Background: Epidemiological studies generally aim to make simple but unbiased measurements of individuals. For this reason measurements of metabolic variables (including blood lipids, glucose and insulin) are usually carried out after a period of fasting. Few studies have examined the extent to which the use of a defined protocol for fasting in epidemiological studies abolishes the influence of fasting duration and time of day on biochemical measurements.
Methods: Cross-sectional survey of British Regional Heart Study participants (4,252 men aged 60-79 years), in which men without diabetes were asked to provide a blood sample after fasting for at least 6 h. Serum total, HDL and LDL cholesterol, triglyceride and insulin, and plasma glucose concentrations were measured between 08:00 h and 18:00 h.
Results: Non-fasting men had lower mean LDL cholesterol and higher glucose, insulin and triglyceride levels than fasting men; these differences were more marked among diabetics. Among fasting men without diabetes, insulin and glucose levels were strongly related to time of day, falling gradually throughout the morning and remaining stable in the afternoon. Because of these relationships and the dependence of fasting duration on time of day, insulin and glucose displayed a periodic relation with fasting duration above 6 h. These associations were largely abolished by adjustment for time of day; associations with time of day were unaffected by adjustment for fasting duration. Triglyceride concentrations fell with increasing fasting duration. This relationship was also mediated through a gradual increase in triglyceride levels throughout the day. Adjustments to compensate for these variations are described.
Discussion: Even after fasting, biochemical measurements may still differ in relation to fasting duration and time of day. In epidemiological studies, it is important to standardize both the period of fasting and the time of day as much as possible, and make adjustments where necessary.