AIM Since the introduction of chemotherapy, survival in localised high-grade osteosarcoma has improved considerably. However, there is still no worldwide consensus on a standard chemotherapy approach. In this systematic review evidence for effectiveness of each single drug and the role of response guided salvage treatment of adjuvant chemotherapy are addressed, whereas in a meta-analysis the number of drugs in current protocols is considered. METHODS A systematic literature search for clinical studies in localised high-grade osteosarcoma was undertaken, including both randomised and non-randomised trials. Historical clinical studies from the pre-chemotherapy era were included for comparison purposes. RESULTS Nine historical studies showed a long-term survival of 16% after only local treatment. Fifty single agent phase II studies showed high response rates for adriamycin (A, 43%), ifosfamide (Ifo, 33%), methotrexate (M, 32%), cisplatin (P, 26%) but only 4% for etposide (E). In 19 neo-adjuvant studies the mean 5-year event free survival (EFS) was 48% for 2-drug regimens and 58% for ⩾3 drug regimens, with a 5-year overall survival (OAS) of 62% and 70%, respectively. Meta-analysis showed that ⩾3 drug regimens including methotrexate plus adriamycin plus cisplatin (plus ifosfamide) (MAP(Ifo)) had significant better outcome (EFS: HR=0.701 (95% confidence interval [95% CI]: 0.615-0.799); OAS: HR=0.792 (95% CI: 0.677-0.926) than 2-drug regimens, but there was no significant difference between MAP and MAPIfo (or plus etoposide). Salvage of poor responders by changing drugs, or intensifying treatment postoperatively has not proven to be useful in this analysis. CONCLUSION Meta-analysis in patients with localised high-grade osteosarcoma shows that 3-drug regimens, for example MAP are the most efficacious drug regimens.