The role played by a hypercoagulable state, either inherited or acquired, in the pathogenesis of upper-extremity deep vein thrombosis (UEDVT) remains a question of debate. We performed a case-control study including 79 patients with a first objectively confirmed episode of UEDVT, 31 secondary and 48 primary, and 165 healthy controls. Nine patients (11.4%) with UEDVT were carriers of the prothrombin G20210A mutation vs. six (3.7%) in controls; P = 0.025, OR: 3.39 (95% CI 1.16 to 9.88). No statistical difference was observed between cases and controls for the factor V Leiden mutation, AT, protein C or protein S deficiency and anticardiolipin antibodies (ACAs). Thirteen (35.1%) UEDVT patients were oral contraceptive (OC) users vs. 12 (16%) controls; P = 0.020, OR: 2.89 (95% CI 1.16-7.21). When secondary UEDVT patients were compared with controls, no differences were observed in any of the risk factors analysed. On the other hand, when primary UEDVT was considered, six (12.5%) patients were carriers of the prothrombin G20210A mutation vs. six (3.7%) controls; P = 0.031, OR: 3.76 (95% CI 1.15-12.26). Regarding ACAs, a borderline statistical significance was observed when primary UEDVT was compared with controls, P = 0.048; OR: 4.88 (95% CI 1.05-22.61). In primary UEDVT, 52% of the fertile women were OC users vs. 16% of controls; P = 0.001, OR: 5.78 (95% CI 2.13-15.67). When the interaction of both factors, i.e. prothrombin G20210A mutation and OC intake, were considered, the risk increased markedly, indicating a synergistic effect as observed with other thrombotic locations. In patients with primary UEDVT screening for antithrombin, protein C and protein S deficiency and APC resistance would not be justified, although it might be reasonable to determine the carrier status of the prothrombin G20210A mutation only in OC users.