To compare the safety and efficacy of continuing zidovudine therapy with that of zalcitabine alone or zalcitabine and zidovudine used together. A randomized, double-blind, controlled trial. AIDS Clinical Trials units and National Hemophilia Foundation sites. 1001 patients with symptomatic human immunodeficiency (HIV) disease and 300 or fewer CD4 cells/mm3 or asymptomatic HIV disease and 200 or fewer CD4 cells/mm3 who had tolerated zidovudine therapy for 6 months or more. Patients were randomly assigned to receive zidovudine, 600 mg/d; zalcitabine, 2.25 mg/d; or zidovudine, 600 mg/d, and zalcitabine, 2.25 mg/d. The primary end point was time to disease progression or death. The median follow-up time was 17.7 months. The estimated 12-month event-free rates were 70%, 67%, and 73%, respectively, for the zidovudine, zalcitabine, and combination groups (P =0.26). A trend analysis showed significantly lower progression rates for combination therapy compared with zidovudine therapy as the pretreatment CD4 cell count increased (P = 0.027). For patients with 150 or more CD4 cells/mm3, those receiving combination therapy were less likely to have disease progression or to die than were those receiving zidovudine (relative risk, 0.51; 95% CI, 0.28 to 0.93; P = 0.029). We observed no difference between the zalcitabine and zidovudine groups (relative risk, 0.74; CI, 0.40 to 1.36; P = 0.33). For patients with 50 to 150 CD4 cells/mm3 or fewer than 50 CD4 cells/mm3, we found no differences among the treatment groups (P = 0.69 and P = 0.57, respectively). Severe toxic effects occurred less frequently among patients with 150 or more CD4 cells/mm3. We found no overall benefits of zalcitabine used alone or with zidovudine. However, a trend analysis suggested a better outcome for combination therapy compared with zidovudine as the pretreatment CD4 cell count increased. *Other members of the AIDS Clinical Trials Group who participated in this study are listed in the Appendix.