OBJECTIVE To evaluate the association between resistance of human immunodeficiency virus type 1 (HIV-1) to zidovudine and clinical progression. METHODS Retrospective analysis of specimens from patients in the AIDS Clinical Trials Group (ACTG) protocol 116B/117, a randomized comparison of didanosine with continued zidovudine therapy in patients with advanced HIV-1 disease who had received 16 weeks or more of previous zidovudine therapy. METHODS Participating ACTG virology laboratories. METHODS 187 patients with baseline HIV-1 isolates. METHODS Zidovudine susceptibility testing and assays for syncytium-inducing phenotype were done on baseline HIV-1 isolates. Relative hazards for clinical progression or death associated with baseline clinical, virologic, and immunologic factors were determined from Cox proportional hazards regression models. RESULTS Compared with other patients, 15% (26 of 170) with isolates showing high-level zidovudine resistance (50% inhibitory zidovudine concentration > or = 1.0 microM) had 1.74 times the risk for progressing to a new AIDS-defining event or death (95% CI, 1.00 to 3.03) and 2.78 times the risk for death (CI, 1.21 to 6.39) in analyses that controlled for baseline CD4+ T-lymphocyte count, syncytium-inducing HIV-1 phenotype, disease stage, and randomized treatment assignment. The clinical benefit of didanosine was not limited to patients with highly zidovudine-resistant baseline HIV-1 isolates. CONCLUSIONS High-level resistance of HIV-1 to zidovudine predicted more rapid clinical progression and death when adjusted for other factors. However, patients with advanced HIV-1 disease may benefit from a change in monotherapy from zidovudine to didanosine whether high-level HIV-1 resistance to zidovudine is present or absent, and laboratory assessment of zidovudine resistance is not necessary for deciding when to switch monotherapy from zidovudine to didanosine.