Non-invasive peripheral perfusion monitoring using photoplethysmography.
| dc.contributor.author | Khan, Musabbir | |
| dc.date.accessioned | 2017-03-12T18:57:37Z | |
| dc.date.available | 2017-03-12T18:57:37Z | |
| dc.date.issued | 2016 | en |
| dc.description.abstract | In the intensive care unit (ICU), cardiovascular disease and sepsis are major drivers of morbidity, mortality, and cost. Microcirculatory dysfunction during sepsis and cardiac failure is known to significantly impair perfusion in major organs and peripheral tissues, resulting in poor patient outcome if not corrected early. Arterial oxygen saturation (SaO2), venous oxygen saturation (SvO2), and their difference, or oxygen extraction (O2E), in association with blood flow, are key parameters necessary to assess tissue perfusion. Although, SaO2 can be reliably estimated by non invasive optical technologies, such as pulse oximetry in terms of SpO2, there are no established, easy, or non-invasive methods to assess SvO2, and thus O2E. Therefore, real-time, low-cost, and non-invasive assessment of SvO2, O2E, and blood flow could significantly advance the diagnosis, management, and of care sepsis or cardiovascular disease in the ICU. This research focuses on the development of a novel pulse oximeter based concept to non invasively and more completely assess tissue perfusion in peripheral extremities, such as the finger. A transmittance pulse oximeter sensor was used as the basis of a device to estimate and monitor peripheral SvO2, O2E, and blood flow changes. Robustness of perfusion assessment methods at various digit skin temperatures was also investigated to improve accuracy and define the limits of sensor accuracy. A custom built pulse oximeter (PO) device and graphical user interface (GUI) were developed to access and acquire all photoplethysmograph (PPG) signals, including the raw PPG signals. A two stage digital filter system was implemented to extract useful signals from the PPG. Time and frequency domain methods, including peak-trough detection algorithms, were developed to analyse the PPG data and estimate parameters, such as SaO2 and heart rate. This system was thus used as the basis for developing SvO2 and blood flow monitoring methods, and also the investigation of temperature effects on PPG measurements. A study was conducted with healthy adult volunteers (n = 20) to investigate the effect of a range of digit skin temperatures on pulse oximeter performance using a transmittance pulse oximeter sensor. It was hypothesized that cold digits can significantly reduce PPG signal quality and the resulting accuracy of SpO2 measurements, due to the likely reduction in blood flow of the periphery at these temperatures. PPG data were recorded at cold, normal, and baseline skin temperatures, using the PO system as well as a commercial pulse oximeter for reference. Results show warm skin temperatures improved PPG signal quality up to 64.4% compared to baseline, and provided SpO2 estimation of 96.5[96.1 – 97] % in the expected range for healthy adults. At baseline conditions, the majority of subjects showed a good PPG signal and expected SpO2 estimate of 95.8[93.2 – 96.8] %. PPG signal quality degraded significantly up to 54.0% at cold conditions and SpO2 estimates of 88.5[87.1 – 92.8] % were unreliable when compared to baseline. The main outcome is a tri-linear model quantifying PPG signal quality as a function of temperature, suggesting warm skin temperature conditions (approximately 33°C) should be maintained for reliable transmittance pulse oximetry. In circulatory dysfunction, peripheral blood flow can be reduced, adding error to assessment of SvO2 and O2E using typical models and assumptions. Relative volumetric blood flow change assessments using the PO sensor were performed in a study using 7 human adult volunteers. The signal amplitude of the high frequency PPG component, which is related to the pulsatile arterial blood flow, was used as an indicator of blood flow change. A vascular Doppler ultrasound sensor was used as a reference measurement. Changes in blood flow conditions were induced by a series of vascular occlusion tests. Good correlation (R2 = 0.69) and trend agreement was obtained between median PPG amplitude and Doppler ultrasound velocity, particularly at normal and clinically important low flow conditions. Thus, PPG amplitude monitoring can be a potential surrogate or alternative to vascular Doppler ultrasound based blood velocity monitoring, and can provide continuous and reliable measurements, ensuring flow conditions can be included in assessing SvO2. A novel artificial pulse generation system (APG) was developed to cause low frequency and low pressure modulations of venous blood in the finger using a pneumatic digit cuff. The APG system used a feedback controlled pressure regulator and solenoid valve to inflate/deflate the pressure cuff. This system was designed to exploit the significant arterial-venous compliance difference and make the peripheral venous blood pulsatile, while having negligible impact on arterial blood flow. Ten healthy human adults were recruited for proof-of-concept testing of this device. Modulation ratio (R) values derived from the artificially modulated PPG signals were used to estimate venous oxygen saturation (SpvO2) using an empirical calibration equation developed for arterial blood. Conventional empirical calibration model estimated arterial and venous saturations of 96.95[96.1 – 97.4] % and 93.15[91.1 – 93.9] % agree with published literature values. Median O2E was 3.6%, with a statistically significant and expected difference (p = 0.002) between pairs of measurements in each subject. The APG system in association with the pulse oximeter device to assess SvO2 was then validated in a clinical study with healthy adult volunteers (n = 8). A range of physiologically realistic SvO2 values were induced using vascular occlusion tests. Gold standard, arterial and venous blood gas measurements were used as reference measurements. Modulation ratios related to arterial and venous systems were determined using a frequency domain analysis of the PPG signals. A strong, linear correlation (R2 = 0.95) was found between estimated venous modulation ratio and measured SvO2, providing a calibration curve relating measured modulation ratio of venous blood to oxygen saturation. Median venous and arterial oxygen saturation accuracies from paired measurements were 0.29% and 0.65%, respectively, showing good accuracy of the pulse oximeter system. Investigations also revealed the empirical calibration model used to estimate SpaO2 cannot be used estimate SpvO2 because of the difference in optical absorption caused by artificial perturbations. In particular, there is a significant difference in gradient between the SpvO2 estimation model (SpvO2 = 111 – 40.5*R) and the empirical SpaO2 estimation model of (SpaO2 = 110 – 25*R). The main outcome of this study presents a novel pulse oximeter based calibration model that can be used to assess peripheral SvO2, and thus O2E, using the device developed in this work. Overall, this thesis successfully develops and demonstrates a non-invasive, pulse oximeter based method to assess peripheral tissue perfusion in terms of SaO2, SvO2, O2E and volumetric blood flow changes. This novel sensor may potentially detect peripheral perfusion alterations in real-time during microvascular and/or overall circulatory dysfunction, such as in sepsis or cardiac failure. Future work will follow the application of the novel sensor in a comprehensive clinical study with a larger, more diverse, cohort. | en |
| dc.identifier.uri | http://hdl.handle.net/10092/13287 | |
| dc.identifier.uri | http://dx.doi.org/10.26021/2402 | |
| dc.language | English | |
| dc.language.iso | en | |
| dc.publisher | University of Canterbury | en |
| dc.rights | All Right Reserved | en |
| dc.rights.uri | https://canterbury.libguides.com/rights/theses | en |
| dc.title | Non-invasive peripheral perfusion monitoring using photoplethysmography. | en |
| dc.type | Theses / Dissertations | |
| thesis.degree.discipline | Bioengineering | |
| thesis.degree.grantor | University of Canterbury | en |
| thesis.degree.level | Doctoral | en |
| thesis.degree.name | Doctor of Philosophy | en |
| uc.bibnumber | 2532796 | |
| uc.college | Faculty of Engineering | en |