To the Editor: Genetic factors interact with modifiable risk factors to influence the risk of cardiovascular disease [1]. They also modulate the effectiveness of interventions for the primary prevention of cardiovascular disease [2]; thus, the identification of the genes involved may help to more efficiently prevent cardiovascular disease [1, 2]. With regard to glucose metabolism, gene–environment and gene–nutrient interactions have previously been demonstrated for the polymorphism in the peroxisome proliferator–activator receptor γ2 gene (PPARG2) that results in the substitution of Pro for Ala at codon 12 [3]. In addition, carriers of the Ala-encoding allele have an improved response to aerobic exercise in terms of glucose metabolism [3]. Because PPARG2 is also considered to play a role in the process of atherosclerosis [3], we tested whether PPARG2 interacts with the effects of lifestyle intervention on cardiovascular traits. Specifically, we studied the effect of this common polymorphism on the increase in flowmediated vasodilation (FMD) and the decrease in serum levels of C-reactive protein (CRP) during a 9 month lifestyle intervention in 166 non-diabetic subjects (58 men, 108 women; age 46±11 years). The population was at increased risk of type 2 diabetes and cardiovascular disease because of one or more of the following risk factors: being overweight (BMI >27 kg/m), being a first-degree relative of a patient with type 2 diabetes, or having IGT or a history of gestational diabetes. After the baseline visit, which included an OGTT and FMD, all subjects started an exercise and dietary lifestyle intervention (Tubingen Lifestyle Intervention Program [TULIP]). The study was approved by the local ethics committee. All subjects gave their written informed consent. This programme includes the goals of the Diabetes Prevention Study [4]: a reduction of body weight by >5%, a reduction of dietary fat intake to 15 g per 4,187 kJ, and an increase in the amount of weekly exercise to >3 h/week. The TULIP study design includes a baseline and two follow-up visits, one after 9 months and the other after 24 months. In this analysis, data from 166 subjects who completed the baseline visit and the first follow-up visit are reported (mean follow-up time [±SD] 264±56 days, range 151– 465 days). At the time of data analysis, less than 50% of these subjects had completed the second follow-up visit, and so these data are not presented. At baseline and at follow-up, FMD of the brachial artery was measured with high-resolution ultrasound (13 MHz) [5]. Subjects with serum levels of CRP of >10 mg/l at baseline or at followup were excluded from the analysis. Genomic DNA was extracted from peripheral blood lymphocytes. Genetic analyses were performed by PCR with subsequent restriction analysis with BstXI, as described previously [6]. The statistical software package JMP (SAS Institute, Cary, NC, USA) was used. Of the 166 subjects, 131 had the Pro/Pro genotype and 35 had an X/Ala genotype. These two genotype groups were not different in terms of baseline FMD (7.2±0.3 vs 7.3±0.6%, respectively; p=0.69) or CRP (1.6±0.1 vs 1.6±0.3 mg/l, respectively; p=0.43). During lifestyle intervention, BMI was reduced (from 29.3±0.4 to Diabetologia (2007) 50:1345–1347 DOI 10.1007/s00125-007-0664-6