| Hypertension, known as 'a silent killer', is an important public health challenge, afflicting approximately 1 billion adults around the world. The monitoring of blood pressure (BP) is vitally important in order to identify hypertension and treat it earlier before serious health problems are developed. The conventional BP measurement provides only intermittent BP and causes circulatory interference if the cuff is inflated frequently. There is an urgent need to develop new devices which are fully wearable and unobtrusive for noninvasive and continuous monitoring of arterial BP in daily life.;Cuffless BP measurement has been proposed as a new concept in recent years to realize the continuous monitoring of BP. This research focuses on the investigation of cuffless BP monitoring technique using heart sound information. Specifically, the thesis proposes a new cuffless technique based on the timing of the second heart sound (S2), which will enable a novel wearable design of BP monitor, for instance, a multifunctional electronic stethoscope.;First, a mathematical model has been developed to investigate the effects of aortic BP on the aortic component (A2) in S2 produced by the vibration of the closed aortic valve. The nonlinear elasticity of aortic wall has been introduced to the model, reflecting the nature of aortic wall tissue and extending the model to the applications involving wide BP variations. The results of simulation show that the fundamental frequency and amplitude of A2 increases as aortic systolic blood pressure (SBP) is elevated, which is able to explain the 'accentuated S2' usually heard in the hypertensives. Nevertheless, the possibility of BP measurement using spectral information of externally recorded heart sounds still needs a careful examination because the frequency characteristics tends to be blurred during sound transmission.;Second, a modified model of heart-arterial system has been proposed in this thesis for describing the timing of aortic valve closure as a result of heart-arterial interaction. A timing parameter, RS2, was defined as the time delay from the peak of ECG R wave to the onset of S2. The study has investigated the relation between RS2 and aortic BP under varying peripheral resistance, arterial compliance, heart rate, cardiac contractility and preload. Based on the simulation results of parametric analysis, it is hypothesized that RS2 bears a significant negative correlation with both SBP and diastolic blood pressure (DBP) as the peripheral resistance, heart rate or cardiac contractility varies.;Third, in order to verify the findings of the model-based study, three experiments were carried out to explore the relationship between RS2 and BP. The alterations of RS2 in the dynamic-exercise experiments are mainly attributable to the interactive effect of the changes in heart rate, cardiac contractility and peripheral resistance, and the effect of heart rate is dominant. In two dynamic-exercise experiments, the timing parameter, RS2, exhibited a close inverse correlation with SBP (r =0.892 and r =0.845, p<0.05 in both experiments) and a moderate inverse correlation with diastolic blood pressure (DBP) (r = 0.687, p0.05). The correlations are comparable to those of PTT-based parameters. However, due to the restricted range of the BP variation, there was no significant correlation observed in long-term rest monitoring experiment. Moreover, the standard deviation (SD) of the errors for SBP and DBP estimated by linear fitting of 1/RS2 is close to that of PTT-based estimation. The results also suggest that the ability of RS2 on BP estimation is as good as that of the PTT based parameters.;Finally, based on the findings on both theoretical and experimental studies, a linear prediction model with a novel calibration scheme has been proposed to estimate the BP using 1/RS2. The proposed method was evaluated in a clinical test on 85 volunteers aged 40+/-13 years, including 18 hypertensives. The average of BP measured by simultaneous ausculatory and oscillometric approaches was used as a reference. The results of clinical test shows that the RS2 based approach can estimate SBP and DBP within the 2.1+/-7.4 mmHg and 0.8+/-6.6 mmHg of the reference respectively, indicating the approach has the potential to achieve the accuracy required for medical diagnosis according to AAMI standard (mean error within +/-5mmHg and SD less than 8mmHg) and BHS protocol.;To summarize, the original contributions of the thesis are: 1. By the introduction of the nonlinear elasticity of aortic wall, a mathematical model for the vibration of the closed aortic valve was improved and extended to the applications involving wide variations of BP To my knowledge, this represents the first study to look into the effects of aortic BP on the frequency characteristic of S2 from the theoretical point of view.;2. A modified model of heart-arterial system was proposed for describing the timing of the second heart sound as a result of the heart-arterial interaction. Simulation results suggest that RS2 bears a significant negative correlation with both SBP and DBP as heart rate, cardiac contractility and peripheral resistance varies. The hypothesis was supported by the experimental data. To our knowledge, it is the first study describing the relation of the timing of S2 to BP by both the model-based study and experimental data.;3. As a preliminary study, a linear predication model using RS2 with a novel calibration scheme was proposed for BP estimation and it has been evaluated in clinical test on 85 volunteers including 18 hypertensives. The results indicate that the approach has the potential to achieve the accuracy required for medical diagnosis. |
