Modeling the circadian variability of ambulatorily monitored blood pressure by multiple-component analysis

The use of a set of new end points derived from ambulatory blood pressure monitoring (ABPM), in addition to the blood pressure (BP) values themselves, has been advocated to improve the sensitivity and specificity in diagnosing hypertension and to evaluate a person's response to treatment. An adequate estimation of rhythmic parameters depends, however, on the ability to describe properly the circadian pattern of BP variability. The purpose of this study was to identify a simple model that could characterize sufficiently well the circadian pattern of BP in normotensive healthy volunteers sampled by ambulatory monitoring. We studied 278 clinically healthy Spanish adults (184 men), 22.7 +/- 3.3 yr of age, without medical history of hypertension and mean BP from ambulatory profiles always below 135/85 mmHg for systolic/diastolic BP, who underwent sequential ABPM providing a total of H 15 series of BPs and heart rates (HRs), sampled on each occasion at 0.5h intervals for 48 h. Subjects were assessed while adhering to their usual diurnal activity and nocturnal sleep routine, without restrictions but avoiding the use of medication. The circadian rhythm in BP and HR for each subject was established by multiple-component analysis. A statistically significant 24h component is documented for 97% of the BP profiles, with a significant second (12h) harmonic documented in 65% of the profiles. Other Ultradian harmonic components were significant in less than 20% of the profiles. A statistically significant increase in the coefficient of determination (percent of overall variability explained by the function fitted to the data) was only obtained after including the periods of 24 and 12 h for BP, and periods of 24, 12, and 6h for HR in the model components. Although other ultradian components can be demonstrated as statistically significant in a small percent of subjects, a rather simple model including only the two first harmonics of the 24h period describes sufficiently well, at the specified sampling rate, the circadian pattern of BP in normotensive subjects. Departure from this model could characterize overt pathology, as recently demonstrated in the diagnosis of preeclampsia.