Interindividual differences in the response to dietary intake are, in some cases, genotype dependent. Moreover, genotype-environment interactions may appear when the impact of lifestyle factors (e.g., diet) on a phenotype (e.g., BMI > 30 kg/m(2)) differs by genotype. A case-control study (obese subjects vs. normal weight controls) was conducted to assess a possible effect modification on obesity risk of the Gln27Glu polymorphism for the beta(2)-adrenoceptor gene depending on dietary intake. The sample included 159 subjects with BMI > 30 kg/m(2) and 154 controls with BMI < 25 kg/m(2). The allele frequency for the Glu27 polymorphism, as assessed by the polymerase chain reaction-restriction fragment length polymorphism methodology, was 0.40 in cases (obese) and 0.37 in controls (lean), which was similar to that of other Caucasian populations. The dietary intake was estimated by using a previously validated food frequency questionnaire. Obesity incidence was not directly affected by the polymorphism [odds ratio (OR) = 1.40; P = 0.246]. However, a significant interaction (effect modification) between carbohydrate (CHO) intake and the presence of the Glu27 variant in the probability of obesity was apparent. Thus, females with the polymorphism and a higher CHO intake [>49% energy (E)] had a higher obesity risk (OR = 2.56, P = 0.051). The product-term introduced in the logistic model to assess effect modification revealed a marginally significant interaction (P = 0.058) between both factors. Furthermore, a high intake of CHO (E > 49%) was associated with higher insulin levels among women carrying the Gln27Glu polymorphism (P < 0.01). This gene-nutrient interaction emphasizes the importance of examining the outcome of some obesity-related mutations depending on lifestyle (including diet) and may explain the heterogeneity of findings from previous studies.