Identifying low-risk febrile neutropenic cancer patients in the West of Scotland

Abstract

18592 Background: Febrile neutropenia (FN) is a potentially life threatening complication of chemotherapy. In-patient treatment using intravenous antibiotics reduces FN related mortality. However, most patients with FN are at low risk of complication. Identifying them could result in new strategies such as out-patient, oral antibiotic, based treatment resulting in improved quality of life and cost savings. The Multinational Association for Supportive Care in Cancer (MASCC) and the Talcott model were developed to identify such patients. Two randomised controlled trials have shown that oral antibiotics in low risk patients are safe and effective. However are such models appropriate in populations with a high incidence of co-morbidity such as the West of Scotland? This study reviews FN admissions to our cancer centre to determine the proportion of patients that would fall into a low risk group according to the MASCC and/or Talcott models and to identify other factors that might predict for low risk. Methods: Review of FN admissions between June–December 2002. Data included: patient demographics, MASCC score, Talcott group, co-morbidity, haematological and biochemical values, prior use of antibiotics and growth factors (GFs) and whether the fever resolved without serious medical complication (FRWMC). Results: 77 episodes of FN involving 68 patients. Mean age = 51 (range 16–79). 94% involved patients with solid malignancies. Commonest tumour type was breast (29%). Patients were classified as MASCC and Talcott low risk in 52% and 31% of episodes respectively. There was a significant association between low risk MASCC score and low risk Talcott score (χ2 28.665, d.f.3, p < 0.001). Low risk MASCC was associated with FRWMC (χ2 4.193, d.f.1, p < 0.05). Multiple logistic regression of risk factors showed that high bilirubin and low albumin were associated with a worse outcome. FN mortality rate was 7.8%. Conclusions: The use of clinical risk models to identify low risk patients can predict for an uncomplicated recovery in our patients. Bilirubin and albumin values at presentation added predicted value for low risk over and above the MASCC model. Future trials may validate this observation. No significant financial relationships to disclose.