Chronic Comorbidities and Chemotherapy-Induced Febrile Neutropenia in Patients with Non-Hodgkin Lymphoma

Chemotherapy induced febrile neutropenia (FN) is a clinically important adverse event as it can impact patient survival by delaying or reducing chemotherapy dose administered. Clinical guidelines recommend granulocyte-colony stimulating factors (G-CSF) be used in cancer patients when febrile neutropenia (FN) risk is >20%. Although the myelotoxicity of the chemotherapy regimen is a key determinant of FN risk, it is now recognized that patient characteristics may increase this risk further. We conducted a retrospective cohort study to evaluate the association of chronic comorbidities to FN in patients with non-Hodgkin lymphoma (NHL).

Incident NHL cases diagnosed between 2000–2009 who received chemotherapy within 12 months of cancer diagnosis were identified from Kaiser Permanente Southern California, a large managed care organization. Those who had prophylactic G-CSFs, dose-dense chemotherapy or bone marrow transplant were excluded. Comorbidities of interest included cardiovascular, liver, renal, metabolic and autoimmune disorders, anemia, previous cancer and HIV infection. History of comorbidity of interest were assessed in the 12 months prior to NHL diagnosis, and identified by ICD-9 codes or disease registries. FN was assessed only in the first chemotherapy cycle, and identified by absolute neutrophil count (ANC), ICD-9 codes for neutropenia and fever, or hospitalization with bacterial/fungal infection. Patients were followed from their first chemotherapy treatment to the start of the second cycle, death or day 28, whichever came first. Logistic regression was used to estimate the propensity score for each comorbidity; these propensity scores included patient characteristics and other comorbidities as covariates. Each comorbidity and propensity score were included in Cox models to determine associations between comorbidities and FN. We also evaluated models that additionally adjusted for cancer stage, baseline ANC, chemotherapy regimen, dose reduction, and radiation treatment prior to chemotherapy.

A total of 2,480 NHL patients who received chemotherapy were included. The mean age was 63.3 years. Fifty-five percent of the cases were male, and the majority of the cases were of white race (65.6%). Sixty percent of the cases received CHOP or R-CHOP. There were 236 (9.5%) patients that had FN in the first chemotherapy cycle. Anemia [adjusted hazard ratio adjusted (HR) =1.6, 95% CI, 1.2–2.2], HIV infection [HR=3.0, 95% CI 1.6–5.3], AIDS [HR=2.4, 95% CI 1.2–4.6], and rheumatoid diseases [HR=2.1, 95% CI 1.2–3.6] were all associated with a statistically significantly increased FN risk (Table 1). In addition, peptic ulcer disease, renal disease and connective tissue disease were also associated with risk of FN, although these associations were not statistically significant. HR estimates did not change materially after also adjusting for cancer stage, ANC, chemotherapy regimen, dose reduction and radiation therapy prior to chemotherapy in the Cox model.

Our findings suggest that several chronic comorbidities may be associated with increased FN risk in the first chemotherapy cycle among NHL patients not already receiving prophylactic G-CSFs. This information, if confirmed by others, may aid clinical decision making with respect to use of prophylactic G-CSF use during chemotherapy treatment.

Chao:Amgen,Inc: Research Funding. Page:Amgen, Inc: Employment, Shareholder Other. Rodriguez:Amgen, Inc: Research Funding. Yang:Amgen, Inc: Research Funding. Huynh:Amgen, Inc: Research Funding. Chia:Amgen, Inc: Employment, Shareholder Other.