Risk factors assessment in fabrile neutropenia.

It is now established that febrile neutropenic cancer patients constitute a heterogeneous population with a variable risk for serious medical complication development. Optimal patient management should take that risk into account by replacing, for instance, the classical, in-hospital administered, broad-spectrum intravenous antibiotics by newer therapeutic approaches including oral and/or outpatient therapeutic strategies for the 'low-risk' patients. The development of such approaches which have been shown safe and feasible, implies the existence of universally accepted, validated and reliable clinical prediction rules for the identification of these low-risk patients. Some prognostic factors predicting the response to the empiric treatment, the development of a bacteremia, and the final outcome of a febrile neutropenic episode have been established (such as duration and profoundness of neutropenia, acute leukemia, administration of chemotherapy for treatment of relapse, high temperature, shock and/or chills, inpatient status at fever onset) and some models combining them have already been proposed, firstly by Talcott and coworkers and more recently by the Multinational Association for Supportive Care in Cancer (MASCC) study section on infections. The sensitivity of these rules as a selection tool for identifying patients at low-risk of complication, however, needs to be improved and we have to assess their clinical usefulness, safety and/or reproducibility better in order to allow a more adequate choice between the therapeutic strategies, to continue to improve patients quality of life and to optimize the cost-effectiveness of the treatments.