Reference intervals and decision limits for plasma lipids and lipoproteins: a practical evaluation of current recommendations.

As part of a longitudinal study--the Kristianstad Survey--we measured plasma cholesterol, HDL- and LDL-cholesterol, triglycerides and lipoprotein (a) in a reference group consisting of 203 men and women aged 20-80, randomly sampled from a well-defined area in the southernmost part of Sweden. The selection of reference individuals and the collection of specimens for assay of the constituents were performed in accordance with current recommendations. The results were subjected to statistical analyses both with and without application of exclusion criteria. Application of the theoretical exclusion criteria resulted in the exclusion of 22% of the participants; however, this procedure had a remarkably weak impact on the results: the mean values and the standard deviations were almost unaltered. The mean (standard deviation) for cholesterol was 5.9 (1.3) mmol l-1, for HDL-C 1.1 (0.3) mmol l-1, and for LDL-C 4.3 (1.2) mmol l-1. Women had higher values than men. Plasma triglycerides were positively skewed; their median and Q3-Q1-values were 1.0 and 0.5 mmol l-1 respectively, men higher than women. There was an increase with age for cholesterol and LDL-C in both sexes and for triglycerides in women. The steepest increase of cholesterol values with age in women coincided with menopause, which have a more clear-cut separation between high and low cholesterol values than did any age limit.