To explore a novel method to use nucleosides in heavily treated patients, 4-week cycles of high-dosage (280 mg/d) stavudine were given to 11 asymptomatic patients who had previously received an average of 6 years of nucleoside reverse transcriptase inhibitors (NRTIs). The stavudine dose was targeted to produce a calculated steady-state concentration of 1.5 microM (336 ng/mL). Participants received stavudine for the first 4 weeks, after which it was discontinued for 4 weeks. Additional 4-week drug cycles were given if plasma HIV-1 RNA levels increased to at least 75% of baseline values. The 11 subjects received 38 cycles (average, 3.5 cycles per subject) during the 44-week study. Stavudine was well tolerated. Considering all data, there was a median 0.65 log(10) reduction in viral load as well as a median increase in CD4 cell count of 110/mm3 at the end of the cycles. However, plasma viremia increased and CD4 cell counts decreased between cycles. Viral load and CD4 cell responses were similar for up to 4 successive cycles. This suggested that increasing viral resistance was not a problem, as reflected by the acquisition of only one new nucleoside reverse transcriptase mutation among the participants. Significant relationships between stavudine exposure and changes in plasma HIV RNA levels were observed. A similar approach might be considered using a more potent regimen for patients in whom resistance to nucleosides is a major reason for therapeutic failure.