This study evaluated the impact of body mass index (BMI) on venous thromboembolism (VTE) site and assessed a possible interaction between BMI and prothrombotic risk factors in patients included in the EDITH (Etude des Déterminants et Interactions de le THrombose veineuse) study. A cross-sectional study was used to compare the site of unprovoked VTE according to BMI categories in 1077 patients and a matched case-control study (732 pairs) assessed the joint effect of BMI and prothrombotic mutations on VTE risk. The cross sectional analysis showed that the proportion of patients with pulmonary embolism was higher in overweight (63%) and obese (63·5%) patients than among patients with a BMI<25kg/m(2) (55%), P=0·02 and P=0·05 respectively. No interaction was found between F5 G1691A (factor V Leiden) and BMI for VTE risk (P=0·90). There was a significant interaction between F2 G20210A and BMI (P=0·02). The risk of VTE associated with BMI was 1·7 [95% confidence interval (CI): 0·8-3·7], 4·36 (95%CI: 1·49-12·78) and 12·03 (95%CI: 1·53-94·29) in patients with BMI<25kg/m(2) , 25≤BMI<30 and ≥30kg/m(2) respectively after adjustment for age and oestrogen use. This study showed that BMI may play a role in determining the site of VTE and may interact with F2 G20210A but not with F5 G1691A for the risk of VTE.