Risk factors for treatment-related death in elderly patients with aggressive non-Hodgkin's lymphoma: Results of a multivariate analysis

Purpose: It has been suggested that age is associated with chemotherapy- related death in patients with non-Hodgkin's lymphoma (NHL) treated with doxorubicin-containing chemotherapy. The purpose of this study was to evaluate the relative influence of increasing age and other clinical parameters on the occurrence of treatment-related death in elderly patients with intermediate- or high-grade NHL treated with cyclophosphamide, doxorubicin, vincristine, and prednisone (CHOP) chemotherapy. Methods: A retrospective study of patients 60 years of age or older with intermediate- or high-grade NHL treated with CHOP chemotherapy in a single cancer center. The following variables were recorded: age (60 to 69, 70 to 79, and 80 to 94 years), histology (Working Formulation [WF] D, E, F, G, and H), Ann Arbor stage, B symptoms, extranodal involvement, bulky disease (> 7 cm), performance status (Eastern Cooperative Oncology Group [ECOG] scale), International Prognostic Index (IPI score), serum lactate dehydrogenase (LDH) level and doxorubicin relative dose-intensity (RDI). The relationship between these features and treatment-related death was assessed in univariate and multivariate logistic regression analysis. Results: From 1982 to 1991, 267 consecutive patients were treated. Median age was 70 years (range, 60 to 94 years). Them were 35 toxic deaths. Sixty-three percent of the deaths occurred after the first cycle. Infection accounted for 82% of the toxic deaths. In the univariate analysis, the features associated with an increased risk of toxic death were ECOG performance status 2 to 4 (relative risk [RR], 7.82), B syptoms (RR, 3.38), diffuse large-cell histology (RR, 3.06), bulky disease (RR, 2.58), serum levels of LDH (RR, 2.53), and IPI score (RR, 2.46). The age groups did not show significance. In the regression model, performance status 2 to 4 was the only independent predictor of treatmentrelated death (RR, 3.52; 95% confidence interval [Cl], 2.98 to 4.06). Conclusion: Our results show that in elderly patients with NHL treated with doxorubicin-based chemotherapy the risk for treatment-related death is associated with poor performance status rather than with increasing chronologic age.