The incidence of systemic non-Hodgkin's lymphoma (NHL) is higher in the human immunodeficiency virus (HIV)-infected population than in the uninfected population. Standard treatment for this cancer involves the administration of systemic chemotherapy. Our objective was to determine the relative risk of opportunistic infection and the relative change in immunologic function in a cohort of patients who had HIV-associated non-Hodgkin's lymphoma(NHL) and who were treated with combination chemotherapy compared with a matched cohort of control subjects who had advanced HIV infection but no signs of NHL. We performed a case-control study in which the clinical course of each patient with HIV-associated NHL (N = 43; case subjects) treated with infusional cyclophosphamide, doxorubicin, and etoposide (CDE) was compared with that of two patients with HIV infection but without lymphoma that were matched for CD4 count and prior opportunistic infection (N = 86; control subjects). Univariate and multivariate analyses were performed to determine whether any of a number of confounding factors (e.g. age, sex, CD4 count, prior opportunistic infection, and prior antiretroviral therapy) could have influenced the risk of developing a first infectious event(defined as opportunistic infection or non-lymphoma death). In the multivariate analysis, being a case subject (RR 2.1, 95% CI = 1.2, 3.6;P <.01), having a low CD4 count (RR 2.1, 95% CI = 1.2, 3.9;P <.05), and being female (RR 3.0, 95% CI = 1.8, 5.6; P <.001) were the only characteristics associated with an increased risk of a first event. In the univariate analysis, other factors associated with an increased risk included prior Pneumocystis carinii pneumonia, any prior opportunistic infection, and prior antiretroviral therapy. When the mean CD4 lymphocyte count at one year was compared with baseline, there was a significantly greater decrease in the CD4 count among case subjects than among control subjects (mean decrease 99/uL ∟+ SD 138/uL| versus 29/uL∟+ SD 100/uL; P = .03). Treatment of HIV-associated NHL patients with a nonsteroid containing chemotherapy regimen was associated with a significant and sustained reduction in the CD4 lymphocyte count and a two-fold increase in the risk of developing opportunistic infection. Oncologists and other physicians who treat patients with HIV-associated NHL should be familiar with the prophylaxis, recognition, and management of opportunistic infection. In addition, there is a need to identify effective strategies for the amelioration of chemotherapy-induced immunosuppression in this population.