Sexual dysfunction is a common problem among HIV-infected patients, in whom a prevalence from 19.5% [1] to 71% [2] has been reported. One of the determinants of this high prevalence seems to be antiretroviral therapy [1,3,4]. A high frequency of sexual dysfunction has been described in patients on highly active antiretroviral therapy (HAART), especially with regimens containing protease inhibitors (PI), which have been found to be an independent risk factor for sexual dysfunction in several studies [1,3–5]. To our knowledge, there are no data so far suggesting a differential effect of particular antiretroviral drugs or regimens on the clinical course of HIV-related sexual dysfunction. We describe two patients with sexual dysfunction who improved significantly after starting treatment with atazanavir/ritonavir. Case 1 A 42-year-old white man, with known HIV infection since 2001 acquired through intravenous drug use, was co-infected with hepatitis C virus and had a past history of hepatic alcohol disease. In May 2003, HAART with lopinavir/ritonavir, tenofovir and lamivudine was prescribed. Soon after starting the antiretroviral regimen, the patient developed sexual dysfunction, with a decrease in sexual interest and desire, and erectile dysfunction. No other associated factors were recognized. Attending to the requests of the patient, who clearly linked sexual dysfunction to HAART, antiretroviral therapy was changed 6 months later, with lopinavir/ritonavir being substituted by efavirenz. After 4 months of therapy, the HIV viral load was undetectable and the CD4 cell count was 125 cells/μl, but no improvement of sexual function had been achieved and the patient complained of cutaneous pruritus. A new HAART regimen including atazanavir/ritonavir, didanosine and zidovudine was started. Three months later, the patient revealed an important sexual function improvement that was first noticed a few days after beginning the new therapy, with full recovery of sexual desire and erectile function over the next few weeks. His CD4 cell count and viral load were 143 cells/ml and less than 50 copies/ml, respectively. Nine months later the patient maintained normal sexual intercourse. Case 2 A 40-year-old white male patient, with known HIV infection since 1988, Centers for Disease Control and Prevention classification C2 was co-infected with hepatitis C virus and had a past history of intravenous drug use receiving substitutive treatment with methadone. The patient had received several HAART regimens since October 1997, including non-nucleoside analogues and PI. In November 2002, while on lopinavir/ritonavir, tenofovir and lamivudine, he started to experience a loss of libido, erectile dysfunction and a decrease in sexual interest. In January 2004, his CD4 cell count was 95 cells/mm3 and his HIV viral load was 5.3 log10 copies/ml, and a new antiretroviral regimen with atazanavir/ritonavir, tenofovir and didanosine was started. One month later he reported spontaneously an outstanding sexual function improvement, with nocturnal erections and a significant augmentation of sexual desire. The CD4 cell count was 137 cells/mm3 and the HIV viral load was 2000 copies/ml. He denied any other particular event that could have accounted for the improvement in sexual dysfunction. Normal sexual function was maintained through one year's therapy with atazanavir/ritonavir. The unexpected occurrence of sexual dysfunction improvement in these two cases prompted us to investigate retrospectively the influence of atazanavir on sexual function in other HIV-infected patients cared for in our clinic who had started atazanavir therapy during the Spanish expanded access programme. In January 2005, we distributed a four-item questionnaire to the 11 male patients receiving atazanavir, eight of whom gave their consent to participate in the investigation. Items were extracted from the International Index of Erectile Dysfunction test [6], and included questions about erectile function, orgasmic function, sexual desire and intercourse excitation. Of the eight patients studied, two had sexual dysfunction before they start receiving the atazanavir-containing regimen, and one of them reported an important improvement in sexual interest and in erection, and a moderate improvement in excitation. Among the six other patients who had normal sexual function before atazanavir therapy, four reported a significant improvement in sexual performance, although there were external circumstances in two of them (a change of the usual partner). This is a clinical observation that precludes any conclusion being drawn. However, the temporal relationship, according to the definition of substance-related impotence, the absence of concomitant medical and non-medical circumstances, and the occurrence of this unusual event in three patients from our cohort on atazanavir therapy, suggest a relationship with the drug. Atazanavir is the latest PI marketed, and clinical experience out of the trials is still limited. As atazanavir has a unique metabolic profile compared with the other PI, it may also have different effects on sexual function. Appropriately designed studies to explore the differential effects of atazanavir and other antiretroviral drugs on the sexual sphere may be warranted.