Risk Factors for Early Death After Rituximab-Based Immunochemotherapy in Older Patients With Diffuse Large B-Cell Lymphoma

Abstract

Background: Older patients with diffuse large B-cell lymphoma (DLBCL) are at risk of severe chemotherapy-related morbidity and mortality. Our objective was to quantify the risk and identify factors associated with death during the first cycle of immunochemotherapy in this population.

Patients and methods: Using Medicare claims linked to the population-based SEER registry (SEER-Medicare), we studied patients with DLBCL aged 65 years and older who received immunochemotherapy containing rituximab, cyclophosphamide, and vincristine, in combination with doxorubicin, mitoxantrone, or etoposide in 2003-2012. Risk factors for death and hospitalization within the first 30 days of treatment were studied in multivariable logistic regression models.

Results: We identified 5,530 patients with a median age of 76 years, of whom 94% received doxorubicin-containing immunochemotherapy. Granulocyte colony-stimulating factor (G-CSF) was administered to 66% of patients during the first treatment cycle. Cumulative incidence of death at day 30 was 2.2%. The risk was significantly higher in patients aged 75 years and older and those who had B symptoms, chronic kidney disease, poor functional status, use of walking aids or wheelchairs, and prior hospitalization or upper endoscopy. The group with 0 to 1 risk factors (56% of patients) had a very low (0.6%) risk of early death, whereas the group with 4 or more risk factors (6% of patients) had a risk of 8.3%. Receipt of G-CSF was associated with a lower probability of early death in the high-risk group. The incidence of hospitalization within the first 30 days was 23.5%, peaking at day 8 of the cycle.

Conclusions: Among older patients with DLBCL who receive contemporary immunochemotherapy, 1 in 45 die during the first month of treatment, and 1 in 4 are hospitalized. Factors identifiable from administrative/electronic records can stratify this risk and could be incorporated into decision support tools. Prophylactic G-CSF is not administered to more than one-third of patients, indicating an opportunity for improved preventive interventions.