Pulse Pressure Analysis

Transcription

1 Pulse Pressure Analysis M. Cecconi, J. Wilson, and A. Rhodes z Introduction Cardiac output monitoring is part of routine practice in the critically ill patient. Recently, there has been increasing interest in continuous cardiac output monitoring, which has seen the development of new devices less invasive than the pulmonary artery catheter (PAC). The insertion of a PAC allows semi-continuous monitoring of cardiac output using the thermodilution technique but these new devices allow continuous monitoring by analyzing the arterial pressure wave. This analysis is known as pulse pressure analysis. This chapter explores the issues associated with pulse pressure analysis and presents the mathematical basis for the devices available. z From Arterial Pressure to Cardiac Output Arterial pressure is one of the most commonly monitored variables in critical care medicine; however, arterial pressure is not in itself enough to assess cardiac output. In fact, arterial pressure is the result of a combination of cardiac output and the resistance of the vasculature: P ˆ CO R Where P =arterial pressure, CO = cardiac output, and R=resistance of the vasculature. It is apparent that the resistance of the vasculature must be known to calculate arterial pressure. Unfortunately, resistance is not constant, but it is possible to relate variations in pressure to variations in stroke volume due to a feature known as compliance. Compliance is the relationship between volume change and pressure change: C ˆ DV=DP Where C=compliance, DV=volume change, and DP=pressure change. Therefore, if we know the value of compliance, we can measure the pressure change (DP) to calculate the volume change (DV). DV ˆ C DP Compliance in the arterial tree is not a linear function (Fig. 1). Moreover, compliance is not constant for one blood vessel. For instance, vasoconstriction causes arteries to become stiffer and this leads to a decrease in compliance. Vasodilation leads to an increase in compliance (Fig. 2).

2 Pulse Pressure Analysis 177 Fig. 1. DV ab and DV cd have the same value. DP ab is lower than DP cd. C ab ist then higher than C cd. DV = change in volume, DP = change in pressure. C = compliance Fig. 2. Compliance curve can shift to different states and the same change in volume can determine very different changes in pressure. C3, C1, C2 = compliance respectively for curve 3, 1, 2 The proximal aorta carries the blood initially pumped from the heart. The aorta is filled according to three forces: z The force of injection of the blood by the pumping of the heart z An opposing force dependent on pulsatile inflow (impedance) z An opposing force depending on the change in volume (compliance). The peripheral resistance is the force that opposes the blood flow to the periphery. Ideally, these forces should be measured at a site as proximal as possible to the aorta. However, as pulse pressure devices have a low level of invasiveness, they will, by their very nature, record the arterial wave peripherally. Arterial pressure measured in the periphery is affected by at least three effects: z The generation of reflection waves in the periphery z Damping of the arterial pressure measurement system z Differences in the flow-pressure relationship centrally and peripherally. Damping is a common problem in clinical practice. The fluid filled tubes used to measure intravascular pressure form a resonant system that can oscillate. The performance of such a resonant system is determined by the frequency of oscillation and by the damping coefficient.

3 178 M. Cecconi et al. If a signal being analyzed has a similar frequency to the resonant frequency of the measuring system, the two oscillations can combine, leading to the phenomenon called dampening. If this combination amplifies the signal, it is known as underdamping, if the combination fades the signal, it is known as overdamping. Both underdamping and overdamping lead to incorrect measurements of arterial pressure. The relationship between flow and pressure depends on the intrinsic characteristics of the cardiocirculatory system. Systole generates two types of wave, pressure waves and flow waves. An arterial transducer will measure pressure waves. In the proximal aorta, the pressure wave and the flow wave will occur at almost the same moment; however, the pressure wave is transmitted to the periphery about twenty times faster than the flow wave. Also, the proximal aorta is filled during systole, resulting in an almost pulsatile flow whereas flow in the periphery is more constant, flowing in both systole and diastole [1, 2]. Therefore, pulse pressure analysis algorithms have been, and are being, developed to relate arterial pressure to cardiac output. Given an initial stroke volume value obtained by calibration, these algorithms use correction factors for the above issues to generate a volume vs. time curve from a pressure vs. time curve. In summary, an ideal algorithm should: z Work independently of the arterial site where pressure is being monitored, given the changes in waveform shape and pressure through the arterial tree from the core to the periphery. z Correct for the non-linear nature of compliance and individual variations in aortic characteristics to give an absolute cardiac output. z Not be affected by changes in vascular resistance caused by increases of arterial pressure due to reflected waves. z Not be reliant on identifying the details of wave morphology. z Be minimally affected by damping effects seen in arterial lines [3]. z History: From Frank to Wesseling The history of pulse pressure dates back more than 100 years. In 1899, Otto Frank developed the Windkessel (air chamber) model to simulate the heart-vessel interaction [4]. This model comprises a circuit in which fluid is pumped in tubes through chambers. The tubes are completely fluid-filled but the chambers contain some air. As the fluid is not compressible but the air is, the behavior of the air mimics aortic distension in blood vessels (compliance). Frank also deduced that by knowing the compliance, the stroke volume could be calculated from the change in pressure. In 1904, Erlanger proposed a correlation between stroke volume and change in arterial pressure and suggested there would be correlation between cardiac output and the arterial pulse contour [5]. This eventually led to the development of algorithms relating the pulse pressure analysis and cardiac output. Only with the recent advent of computer technology has it been possible to develop these algorithms to a level useful for clinical practice. The first algorithm to be used in clinical practice was the Wesseling algorithm in 1983 [6]. This algorithm is based on the hypothesis that the contour of the arterial pressure wave is dependent on stroke volume and this can be estimated from the integral of the change in pressure over the interval from the end of diastole to the end of systole (Asys) (Fig. 3).

4 Pulse Pressure Analysis 179 Fig. 3. Stroke volume is derived by determining the area under the curve of the systolic part of the arterial pressure (Asys) The integral calculation gives the value of the systolic area from the beginning of systole to the end of systole (dicrotic notch). The stroke volume (Vz) is the value of Asys divided by the value of aortic impedance (Zao): Vz ˆ Asys=Zao The stroke volume is adjusted (Vcz) to take account of heart rate (HR, to adjust for the effect of reflection waves in the periphery), mean arterial pressure (MAP) and age: Vcz ˆ Vz 0:66 0:005 HR 0:01 Age 0:014 MAP 0:8 Š Vcz is multiplied by heart rate to give the cardiac output. The system is then calibrated with thermodilution to equalize the pulse contour cardiac output to the measured one: COcz ˆ cal HR Vcz Where COcz is the continuous cardiac output, and cal is the calibration factor derived by the comparison of COcz with cardiac output measured by thermodilution [7]. This was the first algorithm of pulse pressure analysis used to determine cardiac output and from this starting point, new algorithms have been developed. In the Modeflow algorithm, Wesseling et al. refined their original algorithm to take account of compliance and the flow-pressure relationship. Both the Wesseling and Modeflow algorithms have shown good correlation with the gold standard pulmonary thermodilution [8±11] and led to the development of new devices. `Pulse pressure analysis' continues to be an evolving field in hemodynamic monitoring. z Available Devices There are presently four suppliers of devices able to calculate cardiac output from analysis of intravascular arterial pressure: z Pulsion with PiCCO monitor z LidCO with LiDCO TM plus System z FIAB SpA with pressure recording analytical method (PRAM) system z Edwards Lifesciences with FloTrac technology and Vigileo Monitor.

5 180 M. Cecconi et al. These devices and their calculation of continuous cardiac output is briefly described below. PiCCO Line requirements: z Central line for transpulmonary thermodilution z Femoral arterial line (specialized catheter) or radial arterial line (radial arterial long catheter) for transpulmonary thermodilution and pulse contour analysis. The PiCCO (Pulsion, Munich, Germany) monitors cardiac output and several volumes using transpulmonary thermodilution (e.g., intrathoracic blood volume (ITBV), extravascular lung water (EVLW)). Early versions of the device were the first to implement a modified Wesseling algorithm. Later versions use a more robust algorithm that includes analysis of arterial pressure during the diastolic phase to address issues around non-linear compliance and flow-pressure relationships. Both algorithms use transpulmonary thermodilution to calibrate the continuous cardiac output derived from the algorithm to the measured output. After calibration, the continuous cardiac output stroke volume is: First algorithm: SV ˆ cal 163 0:48 HR MAP Asys Second algorithm: SV ˆ cal Asys C p dp=dt dt where SV = stroke volume, cal = calibration factor, HR = heart rate, MAP = mean arterial pressure, C (p) =compliance corrected for arterial pressure, P =pressure, and t=time [6, 12]. The first algorithm uses a very similar approach to the Wesseling algorithm but excludes age. For a long time, continuous cardiac output using the PiCCO was the only pulse pressure analysis device available for use in intensive care and anesthesia. PiCCO continuous cardiac output has been validated against the PAC in several conditions [13±18] and has proven to be reliable, only needing regular recalibration in the event of major hemodynamic changes [19]. LiDCO Line requirement: z Central line or peripheral line for lithium bolus for transpulmonary dilution z Arterial line for lithium transpulmonary dilution and continous cardiac output. The LiDCO system is a new cardiac output monitor that measures cardiac output using lithium transpulmonary thermodilution. The LiDCO TM plus system implements an algorithm for continuous cardiac output monitoring derived from the arterial pressure wave. The LiDCO system is not truly a pulse contour monitor, rather it uses pulse power analysis based on the hypothesis that the change in power in the system (arterial tree) during systole is the difference between the amount of blood entering the system (stroke volume) and the amount of blood flowing out peripherally. It is based on the principle of conservation of mass/power and an assumption that following correction for compliance and calibration there is a linear relationshipbetween net power and net flow [3]. The algorithm overcomes the problem of reflected waves by taking account of the entire beat and uses an `auto-

6 correlation' to determine what proportion of the change in power is determined by the stroke volume. Once this is determined, cardiac output is easily calculated, multiplying stroke volume by heart rate. The first stage of the algorithm transforms the arterial pressure wave into a standardized volume waveform (arbitrary units) using this formula: DV=DP ˆ calibration 250 e k:p Where V=volume, P=blood pressure, k =curve coefficient, and 250 is the saturation value in mls, i.e., the maximum above the starting volume at which atmospheric pressure that the aorta/arterial tree can fill [3]. Autocorrelation uses the volume waveform and derives the period of the beat plus a net effective beat factor, proportional to the nominal stroke volume pumped into the aorta. This nominal stroke volume is then calibrated to be equalized to a measured stroke volume. The calibration is obtained using the lithium dilution technique. The continuous cardiac output of LiDCO has been validated in a number of studies [20±23] and proved to be reliable in surgical and intensive care patients. Studies of the device in different situations are continuing and will likely further refine this method. PRAM Line requirement: z an arterial line. Pulse Pressure Analysis 181 PRAM (Mostcare FIAB SpA) is a new continuous cardiac output monitor. The most innovative feature of PRAM is the lack of a requirement for calibration. The algorithm is based on the principle of perturbations, analyzing the arterial wave using a collecting signal of 1000 Hz. The most important points on the arterial wave for the calculation are the initial point of the pulse wave (diastolic pressure), the highest point (systolic pressure), and the point of closure of the aortic valve, represented by the dichotic notch or incisura dicrota. The PRAM algorithm uses these and other points of perturbance to take into account the interaction of heart contraction, aortic impedance and compliance and peripheral resistance. For each cardiac cycle, the whole area under the curve is measured and analysis of the perturbations gives a factor Z that correlates changes in volume to changes in pressure. The algorithm does not use recorded curves or the age or sex of the patient to calculate compliance but calculates it individually cycle by cycle. PRAM gives the formula for stroke volume as: SV ˆ A= P=t K where A =whole area under the systolic portion of the pressure curve, P =description of the pressure wave profile expressed as the variation in pressure (P) over time (t) during the entire cardiac cycle, and K =factor inversely related to the instantaneous acceleration of the vessel`s cross-sectional area, obtained from the ratio between expected and measured mean blood pressure. The most novel feature of this device is that calibration does not need to be performed to give an absolute cardiac output and a patient can be connected to PRAM as soon as an arterial catheterization is performed. PRAM is still under validation. So far it has been validated against pulmonary thermodilution in animals and in cardiac patients. In the study in cardiac surgery patients, the Bland Altmann plot

7 182 M. Cecconi et al. showed a good agreement between PRAM and standard thermodilution (mean difference l, standard deviation 0.43 l, with limits of agreement ±0.83 and l) [24±26]. FloTrac and Vigileo Line requirement: z an arterial line. Flotrac (Edwards Lifescience) is the name of the transducer incorporated in the Vigileo monitor. As with PRAM, this device does not require calibration and only requires an arterial line. The algorithm is primarily based on the standard deviation of the pulse pressure waveform: CO ˆ f compliance; resistance r p HR where r p is the standard deviation of the arterial pressure, HR is the heart rate, and f (compliance, resistance) is a scale factor proportional to vascular compliance and peripheral resistance. The standard deviation of the arterial pressure waveform is computed on a beatto-beat basis using the following equation: v u 1 X N 1 r p ˆ t 2 Pk P avg N 1 kˆ0 where P(k) is k th pulse pressure sample in the current beat, N is the total number of samples, and P avg is the mean arterial pressure. Compliance and resistance are derived from the analysis of the arterial wave. The hypothesis is that the shape of the arterial pressure wave can be used to calculate the effects of compliance and peripheral resistance on flow. Additional parameters, such as the pressure dependent Windkessel compliance, C w, based on Langwouters' study [27], heart rate and the patient body surface area (BSA) are also included to take other patient specific characteristics into account. This algorithm is now under validation. The two publications available [28, 29] show reasonable bias and precision in comparison with pulmonary thermodilution. Further studies will probably be available shortly. z Conclusions Pulse pressure analysis has an important role in the management of critically ill patients. There are currently several devices available that are less invasive and serve to provide continuous and accurate measurements of cardiac output. PiCCO is the oldest device and has been validated in several clinical situations. LiDCO is more recent and has also been validated, with more studies to follow. PRAM and FloTrac are very new devices that have the advantage of not requiring calibration and being quick and easy to use. There is evidence validating PRAM and results on the use of FloTrac are expected in the near future. There has been much criticism about the use of the PAC recently. It is our vision that pulse pressure analysis will be implemented in clinical protocols to change management and outcome in critically ill patients.

8 Pulse Pressure Analysis 183 References 1. Darovic GO, Vanriper S (1995) Arterial pressure recording. In: Darovic GO (ed) Hemodynamic Monitoring. 2nd ed. WB Sauners, Philadelphia, pp 177± Nichols WW, O'Rourke MF (1990) McDonald's Blood Flow in Arteries. 3rd edition. Lea & Febiger, Philadelphia, pp 251± Rhodes A, Sunderland R (2005) Arterial pulse power analysis: The LiDCOTMplus System. In: Pinsky MR, Payen D (eds) Functional Hemodynamic Monitoring. Springer, Heidelberg, pp 183± Frank O (1899) Die Grundform des Ateriellen Pulses. Z Biol 37: Erlanger J, Hooker DR (1994) An experimental study of blood pressure and of pulse pressure in man. John Hopkins Hospital Records 12:145± Wesseling KH, de Wit B, Weber AP (1983) A simple device for the continuous measurement of cardiac output. Adv Cardiovasc Phys 5:1±52 7. Jansen JRC, van den Berg PCM (2005) Cardiac output by thermodilution and arterial pulse contour techniques. In: Pinsky MR, Payen D (eds) Functional Hemodynamic Monitoring. Springer, Heidelberg, pp 135± Jansen JRC, Wesseling KH, Settels JJ, Schreuder JJ (1990) Continuous cardiac output monitoring by pulse contour during cardiac surgery. Eur Heart J 11:26±32 9. Wesseling KH, Jansen JRC, Settels JJ, Schreuder JJ (1993) Computation of aortic flow from pressure in humans using a nonlinear, three element model. J Appl Physiol 74:2566± Jansen JRC, Schreuder KK, Mulier JP, Smith NT, Settels JJ, Wesseling KH (2001) A comparison of cardiac output derived from the arterial pressure wave against thermodilution in cardiac surgery patients. Br J Anaesth 87:212± Jellema WT, Wesseling KH, Groeneveld ABJ, Stouteribeek CP, Thijs LG, van Lieshout JJ (1999) Continuous cardiac output in septic shock by simulating a model of the artic input impedance. A comparison with bolus injection thermodilution. Anesthesiology 90:1317± Godje O, Hoeke K, Goetz AE, et al (2002) Reliability of a new algorithm for continous cardiac determination by pulse-contour analysis during hemodynamic instability. Crit Care Med 30:52± Buhre W, Weyland A, Kazmaier S, et al (1999) Comparison of cardiac output assessed by pulse contour analysis and thermodilution in patients undergoing minimally invasive direct coronary artery bypass grafting. J Cardiothorac Vasc Anesth 13:437± Zollner C, Briegel J, Kilger E, Haller M (1998) Retrospective analysis of transpulmonary and pulmonary arterial measurement of cardiac output in ARDS patients. Anaesthesist 47:912± Goedje O, Hoeke K, Lichtwarck-Aschoff M, et al (1999) Continuous cardiac output by femoral artery thermodilution calibrated pulse contour analysis: Comparison with pulmonary artery thermodilution. Crit Care Med 27: 2407± Zollner C, Haller M, Weis M, et al (2000) Beat-to-beat measurement of cardiac output by intravascular pulse contour analysis: A prospective criterion standard study in patients after cardiac surgery. J Cardiothorac Vasc Anesth 14:125± Della Rocca G, Costa MG, Pompei L, Coccia C, Pietropaoli P (2002) Continous and intermittent cardiac output measurement: pulmonary artery catheter versus aortic transpulmonary technique. Br J Anaesth 88:350± Della Rocca G, Costa MG, Coccia C, et al (2003) Cardiac output monitorino: aortic transpulmonary thermodilution and pulse contour analysis agree with standard thermodilution methods in patients undergoing lung transplantation. Can J Anaesth 50:707± Rodig G, Prasser C, Keyl C, Liebold A, Hobbhahn (1999) Continuous cardiac output measurement: Pulse contour analysis versus thermodilution technique in cardiac surgical patients. Br J Anaesth 82:525± Hamilton TT, Huber LM, Jessen ME (2002) PulseCO: a less-invasive method to monitor cardiac output from arterial pressure after cardiac surgery Ann Thorac Surg 74:S1408±S Tsutsui M, Mori T, Aramaki Y, Fukuda I, Kazama T (2004) A comparison of two methods for continuous cardiac output measurement: PulseCO VS CCO. Masui 53:929±933

9 184 M. Cecconi et al.: Pulse Pressure Analysis 22. Jonas MM, Tanser SJ (2002) Lithium dilution measurement of cardiac output and arterial pulse waveform analysis: an indicator dilution calibrated beat-by-beat system for continuous estimation of cardiac output. Curr Opin Crit Care 8:257± Pittman J, Bar-Yosef S, Sumping J, Sherwood M, Mark J (2005) Continuous cardiac output monitoring with pulse contour analysis: A comparison with lithium indicator dilution cardiac output measurement. Crit Care Med 33:2015± Romano SM, Pistolesi M (2002) Assessment of cardiac output from systemic arterial pressure in humans. Crit Care Med 30:1834± Giomarelli P, Biagioli B, Scolletta S (2004) Cardiac output monitorino by pressare recording analytical method in cardiac surgery. Eur J Cardiothorac Surg 26:515± Scolletta S, Romano SM, Biagioli B, Capannini G, Giomarelli P (2005) Pressure recording analytical method (PRAM) for measurement of cardiac output during various haemodynamic states. Br J Anaest 95:159± Langewouters GJ, Wesseling KH, Goedhard WJ (1984) The static elastic properties of 45 human thoracic and 20 abdominal aortas in vitro and the parameters of a new model. J Biomech 17:425± Manecke GR, Peterson M, Auger W (2004) Cardiac output determination using arterial pulse: a comparison of a novel algorithm against continuous and intermittent thermodilution. Crit Care Med 32:A43 (abst) 29. McGee W, Horswell JL, Janvier G (2005) Validation of a continuous cardiac output measurement using arterial pressure waveforms. Crit Care 9:S24 (abst)