Dyslipidaemia is very common in patients with HIV infection, but current therapies are often suboptimal. Since niacin may cause insulin resistance and hepatotoxicity, it has generally been avoided in this setting. Non-diabetic male subjects (n=33) who had well-controlled HIV infection on antiretroviral therapy, fasting triglycerides > or =2.26 mmol/l and non-high density lipoprotein cholesterol (non-HDL-C) > or =4.66 mmol/l received escalating doses of extended-release niacin (ERN) up to 2,000 mg nightly for up to 44 weeks. Fourteen subjects (42%) had pre-diabetes at entry. Twenty-three subjects (70%) received the maximum dose, eight (24%) received 1,500 mg. Niacin was well-tolerated. Only four subjects (12%) discontinued study treatment. There were small increases in fasting glycaemia and insulin resistance estimated by the homeostasis model assessment, but insulin resistance measures from the 2-h oral glucose tolerance test only transiently worsened. No subject developed persistent fasting hyperglycaemia; one had persistently elevated 2-h glucose >11.1 mmol/l. There were no significant changes in serum transaminases or uric acid. At week 48, the median change in fasting lipid levels in mmol/l (interquartile range) were: total cholesterol -0.21 (-1.35, -0.05), HDL-C +0.013 (-0.03,+0.28), non-HDL-C -0.49 (-1.37,+0.08) and triglycerides -1.73 (-3.68, -0.72). Favourable changes in large HDL and large very low density lipoprotein particle concentration were observed by nuclear magnetic resonance spectroscopy. ERN in doses up to 2,000 mg daily was safe, well-tolerated and efficacious in HIV-infected subjects with atherogenic dyslipidaemia. Increases in glycaemia and insulin resistance tended to be transient.