The objective of the present study was to investigate the effect of the nucleoside analogue treatment on serum viraemia, CD4+ cell count and disease progression in patients with and without syncytium-inducing (SI) HIV-1 variants. To achieve this in a case-control study, 11 nucleoside-naive patients harbouring SI variants who started treatment with zidovudine or zidovudine plus didanosine were matched with 11 control patients who never formed SI variants during a follow-up of 48 weeks. The matching criteria were age, CD4+ cell count and CDC clinical category at the start of the study and exposure to the same antiretroviral treatment. During the follow-up there were no significant differences in the changes of serum HIV-1 RNA viral load and CD4+ cell counts between the two groups. In contrast, AIDS or new AIDS-defining events were observed in five SI cases but in none of the non-SI controls (P = 0.002). The emergence of a zidovudine-resistant mutation at codon 215 was observed in all the patients harbouring SI strains and in six of the subjects with non-SI variants (P = 0.03). The results of the present study show that in patients carrying SI virus, measurements of CD4+ count or RNA viral burden are neither related to the virulence of the virus strains nor able to predict the clinical course of the disease, at least under anti-retroviral drug conditions. Thus, determination of SI phenotype should be considered in the evaluation and monitoring of HIV-1 therapies.