We systematically reviewed the literature to compare the relative efficacy and safety of extended versus limited duration heparin for perioperative thromboprophylaxis in patients with cancer. We followed the Cochrane Collaboration systematic review methodology and searched MEDLINE, EMBASE, ISI the Web of Science, and CENTRAL. The outcomes of interest included mortality, symptomatic deep venous thrombosis (DVT), pulmonary embolism, and bleeding. We evaluated the quality of evidence by outcome using the GRADE approach. Of 3,986 identified citations, we included three randomized clinical trials using low-molecular-weight heparin (LMWH). The quality of evidence for mortality, DVT, and major bleeding was low. There was no significant difference between extended (4 weeks) and limited duration thromboprophylaxis in terms of death at three months (relative risk [RR] = 0.49; 95% confidence interval [CI] 0.12-1.94), or major bleeding at four weeks (RR = 2.94; 95% CI 0.12-71.85). An extended regimen was associated with a significantly lower risk of asymptomatic DVT (RR = 0.21; 95% CI 0.05-0.94). No data was available for symptomatic DVT. In conclusion, there is limited and low-quality evidence that extended duration LMWH for perioperative thromboprophylaxis reduces DVT in patients with cancer undergoing major abdominal or pelvic surgery. More and better quality evidence is needed to justify extended regimens.