Background: An uncontrolled systematic error in serum biomarkers may be a serious problem when comparing their trends both within and between populations. The aim of the study was to assess which factors are responsible for systematic errors in the measurement of serum triglycerides (Tg) and the effect of fasting on serum triglycerides in Finnish population surveys.
Methods: Data on precision and accuracy during 30 years for serum triglycerides were documented from participation in 492 rounds of five different external quality assessment (EQA) programs. Data on fasting and health status from questionnaires were combined from three population surveys comprising 27,131 participants.
Results: The mean annual accuracy (bias) of the Tg methods from all EQAs during 1978-2007 was -1.54% (95% CI -2.25, -0.83). The mean relative change in triglyceride concentration per fasting hour was -3.7% (95% CI -4.2, -3.1) in all subjects. A minimum serum Tg concentration was seen in men and women who had fasted for at least 8 and 7 h, respectively.
Conclusions: The mean bias in serum Tg analyses has been very small throughout the 30-year period. Fasting has a considerable effect on triglyceride levels, but they can be converted either to fasting or non-fasting levels using specific factors.