Health-based reference values of the Mini-Finland Health Survey: 3. Triglycerides in total serum and in different lipoprotein fractions

Towards common reference intervals in clinical chemistry. An attempt at harmonization between three hospital laboratories in Skåne, Sweden

The population sample of the Kristianstad survey, a reference intervals survey in the county of Kristianstad, was used to establish new reference intervals in clinical chemistry at the laboratories of the Central Hospital in Kristianstad, the University Hospital in Lund and the University Hospital in Mälmo. Three-hundred and fifty nine subjects, male and female, aged 20-80+ years, were invited to participate in the study, with a participation rate of 70%. Up to 70 analyses were performed on each subject, general clinical chemistry parameters in all three laboratories, specialized analyses where available. Separate a priori exclusion criteria were defined for each test. In addition, the test pattern of each individual was evaluated for signs of preclinical disease. Twelve cases of preclinical disease were discovered and clinically confirmed. Details on all test methods are presented along with information concerning instruments used, calibration procedures, methods of calculation and obtained reference intervals. Although the methods were in general calibrated against acknowledged reference materials, in some instances differences were found that made common reference intervals across all laboratories impossible. Problems relating to the practical use of international recommendations and the establishment of reliable reference intervals are discussed.

Poor applicability of the Friedewald formula in the assessment of serum LDL cholesterol for clinical purposes

OBJECTIVE

To determine the accuracy of the estimation of serum low-density lipoprotein (LDL) cholesterol concentration by the Friedewald formula.

METHODS

Modifications of the calculation formula are presented on the basis of ultracentrifugal separation of serum high-density lipoprotein and LDL cholesterol in the specimens collected (n = 1215) in a nationwide health survey.

RESULTS

The formulas obtained from different subject groups differed relatively little from each other. The accuracy of the original Friedewald formula was poor; in about 36% of the subjects the error was more than 5% compared with ultracentrifugally obtained results. By applying the currently recommended coronary heart disease (CHD) risk categorizations, high proportions (5%-28%) of the subjects were classified into wrong CHD risk categories when LDL cholesterol was calculated with any of the formulas. At high serum triglyceride levels, misclassifications were especially common.

CONCLUSIONS

We conclude that even the most accurate LDL cholesterol calculation methods should be used with caution while classifying subjects into categories of CHD risk. In hypertriglyceridemic subjects, the calculation formulas probably should not be used at all.

Health-based reference values of the Mini-Finland Health Survey: 2. Cholesterol in total serum and in different lipoprotein fractions

The reference values for cholesterol concentrations in the whole serum and in its VLDL, LDL and HDL fractions have been produced based on the data obtained in the Mini-Finland Health Survey. The lipoprotein fractions were separated with ultracentrifugation. The aim was to obtain reference values for the apparently healthy, ambulatory population. Two health-derived criteria were used to select subjects for reference populations: those based on the literature available and those based on the recommendations published by the Committee on Reference Values of the Scandinavian Society for Clinical Chemistry and Clinical Physiology, with slight modifications. The 95% inner reference intervals of total serum cholesterol in all the subjects and in the two selection groups were 4.2-9.0, 4.2-8.4, and 4.2-8.5 mmol l-1 for men, and 4.2-9.4, 4.1-8.6, and 4.2-9.0 mmol l-1 for women, respectively. The corresponding medians were 6.3, 6.2 and 6.2 mmol l-1 in men, and 6.4, 6.0 and 6.1 mmol l-1 in women, respectively. Frequency distribution curves showed clear skewness to the right in VLDL cholesterol and slight skewness in LDL and HDL cholesterol. In women there was a clear rise in total, VLDL and LDL cholesterol after the early middle age, whereas HDL cholesterol was lower in the older age groups. In men the age dependency was not as prominent; total cholesterol levels showed lower levels in older men. Of the background lifestyle factors alcohol consumption, smoking, obesity and physical exercise had negligible associations with total, VLDL, LDL and HDL cholesterol reference intervals.