When antiretroviral treatment suppresses HIV RNA levels to below 50 copies/ml, traces of viremia may still be detected with more sensitive assays. In the ARTEMIS trial, 689 antiretroviral treatment-naive patients were randomized to tenofovir/emtricitabine plus either darunavir/ritonavir (n = 343) or lopinavir/ritonavir (n = 346). HIV-1 RNA was evaluated using the Roche Amplicor® Ultrasensitive assay: plasma samples with HIV RNA < 50 copies/ml were classified as either "No HIV RNA detected" (< 5 HIV RNA copies/ml, optical density = background) or HIV RNA detected (5-50 copies/ml). The percentage of patients in each arm with HIV RNA < 5 copies/ml rose progressively from week 2 to week 192. For patients with baseline HIV RNA ≥ 100,000, the percentage with HIV RNA < 5 copies/ml at week 192 was 66% for darunavir/ritonavir and 63% for lopinavir/ritonavir. For patients with baseline HIV RNA < 100,000 copies/ml, the percentage with HIV RNA < 5 copies/ml at week 192 was 79% for darunavir/ritonavir versus 77% for lopinavir/ritonavir. Of the patients on darunavir/ritonavir with HIV RNA < 50 copies/ml, 63% had levels < 5 copies/ml at week 48, versus 80% at week 192. In summary, HIV-1 RNA suppression to < 5 copies/ml is dependent on baseline HIV RNA levels. The HIV RNA levels can remain under quantification limits but still detectable after 2-4 years of antiretroviral treatment.