To the Editor.
—We have noted an error in the article by Dr Cooper and colleagues1 concerning measurement of blood lipid levels. Table 2 of that article is used in laboratory practice and quoted in the medical literature2, yet most of the entries in this table are incorrect because they were calculated using an erroneous formula.Using the notation of the original article by Cooper et al, CVt is the total coefficient of variation (CV, standard deviation divided by the mean) of reported analytical results for an individual, from whom multiple specimens are taken at different times. The total intraindividual CVt can be separated into the intraindividual biological CV (CVb) and the laboratory analytical CV (CVa). Let NS denote the number of time periods when specimens were taken. This was defined as the number of specimens by Cooper et al.1 Let NR denote