Background: Infections in cancer patients receiving myeloablative chemotherapy or stem cell transplantation are potentially life threatening. Both antibiotics and granulocyte colony stimulating factors have shown to be effective in preventing infections. Most current guidelines, however, recommend the use of granulocyte colony stimulating factors (CSFs) as infection prophylaxis when the risk of febrile neutropenia is above 20%. Objectives: To assess factors that may influence the recommendations of guidelines looking at infection prophylaxis in cancer patients. Methods: In the context of a systematic review, randomized clinical trials relating to infection prophylaxis in cancer patients were searched using standard Cochrane methodology. Current and older versions of guidelines relating to infection prophylaxis of medical associations were retrieved. The assessment included the questions of the guidelines, the inclusion criteria and the included trials compared to the results of our literature search, the date of publication with respect to the publication of key trials, and changes in recommendations over time. Results: Over 130 trials reporting infection or fever incidence or on trial mortality were retrieved. Only two trials were a direct comparison of antibiotics and CSFs. All five of the retrieved guidelines relate to only one of the options without highlighting research gaps. Four guidelines report solely on the use of colony stimulating factors. Among the CSF guidelines, missing trials or restrictive inclusion criteria were common, and most missing trials had unfavorable results. The adjustment of the baseline risk for febrile neutropenia from 40% to 20% for use of CSFs was based on few industry funded trials in breast and lung cancer, but the results generalized to all solid cancers and lymphomas. One CSF guideline was funded by a pharmaceutical company. Conclusions: Guidelines looking at infection prophylaxis are titled and oriented towards questions where ample randomized controlled trials and their meta-analyses exist and not towards the clinical reality. Clinical studies and guidelines should focus more on clinical questions and less on single drugs.