Effect of time between measurements on within-subject variability for total cholesterol and high-density lipoprotein cholesterol in women.

A single blood cholesterol measurement may not accurately reflect an individual's true mean concentration. If duplicate blood samples are taken, what number of days between sampling gives the best chance of detecting the maximum within-subject variation? In this study, we analyzed 20 serial blood samples obtained from each of 13 healthy, menstruating women over 35 days. Variability was calculated as the semivariogram, which gives the average squared difference between replicate samples taken over a range of sampling intervals. Data were available for a complete set of intervals from 1 to 26 days. Variability in total cholesterol (TC) increased as the interval between sampling increased from 1 to 12 days. With high-density lipoprotein cholesterol (HDL-C), variability increased from 1- to 7-day intervals. In practice, our results suggest that, irrespective of the time of menstruation, the minimal interval for collecting a second blood sample for TC and HDL-C assays is approximately 2 weeks.