A randomized pilot study comparing alternating and simultaneous regimens of zidovudine and didanosine (ddl) was conducted in 41 patients with AIDS or symptomatic human immunodeficiency virus (HIV) infection. Patients on each regimen received the same overall amounts of zidovudine and didanosine over time. CD4 cell counts in patients on the simultaneous regimen reached a maximum (mean ± SE) of 108 ± 16/mm3 above baseline (two-tailed P.0001) and were significantly higher than in patients on the alternating regimen at all time points during weeks 6-45. At 54 weeks, the CD4 cell counts in the patients on the simultaneous regimen were still 40 ± 19/mm3 above baseline. Patients on the simultaneous regimen also had significantly greater weight gain. While toxicities were generally mild and comparable between the regimens, 1 patient on the simultaneous regimen died of pancreatitis and lactic acidosis. Thus, simultaneous therapy provided more sustained elevations in CD4 cells than alternating therapy over 1 year and may be worth exploring in larger controlled trials.