BACKGROUND: The ability to taste 6-n-propylthiouracil (PROP) may be associated with body composition, but previous findings from observational studies are conflicting and cannot be interpreted causally. The aim of this study was to estimate the causal association between PROP taster status and body composition in a population-based cohort study. METHODS: The study was embedded in a population-based prospective birth cohort study. The TAS2R38 genotype (rs713598) was used as an instrumental variable (IV) to obtain unbiased effect estimates of the relation between PROP taster status and body weight (n=3778). Adiposity measures included body mass index (BMI) and fat mass measured by dual- energy X-ray absorptiometry scan at the child's age of 6 years. Associations were investigated using both ordinary linear regression (OLS) and two-stage least squares regression (2SLS). RESULTS: Non-taster girls had higher BMI standard deviation scores (SDS) and higher body fat as compared with taster girls (results from linear regression BMI SDS: -0.09, P=0.023, body fat mass (%): -0.49, P=0.028). The TAS2R38 genotype predicted PROP phenotype (F=240), indicating a strong IV. The 2SLS effect estimates were imprecise but similar to the observational estimates (-0.08 for BMI SDS and -0.46 for body fat mass %) and were not significantly different from the OLS results (Hausman test: P>0.10). For boys there were no differences observed between tasters and non-tasters. CONCLUSIONS: Our findings suggest a causal relation between PROP taster status and body weight among 6-year-old girls; Mendelian randomization was consistent with conventional estimates. In contrast, body weight among boys appeared to be independent of the PROP taster status. Further research should focus on possible underlying pathways, such as dietary behavior.