Improved automatic analysis of ambulatory blood pressure data based on precise detection of individual night-time from diurnal profile of heart rate

Software programs sold with ambulatory blood pressure monitoring (ABPM) devices are designed to use some set 'typical' night-time (e.g. 2300-0700) to estimate daytime/night-time blood pressure (BP) with limited accuracy. Alternative use of individual periods of sleep/wakefulness from patient diaries is time consuming and subjective. We developed a simple mathematical algorithm for the detection of the 'night-time' as a period of low values in diurnal profiles of heart rate (HR) allowing accurate automatic analysis of daytime/night-time blood pressure. To test this technique we designed a software application allowing automatic analysis of ABPM data based on the different night-time definitions, including the developed algorithm and compared reproducibility of the degree of BP dipping produced by the different methods across two days of 48-h ABPM. A 48-h ABPM study was performed in 33 patients with uncomplicated stage II hypertension. Means and standard deviations (SD) of the differences in the degree of BP dipping between two 24-h periods of 48-h ABPM were obtained separately for three methods of night-time definition: automatic detection from individual HR profiles, fixed 2300-0700 h interval and sleep time from patient diaries. Reproducibility of the BP dip estimation across 2 days of BP monitoring was significantly better for night-time detected from individual HR profiles than for the fixed 2300-0700 h interval or sleep time from diary. The SD of the differences was 6.7/8.2 compared with 13.5/18.3 and 13.0/14.8 respectively (systolic BP/diastolic BP, mmHg). Implementation of the developed method of night-time definition may significantly improve automatic analysis of ABPM data.