The Analysis of Circadian Variation of Blood Pressure by Using Maximum Entropy Method-Least Squares Method (MEM-LSM) : Comparisons among Normotensives, Borderline Hypertensives and Hypertensives, and Their Seasonal Changes

We have previously reported the maximum entropy method-least squares method （MEM-LSM） which provides more accurate measurements of the spectrum than conventional methods. We have used MEM-LSM for analysis of 24-hour human blood pressure （BP） and pulse rate （PR）. In the present study, the MEM-LSM and cosinor method were compared for accuracy in that prediction. Then, to study the characteristics and seasonal changes of 24-hour human BP in normotensives （NT）, borderline hypertensives （BHT） and hypertensives （HT）, MEM-LSM were examined in 59 subjects composed of NT, BHT and HT in summer and 45 subjects in winter. An ambulatory blood pressure monitor （ABPM-630） was attached to each subject, and 24-hour data was obtained and then analyzed by MEM-LSM. In order to examine the approximation, value ε, which is defined as the total summation of the squared differences between the actual measurements and calculated values was used （see NOTE below）. In summer, ε was significantly smaller （p<0.0001） in MEM-LSM than in the cosinor method. In the comparison among 3 groups, there was no significant difference in maximum magnitude between BHT （159±3/101±±2mmHg） and HT （168±3/103±2mmHg）, and in minimum magnitude between BHT （108±3/63±2mmHg） and NT （104±2/61±1mmHg）. Maximum amplitude/midline estimating statistic of rhythm （MESOR） was significantly larger in BHT （38±2/42±2%） than in NT （28±2/31±2%）, and it also had a tendency to be larger （p=0.07） than in HT （30±4/33±4%）. There was no significant difference in the periodic and phasic analysis among these 3 groups. The same tendency was obtained in winter. Turning to seasonal changes, casual BP （138±3/84±2mmHg）, MESOR （127±2/78±1mmHg） and second peak magnitude （146±3/90±2mmHg） were higher in winter than in summer: casual BP （130±3/82±1mmHg, p<0.01/p<0.05）; MESOR （125±2/76±1mmHg, p<0.05/p<0.05）; and second peak magnitude （140±3/86±2mmHg, p<0.01/p<0.05）. But the first peak magnitude and minimum magnitude showed no significant difference between summer and winter. The minimum phase （02:40±00:31/02:41±00:22, systolic blood pressure （SBP）/diastolic blood pressure （DBP）） and the first peak phase （10: 10±00:18/09:55±00:18） in winter was significantly （p<0.01） later than the corresponding values of minimum phase （01:15±00:20/01:25±00:19） and first peak phase （07: 58±00:20/08:31±00:20） in summer. But in the second peak phase, there was no significant difference between summer and winter. From these findings it was concluded that MEM-LSM is a more accurate analysis method than the cosinor method. Using MEM-LSM: （1） BP elevation was more abrupt in BHT in the day time than in NT, but there was a marked BP decrease during the night-time; the night-time BP closely resembled that of NT. （2） Among NT, BHT and HT, there were no chronobiological differences in their 24-hour BP. （3） BP is mainly regulated by the sleep-waking rhythm and partially regulated by outdoor temperature change.