Abstract
It has been reported that that the wrist-brachial index (ratio of systolic blood pressure in the wrist to that in the upper arm) and the finger-brachial index (ratio of systolic blood pressure at the base of the first finger to that in the upper arm) declines with increasing arteriosclerosis ; however, differences can also be shown to emerge due to variations in the conditions under which measurements are taken as well as being due to disease-related complications. The blood pressure of 73 dialysis patients receiving hemodialysis was measured 30 times at the upper arm, wrist, and base of the first finger, and then factors that influenced the ratios among these measurements were evaluated in 2,190 sets of data. These factors were categorized as : 1) factors that varied with every measurement (systolic blood pressure in the upper arm, pulse rate, and first- or second-half of the hemodialysis session) ; 2) arteriosclerotic factors that varied with each patient (age, hemodialysis period, smoking, diabetes mellitus, dyslipidemia, and cardiovascular disease. Hypertension was excluded because of the small number of normotensive patients in the sample) ; and 3) other conditions that varied with each patient (atrial fibrillation and presence of arteriovenous fistula on the measurement side), and then the data sets were analyzed. It was natural to assume independence among these factors and elements, and that any relation between two indices and elements was a linear relationship based on a simple regression analysis when the explanatory variable was on an interval scale or ratio scale, or when there was a difference in the average of 2 groups on an unpaired t-test based on either a nominal scale or an ordinal scale. Therefore, each two-index set was analyzed by multiple regression analysis containing a qualitative scale, the elements of which represented the explanatory variable. The results of the multiple regression analysis were able to explain the 2,190 sets of data well.
