OBJECTIVE To assess safety, pharmacokinetics, and in-vivo virologic activity of five different combination regimens of zidovudine and didanosine compared with zidovudine alone in patients with human immunodeficiency virus type 1 (HIV-1) infection. METHODS Open-label, partially randomized, dose-ranging study. METHODS University-affiliated, medical center clinics. METHODS A total of 69 patients with HIV-1 infection, CD4+ cell counts fewer than 400 cells/mm3, and fewer than 121 days of previous zidovudine treatment. METHODS Fifty-five patients received combination therapy with zidovudine and didanosine, and 14 received zidovudine therapy alone (600 mg/d). Daily dosages in milligrams of zidovudine and didanosine, respectively, in the five combination groups were 150 and 90 mg, 300 and 334 mg, 600 and 334 mg, 300 and 500 mg, and 600 and 500 mg. METHODS CD4+ cell counts, HIV-1 RNA titers in plasma, and toxic effects. RESULTS The combination regimens were associated with higher and more sustained increases in CD4+ cells than zidovudine alone, even after adjustment for initial CD4+ counts and previous zidovudine therapy (P < 0.001). The median increase in CD4+ cell counts was 166 cells/mm3 with combination therapy and 77 cells/mm3 with zidovudine alone (P = 0.001) and did not differ statistically among the five combination regimens. Human immunodeficiency virus type 1 RNA titers in plasma decreased in 15 (83%) of 18 combination-therapy recipients compared with 2 of 7 zidovudine-alone recipients (P = 0.017). No pharmacokinetic interactions were seen between zidovudine and didanosine. Toxicity rates were low among all treatment groups. A greater decrease in hemoglobin levels was seen with the regimen using zidovudine alone (-8 g/L) compared with combination regimens using the same zidovudine dose (-1.5 g/L, P = 0.03). CONCLUSIONS Combination therapy with zidovudine and didanosine produced larger and more sustained increases in CD4+ cell counts, more frequent decreases in plasma HIV-1 RNA titers, and more stable hematologic status than zidovudine therapy alone. The effects of this combination on the progression of HIV disease merit further study, to provide information about clinical outcome, because this was a relatively small study based on surrogate markers of HIV-1 infection.