Validation of a predictive model that identifies patients at high risk of developing febrile neutropaenia following chemotherapy for breast cancer

We have previously developed a predictive model that identifies patients at increased risk of febrile neutropaenia (FN) following chemotherapy, based on pretreatment haematological indices. This study was designed to validate our earlier findings in a separate cohort of patients undergoing more myelosuppressive chemotherapy supported by growth factors. We conducted a retrospective analysis of 263 patients who had been treated with adjuvant docetaxel, adriamycin and cyclophosphamide (TAC) chemotherapy for breast cancer. All patients received prophylactic pegfilgrastim and the majority also received prophylactic antibiotics. Thirty-one patients (12%) developed FN. Using our previous model, patients in the highest risk group (pretreatment absolute neutrophil count < 3.1 x 10(9)/l and absolute lymphocyte count < 1.5 x 10(9)/l) comprised 8% of the total population and had a 33% risk of developing FN. Compared with the rest of the cohort, this group had a 3.4-fold increased risk of developing FN (P = 0.001) and a 5.2-fold increased risk of cycle 1 FN (P < 0.001). A simple model based on pretreatment differential white blood cell count can be applied to pegfilgrastim-supported patients to identify those who are at higher risk of FN.