To evaluate the effectiveness and safety of oral anticoagulants in improving survival of cancer patients. We conducted in January 2007 a comprehensive search for relevant randomized clinical trials (RCTs). We extracted data on methodological quality, participants, interventions and outcomes using a standardized form. Five RCTs fulfilled the inclusion criteria and all compared warfarin to either placebo or no intervention. Their overall methodological quality was acceptable. The effect of warfarin on mortality was not statistically significant at 6 months (RR = 0.96; 95% CI 0.80-1.16), at 1 year (RR = 0.95; 95% CI 0.86-1.05), at 2 years (RR = 0.97; 95% CI 0.87-1.08) or at 5 years (RR 0.91; 95% CI 0.83-1.01). In the subgroup of patients with small cell lung cancer (SCLC), warfarin reduced mortality at 6 months (RR = 0.69; 95% CI 0.50-0.96) but not at 1 year (RR = 0.88; 95% CI 0.77-1.01). This 6 months mortality benefit was statistically significant in the subgroup of extensive SCLC (RR = 0.65; 95% CI 0.45-0.93) but not in the subgroup of limited SCLC (RR = 0.68; 95% CI 0.36-1.28). Warfarin increased both major bleeding (RR = 4.24; 95% CI 1.85-9.68) and minor bleeding (RR = 3.34; 95% CI 1.66-6.74). The evidence suggests a survival benefit from warfarin in patients with extensive SCLC, but not in other patient groups. This survival benefit should be weighed against the increased risk for hemorrhage.