RESUMO
BACKGROUND: Targets for cholesterol reduction are part of the Quality Outcomes Framework and general practitioners have to meet these targets to fulfil their remuneration package. By contrast, there are no targets for the accuracy of cholesterol or other lipid measurements and no recent surveys on performance of these assays. We have assessed the performance of lipid measurement of the available methods in the UK. METHODS: Serum samples collected from individual donors attending the national blood service were distributed after values were obtained from a secondary reference laboratory. Samples were sent to participant laboratories to assess different methods' analytical performance on single donation specimens, on routine external quality assessment pooled specimens, on specimens subjected to a range of freeze-thaw cycles and on frozen-stored specimens. RESULTS: Differences in measured cholesterol were found that were method-dependent and related to triglyceride content. HDL-cholesterol (HDL-C) showed significant positive bias in all assays. Individual donor specimens showed no significant changes with differing numbers of freeze-thaw cycles. Pooled serum was stable for up to six months. CONCLUSIONS: Most cholesterol measurements are accurate but some methods are affected by triglyceride interference. HDL-C methods show significant positive bias. Although there are potential matrix effects introduced as a result of specimen preparation, additional work is needed to show if these effects are present in fresh patient samples.