There are no clinical trials in which the main objective is to compare the efficacy of efavirenz versus ritonavir-boosted protease inhibitor (PI/r)-based initial antiretroviral therapy (ART) in patients with high plasma HIV-1 RNA levels. This study aims to compare these regimens in this patient population in the setting of routine clinical practice. This was a multicentre, observational cohort study, including 596 consecutive treatment-naive patients with plasma HIV-1 RNA>100,000 copies/ml initiating efavirenz or PI/r-based ART between 2000 and 2010. The primary effectiveness end point was the percentage of patients with HIV-1 RNA<50 copies/ml at week 48 by intent-to-treat analysis. Among a total of 596 patients, 57% initiated efavirenz and 43% PI/r-regimens (73% lopinavir and fosamprenavir [62% lopinavir, 11% fosamprenavir]). HIV-1 RNA suppression to <50 copies/ml at week 48 was higher in the efavirenz group (84% versus 74% [difference 10%, 95% CI 3.4%, 16.7%; P=0.002]). The percentage of virological failures was similar (efavirenz 4% versus PI/r 4%; P=0.686), but voluntary discontinuations and toxicity-related treatment changes were higher with PI/r (4% versus 1%; P=0.006 and 11% versus 6%; P=0.069, respectively). However, resistance selection at failure was higher in patients receiving efavirenz (89% versus 50%; P=0.203). Efavirenz was significantly more effective than lopinavir/r or fosamprenavir/r, whereas no significant differences were observed between efavirenz and darunavir/r or atazanavir/r. The high viral suppression in the efavirenz group was also evident in patients with very high viral loads (>500,000 copies/ml) and in those with low CD4(+) T-cell counts. In routine clinical practice, the effectiveness of initial efavirenz-based regimens was at least similar to or even higher than various PI/r-based regimens in HIV-1-infected patients with plasma HIV-1 RNA>100,000 copies/ml.