated with overweight but was related to reduced LDL cholesterol in male offspring. InterpretationProtein intake from animal sources, primarily meat products, consumed during pregnancy seems to increase the risk of overweight in offspring; this association seemed to be stronger for female offspring. Although we cannot exclude potential residual confounding and selection bias, we adjusted for important and relevant confounders. In a subanalysis on selection bias, we found no significant differences for maternal age and BMI, parity, birthweight, gestational age, and maternal protein intake during pregnancy among those with and without outcome data. We were unable to estimate whether part of the association was accounted for by postnatal diet. FundingThe Danish Agency for Science, Technology and Innovation (DSF: 09-067124, DSF: 09-063072, DFS: 2101-06-0005, FSS: 09-065631).