In September, 2008, a 56-year-old man presented to our department with fever, myalgia, pharyngitis, and a maculopapular skin eruption. He reported having unprotected sex with a man 2 months prior to the onset of his symptoms. He was diagnosed with HIV seroconversion illness in view of a high HIV viral load (>14 280 000 copies per ml) and a third generation HIV antibody test that was weakly positive. During routine HIV follow-up in March 2009, his previously normal alkaline phosphatase had risen to 486 IU/L and his alanine aminotransferase was 185 IU/L. His total bilirubin was 15 μmol/L. He had no symptoms except for mild occasional right upper quadrant abdominal discomfort. An abdominal ultrasound scan showed multiple large stones in the gallbladder but was otherwise normal. His antimitochondrial antibody was strongly positive (1:4000 IgG, M2 and M4 positive), consistent with a diagnosis of primary biliary cirrhosis. A liver biopsy showed damaged interlobular bile ducts, with granulomas, and increased amounts of eosinophilic cytoplasm and intraepithelial infl ammatory cells, also typical of primary biliary cirrhosis. In April, 2009, his CD4 count was 360 cells per μL, and he was started on tenofovir 245 mg, emtricitabine 200 mg, and efavirenz 600 mg once daily. In June, 2009, owing to central nervous system side-eff ects, efavirenz was changed to lopinavir 400 mg and ritonavir 100 mg tablets twice daily. Ursodeoxycholic acid was introduced at a dose of 15 mg/kg per day in November, 2009. His liver function tests improved considerably once antiretroviral therapy was introduced (fi gure). In May, 2010, after 13 months of treatment with anti-retroviral therapy, and 6 months of ursodeoxycholic acid, his liver function tests were normal. His antimitochondrial antibodies decreased to 1:250. Using a branched chain hybridisation assay (Quanti-Gene Aff ymetrix, California, USA) we found that serum from our patient was strongly positive for human betaretrovirus RNA. He was last seen in June, 2010, and was asymptomatic, with no clinical or biochemical evidence of cirrhosis. His HIV virus was completely suppressed and his CD4 count increased to 580 cells per μL. In 2003, a human betaretrovirus closely related to mouse mammary tumour virus was cloned from the biliary epithelium from patients with primary biliary cirrhosis. Viral infection of cholangiocytes with human betaretrovirus results in the aberrant expression of the mitochondrial E2 component of the pyruvate dehydrogenase complex. This protein, normally located on the inner mitochondrial membrane, appears on the cell surface in patients with primary biliary cirrhosis, and is thought to trigger the appearance of antimitochondrial antibodies. The relation between betaretrovirus infection and primary biliary cirrhosis is controversial. It has been suggested that randomised controlled trials with antiviral therapy may provide supportive evidence for a causal role of viral infection. Randomised controlled trials using combination lamivudine 150 mg and zidovudine 300 mg twice a day, have shown biochemical improvements in patients with primary biliary cirrhosis, but failed to achieve signifi cant endpoints, possibly because of the development of viral resistance to therapy. Other antiretrovirals that have proven eff ective against betaretrovirus in vitro, include: adefovir, tenofovir, and lopinavir. Here, we report that the combination of tenofovir, emtricitabine, lopinavir, and ritonavir used to manage HIV infection, coincided with marked reduction of cholestatic liver function tests, before the institution of ursodeoxycholic acid. If an association between primary biliary cirrhosis and human betaretrovirus infection can be established, the effi cacy of this highly active antiviral regimen can be tested in future controlled trials in patients with primary biliary cirrhosis unresponsive to ursodeoxycholic acid.