Evaluation of cardiac output using nonuniform hybrid electrical impedance model based on forward lumped parameter and both-hands impedance measurement system Kwangseok Seo The Graduate School Yonsei University Department of Biomedical Engineering
Evaluation of cardiac output using nonuniform hybrid electrical impedance model based on forward lumped parameter and both-hands impedance measurement system A Dissertation Submitted to the Department of Biomedical Engineering and the Graduate School of Yonsei University in partial fulfillment of the requirements for the degree of Doctor of Philosophy Kwangseok Seo Feb 2012
This certifies that the dissertation of Kwangseok Seo is approved. Thesis Supervisor: Prof. Hyungro Yoon Thesis Committee Member: Prof. Kyoungjoung Lee Thesis Committee Member: Prof. Taemin Shin Thesis Committee Member: Prof. Sei-jin Chang Thesis Committee Member: Prof. Byung-su Yoo The Graduate School Yonsei University Feb 2012
Acknowledgements 저와함께계시고순간순간명철한지혜와능력을주신하나님께감사와영광을드립니다. 삶과학문에있어실천으로그깊이와넓음을몸소보여주시고, 늘자상한가르침으로오늘까지이끌어주신윤형로교수님께진심으로감사드립니다. 바쁘신가운데도부족한논문을심사기간내내따뜻하고세심하게지도해주신이경중교수님, 신태민교수님, 장세진교수님, 유병수교수님께머리숙여감사드립니다. 학부와대학원생활동안학업을통해많은가르침으로배움의기회를주신이윤선교수님, 김동윤교수님, 윤영로교수님, 김영호교수님, 김한성교수님, 정병조교수님께도깊이감사드립니다. 특별히학업의끈을놓지않고엔지니어로서의삶을그릴수있도록본이되어주신오건민회장님께진심으로감사드립니다. 학업과인생의대선배로서본이되어주신길문종회장님과김기원사장님께감사드립니다. 연구실에서함께하며성실함을보여주신염호준박사님, 임택균박사님, 홍수용선배님께감사드립니다. 그리고늘친동생처럼아끼고도와주신전대근박사님, 치열함과긍정으로삶에진지할수있도록이끌어주신김응석박사님, 성실함과겸손함을행동으로알려주신박성빈박사님께진심으로감사드립니다. 길지는않았지만연구실생활을통해작은부분까지보살펴주신김동석선배님, 이정우선배님, 강동원선배님, 이찬오선배님께감사드립니다. 특히논문을준비하는동안불평없이한결같이도와준명헌이와연식이, 민용이, 찬솔이, 주홍이에게도고마움을전합니다. 그리고돌아보면너무즐거웠던시간을함께한문재국선배님, 김해관선배님과진술이와홍일점유나에게도감사의맘을전합니다. 늘당당하고자랑스러운친구상돈이와나에게만은항상따뜻한현웅이, 부족한저를항상믿어주는재영이, 언제봐도멋진광재, 끈임없이노력하는정인철박사, 한결같은계형이에게진심으로고마움을전합니다. 그리고 MI FORUM의모든형제들 - i -
에게감사의말씀을드립니다. 지칠때마다따뜻한관심과사랑을주신류기홍박사님과형수님께특별히감사드립니다. 항상도움만받게되는성홍모박사님, 이전박사님그리고이승형선배님께도감사의맘을전합니다. 후배지만배울것이많은조성필박사, 김태균박사, 기수, 덕현, 균정에게도고마움을전합니다. 그리고모든대학원선 후배님들께진심으로감사드립니다. 함께살면서슬픔은나눠서반이되고즐거움은함께해서배가될수있었던영대, 수일, 재원, 희경, 정진, 대연, 호정이와모든 95동기들에게형제애로감사합니다. 가장가까운곳에서큰힘이되어주는한상훈선배님과김성환선배님, 이미동료로서훌쩍커버린기태, 지용, 준섭, 현륭, 동영및동고동락한메디게이트동료여러분께진심으로감사드립니다. 치열한사회에서만났지만형재같이따뜻하게오랜시간함께해주신현계환사장님, 기윤성사장님, 임도진사장님, 성용훈사장님과이상대사장님께도늘감사드립니다. 희망과꿈으로어린시절부터함께해온신혜, 민규, 혜정, 혜숙이와윤영, 문선, 치영, 호정과은진교회여러분께멀리서감사의마음을전합니다. 그리고늘기도와말씀으로도와주시는예원교회김진형목사님과예원교회식구들께도진심으로감사드립니다. 부족한사위를믿고지켜봐주시는장인어른과항상밝은미소와기도로힘이되어주시는장모님, 미국에서공부하시는처형가족특히이쁜주희, 큰처남부부와작은처남까지모두사랑하고감사드리며가까이서돌봐주시는이모부님과이모님에게도감사의마음을전합니다. 그리고항상물심양면으로아낌없이도와주시고격려의말씀을아끼지않으셨던고모님, 고모부님과힘들때마다힘이되어주신외가댁식구들에게도진심으로감사드립니다. - ii -
지금의제가있기까지묵묵히희생으로버팀목이되어준우리재호형, 언제나막내를자랑스러워하고챙겨주는우리수연이누나, 가장으로서남자가가야할길을알려주시고무한한사랑으로모든것을내어주신존경하는아버지, 기도와겸손으로일생을헌신하시는생각만하면눈시울이뜨거워지는우리어머니사랑합니다. 용의해에태어날아직은엄마뱃속에있는소중한튼똘이와작은것에도감사하고기뻐할줄아는내삶에소중함과열정, 그리고목표자체가되어준사랑하는아내하보라에게희망찬미래와함께감사와사랑한다는말을전합니다. 마지막으로넘치는사랑을주신모든분들께빚지는마음으로이작은결실을바칩니다. 2012 년 1 월 서광석드림 - iii -
Table of Contents List of Figures... vi List of Tables... ix Abstract... x Chapter 1 Introduction... 1 1.1 Purpose... 3 1.2 Research hypotheses... 3 1.3 Definition of terms... 4 Chapter 2 Literature Review... 5 2.1 Methods for determining stroke volume and cardiac output... 5 2.1.1 Physiological background... 5 2.1.2 Measurement of cardiac output... 15 2.1.3 limitation of existing method... 20 2.2 Impedance cardiograpy... 22 2.2.1 Measurement principle of impedance cardiography... 22 2.2.2 The method of stroke volume calculation... 26 2.3 Lumped parameter model... 33 2.3.1 Heart model... 33 2.3.2 Artery model... 35 Chapter 3 Methods... 37 3.1 Subjects... 37 3.2 Experimental procedure... 40 3.3 Development system for SV and CO determination... 43 3.3.1 System configuration... 43 3.3.2 Measurement of ICG using both hands... 48 - iv -
3.4 Mathematical analysis of non-uniform hybrid model based on forward lumped parameter... 51 3.5 Determination of SV and CO Algorithm by non-uniform hybrid model based on forward lumped parameter... 69 3.6 Developed system reproducibility... 74 3.7 Statistical analysis... 75 Chapter 4 Results... 76 4.1 SV and CO estimation by development System... 76 4.2 SV and CO predication by non-uniform hybrid electrical impedance model based on forward lumped parameter... 84 4.3 Evaluation of CO by non-uniform hybrid electrical impedance model based on forward lumped parameter and both hands impedance measurement system... 91 Chapter 5 Discussion... 98 Chapter 6 Conclusions... 101 References... 103 Abstract(in Korean)... 113 Appendix 1- IRB Approved... 115 Appendix 2- Clinical Trial Report... 117 Appendix 3- Case Record Form... 125 Appendix 4- Explanation... 128 Appendix 5- Consent... 131 - v -
List of Figures Figure 2.1. The anatomy of heart.... 6 Figure 2.2. The cardiac cycle with four different phases... 7 Figure 2.3. Relationship between ventricular diastolic-end volume of ventricle... 9 Figure 2.4. Effect of sympathetic nerve stimulation of heart on stroke quotient... 10 Figure 2.5. Effect of sympathetic nerves on contraction and relaxation of ventricle... 11 Figure 2.6. Main factors for deciding cardiac output... 12 Figure 2.7. Schematic view of the factors that play an important role in the regulation of the cardiac output.... 14 Figure 2.8. The Fick method to determine the cardiac output.... 16 Figure 2.9. Examples of two indicator dilution curves with recirculation at the end and the area filled with dot is the area under the extrapolated curve... 19 Figure 2.10. Electrode attachment method... 24 Figure 2.11. Effect of distance between electrodes and electrode size on current pass... 25 Figure 2.12. Representing the beating ventricle as a Windkessel with inflow and outflow valves and a time-varying compliance... 35 Figure 3.1. Flow of participants throughout the trial... 39 Figure 3.2. Experimental procedure... 41 Figure 3.3. Experimental environment... 42 Figure 3.4. System configuration... 43 Figure 3.5. Hardware block diagram... 44 Figure 3.6. The result of body impedance calibration... 46 Figure 3.7. Produced system and hand-grip electrode... 48 Figure 3.8. Results of vector plot through modeling... 50 Figure 3.9. Pressure-Volume diagram for either ventricle..... 51 Figure 3.10. A typical blood vessel.... 52 Figure 3.11. Flow of blood from heart to upper and lower parts through aorta... 55 Figure 3.12. Diagram of lossless and lossy transmission equation... 57 - vi -
Figure 3.13. Non-uniform hybrid systemic circulation model based on forward lumped parameter... 60 Figure 3.14. Non-uniform hybrid upper and lower model based on forward lumped parameter.... 61 Figure 3.15. Impedance equivalence model for model presented in this study... 65 Figure 3.16. Flowchart of the proposed algorithm for SV and CO determination... 69 Figure 3.17 Pulse raw data of thoracic and both-hands pulses (a) and Spectra arrived at by FFT analysis of thoracic and both-hands pulses (b)... 71 Figure 3.18. Time domain reconstruction of (a)thoracic and (b)both-hand spectra over firstfour peak frequency..... 72 Figure 3.19. Reconstructed flow waveform by non-uniform hybrid Model based on forward lumped parameter.... 73 Figure 3.20. Reproducibility experiment protocol... 74 Figure 4.1. Measured and estimated SV(ml) in (a) male and (b) female... 78 Figure 4.2. Measured and estimated CO(l) in (a) male and (b) female... 78 Figure 4.3. Scatter plot graphs of relationship between measured and estimated SV (ml) with 95% prediction interval line(orange) and 95% confidence interval line (brown) in (a) male and (b) female... 80 Figure 4.4. Bland-Altman plot with estimated mean bias and 95% limits of agreement for difference between measured and estimated SV(ml), plotted against the mean in (a) male and (b) female.... 81 Figure 4.5. Scatter plot graphs of relationship between measured and estimated CO (l) with 95% prediction interval line(orange) and 95% confidence interval line (brown) in (a) male and (b) female... 82 Figure 4.6. Bland-Altman plot with estimated mean bias and 95% limits of agreement for difference between measured and estimated CO(l), plotted against the mean in (a) male and (b) female... 83 Figure 4.7. Measured and modeled SV(ml) in (a) male and (b) female... 85 Figure 4.8. Measured and modeled CO(l) in (a) male and (b) female... 85 - vii -
Figure 4.9. Scatter plot graphs of relationship between measured and estimated SV (ml) with 95% prediction interval line(orange) and 95% confidence interval line (brown) in (a) male and (b) female... 87 Figure 4.10. Bland-Altman plot with estimated mean bias and 95% limits of agreement for difference between measured and estimated SV(ml), plotted against the mean in (a) male and (b) female... 88 Figure 4.11. Scatter plot graphs of relationship between measured and estimated CO (l) with 95% prediction interval line(orange) and 95% confidence interval line (brown) in (a) male and (b) female... 89 Figure 4.12. Bland-Altman plot with estimated mean bias and 95% limits of agreement for difference between measured and estimated CO(l), plotted against the mean in (a) male and (b) female... 90 Figure 4.13. Modeled and estimated SV(ml) in (a) male and (b) female... 92 Figure 4.14. Modeled and estimated CO(l) in (a) male and (b) female... 92 Figure 4.15. Scatter plot graphs of relationship between modeled and estimated SV (ml) with 95% prediction interval line(orange) and 95% confidence interval line (brown) in (a) male and (b) female... 94 Figure 4.16. Bland-Altman plot with estimated mean bias and 95% limits of agreement for difference between modeled and estimated SV(ml), plotted against the mean in (a) male and (b) female... 95 Figure 4.17. Scatter plot graphs of relationship between modeled and estimated CO (l) with 95% prediction interval line(orange) and 95% confidence interval line (brown) in (a) male and (b) female.... 96 Figure 4.18. Bland-Altman plot with estimated mean bias and 95% limits of agreement for difference between modeled and estimated CO(l), plotted against the mean in (a) male and (b) female.... 97 - viii -
List of Tables Table 3.1. Baseline Characteristics of Included Participants... 38 Table 3.2. Properties of each organ (at 100kHz)... 49 Table 3.3. Reproducibility test... 74 Table 4.1. Description of physioflow SV, CO and developed system SV (ml), CO(l)... 77 Table 4.2. Paired Samples Test and Wilcoxon Signed Ranks Test results between physioflow and developed system SV, CO in males (N=54)... 79 Table 4.3. Paired Samples Test and Wilcoxon Signed Ranks results between measured and estimated SV, CO in females (N=18)... 79 Table 4.4. Description of physioflow SV, CO and modeled SV (ml), CO(l)... 84 Table 4.5. Paired Samples Test and Wilcoxon Signed Ranks Test results between physioflow and modeled SV, CO in males (N=54)... 86 Table 4.6. Paired Samples Test and Wilcoxon Signed Ranks Test results between physioflow and modeled SV, CO in females (N=18)... 86 Table 4.7. Description of modeled SV, CO and developed system SV (ml), CO(l)... 91 Table 4.8. Paired Samples Test and Wilcoxon Signed Ranks Test results between modeled and developed system SV, CO in males (N=54)... 93 Table 4.9. Paired Samples Test and Wilcoxon Signed Ranks Test results between modeled and developed system SV, CO in females (N=18)... 93 - ix -
Abstract Evaluation of Cardiac output by non-uniform hybrid electrical impedance model based on forward lumped parameter and both hands impedance measurement system Kwangseok Seo Dept. of Biomedical Engineering The Graduate School Yonsei University In this dissertation, cardiac output using non-uniform hybrid electrical impedance model, which is based on the forward lumped parameter and the both-hands impedance measurement system, is proposed. This noninvasive method for cardiac output monitoring has been clinically accepted as a replacement for thermo dilution, the gold standard in cardiac output measurement. Alternatively, measurement using impedance cardiogram, which has several distinct advantages, has been identified as a promising method for cardiac output measurements. The thoracic impedance cardiogram (ICG) has been proposed as a noninvasive, continuous, operator-independent, and cost-effective method for cardiac output monitoring. However, this method is generally regarded to be restrictive because measurements are performed using a band or spot-type electrode adhered to the body. Traditionally, lead has been used for such measurements, thus rendering the entire system highly inconvenient because the assistance of - x -
a specialist is required. Further, the development and attachment of the lead electrode, used with the traditional system, is both expensive and complicated. In this dissertation, we evaluate the effectiveness of the proposed non-uniform hybrid model, which is based on the forward lumped parameter. This system seeks to combine the existing lumped parameter method and the non-uniform hybrid model to create a coherent system capable of leveraging the advantages of both approaches. For developing an effectiveness rating for cardiac output measurements using both hands, the presented model was mathematically interpreted and the relevant results were compared and analyzed against the stroke volume and the cardiac output of the thoracic impedance measurements (Physio FlowR-PF104D, Manatec Biomedical, France). To develop the non-uniform hybrid electrical impedance model, based on the forward lumped parameter and the both-hands impedance measurement system, 80 subjects (58 male, 22 female) from Yonsei University and the surrounding areas, aged 18 74 years, participated in this study. All participating subjects completed stroke volume and cardiac output tests through PhysioFlow and the developed system. In the developed system, electrodes are used to gripping on both-hands instead of attaching to the chest. Similar to previously adopted noninvasive cardiac output tests, the developed system measures stroke volume through impedance changes over each cardiac cycle. Additionally, this study compares cardiac output measurements in the thorax and in both hands. These measurements and comparisons were verified using the presented non-uniform hybrid model. To verify the proposed approach, statistical methods such as correlation analyses, paired T-test, and the Bland-Altman plot were used. For verification of the non-uniform hybrid electrical impedance model, the presented value of r, scatter plot, and the Bland-Altman plot of measured and estimated SV and CO were used. - xi -
The results were as follows: 1) The SV/CO obtained from the PhysioFlow and the proposed approach (developed system) showed significant correlation in both male and female SV (r = 0.715, P < 0.001; r = 0.704, P < 0.001, respectively) and CO (r = 0.826, P < 0.001; r = 0.804, P < 0.001, respectively). 2) The SV/CO obtained from the PhysioFlow and the proposed approach (non-uniform hybrid electrical impedance model based on the forward lumped parameter) demonstrated significant correlation in both male and female SV (r = 0.735, P < 0.001; r = 0.827, P < 0.001, respectively) and CO (r = 0.767, P < 0.001; r = 0.853, P < 0.001, respectively). 3) The SV/CO obtained from the non-uniform hybrid electrical impedance model and the development system showed significant correlation in both male and female SV (r = 0.788, P < 0.001; r = 0.812, P < 0.001, respectively) and CO (r = 0.802, P < 0.001; r = 0.823, P < 0.001, respectively). From these results, it can be concluded that SV and CO can be measured using the bothhands cardiac output measurement method at low cost and convenient without the help of a specialist. Furthermore, this system was verified by using the developed model as a substitute for the existing method. Key words: Cardiac output (CO), Stroke volume (SV), Impedance cardiogram (ICG), Hand grip electrode, PhysioFlow(PF104D) - xii -
Chapter 1 Introduction For many patients in the intensive care unit, emergency medicine unit, or those being investigated for some cardiovascular complaint, simply measuring heart rate (HR) and blood pressure does not provide adequate data on their hemodynamic state. Together with an electro cardiogram (ECG) and blood pressure (BP) monitoring, the measurement of cardiac output (CO) and stroke volume (SV) can play a major role in the diagnosis and therapy of chronic cardiac conditions such as heart failure, hypertension, coronary artery disease, pericardial disease, obstructive lung, pleural disease, and renal disease/dialysis [1 3]. CO determination is an important procedure in interventional cardiology and has also been used in cardiothoracic surgery [4]. CO and SV are the functional expressions of cardiovascular performance and can be used to confirm the need for, or efficacy of, various treatment options. CO and SV are reliable indicators of cardiac performance, the measurement of which is essential for the monitoring and assessment of cardiac disorders or other conditions of hemodynamic compromise. These two parameters essentially define the average blood flow in the entire cardiovascular system. In light of the intended use of cardiac output as a medical diagnostic tool, the associated measurement technique must be clinically acceptable and reliable. In clinical practice, several, invasive CO- and SV-estimation methods are available, such as Fick s method, dye-dilution, and thermo-dilution. Currently, thermo-dilution is the most commonly used method for measuring cardiac output. It is still the gold standard in CO and SV estimation. These require catheterization of the patient, which itself adds to the morbidity and sometimes mortality of the patient. When catheterization is performed, the thermodilution method may be utilized to monitor cardiac output and make appropriate decisions regarding flow modulation through drug therapy. Outside intensive care units, however, - 1 -
catheters are usually removed, and the thermo-dilution method is abandoned. It gives only intermittent measurement of cardiac output of the patient. Consequently, little information is available regarding cardiac output after the critical stages of recovery. To date, noninvasive methods have been clinically accepted as replacements for the thermo-dilution method, which was traditionally the gold standard for cardiac output measurement [5, 6]. Alternatively, the impedance cardiogram method, with its numerous advantages, is a promising method for measuring cardiac output [7-9]. A thoracic impedance cardiogram (ICG) approach has been proposed as a noninvasive, continuous, operatorindependent, and cost-effective method for cardiac output monitoring. Currently existing impedance based cardiac output monitors operate by emitting a low voltage (2.5 to 4 ma), high-frequency (50 to 100 khz), and alternating electric current through the thorax via spot or band electrodes. The electrical impedance changes, caused by changes in the volume and velocity of blood flow in the thoracic aorta (within the thorax), are detected via the electrodes as pulsatile decreases in impedance (dz). The impedance can also be further expressed as its derivative (dz/dt). This derivative has been shown to be proportional to the stroke volume. Along with heart rate, stroke volume can indicate the CO of the patient. Charloux et al. researched the effects of thoracic hyperinflation through ICG and Koobi et al. used full-body ICG to compare thermo dilution and the direct Fick s method in order to show that ICG reliably and reproducibly estimates CO in sedated preoperative patients without marked valvular disease [10]. However, this method is considered restrictive, because it requires the application of band or spot-type electrodes on the body. As previously mentioned, the inclusion of lead in the attachment process materializes into compounded expense and inconvenience because specialists must be involved. - 2 -
1.1 Purpose This study presents a system for detecting clinical indicators of Stroke Volume (SV) and Cardiac Output (CO), which reflect hemodynamic function of cardiovascular activity through electric impedance measurement method using both hands, rather than measurement of Cardiac Output using previously accepted invasive methods, ultrasound, and thorax-related measurements. The purposes of this study were to (1) develop a system for detecting SV and CO through the noninvasive and convenient both-hand ICG measurement method, (2) develop the nonuniform hybrid model based on the forward lumped parameter and present the advantages of the previous lumped parameter method and non-uniform hybrid model to evaluate the effectiveness of a combined system, (3) conduct mathematical interpretation of the presented model and compare the SV/CO results for the thoracic impedance to verify the effectiveness of the both-hand CO measurement presented in this study. 1.2 Research hypotheses It was hypothesized that: 1) During cardiac systole, blood is released through the main artery and the electrical impedance method can detect the discharge of blood. 2) The increasing number and decreasing size of blood vessels trends from the central to the peripheral blood vessel system. It shows the biggest change in discharge near the center of the body. If the injected current, passing through both hands, contains discharge from the left ventricle, it can draw the same result as the original cardiac output measurement. - 3 -
3) According to the blood dynamic methods, through the use of the direct parameter model of Systemic Simulation, we can effectively estimate the indirect cardiac output and thus compile a proposed system model. 1.3 Definition of terms 1) Stroke Volume (SV) : Amount of blood pumped by the left ventricle each heart beat[ml] 2) Cardiac Output (CO): Amount of blood pumped by the left ventricle each minute[l/min] CardiacOutput Heart Rate X Stroke Volume 3) Impedance Cardiogram (ICG): Harmless current is exerted to measure impedance change in artery according to the amount of blood during cardiac impulse in voltage form. This method achieves noninvasive acquisition of data according to cardiac impulse and monitors dynamic function of heart, such as cardiac output, Stroke Volume, and contractile power of heart muscle. 4) Z 0 (Body impedance): Basic impedance of the body segment limited by receiving electrodes [Ω] [Ω] 5) Z(ICG): Changes of the impedance of the segment limited by receiving electrodes 6) dz/dt(1'st derivation of ICG): The maximum of the first derivative of the impedance signal (Ω/sec) 7) Bland-Altman plot: A statistical method of data plotting used in analyzing the agreement between two different assays. - 4 -
Chapter 2 Literature Review This chapter consists of three parts: 2.1. Methods for determining stroke volume and cardiac output describes the physiological background of the cardiovascular system and explains the existing method for measuring SV and CO and relevant limitations. 2.2 Methods for determining impedance cardiogram, uses an impedance cardiogram to explain principles for determining CO and SV, and introduces various calculation methods. 2.3 Lumped Parameter Model explains the approach method and trend of previous model-based studies. 2.1 Methods for determining stroke volume and cardiac output 2.1.1 Physiological background 1) The heart The heart is enclosed in a double-walled sac, which is called the pericardium. The pericardium protects the heart, anchors it to the surrounding structure and prevents over filling of the heart with blood. The heart wall consists of three layers. The outer wall is called the epicedium, the middle layer the myocardium and the inner layer, the endocardium. The pericardium is often infiltrated with fat, especially in older people. The myocardium is the thickest of the three layers and is the layer that actually contracts. The endocardium consists of a white sheet of endothelium resting on a connective tissue layer and covers the connective tissue skeleton of the valves. The heart functions as a pump and can be divided in four chambers, the left atrium, the left ventricle, right atrium and right ventricle, see Figure 2.1. - 5 -
The left atrium and the left ventricle are separated by the mitral valve. From the left ventricle the blood flows into the circulatory system through the aortic valve into the aorta. The blood enters the right atrium through three veins. The superior vena cava returns blood from above the diaphragm, second the inferior vena cava returns blood from below the diaphragm and the last returning blood flow is from the coronary veins of the heart. The right atrium and right ventricle are separated by the tricuspid valve. The blood enters the pulmonary artery through the pulmonary valve. Then the blood flows to the right lung and left lung through the right pulmonary artery and left pulmonary artery. Figure 2.1. The anatomy of heart. 2) The cardiac cycle The cardiac cycle can be divided in four phases: the diastolic phase, the isovolumic contraction phase, the ejection phase and the isovolumic relaxation phase, see Figure 2.2. In the first part of the diastolic phase when the mitral valve is opened the ventricle is filled with - 6 -
blood. In the last part of the diastolic phase an action potential is generated by the sinus node, located in the right atrium and causes an additional filling of the left atrium. The action potential will travel rapidly through both atria and through the A-V bundle and the conducting system and causes the initiation of the contraction of the ventricles. The ventricular pressure will increase and causes the mitral valve closure and marks the beginning of the isovolumic contraction phase. In this phase the ventricular volume remains constant, but the ventricular pressure increases. When the ventricular pressure rises above the pressure in the aorta, the aortic valve will open and the ejection phase begins. During the ejection phase, the aortic and ventricular pressure increase to its maximum and then decreases, at the point where the ventricular pressure is less than the aortic pressure, a slightly aortic back flow occurs which results in the closure of the aortic valve. This marks the beginning of the isovolumic relaxation phase. In this phase the ventricular volume remains constant and the pressure of the ventricle will decrease. As soon as the ventricular pressure drops below the atrial pressure, the mitral valve will open and the cardiac cycle begins again. Figure 2.2. The cardiac cycle with four different phases, diastolic, isovolumic contraction, ejection, and isovolumic relaxation and the time course of left ventricular pressure (plv), aortic pressure (pao), left atrial pressure (pla) and left ventricular volume (Vlv). Typical volumes of left ventricle at two time points are the end-diastolic volume EDV and end-systolic volume ESV. - 7 -
3) The cardiovascular system The function of the circulation, which consist of the heart and the blood vessels, is to supply the tissues in the body with oxygen and nutrients and to transport waste products away. The regulation of the circulation to satisfy the oxygen demands through the body is controlled by the autonomic nervous system. The autonomic nerves system can be divided in the sympathetic nervous system and the parasympathetic nervous system. The sympathetic nervous system is activated in stressful, emotional situations or by physical activity and the parasympathetic nervous system is more active in rest, and for example stimulates the organs for digesting food. When sympathetic stimulation excites the blood flow to a particular organ, often parasympathetic stimulation inhibits it. That is, the two systems occasionally act reciprocally to each other. However, the blood flow to most organs is mainly controlled by one of the two systems. For the control of blood flow, the effect of the two systems on arterial pressure, on the blood vessels, and on the heart are most significant. Blood pressure is regulated by means of the baroreceptor reflex. The baroreceptor reflex is the most powerful tool in the control of systemic arterial pressure. Baroreceptors are lying in the walls of the carotid sinus and the aortic arch. As soon as the blood pressure falls and the baroreceptors are less stimulated, the sympathetic nervous system is activated and the parasympathetic activity is decreased. As a consequence, heart rate and cardiac contractility increase and the small arteries and large arterioles are constricted. This way the arterial blood pressure is regulated towards steady state again. Most blood vessels are constricted by sympathetic stimulation. Sympathetic constriction of the small arteries and the large arterioles increases the resistance and therefore reduces the blood flow through the vessels. Sympathetic stimulation of the veins decreases the volume of these vessels and therefore translocates the blood into the heart. Parasympathetic stimulation has little or no effect on blood vessels. It merely dilates vessels in certain restricted areas, such as in the blush area of the face. The heart is controlled by both systems. Sympathetic stimulation increases the heart rate and enhances cardiac contractility. Parasympathetic stimulation causes mainly the opposite effects, it decreases the heart rate and also slightly decreases contractility. In short, sympathetic activity increases the effectiveness - 8 -
of the heart as a pump whereas parasympathetic stimulation decreases the pumping capability of the heart. 4) End diastolic volume (Frank-Starling s law) End-diastolic volume (EDV) is the intraventricular blood volume directly before contraction. This is also called pre-load as it is the work quantity imposed on the ventricles of heart before contraction. The stroke volume per session is in direct proportion with the preload. Heart ventricles perform stronger contraction during the contraction period if more blood is filled during diastole. Thus, the stroke quotient increases with the end-diastolic capacity under identical conditions. This relation is presented on the ventricular function curve (see Figure 2.3). The relation between stroke quotient and end-diastolic capacity is called the Frank-Starling mechanism. Under stable conditions, the length of the heart muscle is not optimal for contraction like skeletal muscle but is positioned in the up-phase of the curve. Thus, the length of the myocardial fiber is increased with the increase in blood within ventricles along with an increased contractile force. Figure 2.3. Relationship between ventricular diastolic-end volume of ventricle (Frank- Starling s mechanism) - 9 -
In Frank-Starling s mechanism, end-diastolic capacity increases with higher venous return, thus the quantity of blood returning to the heart through veins, during an identical pulse rate. Furthermore, cardiac output is automatically increased with increased pulse rate. 5) Contractility of the myocardium The sympathetic nerve is diffused throughout the entire heart muscle, and the norepinephrine secreted from sympathetic nerves combines with the beta-adrenaline operational acceptor to increase ventricle contractility. Ventricle contractibility refers to the contraction in stable end-diastolic capacity. Plasma epinephrine also combines with the same acceptor to increase ventricle contractility. Figure 2.4. Effect of sympathetic nerve stimulation of heart on stroke quotient As shown in Figure 2.4, ventricle contractility is increased in the Frank-Starling mechanism under identical end-diastolic capacity, and thus identical length of heart muscle, - 10 -
due to the stimulation of the sympathetic nerves. In other words, an increase in contractibility leads to greater emission of blood within the end-diastolic ventricle. Increased stimulation of sympathetic nerves for a ventricle not only increases contractibility but also expedites contraction and relaxation of the ventricle to increase pulse rate (see Figure 2.5). Increased pulse rate reduces the ventricular filling time at diastole. However, this problem is partially compensated as a large part of the cardiac cycle can be used to fill up the ventricle with accelerated contraction and relaxation speed by sympathetic nerves. Figure 2.5. Effect of sympathetic nerves on contraction and relaxation of ventricle 6) Afterload Afterload is the resistance force generated in the ventricular wall during blood eruption in the left ventricle. Increased afterload reduces stroke volume. This is because arterial blood pressure acts as load for contracted ventricular muscles like skeletal muscle. Increased load in ventricular muscles reduce contraction of heart muscle. Thus, increase in arterial blood pressure reduces stroke quotient. Figure 2.6 presents a diagram of the effect on CO by main factors that decide stroke quotient and heart rate. - 11 -
Figure 2.6. Main factors for deciding cardiac output 7) Cardiac output Cardiac output is the amount of blood that is pumped by the heart into the aorta each minute. It equals the product of heart rate and stroke volume. With a heart rate at rest of 70 beats/min and a stroke volume of 70ml the heart pumps about 4:9L=min and this amount can increase to about four to seven times during heavy exercise. Stroke volume (SV) represents the difference between volume of blood in the ventricle at the end of the diastolic phase, the end-diastolic volume (EDV) and the volume of blood that remains in the ventricle after its contraction, the end-systolic volume (ESV): SV[ ml] EDV[ ml] ESV[ ml] (2.1) - 12 -
The pumping ability of the heart depends on contractility, preload, afterload and heart rate. The most important factors that affect the SV by causing changes in EDV or ESV are the contractility, the force of contraction of the cardiac muscle cells, preload, which is the degree of stretch of the cardiac muscle cells before contraction and the afterload, the pressure that must be overcome for the ventricles to eject blood from the heart. Afterload influences stroke volume by affecting the velocity of contraction. The intact heart can increase its contractility with the help of the Frank-Starling mechanism. The Frank-Starling mechanism means the intrinsic ability of the heart to adapt to changing loads of inflowing blood. The heart pumps all the blood that comes to it into the aorta without allowing excessive damming of blood in the veins. If the amount of blood returning to the heart is increased (larger EDV), causing the preload to increase, the cardiac muscle is stretched more and, in turn, contracts with increased force. The increased force of contraction is probably caused by the fact that the contractile proteins become more sensitive for calcium when they are stretched. Also the contractility can be increased by extrinsic control by the sympathetic nervous system activity. The contractility increase caused by the activation of the sympathetic nervous system is independent of the stretch of the cardiac muscle fiber and the EDV. The sympathetic stimulation is responsible for the increases in heart rate. In contrast the parasympathetic nervous system reduces the heart rate. The factors that are involved in the regulation of the cardiac output are shown in Figure 2.7. - 13 -
Figure 2.7. Schematic view of the factors that play an important role in the regulation of the cardiac output - 14 -
2.1.2 Measurement of cardiac output 1) Fick method The cardiac output can be calculated by using the Fick principle. The Fick principle is based on the uptake of oxygen by blood as it flows through the lungs. It is assumed that all oxygen molecules in the pulmonary vein has its origin from the blood in the pulmonary artery or from the oxygen transported from the lung to the blood. The oxygen that enters the blood is reflected by the difference in oxygen content between the pulmonary vein and the pulmonary artery, assuming that no oxygen is consumed by the tissues between the pulmonary artery and vein. Because the entire output of the right heart passes through the lungs, assuming that there are no shunts across the pulmonary system, the blood flow through the lungs is equivalent to cardiac output. This can be written as [11]: F O2 CO CO C C ao2 ao2 F O2 C CO C vo2 vo2 (2.2) with F O2 the oxygen flow from lung to blood in mlo 2 / min, CO the cardiac output in L / min and C vo2 and C ao2 the oxygen contents of vein and artery in mlo 2 / L. In Figure 2.8 an example of the Fick method is shown. Because it is impossible to measure the rate at which oxygen is taken up by the capillaries, the rate of oxygen uptake must be measured at the mouth. Then F O2 is often expressed as V O 2. The error introduced by measuring oxygen at the mouth is unimportant if the period of measurement is much longer than the time of a single breath. [12] The VO 2 can be measured by making use of spirometry. The values of C vo2 and C ao2 can be determined indirectly by taking blood samples at the appropriate location. Because the concentrations fluctuate due to the pulsations caused by respiration and circulation the values has to be averaged for a sufficient time. The arterial oxygen content of - 15 -
blood can be sampled at any convenient location in the large arteries, since oxygen transfer to the tissues only take place in the capillaries. Figure 2.8. The Fick method to determine the cardiac output In contrast, the venous oxygen content in blood can very significant between measuring sites, because it depends on how much oxygen the organs have extracted. Only where the different oxygen contents from the veins merge, like the right atrium, right ventricle or pulmonary artery, the so called mixed venous oxygen content, can be measured. The oxygen in the blood is primarily bounded to hemoglobin as oxyhemoglobin. The O 2 carrying capacity of hemoglobin Hb is 1.34 (ml O 2 / gram Hb). The oxygen content of the blood can then be calculated by: C C ao2 vo2 Hb1.34 S Hb1.34 S ao2 vo2 10 10 2 2 0.031P 0.031P ao2 vo2 (2.3) - 16 -
with Hb the concentration of hemoglobin in blood in gram / L. S ao2 and S vo2 are the oxygen saturation of blood in %. The oxygen saturation at the arterial site, S ao2 can also be measured continuously with a pulse oximeter. The mixed venous oxygen saturation, S vo2, can be measured continuously with a pulmonary artery catheter. P ao2 and P vo2 represents the partial pressure of oxygen at the arterial and venous site in mmhg and 0.031 is the solubility coefficient of oxygen in blood in ml O 2 / L mmhg. For the continuously measurement of the cardiac output the partial oxygen pressures are often ignored, because it has only a small contribution. 2) Indicator dilution methods The indicator dilution method is very similar to the Fick method, but instead of the measurement of oxygen, the concentration of an indicator is measured. In indicator dilution methods a bolus of an indicator is brought into the blood stream and the concentration of the indicator is measured downstream. There are many indicators like chemicals, inert gases, radioactive isotopes, dyes and heat. [11] The indicator dilution method is based on the following theory. If the concentration of a small known bolus that is uniformly dispersed in an unknown volume V is determined, and the volume of the injected indicator is known, then the unknown volume can be determined too [13]: dv( t) ( t) dt (2.4) where (t) and (t) V are the instantaneous flow and volume of the carrier. Then the next expression counts too: c( t) dm dv (2.5) - 17 -
with m and c (t) the mass of the tracer and its concentration at time t. Then the following equation can be derived: dv ( t) ( t) dt 1 c( t) dm dt (2.6) Because only the averaged flow determines how much indicator is transported and not the fluctuations of the flow the previous equation can be rewritten into [19, 22]: m 0 CO ( t) c( t) dt CO 0 m c( t) dt 0 c( t) dt (2.7) The instantaneous flow (t) can be expressed as the cardiac output CO and put out of the integral. Before the concentration decreases to zero, some of the indicator has already circulated and passes the measurement site for the second time, this is also called recirculation, see Figure 2.9. Because of this phenomenon an extrapolation is necessary for the calculation of the area of the curve without circulation. The basic assumption of indicator dilution techniques for the injection of a bolus are that the blood flow is constant during the measurement, there is no loss of the indicator, and the mixing of the indicator is uniform and the rapid injection can be modeled as an impulse. - 18 -
Figure 2.9. Examples of two indicator dilution curves with recirculation at the end and the area filled with dot is the area under the extrapolated curve. 3) Doppler ultrasound Measurement of stroke volume using trans-esophageal echocardiography can be achieved by measuring blood quantity flowing through the left ventricular outflow tract, aorta, or pulmonary artery. In this case, blood flow is laminar flow rather than turbulent flow, the conduit in which blood flow passes through continuously maintains a circle form and the area is assumed to be π (radius) 2. A continuous wave Doppler or pulsed wave Doppler is used to acquire the following expression: CO VTI CSA cos HR VTI : Velocity Time Interval CSA : Cross Sectional Area (2.8) - 19 -
4) Impedance cardiogram Capacity change in the thorax triggers changes in the thoracic electrical bio-impedance. If changes in thoracic resistance are measured according to the ventricular depolarization, stroke volume can be continuously measured. The alternating current resistance cardiac impulse method is a noninvasive method in which four sets of ECG electrodes are attached to the thorax to emit sample micro-currents and detect bio-current resistance on both sides of the thorax. To measure bio-current resistance of the thorax, low-voltage, high-frequency pulses of alternating currents are applied and detection is simultaneously performed in two sets of electrodes located near the neck and xiphoid process. Owing to recent technical and program advances, high correlation was presented between the two methods in several studies conducted with regard to the use of the thermo-dilution method on healthy adults, critical patients, and surgical patients (excluding heart surgery patients) [14, 15]. 2.1.3 Limitation of existing method 1) Thermal indicator dilution method By using the Swan-Ganz catheter to measure cardiac output, the thermo-dilution method can simultaneously measure systematic circulation resistance, pulmonary arterial pressure, and pulmonary artery wedge pressure to be useful for diagnosing hemodynamically unstable patients. However, measurement is inaccurate for cases with low cardiac output. 2) Dye dilution method Using indocyanine green as the indicator of cardiac output, dye-dilution was the most commonly used method before the thermo-dilution method. This method is disadvantageous in that it presents difficulties in calculating the area under the primary recurring curve and in that, arterial blood must be collected. - 20 -
3) Lithium dilution method This method calculates cardiac output by measuring the concentration of lithium through the detection device connected to the peripheral artery catheter after injecting lithium chloride through the central venous catheter or peripheral vein catheter. This method can achieve quick, easy measurement of cardiac output by using the previously inserted central venous catheter and arterial catheter without inserting the pulmonary artery catheter. Lithium is excreted through urine without metabolism within the body and is not combined with plasma or protein. Although clinical utility has been verified among young critical patients and in the intensive care unit after coronary artery surgery, data on the clinical utility in a surgical environment remains insufficient as sudden hemodynamic changes are presented with the frequent collection of arterial blood. 4) Doppler ultrasound method The accuracy of the cardiac output calculated using this method depends on the accuracy of the blood vessel diameter measurement and the parallelism of the ultrasound beam and the direction of blood flow. Although this method presents high accuracy and clinical credibility, it is disadvantageous in that it is dependent on the operator and has a long measurement time. 5) Impedance cardiogram method This method is disadvantageous in that it is easily affected by electric interference, the attachment status of the ECG electrode, and can exhibit low resistance to alternating current in relatively heavy patients. In addition, the accuracy of this method is low for patients that have received heart surgery, and in patients with pulmonary edema or aortic valve disease [16]. - 21 -
2.2 Impedance Cardiogram 2.2.1 Measurement principle of impedance cardiography 1) The impedance signal The thoracic impedance consists of three components. The largest component, the baseline impedance Z 0, is the electrical impedance of the total thoracic mass, which include the different tissues, fluid and air. The second component corresponds with the changes due to respiration, Z r (t). The third component is related to the changes caused by the cardiac cycle, Z c (t). This gives the next equation [17]: Z(t) = Z 0 + Z r (t)z c (t) (2.9) The values of Z 0 is about 25Ω for healthy men. The changes of the impedance signal induced by respiration is about 1Ω. The third and smallest variation due to the cardiac cycle in the impedance signal is approximately 0.1Ω to 0.2Ω. The contribution to the changes in the thoracic impedance signal, especially the cardiac related changes Z c, has its origin for about 61% from the lungs, 23% from the large arteries and about 13% from the skeletal muscles. [18] The amplitude caused by the respiratory component is much larger than the amplitude of the cardiac component, but the frequency of the cardiac component is higher than at the respiratory component. Therefore, in the first derivative of Z, the thoracic impedance change from the respiratory and cardiac component together, strongly reffects the signal of the cardiac component.[19] 2) Frequency and current value Impedance (Z; ohm) is composed of a complex amount because it generates phase-shift in biomaterial Impedance (Z: ohm) is a complex quantity because it generates phase shift in the biomaterial and in the time domain. The skin is composed of small inosculations of cells and - 22 -
small cell membranes. As the skin includes a capacitor component, lower impedance is presented with higher frequency. The result is that the skin can be regarded as an insulator due to such a capacitor substance. Because skin impedance is extremely high at low frequencies (<1 khz), the electrical impedance can be ignored and the results obtained in the deep skin layer can be determined using high frequencies (>100 khz) [20]. There is general agreement in the medical community that frequencies between 20 100 khz should be used and that the amplitude of the sinusoidal current curve, ought to rest somewhere between 1 and 5 ma [21]. However, some designers use a lower-than-suggested value of current. The lower boundary application is the suggestion leading to obtain the sufficient signal-to-noise ratio. A 1-mA current can create muscle excitation below 20 khz frequency and the skin-electrode impedance at 100 khz is approximately 100 times lower than that at low frequencies. This helps to diminish the unwanted impact of the changes in the skinelectrode impedance, occurring during motion, into measurable cardiac signals. However, applied currents of frequency greater than the 100 khz threshold result in stray capacitance. 3) Electrode types and Topography There are two main methods for measuring bioimpedance: bipolar and tetrapolar. In the bipolar method there are two electrodes that play a critical role in application and receiving. The current density in the regions near the electrodes is higher than in other parts of the tissue, thus influencing the overall impedance measurement in a non-uniform manner. The total impedance signal is a superposition of two components: the skin-electrode impedance (modified by blood flow-induced movement) and the original signal (caused by blood flow). In practice, it is difficult to separate the two variables. The scheme of the bipolar impedance measurement is presented in Figure 2.10(a). In a four-electrode (tetrapolar) configuration, the application electrodes and receiving are separated. Figure 2.10(b) presents the scheme of the tetrapolar impedance measurement and the typical means of obtaining the impedance cardiography signal. The constant amplitude current oscillates between the application electrodes; from this the voltage changes are detected on the receiving electrodes. Because the - 23 -
amplitude of the current constant, this voltage is proportional to the impedance of the tissue segment limited by the band electrodes. The voltage changes are proportional to the impedance changes between the receiving electrodes. The main advantage of this method over the bipolar method is that the current density distribution is more uniform. Another advantage is that the disturbances caused by electrode impedances are minimized. (a) Bipolar (b) Tetra polar Figure 2.10. Electrode attachment method 2-electrode methods are affected by the contact impedance of the electrodes and by the impedance frequency dependence observed during the impedance measurement. However, the 4-electrode method is a method for indirectly measuring impedance to remove electrode impedance and contact resistance. As presented in Figure 2.11, current enters deeply into the skin with greater distance between current-injected electrodes. Thus, the arrangement or distance of the electrodes affects the measurement of stroke volume and cardiac output. - 24 -
Furthermore, the electrode form can also affect cardiac output. In the study conducted by Jennifer J. Mcgrath, band-type and spot-type electrodes were used to measure ICG to present approximately 2x differences between the results of the targeted cardiac output. When compared with the use of band type electrodes, the use of spot electrodes can increase the signal quality, because it minimizes the motion artifact. Spot electrodes have approximately 45% higher signal-to-noise ratio than band electrodes, and present a low Z 0 value according to the degree of skin contact. These factors generate difference in results during the use of different types of electrodes [22]. Figure 2.11. Effect of distance between electrodes and electrode size on current pass - 25 -