Eligible studies included patients with cancer receiving chemotherapy for solid tumours or malignant lymphoma randomly allocated to granulocyte-colony stimulating factor in addition to chemotherapy or chemotherapy alone. Follow-up of at least two years was required. Studies in which patients received stem cell or bone marrow transplants or had an initial diagnosis of leukaemia were excluded. Primary outcomes were risk of acute myeloid leukaemia, myelodysplastic syndrome and overall survival, disease free survival or all cause mortality.
Patients in included studies had Hodgkin's disease, non-Hodgkin's lymphoma, breast, germ cell, endometrial, urothelial or non-small cell lung cancer; they were treated with a range of chemotherapy regimens. Granulocyte-colony stimulating factor used was filgrastim in most studies; lenograstim was used in the remainder.
The number of reviewers performing study selection was unstated.