Objective:To evaluate the antiretroviral effect of a combination of zidovudine (ZDV) and didanosine (ddI) on plasma, peripheral blood mononuclear cells (PBMC) and lymph nodes after 24 weeks.Methods:Eight patients naive of antiretroviral therapy were followed by monthly blood samples and two surgical lymph-node biopsies taken at baseline and after 24 weeks. CD4+ T cells were counted monthly by flow cytometry. Plasma HIV-1 RNA was measured monthly by polymerase chain reaction (PCR). Infectious cellular viraemia was measured monthly by a culture technique. Proviral DNA titres in PBMC were measured by endpoint dilution PCR at baseline and 24 weeks. Infectious HIV-1 and proviral DNA titres were measured in the lymph-node mononuclear cells (LNMC). The total HIV-1 RNA content of lymph nodes was measured by PCR. In some cases, phenotypic resistance to ZDV was measured, and codon 215 and 74 mutations in PBMC and LNMC were analysed.Results:A mean increase in CD4 cell count of 122 × 106/l, a mean decrease in HIV-1 RNA of 1.47 log10in plasma and a mean decrease in HIV-1 DNA titre of 0.63 log10were found after 24 weeks of therapy. Nevertheless, there were no statistically significant changes in the mean infectious HIV-1 titre in PBMC and LNMC, in the HIV-1 DNA titre in LNMC or in the total lymph-node HIV-1 RNA burden at week 24. Phenotypic or genotypic markers of drug resistance were rarely found in PBMC at week 24, although they were detected in LNMC from some patients.Conclusion:A discrepancy in the therapeutic effect can be observed between lymphoid organs and blood after 24 weeks of therapy with ZDV and ddI. This difference could be explained by the insufficient antiretroviral potency of this combination facing the significant viral burden present in lymph nodes. Development of drug resistance in this compartment prior to blood can be demonstrated in some cases, although other mechanisms remain to be investigated in future studies to explain this difference.