ABSTRACT
BACKGROUND: For practical reasons it is not easy to obtain fasting samples in large population health surveys. Non-fasting triglyceride (Tg) values are difficult to interpret. The authors compared the accuracy of statistically corrected non-fasting Tg values with true fasting values and estimated the misclassification of subjects with high low-density lipoprotein cholesterol (LDL-C) and the metabolic syndrome. METHODS: Non-fasting blood was obtained from a population-based sample of 4282 individuals aged 24-75 years in the National FINRISK 2007 Study. Fasting blood samples were drawn from the same persons 3 months later. Non-fasting serum Tg values were converted into fasting values using previously published formula. LDL-C was calculated and classification of the metabolic syndrome was carried out according to three different latest guidelines. RESULTS: The median (25th, 75th percentile) non-fasting serum Tg concentration was 1.18 (0.87, 1.72) mmol/L and after postprandial correction 1.06 (0.78, 1.52) mmol/L. The true-fasting serum Tg concentration was 1.00 (0.75, 1.38) mmol/L (P < 0.001) vs. non-fasting and corrected value. Bias of the corrected value was +5.9% compared with the true-fasting Tg. Of the true fasting subjects, 56.4% had LDL-C ≥ 3.00 mmol/L. When calculated using non-fasting serum Tg, the prevalence of high LDL-C was 51.3% and using statistically corrected Tg it was 54.8%. The prevalence of metabolic syndrome was 35.5% among fully fasted persons and among non-fasting subjects 39.7%, which after statistical correction of Tg decreased to 37.6% (P < 0.001 for all comparisons). CONCLUSIONS: Correction of non-fasting serum Tg to fasting values plays a minor role in population studies but nevertheless reduces misclassification of calculated high LDL-C from 5.1 to 1.6% and the metabolic syndrome from 4.2 to 2.1%.