抄録
Data measured for parameters of biological functions and metabolism always present some variations, arising from physico-biochemical fluctuations within the body, technical errors in measurements, instability in instruments, and limitation of the methods employed. Statistically, these variations in data are evaluated by such indices as standard deviation (SD), standard error (SE), coefficient of variation (CV) and logarithmic (geometric) standard deviation (sigma-x). When we use sigma-x to analyze variations in data gained from populations (n<100), it shows unique characteristics ; i) it indicates the extent of variations shown by individuals in a population; ii) it is an anonymous value without dimension of measured unit; iii) it is not influenced by the value of the mean data and iv) each parameter has its own fixed sigma-x value independent of the population studied, gender concerned, and method of measurement used. Therefore, it is said to be a quasi-constant value for each parameter. To ascertain the e characteristics of sigma-x, we examined a total of 20, 200 subjects (10, 599 males and 9, 601 females) from 8 populations (Muroran-Noboribetsu 1, 787, Iwate 5, 326, Shibuya-Tokyo 1, 384, Mitaka-Tokyo 1, 998, Niigata 2, 230, Wakayama 1, 878, Anan-Tokushima 1, 887, and Miyazaki 3, 710) for the sigma-x of total cholesterol, triglycerides and HDLcholesterol. Sigma-x of total cholesterol was 0.08, triglycerides 0.22, and HDL-cholesterol 0.11. These sigma-x values were nearly constant throughout subpopulations, did not differ between males and females, and were independent of methods used for measurements. Thus, individual variabilities of the parameters of lipid metabolism demonstrated their own quasi-constant values.
