HIV infection and its treatment are associated with dyslipidaemia and increased risk of cardiovascular disease. Accurate high-density lipoprotein (HDL) cholesterol values are necessary for the management of these abnormalities, but current methods have not been properly assessed in these patients. The aim of this study was to assess in HIV-infected patients the consistency and accuracy of a synthetic polymer/detergent homogeneous assay used to measure HDL cholesterol concentrations and to evaluate the impact of storage.HDL cholesterol was measured using a synthetic polymer/detergent homogeneous method in samples from HIV-infected patients and healthy subjects for each of the storage regimens: baseline, after 1 week at 4 degrees C, and after 12 months at -80 degrees C. The ultracentrifugation and precipitation assays were used for comparison.Three out of every 20 samples from HIV-infected patients had discrepant HDL cholesterol values with respect to the ultracentrifugation method. Overestimation was associated with high C-reactive protein concentrations and underestimation with plasma gamma-globulin concentrations, an effect that was amplified by any of the storage conditions tested.Caution is needed when using the synthetic polymer/detergent homogeneous method for direct measurement of HDL cholesterol concentrations in HIV-infected patients. This assay is of limited use in clinical trials in which frozen samples are analysed.