Session M1.6

Automated Blood Pressure Measurement: Reasons for Measurement Variability Uncovered

D Zheng*, JN Amoore, S Mieke, FE Smith,
ST King, A Murray

Newcastle University
Newcastle upon Tyne, UK

Blood pressure is an important clinical measurement. It is measured frequently by healthcare providers and by the general public at home, but the accuracy of many measurements is still in doubt. The most common automated blood pressure devices use the oscillometric technique. In this technique, systolic, diastolic and mean arterial blood pressure (SBP, DBP and MAP) are determined by analysing the oscillometric pulse waveform. However, there is little information available on how these pressures relate to features extracted from oscillometric waveforms. The aim of this study was to assess the variability of the amplitude of oscillometric pulse waveform characteristics at SBP, DBP and MAP.
Sixty oscillometric waveforms from twenty healthy subjects were studied. The cuff pressure was deflated from 150 mmHg to approximately 40 mmHg over the duration of 40 s. SBP and DBP were obtained with a manual sphygmomanometer, from which manual MAP was calculated from the standard empirical equation. The cuff pressure corresponding to the maximum oscillometric pulse was taken as the automated MAP. The oscillometric pulse amplitudes, normalised to the maximum oscillometric pulse, at SBP, MAP and DBP were then determined. The cuff pressures associated with the oscillometric pulses having half maximum amplitude in each subject were also measured.
The overall means and SDs of SBP and DBP were 118±11 and 76±9 mmHg. The normalised oscillometric pulse amplitudes were 0.45±0.10 (at SBP), 0.89±0.09 (at manual MAP) and 0.80±0.12 (at DBP). They were significantly different between SBP and DBP, and between manual and automated MAP (both P<0.001).
The cuff pressures at half the maximum oscillometric pulse amplitude were 117±12 mmHg and 66±10 mmHg. The paired differences between the higher value and SBP was 1±5 mmHg (close, but significantly different, P<0.05), and between the lower value and DBP was 10±7 mmHg (P<0.001). Significant differences between manual and automated MAP were also observed (90±8 mmHg compared with 87±10 mmHg, with a paired difference of 3±6 mmHg, P<0.001).
In conclusion, there are large variations in the pulse characteristics at SBP, DBP and MAP. This complicates a reliable automatic estimation of these values.

(Abstract Control Number: 188)