OBJECTIVE To examine tolerance and efficacy of a zidovudine plus lamivudine combination in HIV-infected children without previous exposure to antiretroviral drugs. METHODS Thirteen vertically infected children (aged 4 months to 10 years) were treated with zidovudine (approximately 100 mg/m2 three times daily) and lamivudine (4 mg/kg twice daily). CD4 T-cell count, plasma HIV RNA concentration, complete blood count and blood chemistry profile were monitored before treatment and at months 1, 3 and 6. RESULTS In general, treatment was well tolerated. One child developed slight neutropenia in the presence of antineutrophil antibodies. CD4 cell count increased from 851+/-621 x 10(6)/l at baseline to 1073+/-945 x 10(6)/l at month 3 (P < 0.05) and to 1133+/-728 x 10(6)/l at month 6 (P = 0.01). CD4+ cell count increased in 10 patients after 3 months and in 11 patients treated for 6 months. One child showed a continuous decrease of CD4 cells despite treatment. Before treatment the plasma HIV RNA concentration was elevated in nine children (> 4.0 log10 copies/ml) and decreased in all of them: by month 1, the mean reduction was -1.16 log10 copies/ml; by month 3, -1.38 log10 copies/ml; and by month 6, -1.53 log10 copies/ml compared with baseline. However, one child showed steadily increasing viral load from 2.7 log10 copies/ml to a maximum of 4.52 log10 copies/ml, surprisingly in association with increasing numbers of CD4 cells. This child was switched to a new combination regimen after 6 months of treatment. Plasma HIV RNA levels below limit of detection were reached in six patients: after 1 month of treatment in one patient, after 3 months in five patients, and after 6 months in six patients. There was a mean reduction of viral load from 4.56 log10+/-4.63 log10 copies/ml (n = 13) to 3.8 log10+/-3.9 log10 copies/ml (P < 0.05; n = 9) after 1 month, to 3.67 log10+/-3.88 log10 copies/ml (P < 0.01; n = 13) after 3 months, and to 3.64 log10+/-3.95 log10 copies/ml after 6 months of treatment (P < 0.001; n = 13). CONCLUSIONS This pilot study demonstrates the feasibility of zidovudine-lamivudine combination in children not previously exposed to antiretroviral drugs. This promising combination should therefore be evaluated in larger trials.