Therapy with two nucleoside reverse transcriptase inhibitors (NRTI), the backbone of triple combinations, is still widely used in early stages of HIV-1 disease. However, factors influencing virologic response need to be further analyzed, to test the hypothesis that the reduction of plasma RNA viremia with NRTI may be greater in patients with higher baseline viral load (BVL) and to analyze the predictive factors of viral load drop below detection (200 HIV RNA copies/ml of plasma). Selected for the study were 169 HIV-1-infected antiretroviral therapy-naive patients with CD4+ T-lymphocyte counts ranging from 6 to 1040/microl coming from three randomized studies comparing the efficacy of monotherapy (zidovudine [ZDV], 250 mg every 12 hours; N=40) versus two-drug therapy consisting of ZDV (250 mg every 12 hours) with dideoxycytidine (ddC, 0.75 mg every 8 hours) or didanosine (ddI, 200 mg every 12 hours; N=129). Viral load was measured at 1, 3, and 6 months by polymerase chain reaction (PCR). A linear regression model was used to analyze the relation between BVL and the peak reduction of plasma RNA viremia. The variables included in a logistic regression model to determine the likelihood of VLs dropping below detection levels were age, gender, risk group for HIV-1 infection, baseline CD4+ lymphocyte count, BVL, clinical status (AIDS versus non-AIDS), HIV-1 phenotype (syncytium-inducing [SI] versus non-syncytium-inducing [NSI]) and type of treatment (monotherapy versus double therapy). The peak reduction of VL was related to baseline level following a linear model in both monotherapy and double-therapy regimens. In the subgroup of patients treated with two NRTIs, the regression line that fitted best with the data was log10 (peak reduction)=1.8-0.36 log10 (BVL) (F=23.5; p < .0001). This indicates that for every increase of 1 log10 of BVL, the peak reduction would be of 0.64 log10 greater. Forty-nine (29%) of the 169 patients dropped to <200 copies/ml. The likelihood of dropping below detection level was significantly greater in those receiving double therapy versus monotherapy (odds ratio [OR]=16.1; 95% confidence interval [CI], 2-128), in those with baseline CD4+ lymphocyte count >350/microl (OR=2.28; 95% CI, 1.1-4.9) and in those with BVL <10,000 HIV-1 RNA copies/ml (OR= 2.25; 95% CI, 1.1-6.1). None of the 13 patients with an SI phenotype at baseline dropped below detection levels. The reduction of VL in response to two NRTIs was greater in those patients with a higher level of BVL. In conclusion, peak reduction below detection in response to NRTI can be predicted and is associated with double therapy, with a baseline CD4+ cell count >350/microl and with a BVL <10,000 RNA copies/ml.