Abstract
PURPOSE The overall results of chemotherapy in human immunodeficiency virus (HIV)-associated non-Hodgkin's lymphoma (NHL) have been poor. To define a subgroup of patients who may have a better outcome, an analysis of prognostic factors was performed of patients treated in AIDS Clinical Trials Group (ACTG) protocol 142, a phase III randomized trial of low-dose versus standard-dose methotrexate, bleomycin, doxorubicin, cyclophosphamide, vincristine, and dexamethasone (m-BACOD) plus granulocyte-macrophage colony-stimulating factor (GM-CSF) for the treatment of patients with newly diagnosed HIV-associated NHL. MATERIALS AND METHODS The following baseline variables were included as potential predictors of survival among 192 patients who received treatment: age; intravenous drug use (IVDU); specific type of sexual contact as risk factors (homosexual, bisexual, or heterosexual contact); prior AIDS diagnosis; CD4 cell count; serum lactic acid dehydrogenase (LDH); histology; Karnofsky performance status (KPS); stage; B symptoms; race (white/nonwhite); nodal involvement; extranodal involvement; number of extranodal sites; specific sites: bone marrow, liver, kidney, lung, or gastrointestinal tract; and treatment arm (standard-dose m-BACOD/low-dose m-BACOD). RESULTS Age greater than 35 years, IVDU, stages III/IV, and CD4 cell counts less than 100/microL were adverse prognostic factors in multivariate analyses using the Cox proportional hazards model. The median overall survival for patients with none or one of the adverse factors was 46 weeks, with two was 44 weeks, and with three or four was 18 weeks. At 144 weeks, 29.5% of patients with none or one, 16.9% with two, and 0% with three or four factors were alive (P < .001). CONCLUSION Long-term survival can be achieved in approximately one third of patients with HIV-associated NHL with favorable characteristics.