Accuracy and precision of analyses for total cholesterol as measured with the Reflotron cholesterol method

We compared plasma cholesterol measurements made with the Boehringer Mannheim Reflotron reflectance photometric analyzer in 1298 capillary blood samples with measurements made in venous blood samples collected at the same time and analyzed in four standardized Lipid Research Clinics laboratories. The Reflotron measurements averaged 0.8% to 7.8% lower than the laboratory values. Correlations (r) between the two sets of measurements ranged from 0.92 to 0.96. In some samples, however, the Reflotron values differed from the laboratory values by greater than or equal to 12%; the cholesterol concentrations in these samples tended to be higher than in those for which better agreement was observed. The smaller negative biases were observed when test strips were used that were calibrated with reference to the Centers for Disease Control Reference Method for cholesterol. The agreement between sequential Reflotron values averaged less than or equal to 4.3%. There was an average difference of less than or equal to 1.0% between Reflotron measurements made in each of two sequential capillary blood samples taken from a single finger puncture.