[Linear discrimination analysis in blood pressure monitoring]

The 24-hours blood pressure monitoring (BPM) is a reliable technique largely used for hypertension diagnosis in clinical practice. Presence of the lack of a nocturnal decrease in blood pressure values is one differentiating criterion between essential and secondary hypertension. New useful method of mathematical and statistical processing of BPM data using the linear discriminant analysis (LDA) was described. This statistical analysis was applied for 123 BPM measurements from patients with essential hypertension without and with different kidney diseases including hemodialyzed patients. The control group was normotensive healthy volunteers. The LDA method successfully separate group of patients with essential hypertension from patients with secondary renal hypertension but the error of return ranking to these groups was 28.5%. The recalculation for groups-controls, hypertensive patients and hypertensive hemodialyzed patients leads to decreasing of the error of return ranking to 5.7%. This retrospective study of nonhomogeneous groups could not clearly differentiate groups of hypertensive patients. The new prospective study on accurately definite groups of patients can eliminate these problems.