The aim of this thesis is to address different aspects of paediatric cardiopulmonary bypass in detail and ...... Mc Graw-Hill Book Company;. 1961. 6. Guiot C. et al.
RIJKSUNIVERSITEIT GRONINGEN
Strategies for optimisation of paediatric cardiopulmonary bypass
PROEFSCHRIFT
ter verkrijging van het doctoraat in de Medische Wetenschappen aan de Rijksuniversiteit Groningen op gezag van de Rector Magnificus, dr. F. Zwarts, in het openbaar te verdedigen op woensdag 12 februari 2003 om 16.00 uur
door
Filip Maria Jan Jozef De Somer geboren op 10 mei 1960 te Aalst (België)
Promotores:
Prof. dr. T. Ebels Prof. dr. G. Van Nooten
Co-promotor:
Prof. dr. P. Verdonck
Beoordelingscommissie:
Prof. dr. R.Berger Prof. dr. H.J.Busscher Prof. dr. M. Hazekamp
ISBN
90-423-0210-0
Voor Caroline en Casper Voor mijn ouders
© Copyright Shaker Publishing 2002 All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted, in any form or by any means, electronic, mechanical, photocopying, recording or otherwise, without the prior permission of the publishers. Printed in The Netherlands. ISBN 90-423-0210-0 Shaker Publishing BV St. Maartenslaan 26 6221 AX Maastricht Tel.: 043-3500424 Fax: 043-3255090 http:// www.shaker.nl
Contents
Contents Chapter 1:
Introduction
3
Chapter 2:
Vascular access for total body perfusion
9
Chapter 3:
Circuit design
21
Chapter 4:
Oxygenation by artificial lung systems
33
Chapter 5:
Systemic inflammatory response
43
Chapter 6:
Summary and new prospectives
59
Appendix 1
Evaluation of different paediatric venous cannulas using
63
gravity drainage and VAVD: an in vitro study Perfusion, 2002; 17(5): 321 – 326 Appendix 2
Hydrodynamical Comparison of Aortic Arch Cannulae
83
Int. J. Art. Organs, 1998; 21(11): 705 – 713 Appendix 3
Comparison of two dissimilar designs of paediatric aortic
115
cannulae Int. J. Art. Organs, 2002, 25(9): 867 – 874 Appendix 4
D-901 Neonatal oxygenator: a new perspective
141
Perfusion 1994; 9: 349 – 355 Appendix 5
Low extracorporeal priming volumes for infants: a benefit?
159
Perfusion 1996; 11: 455 – 460 Appendix 6
Hydrodynamic characteristics of artificial lungs
175
ASAIO, 2000; 46(5): 532 – 535
1
Contents
Appendix 7
Impact of oxygenator design on hemolysis, shear stress,
195
white blood cell and platelet count J. Cardiothor.Vasc. Anesth. 1996; 10: 884 - 889 Appendix 8
Can an oxygenator design potentially contribute to air
219
embolism in CPB. A novel method for the determination of the air removal capabilities of neonatal oxygenators Perfusion, 1998; 13: 157 – 163 Appendix 9
In vivo evaluation of a phosphorylcholine coated
241
cardiopulmonary bypass Journal of Extra-corporeal technology, 1999; 31 (2): 62-67 Appendix 10 Phosphorylcholine coating of extracorporeal circuits
261
provides natural protection against blood activation by the material surface European Journal of Cardio-Thoracic Surgery, 2000; 18(5): 602 – 606 Appendix 11 Tissue factor as main activator of the coagulation system
283
during cardiopulmonary bypass The Journal of Thoracic and Cardiovascular Surgery, 2002; 123: 951 – 958 Nederlandse samenvatting
309
Dankwoord
315
Curriculum vitae
317
2
Chapter 1
Chapter 1 Introduction and aim of the thesis The mortality associated with the repair of congenital heart defects in early life has decreased considerably over the years. However improved survival has unmasked a whole spectrum of morbidity associated with the practice of cardiopulmonary bypass [1]. As a general concept, cardiopulmonary bypass will temporarily bypass heart and lungs. This is achieved by introducing one or two venous cannulas in the venae cavae that direct venous return of the patient, by means of plastic tubing, into a reservoir. This reservoir replaces the compliance of the veins. From the reservoir blood is pumped through an artificial lung or oxygenator. The oxygenator heats or cools the blood and maintains physiologic blood gases. Subsequently the oxygenated blood is guided through an arterial filter and re-infused by means of an arterial cannula into the aorta. All these components need to be primed before cardiopulmonary bypass can be started. Apart of this life support, the circuit is designed to meet specific surgical needs. Most systems have one or more aspiration lines for the recuperation of blood losses in the surgical field, the unloading of the left ventricle and aspiration of blood from additional blood vessels such as a left superior vena cava or collateral blood vessels. In many institutions the cardioplegia delivery is also integrated into the cardiopulmonary bypass circuit. During conduct of paediatric cardiopulmonary bypass quite drastic changes occur. Due to haemodilution by priming solutions and cardioplegia, the haematocrit varies between 20 – 35%. Most operations require a certain
3
Chapter 1 amount of hypothermia. Depending on the specific procedure the actual blood temperature might vary between 15 and 38° C. As a consequence of these temperature and haematocrit changes, viscosity will change and thus influence tissue perfusion. Also blood flows will change depending on the surgical procedure from circulatory arrest to high flow (up to 150 mL/kg) in the rewarming phase. It is often assumed that a paediatric cardiopulmonary bypass circuit is a miniaturised adult system. This is not correct. In contrast to adults the priming volume of even the smallest paediatric circuits will equal or exceed the total blood volume of a baby. At the same time blood of the child will be exposed to at least four times more foreign surface relative to an adult. The unique physiology of the neonate and his sometimes aberrant anatomy, leads to technical limitations and, therefore, makes the design and conduct of a dedicated paediatric cardiopulmonary bypass complicated. The combination of a new-born at one hand and open-heart surgery and cardiopulmonary bypass at the other hand is quite challenging. The new-born is a fast developing organism with immature organs within which the organic systems are developing or maturing at different rate. Open-heart surgery and cardiopulmonary bypass represent an extreme stress to the functioning of these developing systems. Moreover, the response of those organs to this stress will be different from what is reported in adults. Children are definitively more prone to inflammatory response. Also neurological consequences of the developing brain are different from those observed in the developed or degenerating brain.
4
Chapter 1 The small size of vascular and cardiac structures not only challenges surgical skills but also limit the possibilities for obtaining an optimal vascular access and a bloodless surgical field. Due to this unique anatomical and physiological environment specially designed components have been developed. This research and development is expensive and will often reach the end spectrum of technical know how. Unfortunately, most of the time some industries are reluctant to invest in the paediatric domain because of the small numbers compared to the huge amount of adult cardiac procedures performed yearly. Further research is also required to investigate the long and short-term influence of different surgical strategies and techniques for conducting cardiopulmonary on the different organ systems. Recent research clearly demonstrates a correlation between conduct of cardiopulmonary bypass and morbidity [2-6]. However, as pointed out by Jonas and Elliott [1], the consequences of a badly conducted paediatric cardiopulmonary bypass should not be underestimated as it may impact several decades. The child’s quality of life is likely to be markedly diminished. Yet that is only part of the potential disaster. Children have parents and relatives. Each will be affected by the poor outcome of cardiopulmonary bypass. One bypass disaster can ruin many lives.
5
Chapter 1
Aim of the thesis The aim of this thesis is to address different aspects of paediatric cardiopulmonary bypass in detail and to propose modifications in order to reduce cardiopulmonary bypass related morbidity and by doing so, improve patient outcome. We will focus on four major items: (1) vascular access, (2) mass transfer and fluid dynamics of oxygenators, (3) circuits and (4) whole body inflammatory reaction. •
The small vascular structures of the new-born demand a better design description of the geometry and fluid dynamic characteristics of cannulas. There is not only a need for a better validation of today’s cannulas but also for research into the relation between the hemodynamic characteristics of these cannulas and possible damage to blood elements.
•
The oxygenator is prone to less optimal flow, due to its tortuous flow path, its large foreign surface area and the rapid changes in blood velocity resulting in non-optimal mass transfer and activation of the whole body inflammatory response. Additionally, most oxygenators have a priming volume that is too high compared to the total blood volume of a new-born. There is an urgent need for smaller, more blood compatible oxygenators, with optimisation of their fluid mechanics and gas exchange in order to fit the paediatric needs. These needs will include the capability for achieving subnormal
arterial
oxygen
tensions
in
cyanotic
children
without
compromising the high oxygen consumption of children during rewarming. •
Most circuits today have been designed based on empirically derived data. This results in large volumes in the arterial and venous lines as well as in
6
Chapter 1 the aspiration lines. The use of an arterial line filter is highly recommended although it is not used in an appropriate way in most institutions. •
Finally, the use and conduct of a paediatric cardiopulmonary bypass will end in a mild or more pronounced whole body inflammatory reaction. The strength of this reaction will vary from child to child, the equipment used, and the conduct of the bypass.
We will propose techniques and strategies to overcome or to reduce these problems and by doing so to ameliorate the cardiopulmonary bypass related morbidity.
References 1. RA Jonas, MJ Elliott. Cardiopulmonary bypass in neonates, infants and young children. Butterworth-Heinemann, Oxford 1994. 2. S Daniel. Review of the multifactorial aspects of BioInCompatibility in CPB. Perfusion, 1996; 11: 246-255. 3. DT Pearson, RF Carter, MB Hammo, PS Waterhouse. Gaseous microemboli during open heart surgery. In: Towards safer cardiac surgery. Ed. DB Longmore. Lancaster, MTP Press, 1981: 325-354. 4. JM Pearl, DW Thomas, G Grist, JY Duffy, PB Manning. Hyperoxia for management of acid-base status during deep hypothermia with circulatory arrest. Ann Thorac Surg 2000; 70: 751-755. 5. RA Jonas, DC Bellinger, LA Rappaport et al. Relation of pH strategy and development outcome after hypothermic circulatory arrest. J Thorac Cardiovasc Surg. 1993; 106: 362-368.
7
Chapter 1 6. T Shin’oka, D Shum-Tim, PC Laussen et al. Effects of oncotic pressure and haematocrit on outcome after hypothermic circulatory arrest. Ann Thorac Surg 1998; 65: 155-164.
8
Chapter 2
Chapter 2 Vascular access for total body perfusion
2.1. Introduction
This chapter introduces the limitations and boundary conditions of vascular access in paediatric cardiopulmonary bypass. The different requirements for venous and arterial access are reviewed. Finally, the hydrodynamic characteristics and different evaluation methods are presented and discussed. Recommendations for an optimal communication between manufacturer and clinician are given. 2.1.1. Problems related to vascular access Unsuccessful cannulation may lead to cerebral complications [1-3] A malpositioned aortic cannula may obstruct cerebral blood flow, or it may cause a preferential flow into the descending aorta and “steal” blood from the brain’s circulation [3]. Alternatively, obstruction by the superior vena caval cannula may decrease cerebral venous drainage and potentially lead to brain dysfunction [3]. A direct correlation between age and cerebral alterations (low cerebral blood flow velocity and EEG slowing) caused by malpositioning of the cannulas has been reported [3]. 2.2. Venous access
Cannulation of the venous side of the circulation aims at draining the venous blood from the central veins or right heart cavities in a laminar flow without inducing any marked change of the pressure within the large veins. Only then an adequate forward flow can be established. The entire venous return to the
9
Chapter 2 heart should be able to pass through the chosen cannulas without obstruction and without damaging the blood vessel [4]. An essential problem of venous drainage is a compliance and geometric mismatch. Wide, low-resistance, collapsible vessels are connected to smaller, less compliant, artificial conduits. When suction is applied to the venous reservoir, flow starts to increase linearly, but once the vessel starts to collapse, the flow will stagnate. Increase in suction force beyond a critical level, therefore, cannot increase the amount of venous drainage. Additionally, high resistance in the drainage tube necessitates higher degrees of suction than is needed with short, wide tubing. Maintenance of a positive pressure at the tip of the cannula broadens the range of flow regulation because it prevents venous collapse [5]. Reduced venous drainage may be due to reduced venous pressure, inadequate height of the patient above the venous reservoir, malposition of the venous cannulas or obstruction or excess resistance of the lines and cannulas. Venodilation or hypovolaemia may cause inadequate venous pressure. 2.3. Arterial access
Cannulation of the arterial side of the circulation must provide an adequate forward flow of blood to the patient. The cannula and its placement must not be non-obstructive and flow must be directed to the distal aorta in order to perfuse all areas of the body. The ideal cannula will generate sufficient flow without obstructing or damaging the blood vessel.
10
Chapter 2 2.4. Cannula characteristics
2.4.1. Design related problems The choice of the best cannula for a given procedure is not simple. In general, manufacturers do not mention in their information brochures the internal diameter of a cannula but only the outer diameter. Depending on the production process, the wall thickness of comparable cannulas can be quite different although their respective manufacturers measured identical outer diameters [6]. Additionally, production tolerances result in important differences in internal diameter even between cannulas of identical size and manufactured by the same company. Since the pressure-flow relation highly depends on the inner diameter and cannulas standard used in paediatric cardiopulmonary bypass have small diameters, this results in significant deviations of the mean values given by the manufacturer. Another difficulty is related to the fact that the pressure-flow characteristic of a cannula is always measured for water (low viscosity and Newtonian fluid). Unfortunately, it is difficult to extrapolate water values towards blood (higher viscosity and non Newtonian fluid) flow conditions. 2.4.2. Available data for clinicians Manufacturers only report the polynomial regression of the water data of a certain number of cannulas (Figure 1). Thus, the user has no information about of the possible variability range. This is demonstrated in Figure 1 where both the polynomial regression (full line) as given by the manufacturer and the measured data of ten cannulas (dots) are depicted.
11
Chapter 2 Figure 1: Pressure-flow relationship for two paediatric arterial cannulas
Pressure drop [mmHg]
DLP 77108
DLP 75008
300
300
200
200
100
100
0
0
0.0
0.5
1.0
1.5
0.0
0.5
1.0
1.5
Water flow [L/min]
2.4.3. Theoretical relationship For a horizontal straight tube the relation between pressure and flow can be described by Poiseuilles formula: 8µL ∆P = 4 Q πR 32 µL ∆P = 2 U D where µ = dynamic viscosity [N/m².s], L = length [m], R = radius [m], Q = average flow [m³/s], U = mean velocity [m/s], D = diameter [m]. For cannulas this formula cannot be used since most cannula are not straight tubes. 2.4.4. Practical characterisation Several attempts have been described to predict the clinical performance of cannulas. 12
Chapter 2 (1) Montoya et al. propose a system in which any vascular access device can be characterised by a single number denoted as “M” which may be determined from the geometry and/or from simple in vitro pressure-flow measurements [7-9]. M is defined as log (LDC-4.75) where L represents the length and DC the characteristic diameter of the cannula. The Dc is also known as hydraulic diameter for non-circular ducts representing the diameter of a corresponding circular orifice. The method can be used to choose the best possible cannula when a given diameter or pressure may not be exceeded during the procedure. Unfortunately, the method has some disadvantages. In order to obtain the Mnumber on a non-uniform design, such as a cannula, one has to do in vitro measurements. The M-number also assumes that the flow regimen is turbulent. However the obtained value is not useable in clinical practice, especially if it is obtained by water measurements. Water measurements tend to lie in the turbulent region while the blood flows used during clinical use are in the laminar region. The latter limits its use in open-heart surgery [10].
(2) Another approach is based on the theory of dynamic similarity [6,11-12]. Flows become identical if the Reynolds number, a measure of the ratio between inertial and viscous forces, is identical for both fluids [6] in the experimental set-up (e.g. water) and in the clinical situation (blood). Re =
UD 4ρQ µ = with ν = µπD ν ρ
Where Q = flow [m³/s], ρ = density [kg/m³], μ = dynamic viscosity [N/m² s], D = diameter [m], ν =kinematic viscosity [m²/s], U = mean velocity [m/s].
13
Chapter 2 For Reblood = Rewater : Qblood = Qwater
νblood νwater
The pressures for a given water flow can be transformed to those of blood in an analogue way by using the Euler number, a measure of the ratio between pressure and inertial forces: Eu =
P π 2 D 4 ∆P = ρU ² 16 ρQ 2
Where P = pressure [Pa] For Eublood = Euwater: Pblood = Pwater
ρblood ρwater
ρblood µblood Ublood P blood = Pwater so that Uwater ρwater µwater 2
2
The dimensionless numbers Reynolds and Euler are independent of the fluid physical properties. This allows converting directly flow rates and pressures. In order to apply this technique one has to know the rate of the densities and the rate of the dynamic and kinematic viscosity of both fluids. Since water tests are performed at room temperature water density is approximately 1000 kg/m³ (998.2019 kg/m³) and water kinematic viscosity 1 10-6 m²/s (1.0038 10-6 m²/s). If we compare water data with blood at a temperature of 37°C and a haematocrit of 33.5% we obtain the following pressure and flow conversion factors presented in Table 1. The factors in table 1 are calculated using the formulas presented in section 3.1.2.3.
14
Chapter 2 Table 1. Pressure and flow conversion factors Qblood/Qwater
Pblood/Pwater
T = 37°C
2.43
6.21
T = 20°C
3.40
12.19
Flows and pressures measured during water tests are multiplied with these factors to obtain corresponding blood flows and pressures.
(3) A third method rescales the coefficients of the fitted parabolic equation between pressure drop (∆P) and flow rate (Q) ∆Pwater = awaterQwater 2 + bwaterQwater to blood ∆Pblood = abloodQblood 2 + bbloodQblood For a given awater, bwater and the relationship between pressure and flow one can determine ablood and bblood as: ablood =
ρblood awater ρwater
bblood =
µblood bwater µwater
Table 2. Conversion factors for coefficients a and b ablood/awater
bblood/bwater
T = 37°C
1.055
2.56
T = 20°C
1.055
3.59
The factors in Table 2 are derived from Table 1 taking into account a
ρblood ρwater 15
Chapter 2 ratio of 1.03. In Figure 3 a comparison of both methods (calculation based on dynamic similarity and the parabolic method) is presented. There is still a deviation from the measured data but it gives an estimate of what can be expected under given conditions. The deviation is due to the low accuracy of water measurements caused by the error range on pressure transducers and flow meters. These errors are subsequently multiplied with the conversion factors resulting in even larger deviations. This also explains why the deviation of the calculated data is smaller at 37°C than at 20°C. Use of water-glycerin solutions by manufacturers for validation of their cannulas instead of water will reduce the error.
DLP 77008 20°C - Hct 33.5%
300
37°C - Hct 33.5%
300
Dynamic similarity Measured Parabolic method
Dynamic similarity Measured Parabolic method
250
Pressure drop [mmHg]
225 200
150
150
100 75 50
0 0.0
0.2
0.4
0.6
0.8
1.0
0 0.0
0.2
0.4
0.6
0.8
1.0
Blood flow [L/min]
16
Chapter 2 2.4.5. Quantification of blood damage Pressure-flow relationships do not give direct information regarding the possible damage of blood elements when a given cannula is used. It is not necessarily the cannula with the highest pressure drop that will generate most damage. The exerted shear rate and specifically shear stress in combination with the duration of these forces (residence time) are far more important factors for blood cell damage [13]. Shear stress equals fluid dynamic viscosity multiplied by shear rate. τ = µ δu δr
with u the axial velocity component and r the radial variable
or τw = ∆P
R 2L
where τw = shear stress [N/m²], R = radius [m], L = length [m]
As tube length is usually several orders of magnitude greater than radius, pressure is generally orders of magnitude greater than shear stress [14]. Physiological values of shear stress range from 1 – 50 dynes/cm² 1[14]. Most actual cannulas will easily generate shear stresses of several hundred dynes/cm² [15], which is far above the trigger values of 75 and 100 dynes/cm² [14,16] needed to activate white blood cells and platelets, respectively.
1
10
dyne N =1 cm² m² 17
Chapter 2 2.5. Conclusions
Vascular access in neonates and small infants remains a major challenge for adequate paediatric cardiopulmonary bypass. Small vascular structures, congenital malformations and technical limitations in the manufacturing of cannulas give rise to specific problems. A better documentation of the pressure-flow relationship of a cannula in combination with its shear stress data will help the clinician in choosing the best cannula for a given procedure. Thus manufacturers should provide more adequate information regarding the pressure-flow characteristics and both the inner and outer diameter of their products.
References 1. FH Kern, PR Hickey. The effects of cardiopulmonary bypass on the brain. In: Cardiopulmonary bypass in neonates, infants and young children. Eds: RA Jonas, MJ Elliott. Butterworth-Heinemann, Oxford 1994: 263-281 2. RA Rodriguez, G Cornel, L Semelhago, WM Splinter, NA Weerasena. Cerebral effects in superior vena caval cannula obstruction: the role of brain monitoring. Ann Thorac Surg 1997; 64: 1820-1822. 3. RA Rodriguez, G Cornel, WM Splinter, NA Weerasena, CW Reid. Cerebral vascular effects of aortovenous cannulations for pediatric cardiopulmonary bypass. Ann Throac Surg 2000; 69: 1229-1235. 4. M Elliott. Canulation for cardiopulmonary bypass for repair of congenital heart disease. In: Cardiopulmonary bypass in neonates, infants and young children. Eds: RA Jonas, MJ Elliott. Butterworth-Heinemann, Oxford 1994: 128-140. 18
Chapter 2 5. PM Galletti, GA Brecher. Connection of the vascular system with an extracorporeal circuit. In: Heart lung bypass; principles and techniques of extracorporeal circulation. New York: Grune and Stratton; 1962: 171-193. 6. JF Douglas, JM Gaiorek, JA Swaffield, Part III Dimensional Analysis and Similarity in Fluid Mechanics, 3rd ed., Longman Scientific & Technical, Harlow, UK; 1985. 7. Delius RE, Montoya JP, Merz SI, McKenzie J, Snedecor S, Bove EL, Bartlett RH. New method for describing the performance of cardiac surgery cannulas. Ann Thorac Surg. 1992 Feb;53(2):278-81. 8. Sinard JM, Merz SI, Hatcher MD, Montoya JP, Bartlett RH. Evaluation of extracorporeal perfusion catheters using a standardized measurement technique--the M-number. ASAIO Trans. 1991 Apr-Jun;37(2):60-4. 9. Montoya JP, Merz SI, Bartlett RH. A standardized system for describing flow/pressure relationships in vascular access devices. ASAIO Trans. 1991; 37(1):4-8 10. Kim WG, Park SS. Clinical application of the M-numbers of aortic cannulas during hypothermic cardiopulmonary bypass in pediatric patients. Artif Organs. 1999 Apr;23(4):369-72. 11. Uyttersprot N. “Stromingseigenschappen en bloedcompatibiliteit van kindercanules.” Master of Science in Engineering, Thesis in Dutch, Ghent University, 1999. 12. Verdonck P, Siller U, De Wachter D, De Somer F. Hydrodynamical comparison of aortic arch cannulae. Int J Artif Organs, 1998; 21:705-713.
19
Chapter 2 13. LJ Wurzinger, R Opitz, P Blasberg, H Schmid-Schönbein. Platelet and coagulation parameters following millisecond exposure to laminar shear stress. Thrombosis and Haemostasis. 1985; 54: 381-386. 14. SM Slack, VT Turitto. Fluid dynamic and hemorheologic considerations. Cardiovasc Pathol 1993; 2(3): 11S-21S. 15. F De Somer, L Foubert, M Vanackere, D Dujardin, J Delanghe, G Van Nooten. Impact of oxygenator design on hemolysis, shear stress, white blood cell and platelet count. J. Cardiothor.Vasc. Anesth. 1996; 10: 884889 16. LV McIntire, RR Martin. Mechanical trauma induced PMN leukocyte dysfunction. In The Rheology of Blood Vessels and Associated Tissues Eds Gross DR, Hwang NHC.. Alphen aan den Rijn: NATO Advanced Study Institute Series - E, No 41, Sijthoff & Noordhoff, 1981
20
Chapter 3
Chapter 3 Circuit design The cardiopulmonary bypass circuit consists basically of venous and arterial (often including an arterial filter) tubing lines and an oxygenator with integrated heat exchanger. This chapter deals with the hydrodynamic design of the tubing and arterial filter. The artificial lung or oxygenator is discussed in chapter 4.
3.1. Tubing
3.1.1. Priming volume Once cardiopulmonary bypass is started, the volume in the arterial and venous line as well as the priming volume of the oxygenator enlarges the total circulating blood volume of the baby. Additionally, suction and vent lines that are empty before starting cardiopulmonary bypass, remove an important amount of blood out of the circulation once in use. Subsequently this blood is returned into the circulation just before weaning cardiopulmonary bypass. As a result important and rapid changes in circulating blood volume occur during cardiopulmonary bypass. Because of this it is important to keep volumes in the complete extracorporeal circulation as small as possible without jeopardising flow requirements of the given lines. Its length and diameter (Table 1) determine the volume of a line
21
Chapter 3 Table 1: Priming volumes for different tubing diameters Tubing diameter 1
Priming volume per 10 cm of
Inch
mm
length (mL)
1/8
3.17
0.792
3/16
4.76
1.781
1/4
6.35
3.167
3/8
9.53
7.126
1/2
12.70
12.668
3.1.2. Dimensions of the tubing 3.1.2.1. Introduction The dimensions of the venous and arterial lines depend on the desired blood flow rate and the height difference between table and oxygenator. When gravity drainage is used a height difference between 30 and 40 cm is generally accepted [1]. In many institutions sizing of tubing is established in an empirical way. A more objective way is to decide based on fluid dynamic parameters [2], thus limiting the dead volume in the aspiration lines to an absolute minimum. The resulting reduction in priming volume results in less homologous blood product utilisation [3,4]. 3.1.2.2. Laminar or turbulent flow Two types of steady flow of real fluids exist: laminar flow and turbulent flow with a transition zone in between. Different fluid dynamic laws govern the two types of flow.
1
1 inch = 25.4 mm 22
Chapter 3 In laminar flow, fluid particles move along straight, parallel paths in layers. Magnitudes of velocities of adjacent layers are not the same. The viscosity of the fluid is dominant and thus suppresses any tendency for turbulent conditions due to the inertia of the fluid. In turbulent flow, fluid particles move in a haphazard fashion in all directions. The critical velocity is the velocity below which all turbulence is damped out by the viscosity of the fluid. It is found that a Reynolds number of about 2000 represents the upper limit of laminar steady flow of practical interest. The Reynolds number is a dimensionless number, representing the ratio of inertia forces to viscous forces, in circular pipes [2]. Re =
UD ν
U = mean velocity [m/s], D = diameter [m], ν =kinematic viscosity [m²/s] with ν=
µ ρ
where ρ = density [kg/m³], µ = absolute blood viscosity [N/m² .s] 3.1.2.3. Blood viscosity Dynamic viscosity of a fluid (µ) is either determined from literature data or measured in a viscosity meter. Blood viscosity can be described by exponential formula with: 1800 exp− 5.64 + (T + 273) µplasma = 100 µ = µplasma exp(2.31Hct ) ρ = [1.09 Hct + 1.035(1 − Hct )]
23
Chapter 3 µplasma = plasma viscosity [N/m².s], T = absolute temperature [°C], Hct = haematocrit [expressed as fraction] Figure 1: Relationship between haematocrit, temperature and kinematic blood viscosity
Blood viscosity calculation 4.0
Hct 36% Hct 34% Hct 32% Hct 30% Hct 28% Hct 26% Hct 24% Hct 22% Hct 20%
Blood viscosity [x 10-6 N/m².s]
3.5
3.0
2.5
2.0
1.5 20
22
24
26
28
30
32
34
36
38
Blood temperature [°C]
Based on these calculations a nomogram can be constructed for a quick estimate of blood viscosity when haematocrit and temperature are known (Figure 1).
3.1.2.4. Pressure-flow relationship In general the pressure drop can be calculated in function of diameter, length, blood viscosity and height difference between patient and heart-lung machine, using the equation:
24
Chapter 3 ∆P = f
L U2 D 2g
where f = friction factor, g = gravitational acceleration [m/s²] and f =
64 when flow is laminar. Re
However when the flow regimen is turbulent f is calculated using the Colebrook equation: ε 2.51 = −2 log + 3.7 D Re f f
1
with ε the roughness parameter. Besides the Colebrook equation the Blasius formula is valid for smooth pipes and low Reynold numbers. The friction factor becomes independent of the roughness of the tube f = 0.316 Re
−
1 4
By using these equations flow diagrams can be calculated for venous and arterial lines in function of length, diameter, required blood flow, viscosity and desired pressure drop. 3.1.2.5. Case study If a baby needs cardiopulmonary bypass support one can calculate what should be the appropriate diameter for both arterial and venous line. In our example, the cardiopulmonary bypass circuit has an arterial and venous line of 150 cm. The surgeon wants for this specific case a haematocrit of 30% and no hypothermia during cardiopulmonary bypass. The maximum blood flow to ensure adequate tissue perfusion is 700 mL/min.
25
Chapter 3 From Figure 2 we learn that both 3/16 and 1/4 inch arterial lines generate laminar flow (shaded zone) for the given conditions. However, the pressure loss over the arterial line will be approximately 20 mmHg higher if a 3/16 inch diameter is chosen. This difference is acceptable so a 3/16 inch line gives the best compromise between priming volume and pressure-flow characteristics. Figure 2. Flow regimen in paediatric arterial lines
Characteristics of 3/16" and 1/4" arterial lines. 3/16
Length: 150 cm 3/16
Tem perature: 37° Celsius
150
Haem atocrit: 30%
3/16
Pressure drop [mmHg]
3/16 3/16 3/16
100 3/16 3/16 3/16 3/16 3/16
50 3/16 3/16 3/16 3/16 3/16
0
3/16 3/16 1/4 1/4 1/4 3/16 1/4 3/16
3/16 1/4
1/4
1/4
1/4
1/4
1/4
1/4
1/4
1/4
1/4
1/4
1/4
1/4
1/4
1/4
1/4
1/4
Reynolds < 2000
0.1
0.3
0.5
0.7
0.9 1.1 Blood flow [L/min]
1.3
1.5
1.7
1.9
Suppose it is decided to use a 3/16 inch venous line in the above described case and the height difference between the operating table and the oxygenator is 35 cm H20. We can determine the limitations of this choice by using Figure 3. On the right Y-axis we notice that the Reynolds number (squares), when using a haematocrit of 30% (X-axis) and a blood temperature of 37°C, is below 2000 for a blood flow of 700 mL/min. The maximum blood flow we can drain for these conditions (circles) is 770 mL/min (left Y-axis).
26
Chapter 3 This is approximately 10% higher than the maximum flow we anticipate. Thus, a 3/16 inch venous line is a correct choice for this particular case.
Figure 3. Flow characteristics of a 3/16 inch venous line
Characteristics of a 3/16" venous line Blood flow at 37°C Blood flow at 20°C Reynolds number at 37°C Reynolds number at 20°C
1.0
2500
0.9
0.8 1500 0.7
Reynolds number
Blood flow [L/min]
2000
1000 0.6 Tubing length: 150 cm Height difference between oxygenator and patient: 35 cm H2O
0.5
500 20
22
24
26
28
30
Hematocrit [%]
It is important to notice that in Figure 2 and 3 paediatric cardiopulmonary bypass blood flow is laminar up to 1 L/min, this in contrast to adult cardiopulmonary bypass where blood flow is mostly turbulent. As a consequence pressure losses will be smaller in paediatric cases and less energy will be needed for generating a given blood flow. 3.3. Arterial filter
Arterial filters were introduced during the era of bubble oxygenators. Those elderly generation oxygenators were well known sources of gaseous
27
Chapter 3 microemboli. At the end of the eighties membrane oxygenators became the standard resulting in almost no gaseous microemboli. The removal of gaseous microemboli by arterial filters is based on the concept of the bubble trap and the bubble barrier. The bubble trap concept exploits the tendency of bubbles to rise in a liquid if given the opportunity. This can be accomplished by reducing the velocity of the incoming blood so that the natural buoyancy of the bubbles becomes the dominant force. If an escape path is provided these bubbles can be eliminated. This technique can remove bubbles of 300 µm or more in diameter. Gas separation based on the surface tension phenomena at a wetted screen is employed for the removal of bubbles less than 300 µm. The mechanism takes advantage of the surface tension of the liquid. In simple terms the pressure applied across a pore of the filter screen, must be sufficient to disrupt the surface tension and only then air can be driven through the pore (Figure 4). The critical pressure or bubble point pressure, below which no air can pass the pore, is calculated by the equation: P=
4γ cos Θ D
where P is bubble point pressure [mmHg], γ is the surface tension [dynes/cm], D is the diameter of the pore [cm], Θ is the wetting angle. For most filters, Θ approaches 0 and thus cos Θ = 1.
28
Chapter 3 Figure 4 Equilibrium position Pore size [D]
wetting angle
P2
Hydrophylic material of filter screen
Θ
P1
Direction of fluid flow circumference of pore π D
γ surface tension
Θ
γ cos Θ surface tension acts at contact with pore) surface of gas bubble
For a typical system γ = 50 dynes/cm and D = 40 µm, resulting in a bubble point pressure of 37 mmHg. The pressure drop over a clean 40 µm screen is about 3 mmHg at a blood flow of 5 L/min, the wetted screen can act as a barrier to gas micro-emboli until the bubble point is reached. Any increase in pressure drop above the bubble point pressure will result in passage of the bubble, any decrease in pressure drop over the filter screen will the bubble retract from the pore.
Unfortunately in paediatric cardiopulmonary bypass the gas escape path of the arterial filter, the vent at the top of the filter, cannot be opened continuously since this will create an important arterio-venous shunt. As a consequence the arterial filter in combination with its bypass line will enlarge the circuit volume and thus the circulating blood volume of the child with
29
Chapter 3 approximately 50 mL. This volume increase represents approximately 25% of the total circuit volume.
However, the microporous fibres of the membrane can actively remove gaseous microemboli. When blood enters the oxygenator its velocity will be reduced, in the same manner as in an arterial filter, due to the larger open area for blood flow. When gas comes into contact with the microporous fibres it will be transported through the micropores due to the pressure difference between the blood and gas side. This process is in function of pressure drop, contact area and the availability of gas exchange fibres at the entrance of the oxygenator.
3.4. Conclusions
The use of hydrodynamic formulas for the calculation of tubing length and diameter allows the surgical team to define the best possible solution for a given clinical situation based on desired pressure drop and flow pattern. The use of an arterial line filter is debatable since it is a passive device that cannot operate with open vent line during paediatric cardiopulmonary bypass. The exclusion of the arterial filter in combination with an adequate choice of tubing will result in an important reduction of dead volume and less haemodilution, leading to a reduced use of homologous blood products.
30
Chapter 3 References 1. JE Brodie, RB Johnson. In The manual of clinical perfusion. Augusta, Glendale Medical Corporation, 1994, 9-14. 2. P Dierickx, D De Wachter, P Verdonck. Fluid mechanical approach of extracorporeal circulation. Course notes Institute Biomedical Technology, Hydraulics laboratory Ghent University, 1998. 3. Elliot M. Minimizing the bypass circuit: a rational step in the development of pediatric perfusion. Perfusion 1993; 8: 81-86 4. Tyndal M, Berryessa RG, Campbell DN, Clarke DR. Micro-Prime Circuit Facilitating Minimal Blood use during Infant Perfusion. J. Extra-Corpor. Technol. 1987, 19: 352-357
31
Chapter 3
32
Chapter 4
Chapter 4 Oxygenation by artificial lung systems The artificial lung or oxygenator is the most technical part of the cardiopulmonary bypass circuit. The design objectives of the “ideal” oxygenator are still the same as in 1962 when Galletti and Brecher [1] described, the “ideal” oxygenator as one that provided: oxygenation of venous blood, carbon dioxide elimination, minimum blood trauma, small priming volume and safety. Today almost 100% of the oxygenators used are membrane oxygenators [2]. Meaning that a membrane separates the gas and the blood phase. The majority of devices use a microporous hydrophobic membrane. Beside the function of gas exchanger most devices incorporate a heat exchanger and a reservoir. Thus the oxygenator performs all major functions of the natural lungs except for their endocrine function, which can be suspended for a short time without major ill effects. 4.1. The venous reservoir
There are two basic types of venous reservoirs: closed and open. The closed system consists of a PVC bag with an in and outlet and one or more venting ports for the evacuation of air. Advantages of the closed system are almost no blood-air interface, small foreign surface area; collapse of the outlet when the reservoir is suddenly emptied; quick indication of fluid changes and the ability of volume controlled weaning from cardiopulmonary bypass. Disadvantages are a more difficult air removal, when air accidentally enters the system, and the need for an additional cardiotomy reservoir.
33
Chapter 4 The open system is in essence a reservoir open to the atmosphere with incorporated cardiotomy reservoir. This system is somewhat easier to set-up than a closed system and avoids the use of an additional cardiotomy reservoir. When accidentally large amounts of air enter the reservoir, this can be faster removed than in a closed system. The major disadvantages are the large foreign surface area, the hold-up of volume in filter and defoamer and the risk of inadvertently pumping air. 4.2. The heat-exchanger
The working principle of a heat exchanger is based on the principles of conduction and forced convection. Water is used to control the temperature of the blood. A common misconception is that the blood side is the determining factor for performance. The water side is as important because it is desirable to have high flows and turbulent flow to promote conductance. On the blood side it is important to maintain laminar flow to minimise blood component damage, but also to keep the total cross sectional area for blood flow as small as possible to increase conductance [3]. The material used for the separation between the blood and water flows should be as thin as possible for the highest conductance, with a very high thermal conductivity, yet still have the integrity to withstand the expected water and blood side pressures without failure. Unfortunately, the most haemocompatible materials used in extracorporeal blood handling devices have very poor thermal conductivities (k)(Table 1). There is a trend to use more polymeric heat exchangers since these can be
34
Chapter 4 more easily coated or surface modified compared to metal heat exchangers [3], what makes them more haemocompatible. Table 1: Thermal conductivity of different materials Material
K (W/m.K)
Stainless steels
15.1
Aluminium
237
Polycarbonate
0.2
Silicone
0.2
Epoxy
0.2
Polyurethane
5
4.3. The gas exchanger
The intrinsic physico-chemical and transport characteristics of the membrane, the fluid dynamics and the haemocompatibility of the membrane module will all determine its final mass transfer. As soon as blood gets into contact with the hydrophobic polymeric surface, the material will adsorb proteins. The amount of the protein layer and the nature of the proteins that are adsorbed will depend on the physico-chemical characteristics of the membrane and on the fluid dynamics in the membrane module. Poor fluid dynamics in the blood flow channel of the membrane module will affect dramatically its performance [4] because of:
35
Chapter 4 1. High blood boundary layer resistance to mass transport. This remains extremely important since the resistance to mass transfer in a microporous membrane oxygenator is fluid bound. 2. Poor haemocompatibility. High shear rates, eddy formation and stagnation will favour the occurrence of clotting [5] 3. Large membrane surface. This will is needed for obtaining enough mass transfer but will on the other hand cause activation of the complement system [6,7] In order to obtain the best possible fluid dynamics, most manufacturers use today extra luminal flow (ELF) designs. In this design blood is flowing outside regularly spaced hollow fibres. The hollow fibres are delivered knitted together in a double layer mat. The membrane module is manufactured by wrapping a double layer hollow fibre mat around a solid core, which is then inserted into a cylindrical shell. In these modules blood flows through the membrane mesh while gas flow is fed counter-currently into the hollow fibres. Since flow through the membrane mesh will be forced to flow partly along and partly around each hollow fibre secondary flows will be generated. This particular membrane arrangement induces mixing in every section of the membrane module to an extent that will depend on the membrane angle with respect to the main direction of blood flow [4]. The efficient destruction of boundary layers by this “static mixer” configuration leads to reduced resistance to mass transfer [8] and yields high transfer rates across the membrane. Aside of the better mass transfer this design has also lower pressure drops at the blood side and no sharp edges in the blood flow path resulting in a better haemocompatibility.
36
Chapter 4 The introduction of this ELF design in paediatric oxygenators has resulted in an important reduction of the total priming volume of the paediatric cardiopulmonary bypass (Figure 1). Figure 1. Evolution of priming volume in the University Hospital in Gent 500 Priming volume [mL]
450 400 350 300 250 200 150 100 50 0 1990
1991
1993
1994
1995
2002
Nevertheless the priming volume and the total amount of foreign material remains large in even the smallest circuits (Figure 2). Each millilitre of blood in such a small paediatric cardiopulmonary bypass is still exposed to more than three times the amount of foreign surface compared to an adult circuit. This will have a major impact on the inflammatory response [7].
25000
16 14
20000
12 10
15000
8 10000
6
cm²/ mL blood
Blood volume (mL) - Surface area (cm²)
Figure 2. Relationship between blood volume and foreign surface
4
5000
2 0
0 Adult (70 Kg) Blood volume
Baby (5 Kg) Surface
Ratio
37
Chapter 4 4.4. Fluid dynamics and shear stress
As pointed out when describing the gas exchanger module, fluid dynamics is an important item for obtaining optimal mass transfer. However an oxygenator does consist of different components which must be connected. At the same time, blood has to be evenly distributed over the heat exchanger and through the membrane mesh by manifolds. As a result blood velocity will change when blood passes through the oxygenator and this may result in zones of stasis, eddy formation and or high shear. The average shear stress at the wall in a membrane oxygenator can be calculated by starting with the general macroscopic force balance for flow in a tube [9]. However, tube flow does not accurately represent the complex flow in an ELF oxygenator. Flow through an oxygenator can be considered as flow through a porous medium. According to Bird [10] the shear stress in each oxygenator was calculated by considering the flow equivalent to the flow in a packed column governed by: τ=
R h∆ P L
where: ∆P = pressure drop [N/m²], L= blood path length [cm], R h = hydraulic radius [cm]
Rh =
Q (25 / 6) µL (∆P)ε ( Ae)
where: ε = porosity of membrane area that fills that cross section Q = volumetric pump flow [L/min] µ = dynamic fluid viscosity [N/m².s] Ae = cross sectional area for flow [m²] 25/6 = experimental derived factor.
38
Chapter 4
Mockros proposed a different formula for calculating shear in an oxygenator [11,12]. 1 ∆PQµ 2 τ = V where: V = volume oxygenator [L] The average shear for two different neonatal oxygenators calculated by both formulas are given in table 2. Table 2: Characteristics of two neonatal oxygenators
Parameter
Dideco D901
Polystan Safe Micro
Membrane Surface area, m² Heat Exchanger Area, m² Void volume Total priming volume, cm³ Blood Pressure Drop @ 0.8 lpm (mmHg) Blood Pressure Drop @ 0.6 lpm (mmHg) Blood Pressure Drop @ 0.4 lpm (mmHg) Blood Pressure Drop @ 0.2 lpm (mmHg) Blood Path Length oxygenator, cm Average Cross sectional area for flow, cm² 1 τ oxygenator (Ben Brian) [dynes/cm²] 1 τ oxygenator (Mockros) [dynes/cm²] τ membrane compartment (Ben Brian) 1 [dynes/cm²] τ membrane compartment (Mockros) 1 [dynes/cm²]
0.34 0.02 0.58 60 95 65 40 20 30 12 18 25 31
0.33 0.05 0.48 52 51 35 21 9 15.3 7.62 17 20 19
7
5
Although the calculated values are comparable with those in blood vessels (see chapter 2), these values are average values and do not exclude that at certain points in the design shear stress is above the critical level of 75 – 100 dynes/cm² needed to activate white blood cells and platelets.
1
10
dyne N =1 cm² m² 39
Chapter 4 Every extracorporeal device will have a flow window with “ideal shear”. If shear is to high platelets and blood elements will be damaged but when shear is too low platelets will be more easily adsorbed by the material. As explained earlier not only the magnitude of shear stress is important but also the exposure time to this absolute value. It is well known that high shear for a short time period is better tolerated than average shear during a long exposure time [13]. In order to define spots with high or very low shear stress in a design computational fluid dynamics are used [14-16].
4.5. Conclusions
Major improvements in oxygenator design has led to a large reduction in foreign surface area, better haemocompatibility and enhanced mass transfer. Although fluid dynamics have improved more work should be done to locate risk zones at micro level. Computational fluid dynamics might offer the tool for obtaining this goal. Finally this may lead to the ideal paediatric oxygenator that will combine optimal fluid dynamics and thus mass transfer with a small priming volume and foreign surface area.
References 1. PM Galletti, GA Brecher. Bubble oxygenation and membrane oxygenation. In: Heart lung bypass; principles and techniques of extracorporeal circulation. New York: Grune and Stratton; 1962: 108-120.
40
Chapter 4 2. Giovanni Cecere, Robert Groom, Richard Forest, Reed Quinn, Jeremy Morton. A 10-year review of pediatric perfusion practice in North America. Perfusion 2002; 17: 83-89. 3. RL Rigatti, R Stewart. Heat exchange in extracorporeal systems. In: Cardiopulmonary bypass Principles and techniques of extracorporeal circulation. Ed. CT Mora. New York: Springer Verlag; 1995: 247-256. 4. G Catapano, A Wodetzki, U Baurmeister. Blood flow outside regularly spaced hollow fibers: the future concept of membrane devices. The Int J Artif Organs 1992; 15: 327-330. 5. HL Goldsmith. The effects of flow and fluid mechanical stress on red cells and platelets. Trans ASAIO 1974; 20: 21-26. 6. A Mahiout, H Meinhold, M Kessel, H Schulze, U Baurmeister. Dialyzer membranes: effects of surface area and chemical modification of cellulose on complement and platelet activation. Artif Organs 1987: 11: 149-154. 7. J Sonntag, I Dähnert, B Stiller, R Hetzer, PE Lange. Complement and contact activation during cardiovascular operations in infants. Ann Thorac Surg 1998; 65: 525-531. 8. WJ Dorson, KG Larsen. Secondary flows in membrane oxygenators. In Mechanical devices for cardiopulmonary assistance. Eds. RH Bartlett, PA Drinker, PM Galletti Adv. Cardiol., vol 6, pp 17-39 Karger, Basel 1971. 9. BF Brian. Comparative analysis of shear stress and pressure drop in membrane oxygenators. White paper. Cobe Laboratories, Inc. 1995. 10. RB Bird, WE Stewart, EN Lightfoot. In: Transport phenomena. John Wiley & Sons, NY, 1960.
41
Chapter 4 11. JM Ramstack, L Zuckerman, LF Mockros. Shear induced activation of platelets. J Biomech 1979; 12: 113-125. 12. M Bluestein, LF Mockros. Hemolytic effects of energy dissipation in flowing blood. Med Biol Eng 1969; 7: 1-6. 13. VT Turitto, CL Hall. Mechanical factors affecting hemostasis and thrombosis. Thromb Res 1998; 15: S25-31. 14. MS Goodin, EJ Thor, WS Haworth. Use of computational fluid dynamics in the design of the Avecor Affinity oxygenator. Perfusion 1994; 9: 217-222. 15. PW Dierickx, F De Somer, DS De Wachter, G Van Nooten, PR Verdonck. Hydrodynamic characteristics of artificial lungs. ASAIO Journal, 2000; 46(5): 532-535. 16. Peter W Dierickx, Dirk S De Wachter, Filip De Somer, Guido Van Nooten, Pascal R Verdonck. Mass Transfer Characteristics of Artificial Lungs. ASAIO Journal, 2001; 47(6): 628-633.
42
Chapter 5
Chapter 5 Systemic inflammatory response At the moment cardiac surgery starts; the baby is aggressed by many factors. This agression by both surgery and cardiopulmonary bypass results in an inflammatory response. There is little doubt that this inflammatory response is responsible for a proportion of the mortality and morbidity associated with cardiac surgery. Certain organs and tissues are at higher risk of developing deranged function after the perfusion and in the postoperative period. At the greatest risk are the formed elements in the blood, the platelet and white cell, resulting in clotting problems and abnormal organ and tissue functions. In particular the pulmonary system, heart and myocardium, kidney and splanchnic bed, and the brain and cerebral circulation are specifically affected and thus contribute to early postoperative morbidity and mortality [1]. Small babies are even more at risk due to the larger volume and foreign surface area of the extracorporeal circuit in combination with the immaturity of many organs systems and the large amount of blood that after contact with tissue is returned into the systemic circulation. The bio-incompatibility of cardiopulmonary bypass is multifactorial (Figure 1) and can be divided in two major groups: material independent and material dependent [2].
43
Chapter 5 Figure 1.Bioincompatibility of paediatric cardiopulmonary bypass is multifactorial
Pathophysiology and bioincompatibility of CPB
Ca lati nu on -
Temp.
Shear Stres s
Shed and/or Suctione Blood B d lood. Air
Sterility
Tissu Factor F e actor
Surgery Coag relate d CardioDrugs
plegia
Open vs Close d
Materials relate d Surfac Area(s) eArea(s)
Debris
Emboli
Circuit relate d Roller vs Centrif .
Stasi Point P s oint s
Pulse vs Non
Patient Genetic relate d
5.1. Material dependent
Under normal conditions, when blood is in a blood vessel with intact endothelium, no activation of blood proteins or elements will occur. However, the moment blood leaves this protected environment and comes into contact with damaged endothelium, other tissue or artificial surfaces several cascades of reactions will start. At the same time the shear stresses that work beneficial when applied on endothelium by releasing mediators such as nitric oxide, will now in absence of the endothelium activate blood elements. Aspects that contribute to this activation cascade are the surface characteristics [3], the sterilisation method and the chemical composition of the surface of the polymer. It is important to notice that there can be major chemical differences between the bulk material and the surface.
44
Chapter 5 5.2. Blood interactions with polymers
5.2.1. Protein adsorption and complement activation As soon as blood comes in contact with the hydrophobic polymer surfaces of the cardiopulmonary bypass the latter will be almost immediately covered with proteins. The formation of this protein layer is followed by the adherence of platelets. In addition to fibrinogen, γ-globulin preadsorbed to artificial surfaces enhances the platelet release reaction in vitro. In contrast, serum albumin passivates the surface towards platelet adhesion [4]. Glycosyl transferase reactions involving incomplete terminal oligosaccharide units were postulated as mediators for these platelet-protein interactions. These groups are present in fibrinogen, γ-globulin and many other glycoproteins in plasma, but are absent in albumin [5]. The highest concentration of fibrinogen on the material is realised after 15 minutes [5]. Fibrinogen adsorption has been used as a measure of thrombogenicity of materials. Aside from its role in the fibrin formation it will bind blood platelets via their surface glycoproteins IIb/IIIa and Gib [6]. However the platelets do not seem to interact with the material directly but through the adsorbed protein layer. In high flow rate conditions it seems that the platelet response is a major determinant of blood incompatibility with artificial surfaces [2]. At this point it is important to put in perspective the effects of shear stresses near the wall of the hydrophobic polymers since this will contribute to leukocyte and platelet activation and in exceptional situations red blood cell lysis. Although complement is activated to a large degree via the alternative pathway, it is only to a minimal extent, in adult surgery, linked to the foreign
45
Chapter 5 materials. Other pathways must be playing a role in the complement activation such as factor XIIa, kallikrein and tissue factor [7]. Likewise, C3a and C5a anaphylatoxins may appear to be reduced in plasma where in reality they are adsorbed by the protein layers and thus are measured in lower amounts [8]. 5.2.3. Contact activation The intrinsic coagulation cascade as well as the fibrinolysis system are both initiated by the contact activation phase. Four proteins are activated during the contact phase: factor XII, high molecular weight kininogen (HMWK), prekallikrein and factor XI [9]. Adsorption of factor XII in presence of prekallikrein and HMWK produces active proteases, factors XIIa and XIIf [10]. In a feedback loop, factor XIIa cleaves prekallikrein to produce kallikrein and HMWK to produce bradykinin, a short acting vasodilator. Factor XIIa in the presence of kallikrein and HMWK also activates factor XI to factor XIa activates the intrinsic coagulation cascade, which proceeds through factor IX to activate factor X and form thrombin [10]. Electrical charge (cationic or negatively charged surface) and the hydrophobicity of the artificial surface can also promote this initial contact activation with foreign material. The contact activation phase, as seen previously by factor XII and kallikrein, will also directly activate the complement system and initiate the plasminogen/plasmin formation. Contact activation may be more prominent at low flow than high flow conditions. Interestingly, recent research [11-12] shows a much lower activation of the intrinsic pathway but on the other hand the activation pathway with KK and FXII on leukocytes may be more that what has been shown so, far.
46
Chapter 5 5.3. Material independent
Other factors that influence the degree of inflammatory response do not depend on the material but are equally or more important for the initiation of an inflammatory response. A very aggressive activator is the cardiotomy suction. Especially in paediatric surgery the amount of blood recuperated by the cardiotomy reservoir can be quite large due to additional blood vessels (e.g. left vena cava superior), flow through collateral vessels etc. This aspirated blood is contaminated with tissue factor, tissue and fat fragments, free plasma haemoglobin, thrombin, tissue plasminogen activator and fibrin degradation products. All these elements in combination with the turbulent flow and the blood-air mixing in the aspiration lines will activate, through blood platelets and leukocytes, both coagulation and complement cascades. At the same time the aspirated fat emboli are an important source of cerebral embolisation [13] which, unfortunately cannot be prevented by the use of venous or arterial filters [14-15]. Important is also the presence of high amounts of S100BB in aspirated blood originating from fat, muscle and marrow in the mediastinal blood [16]. Since it had always been postulated that S100BB was a specific marker for brain damage and that the elevated plasma levels found after cardiopulmonary bypass were caused by damage of the brain. A second factor is flow dynamics and fluid mechanical stresses (See also chapters 2 & 4). Especially stasis and eddy formation has an important impact on protein adsorption and thus on the formation of thrombi. Also shear stress is an important activator of primarily platelets and leukocytes. The magnitude and duration of shear stress will dependent from component to component 47
Chapter 5 and the blood flow characteristics in a given cardiopulmonary circuit, but will always be present to some extent. A high value with a short duration will be found in arterial cannulas while different magnitudes of shear but with longer duration are found in oxygenators and reservoirs [17,18]. Shear stress induced platelet activation is mediated by von Willebrand factor binding to platelet membrane receptors GPIb and GPIIb/IIIa [2]. Shear stress as small as 100 dynes/cm² will induce platelet and leukocyte activation [19]. A third factor is related to the use of homologous blood products and haemodilution. The risks of homologous blood transfusion such as immunobiological disorders [20] and transmission of infections are well documented [21]. Because of their young age infections caused to the use of homologous blood products should be avoided in every extent. Open-heart surgery without the use of homologous blood products is commonly performed in adults, but still difficult in small children because priming volume of the cardiopulmonary bypass circuit results in extreme haemodilution [22]. A fourth factor is related to the use of drugs. Best documented is the activation of the classical pathway of the complement system by the heparinprotamine complex. This will lead to monocyte and neutrophil activation [2, 23-24]. Finally also conduct of cardiopulmonary bypass as well as the genetic footprint of the child will play a role. The use of open or closed system [25], the oxygen tension used during cardiopulmonary bypass [2, 26-27], the cooling protocol [27, 28] and haemoglobin content when using deep hypothermic circulatory arrest [29] have all been put forward as variables that can influence inflammatory response. Of course every child is unique in his
48
Chapter 5 genetic footprint and this can interact in the way they biologically will react on the damage caused by the surgery and cardiopulmonary bypass. The haptoglobin phenotype for example will determine the capacity for binding free plasma haemoglobin [30] and might have an impact on the immune response [31]. While the different platelet PLA allelic frequency have been associated with a predisposition for increased thrombogenicity [32], increased release of IL8 and TNF after cardiopulmonary bypass [33] and more pronounced neurocognitive decline after cardiopulmonary bypass [34]. Beside these genetic factors also the pathology might influence the activation of the different cascade. The higher incidence of fibrinolysis in cyanotic children is a perfect example of the latter. Inflammatory response to cardiopulmonary bypass is considerably more complex than it seemed a decade ago. In children the analysis of inflammatory response is even more complex due to different response of neonates and children to cardiopulmonary bypass [35]. Nevertheless, it is possible, based on our present knowledge to attenuate inflammatory response. The large foreign surface area of the paediatric cardiopulmonary bypass circuit, almost 4 times more than an adult circuit, remains an important issue [5]. Changing this surface into a more blood compatible surface looks promising. The aims of such a re-engineering should be elimination or reduction of [2]: 1. Plasma protein adsorption in order to reduce cellular activation 2. Coagulation activation 3. Complement activation 4. Leukocyte activation
49
Chapter 5 While at the same time the physical properties of the various bulk polymers are preserved. Different approaches have been published in order to achieve these goals. Best known is heparin coating of polymers. In adults non-uniform results have been published over the years [36]. This might be related to the fact that in most clinical studies aspirated blood, recognised as one of the most injurious components [37], is still re-used. In paediatric open heart surgery this aspect will even gain in importance due to the larger amounts of aspirated blood. Nevertheless, lower inflammatory response is reported with heparin coated paediatric cardiopulmonary bypass [38-41], although not for all markers [42]. Also the use of phosphorylcholine coating was reported to be beneficial [43]. The attractive idea of combining surface amelioration with separation of aspirated blood for further reduction of the inflammatory cascade has not been realised yet due to technical limitations. More controversial is the use of ultrafiltration for removal of inflammatory mediators [44-45] especially when compared to cardiopulmonary bypass circuits with a low priming volume and reduced foreign surface area. A last method to control inflammatory response is by pharmacological interaction. Aprotinin has been reported to attenuate cellular and humoral response to cardiopulmonary bypass both in adult [46] and paediatric [47-49] populations. Also the use of some inhibitors [50] looks promising, but larger study cohorts are necessary to confirm these data.
50
Chapter 5 5.4. Conclusion
The inflammatory response to cardiopulmonary bypass is considerably more complex than it seemed a decade ago. The acute phase response to trauma may be an integral part of this process. Our expanding knowledge of inflammatory mediators will allow a better understanding of cardiopulmonary related morbidity and may hopefully lead to improvement of biocompatibility of cardiopulmonary bypass resulting in less injurious systemic responses and diminished organ and tissue damage.
References 1. D Royston. Systemic inflammatory responses to surgery with cardiopulmonary bypass. Perfusion, 1996; 11: 177-189. 2. S Daniel. Review of the multifactorial aspects of BioInCompatibility in CPB. Perfusion, 1996; 11: 246-255. 3. NP Ziats, DA Pankowsky, BP Tierney et al. Absorption of Hagemann factor and other human plasma proteins to biomedical polymers. J Lab Clin Med 1990; 116: 687-696. 4. V Videm, E Fosse, JL Svennig. Platelet preservation during coronary bypass surgery with bubble and membrane oxygenators: effect of albumin priming. Perfusion 1993;8: 409-415. 5. BR Young, LK Lambrecht, SL Cooper. Plasma Proteins: Their Role in Initiating Platelet and Fibrin Deposition on Biomaterials. In Cooper SL, Peppas NA, eds. Biomaterials: Interfacial Phenomena and Applications.. Washington DC: Advances in Chemistry Series 199, American Chemical Society, 1982; 317-350 51
Chapter 5 6. BR Young, LK Lambrecht, RM Albrecht et al. Platelet protein interactions at blood-polymer interfaces in the canine test model. Trans ASAIO, 1983; 29: 442-446. 7. YJ Gu, MA Mariani, PW Boonstra, JG Grandjean, W van Oeveren. Complement activation in coronary bypass grafting patients without cardiopulmonary bypass. The role of tissue injury by surgical incision. Chest 1999; 116: 892-898. 8. H Nishida, S Aomi, Y Tomizawa et al. Comparative study of biocompatibility between the open and closed circuit in cardiopulmonary bypass. Artificial Organs 1999; 23: 547-551. 9. RW Colman. Surface mediated defense reactions. The plasma contact activation system. J Clin Invest 1984; 73: 1249. 10. LH Edmunds, N Stenach. Blood-surface interface. In: GP Gravlee, RF Davis, M Kurusz, JR Utley eds. Cardiopulmonary bypass. Principles and practice Philadelphia: Lippincott Williams & Wilkins, 2000: 150-166. 11. W Gil. Inflammo-coagulatory response, extrinsic pathway thrombin generation and a new theory of activated clotting time interpretation. Perfusion 2001;16(1):27-35. 12. C Baufreton, JL de Brux. Les traitements de surface en circulation extracorporelle. RBM 1999;21 suppl 1: 20-25. 13. RF Brooker, WR Brown, DM Moody, et al. Cardiotomy Suction: A Major Source of Brain Lipid Emboli During Cardiopulmonary Bypass. Ann Thorac Surg 1998;65:1651-55. 14. DA Stump. Emboli: Their source and significance in neurological outcome. 25th Anniversary NeSECC Journal 2001; 26: 15-19
52
Chapter 5 15. EH Kincaid, TJ Jones, DA Stump et al. Processing scavenged blood with a cell saver reduces cerebral lipid microembolization. In: Annual meeting of the Southern Thoracic Association; Puerto Rico, 1999. 16. RE Anderson, LO Hansson, O Nilsson, J Liska, G Settergren, J Vaage. Increase in serum S100A-1B and S100BB during cardiac surgery arises from extracerebral sources. Ann Thorac Surg 2001; 71: 1512-1517. 17. F De Somer, L Foubert, M Vanackere, D Dujardin, J Delanghe, G Van Nooten. Impact of oxygenator design on hemolysis, shear stress, white blood cell and platelet count. J. Cardiothor.Vasc. Anesth. 1996; 10: 884889 18. YJ Gu, PW Boonstra, R Graaff, AA Rijnsburger, H Mungroop, W van Oeveren. Pressure drop, shear stress, and activation of leucocytes during cardiopulmonary bypass: A comparison between hollow fiber and flat sheet membrane oxygenators. Artificial Organs 2000; 24: 43-48. 19. JD Hellums, RA Hardwick. Response of Platelets to Shear Stress - a Review. In The Rheology of Blood Vessels and Associated Tissues Eds Gross DR, Hwang NHC.. Alphen aan den Rijn: NATO Advanced Study Institute Series - E, No 41, Sijthoff & Noordhoff, 1981 20. A Salama, EC Mueller. Delayed hemolytic transfusion reactions. Evidence for complement activation involving allogeneic and autologous red cells. Transfusion 1984; 24: 188-193. 21. JW Rasenack, HJ Schlayer, F Hettler, T Peters, AS Preisler, W Gerok. Hepatitis B virus infection without immunological markers after open-heart surgery. Lancet 1995; 345: 355-357.
53
Chapter 5 22. LA Chambers, DM Cohen, JT Davis. Transfusion patterns in pediatric open heart surgery. Transfusion 1996; 36: 150-154. 23. NC Cavarocchi, HV Schaff, TA Orszulak, HA Homburger, WA Schnell, JR Pluth. Evidence for complement activation by protamine-heparin interaction after cardiopulmonary bypass. Surgery 1985; 98(3): 525-531 24. S Ashraf, Y Tian, D Cowan et al. “Low-dose” aprotinin modifies hemostasis but not pro-inflammatory cytokine release. Ann Thorac Surg 1997; 63: 68-73. 25. H Nishida, S Aomi, Y Tomizawa et al. Comparative study of biocompatibility between the open and closed circuit in cardiopulmonary bypass. Artificial Organs 1999; 23: 547-551. 26. DT Pearson, RF Carter, MB Hammo, PS Waterhouse. Gaseous microemboli during open heart surgery. In: Towards safer cardiac surgery. Ed. DB Longmore. Lancaster, MTP Press, 1981: 325-354. 27. JM Pearl, DW Thomas, G Grist, JY Duffy, PB Manning. Hyperoxia for management of acid-base status during deep hypothermia with circulatory arrest. Ann Thorac Surg 2000; 70: 751-755. 28. RA Jonas, DC Bellinger, LA Rappaport et al. Relation of pH strategy and development outcome after hypothermic circulatory arrest. J Thorac Cardiovasc Surg. 1993; 106: 362-368. 29. T Shin’oka, D Shum-Tim, PC Laussen et al. Effects of oncotic pressure and haematocrit on outcome after hypothermic circulatory arrest. Ann Thorac Surg 1998; 65: 155-164. 30. J Delanghe, K Allcock, M Langlois, L Claeys, M De Buyzere. Fast determination of haptoglobin phenotype and calculation of hemoglobin
54
Chapter 5 binding capacity using high pressure gel permeation chromatography. Clin Chim Acta 2000; 291: 43-51. 31. M Langlois, JR Delanghe. Biological and clinical significance of haptoglobin polymorphism in humans. Clin Chem 1996; 42: 1589-1600. 32. EJ Weiss, PF Bray, M Tayback et al. A polumorphism of a platelet glycoprotein receptor as an inherited risk factor for coronary thrombosis. N Eng J Med 1996; 334: 1090-1094. 33. N Drabe, G Zünd, J Grünenfelder et al. Genetic predisposition in patients undergoing cardiopulmonary bypass surgery is associated with an increase of inflammatory cytokines. Eur J Cardiothorac Surg 2001; 20: 609-613. 34. JP Mathew, CS Rinder, JG Howe et al. Platelet PlA2 polymorphism enhances risk of neurocognitive decline after cardiopulmonary bypass. Ann Thorac Surg 2001; 71: 663-666 35. SS Ashraf, Y Tian, S Zacharrias, D Cowan, P Martin, K Watterson. Effects of cardiopulmonary bypass on neonatal and paediatric inflammatory profiles. Eur J Cardiothorac Surg 1997; 12: 862-868. 36. HP Wendel, G Ziemer. Coating-techniques to improve the hemocompatibility of artificial devices used for extracorporeal circulation. Eur J of Cardiothoracic Surg 1999;16:342-50. 37. de Haan J, Boonstra PW, Monnink SHJ, Ebels T, van Oeveren W. Retransfusion of Suctioned Blood During Cardiopulmonary Bypass Impairs Hemostasis. Ann Thorac Surg 1995;59: 901-7.
55
Chapter 5 38. HH Schreurs, MJ Wijers, J Gu et al. Heparin-coated bypass circuits: effects on inflammatory response in paediatric cardiac operations. Ann Thorac Surg 1998; 66: 166-171. 39. EA Grossi, K Kallenbach, S Chau et al. Impact of heparin bonding on pediatric cardiopulmonary bypass: a prospective randomized study. Ann Thorac Surg 2000; 70: 191-196. 40. T Ozawa, K Yoshihara, N Koyama, Y Watanabe, N Shiono, Y Takanashi. Clinical efficacy of heparin-bonded bypass circuits related to cytokine responses in children. Ann Thorac Surg 2000; 69: 584-590. 41. K Miyaji, RL Hannan, J Ojito, JP Jacobs, JA White, RP Burke. Heparincoated cardiopulmonary bypass circuit: clinical effects in pediatric cardiac surgery. J Card Surg 2000; 15: 194-198. 42. SB Horton, WW Butt, RJ Mullaly et al. IL-6 and IL-8 levels after cardiopulmonary bypass are not affected by surface coating. Ann Thorac Surg 1999; 68: 1751-1755. 43. F. De Somer, K. François, W. van Oeveren et al. Phosphorylcholine coating of extracorporeal circuits provides natural protection against blood activation by the material surface. Eur J Cardiothorac Surg 2000; 18(5): 602-606. 44. RM Ungerleiter. Effects of cardiopulmonary bypass and use of modified ultrafiltration. Ann Thorac Surg 1998; 65: S35-39. 45. MS Chew, I Brandslund, V Brix-Christensen et al. Tissue injury and the inflammatory response to pediatric cardiac surgery with cardiopulmonary bypass. Anesthesiology 2001; 94: 745-753.
56
Chapter 5 46. D Royston. Preventing the inflammatory response to open-heart surgery: the role of aprotinin and other protease inhibitors. Int J Cardiol 1996; 53: S11-S37. 47. H Mössinger, W Dietrich. Activation of hemostasis during cardiopulmonary bypass and pediatric aprotinin dosage. Ann Thorac Surg 1998; 65: S4551. 48. J Boldt. Endothelial related coagulation in pediatric surgery. Ann Thorac Surg 1998; 65: S56-59. 49. CF Wippermann, FX Schmid, B Eberle et al. Reduced inotropic support after aprotinin therapy during pediatric cardiac operations. Ann Thorac Surg 1999; 67: 173-176. 50. B Stiller, J Sonntag, I Dähnert et al. Capillary leak syndrome in children who undergo cardiopulmonary bypass: clinical outcome in comparison with complement activation and C1 inhibitor. Intensive Care Med 2001; 27: 193-200
57
Chapter 5
.
58
Chapter 6
Chapter 6 Summary and new prospectives Since the first use of a heart-lung machine for total cardiopulmonary bypass on April 5, 1951 major changes took place. While the first two patients did not survive, today cardiopulmonary bypass related mortality is almost nihil. The huge circuits with a bubble or film oxygenator requiring several litres of prime have been replaced with small membrane oxygenators and circuits that only require a few hundred millilitres of prime. As more procedures were done the knowledge and long term follow-up of patients with congenital heart disease increased. Based on these new insights, more and more children are operated in the first days or weeks of their life since this seems to have a major impact on long term survival. However, this approach confronts the clinician with a lot of technical limitations when he has to place a neonate of 2 Kg on cardiopulmonary bypass. The often still immature organs demand for further research in order to keep bypass related damage to an absolute minimum. A first major problem is that of vascular access. The small blood vessels of the child need to be cannulated without obstructing blood flow or damaging the vessel wall. What is the best design for obtaining this goal? How can one be sure that all organs are perfused, that the native heart will not be challenged by an additional afterload and that total venous return is directed towards the cardiopulmonary bypass? Appendix 1 focuses on the limitations and advantages of vacuum assisted venous return (VAVD) in small babies. VAVD makes it possible to enhanced venous return with approximately 10% mainly due to a larger pressure difference. Additionally, VAVD also allows to 59
Chapter 6
use smaller cannulas resulting in less obstruction of the blood vessel and less damage to the blood vessel wall. The combination of smaller cannulas with VAVD might result in a larger operating field for the surgeon with less back flow. For arterial re-infusion the design of a cannula is of main importance. Appendix 2 explains how arterial cannula design will affect jet formation while Appendix 3 points to the limitations of existing paediatric arterial cannulas. Large differences for their pressure flow characteristics were found based on deviations in internal diameter and design.
In paediatric cardiopulmonary bypass the oxygenator remains a problem because of his priming volume, large foreign surface area and not always optimal fluid dynamics. These problems are partly due to the fact that most if not all paediatric oxygenators are “downscaled” adult oxygenators and not specifically adapted for neonatal procedures. Appendix 4 represents the clinical
benefits
of
an
oxygenator
specially
designed
for
neonatal
cardiopulmonary bypass. The use of such a neonatal oxygenator makes it possible to construct much smaller circuits resulting in less haemodilution. Appendix 5 gives the clinical impact on blood products when using a neonatal oxygenator in combination with a small circuit. The fluid dynamics in an oxygenator are important for
achieving
optimal mass transfer
and
haemocompatibility. Appendix 6 presents a new technique for the comparison of the pressure flow relationship in oxygenators with a different design. This approach makes it possible to make more objective decisions when
60
Chapter 6
comparing different products. The impact of the new ELF membrane oxygenators on blood elements was studied in appendix 7. One can question the use of an arterial filter in a paediatric circuit as it will enlarge total priming volume without adding any additional safety. Appendix 8 suggests that the hollow fibre stack of the membrane compartment might be an acceptable alternative since it will act as a depth filter and it is able to remove gaseous emboli. This alternative will reduce priming volume without jeopardising safety.
Control of the inflammatory response is a major goal for the paediatric team. One approach coating all artificial surfaces with a coating that biomimicks the outer layer of the cell membrane leads to a reduction in complement activation and a better platelet preservation. This is reported in a dog model in appendix 9 and confirmed in the clinical setting in appendix 10. Unfortunately this coating does inhibit the inflammatory response completely and this might be explained by the findings of appendix 11 that blood coming from structures not covered with endothelium such as the pericardium and pleural cavities does activate the coagulation system. By doing so it will also activate the complement system and promote capillary leak.
Clinical implications and possible future directions
More and more new-borns with congenital heart disease are operated within the first days or weeks of life. As a result body weight can be very low and the anatomical structures will be small. Institution of cardiopulmonary bypass 61
Chapter 6
under such conditions asks for dedicated cannulas with minimal deviation of the inner diameter. In order to achieve optimal venous drainage and arterial re-infusion under all circumstances, more designs and diameters should be developed. Pressure-flow diagrams based on viscous solutions such as water-glycerine should accompany these new designs as well as existing designs. Vacuum assisted venous return in combination with dedicated venous cannulas will further reduce the total priming volume of the cardiopulmonary circuit and more importantly also reduce the “dead volume” in aspiration lines. As a result blood will be exposed to a lower amount of foreign material and less haemodilution of coagulation proteins and blood elements will occur. Due to the lower haemodilution less homologous products are needed and exposure to multiple blood donors can be avoided. The treatment of all foreign material with a biocompatible coating will reduce inflammatory response. Future developments should focus on 1. Membrane technology: microporous versus diffusive 2. Surface treatment of all foreign surface 3. Integration of components and miniaturisation of the cardiopulmonary bypass for further reduction of priming volume and foreign surface 4. Fluid mechanics of the complete cardiopulmonary bypass circuit combined with extensive modelling of the fluid mechanics in each component 5. Cannulas in combination with the physical and biological aspects of vascular access in general 6. Selective blood treatment for activated blood 62
Appendix 1
Evaluation of different paediatric venous cannulas using gravity drainage and VAVD: an in vitro study
F. De Somer, D. De Wachter, PR Verdonck, G. Van Nooten, T. Ebels
Perfusion, 2002; 17(5): 321-326
63
Appendix 1
Abstract
Six different commercially available paediatric venous cannulas together with a special constructed cannula were tested in vitro for their pressure-flow relationship. With the cannulas placed in an open reservoir, flow increased with larger diameters and higher pressures. At a pressure of 30 cm H 2O flows were 219 ± 20 mL/min, 285 ± 13 mL/min, 422 ± 11 mL/min, 728 ± 4 mL/min, for the 12 Fr, 13.2, 14 Fr and 16 Fr, respectively. No differences were founds between angled and straight cannulas. When the cannulas were tested in a latex model simulating the right atrium and venae cavae, the highest flow obtained by gravity was 164 mL/min using an angled 14 Fr cannula. When vacuum was applied to augment venous return a maximum flow of 179 mL/min was measured using an angled 14 Fr cannula. Collapse can occur when the pressure difference becomes too high in the test system. This is important since most children are selectively cannulated in both major veins. Monitoring of the intravascular pressure might help to prevent collapse. A larger diameter venous cannula does not always produce the highest flow when placed in a vein. This is most obvious when augmenting venous return. The design of the cannula tip in combination with VAVD can affect the venous return.
64
Appendix 1
Introduction
Vascular access remains an important aspect of cardiopulmonary bypass (CPB) in paediatric cardiac surgery. Bi-caval cannulation with straight or angled cannulas, using gravity siphon drainage, is most often used. In the last decade major improvements have been made to decrease the extracorporeal blood volume [1], as a result of which the volume of the tubing becomes more important. Once the diameter required for a calculated flow be chosen, the only way further to decrease this volume is to shorten the length of the lines. This can be achieved by using active vacuum augmentation of the venous return, which allows the user to place the oxygenator closer to the patient [2]. Vacuum assisted venous drainage gained renewed interest since the start of minimally invasive cardiac procedures [3,4], and for reduction of priming volume in paediatric circuits [5-8]. Veins are compliance vessels and will collapse at negative pressures between minus 5 – 10 mmHg [2]. When using gravity drainage, the pressure in the vein(s) will be more or less constant during the procedure. However, when vacuum is applied as driving pressure much lower pressures can be achieved. As a result the veins can collapse and instead of an increase a reduction in flow, due to partial obstruction of the open area for flow by the vein, will result. In this study we investigate the influence of vein collapse, cannula diameter and exerted negative pressure on the venous return in vitro.
65
Appendix 1
Methods
Single stage paediatric venous cannulas (Medtronic, Brussels, Belgium) in three sizes (12, 14 and 16 French, wall thickness 0.025”) and two configurations (straight (DLP 661xx) and right angled (DLP 675xx)) were tested, together with a special constructed cannula. The latter consists of a plastic helix with a diameter of 15 Fr mounted on 24 cm of 1/8 inch tubing. This 1/8 inch tubing had an inner diameter of 9.6 Fr and an outer diameter of 13.2 Fr (Figure 1). Due to the use of PVC tubing as connection between the tip and the venous line, the wall thickness of the specially constructed cannula is much thicker than that of the commercially available cannulas. This cannula was only available in straight configuration and was used to investigate the potential benefit of a design less prone to obstruction in case of collapse of the vein. In the text this cannula will be referred to as 13.2 Fr. A first group of measurements validated all cannulas for their pressure flow relationship. The test fluid was a 30% glycerine solution with a kinematic viscosity of 2.5 mm2/s, which is similar to blood. The test cannula is placed horizontally in a reservoir, while the level in the reservoir is kept constant by means of an overflow. The flow rate through the cannula is regulated by the height of the collecting chamber, which could be placed as low as 30 cm below the cannula. Pressure is measured at the tip and the end of the cannula by means of a differential pressure transmitter (Fuji Electric, Erlangen, Germany). The flow rate is obtained gravimetrically by a timed fluid mass collection (Figure 2A).
66
Appendix 1
For a second experiment a model of the right atrium including both caval veins was constructed in latex. The dimensions of the model were based on the echocardiographic measurements of the right atrium and caval veins in 10 babies. The average weight of the 10 children was 5.5 ± 0.7 kg. The average diameter of the superior and inferior caval vein was 4.9 ± 0.4 mm and 5.3 ± 0.4 mm, respectively. For ease of construction both veins in the model had a diameter of 5 mm. This results in a cannula-vein diameter ratio of 0.79 for the 12 Fr, 0.87 for the 13.2 Fr, 0.92 for the 14 Fr and 1.05 for the 16 Fr cannula. For the measurements only the inferior caval vein was cannulated and a purse string was used to prevent back flow into the right atrium. If both caval veins had been cannulated, validation of the exact flow in each of the two cannulae would have been difficult. The compliance of both vessels was designed in such a way that they could collapse at a pressure of approximately minus 10 - 15 mmHg. Pressures were recorded at the tip and the end of the cannula. Flow was measured by a Transonic flow meter (Transonic®, Ithaka,NY, USA). In a first approach gravity drainage was applied with a height difference of 30 cm H2O. In a second approach a VAVD controller (Polystan AS, Vaerlose, Denmark) was used to assist venous return (Figure 2B), with the reservoir fluid level situated 18 cm below the model.
67
Appendix 1
Results (Table 1)
Pressure – flow relationship The maximum flow obtained at 30 cm H 2O with the straight cannulas was 219 ± 20 mL/min, 285 ± 13 mL/min, 422 ± 11 mL/min, 728 ± 4 mL/min, for the 12 Fr, 13.2, 14 Fr and 16 Fr, respectively (Figure 3). For the angled canulae flows of 216 ± 13 mL/min, 454 ± 7 mL/min and 727 ± 35 mL/min were obtained at 30 cm H 2O for the 12 Fr, 14 Fr and 16 Fr cannulas (Figure 3). First experiment: gravity drainage With the straight cannulas, the maximum flow before collapse occurred was 136 mL/min, 142 mL/min, 142 mL/min and 149 mL/min for the 12 Fr, 13.2 Fr, 14 Fr, and 16 Fr, respectively. With the angled cannulas flows of 131 mL/min, 164 mL/min and 151 mL/min were obtained with the 12 Fr, 14 Fr and 16 Fr cannula (Figure 4). The mean pressure at which collapse of the vessel occurred was 9.9 ± 1.2 mmHg. Second experiment: VAVD With the straight cannulas, the maximum flow before collapse occurred was 155 mL/min, 163 mL/min, 129 mL/min and 143 mL/min for the 12 Fr, 13.2 Fr, 14 Fr and 16 Fr, respectively. With the angled cannulas flows of 156 mL/min, 179 mL/min and 165 mL/min were obtained with the 12 Fr, 14 Fr and 16 Fr cannula (Figure 5). The mean pressure at which collapse of the vessel occurred was 12.4 ± 1.1 mmHg.
68
Appendix 1
Discussion
Very early in the development of cardiopulmonary bypass techniques for obtaining maximal venous return were investigated [2]. Several variables have been put forward as important, cannula/vein ratio, design of the cannula, cannula position, characteristics of the connecting system and techniques for augmenting venous return. Today 18% of all paediatric cardiac surgery is performed in the first month of life, while over 50 % is performed in children less than 1 year [3]. As a consequence of the small vascular structures an optimal venous return is mandatory for a bloodless surgical field. In general the resistance of cannulas is low in babies because of the relatively large cannula/vein ratio compared to adults. In spite of the surgical trend, it is difficult to find venous cannulas with an external diameter smaller than 12 Fr. In our series not surprisingly flow in a reservoir increased with larger diameters and higher pressures. Our pressure-flow results support previous findings showing no correlation between filling pressure and maximum flow when the cannula is placed in an open reservoir [3]. However, with the cannula placed in the model, filling pressure and cannula size does influence flow rates. When gravity drainage was applied only 5 % less flow was obtained with a straight 12 Fr cannula compared to the larger cannulas. However, the combination of an angled 12 Fr cannula with gravity drainage generated less flow compared to the larger diameter cannulas. The difference between straight and angled cannulas might be explained by the fact that a straight 69
Appendix 1
cannula tip will be easier pushed towards the wall of the vessel, especially with the large cannula vein diameter ratios, than an angled cannula when traction or manipulation is exerted. The small differences in flow found between the three cannulas might be due to the fact that the cannula vein diameter ratio was exceeding in every case 0.5. This has been shown to compromise flow [3]. Based on the pressure flow characteristics in the open reservoir one probably would have chosen a 14 Fr cannula whereas the 12 Fr performed almost as good in our experimental model. The 13.2 Fr did not improve return compared to the 12 and 14 Fr cannulas. Augmenting venous return by vacuum assist resulted in higher flows compared to gravity drainage. The difference is most pronounced in the 12 Fr and 14 Fr angled cannulas as well as with the 13.2 Fr. Due to the fact that the 16 Fr cannula has a diameter equal or even somewhat larger then the vein diameter (cannula vein ratio: 1.05), the vein will become the limiting factor. In our experimental model the smallest cannulas, being the 12 Fr and 13.2 Fr, therefore had the best performance. This might be explained by the fact that with a small cannula the greater pressure gradient will be between the reservoir and the tip of the cannula instead of between the tip of the cannula and the patient’s venous system. This might prevent severe ‘fluttering’ of the walls of the IVC around the end of the venous cannula [3,9]. The helix design resulted in excellent flow rate most probably related to its large open area for flow. Monitoring of the vein pressure is of major importance for preventing ‘fluttering’ and collapse of the vein, which will result in flow reduction and haemolysis due to high shear stress at the cannula entrance [10]. 70
Appendix 1
Unfortunately, in most articles the authors only report the pressure measured on top of the venous reservoir [3-8,10]. However this pressure will be the sum of the pressure at the tip of the cannula, the hydrostatic pressure and the pressure loss over the tubing and cannula. The latter makes it very difficult to compare results of different studies since most authors do not mention tubing length and hydrostatic pressure. The flow regimen in a piece of 3/16 or ¼ inch will be laminar with the blood flows, blood temperatures and haematocrit conditions used during CPB on a baby of 5 kg. Using the Hagen-Poiseuille equation (See appendix) one can calculate the contribution of the tubing in the pressure difference measured on top of the venous reservoir. When a baby of 5 kg, with a haematocrit of 25% and a blood temperature of 25°C, is perfused with a flow of 120 mL/kg, the pressure difference for each meter of 3/16 or ¼ inch tubing will be 15 mmHg and 5 mmHg, respectively. When the oxygenator is not at the same level of the right atrium we also have to add the hydrostatic pressure head to the total pressure difference. This is approximately 8 mmHg for each 10 cm height difference. In summary, we found that collapse can occur when the venous pressure becomes too low in the test system. This is important since most children are selectively cannulated in both major veins. Measuring vacuum at the top of the venous reservoir is not a good indicator of the pressure in the vein. Monitoring of the intravenous pressure might help to prevent collapse of the vein. A larger diameter venous cannula does not always produce the highest flow when placed in a vein. This is most obvious when augmenting venous return. The design of the cannula tip in combination with VAVD can affect the
71
Appendix 1
venous return. Development of smaller cannulas with tips adapted for the use of VAVD should be stimulated.
Limitations of the study
Although major efforts have been taken to mimic the anatomical and physiological situation, it is impossible to simulate all surgical events and their impact on venous return. The absence of vascular tonus in the model might also influence our results. For those reasons interpretation of the results must be done with caution.
Acknowledgements
The authors received sample cannulas free of charge from Medtronic (Medtronic, Brussels, Belgium) for testing purposes. Polystan (Polystan, Oelegem, Belgium) kindly offered the vacuum controller for the duration of the experiment. The authors thank Mrs. Oancea for her technical assistance.
72
Appendix 1
Appendix
The Reynolds number, which is dimensionless, represents the ratio of inertia forces to viscous forces and is calculated by: V⋅d
Re =
ν
V = velocity [m/s], d = diameter [m],ν = kinematic viscosity [m²/s] The kinematic viscosity,ν for blood is calculated according to following formula: ν =
η ρ
ρ = density [kg/m³],η = absolute blood viscosity [Pa.s] 1800 exp −5.64 + ( T + 273) η plasma = 1000 η = η plasma⋅exp( 2.31 ⋅Hct) ρ = [ 1.09 ⋅Hct + 1.035 ⋅( 1 − Hct) ] ⋅10 If the Reynolds number is below 2000 flow is considered to be laminar. For laminar flow, pressure drop can be calculated in function of diameter, length, blood viscosity and height difference between the patient and the heart-lung machine, using the Hagen-Poiseuille equation: ∆P =
128 ⋅η ⋅L⋅V 4
π ⋅D
+ ρ ⋅g⋅H
Where L = lenght [m], Q = blood flow [m³/s], D = diameter [m], ∆ P = pressure drop [Pa], H = height [m], g = gravity constant [m/s²]
By using these equations pressure drop can be calculated for a venous line in function of length, diameter, required blood flow, viscosity and desired
73
Appendix 1
pressure drop. If we have a venous line of 1 meter, a haematocrit of 25% and a temperature of 25°C, we would obtain following values:
Blood [mL/min] 0 100 200 300 400 500 600 700 800 900 1000
flow Reynolds number 3/16 inch 0 176 352 527 703 879 1055 1230 1406 1582 1758
∆P [mmHg] 3/16 inch 0.0 2.6 5.3 7.9 10.5 13.2 15.8 18.4 21.1 23.7 26.3
Reynolds number 1/4 inch 0 234 469 703 937 1172 1406 1641 1875 2109 2344
∆P [mmHg] ¼ inch 0.0 0.8 1.7 2.5 3.3 4.2 5.0 5.8 6.7 7.5 8.3
74
Appendix 1
References
1. De Somer F, Foubert L, Poelaert J, Dujardin D, Van Nooten G, François K. Low extracorporeal priming volumes for infants: a benefit? Perfusion 1996; 11: 455-460. 2. Galetti PM, Brecher GA. Connection of the vascular system with an extracorporeal circuit. Heart-lung bypass, principles and techniques of extracorporeal circulation. New York: Grune & Stratton, 1962: 171-93. 3. Kurusz M, Deyo DJ, Sholar AD, Tao W, Zwischenberger JB. Laboratory testing of femoral venous cannulae: effect of size, position and negative pressure on flow. Perfusion 1999; 14: 379-387. 4. Münster K, Andersen U, Mikkelsen J, Petterson G. Vacuum assisted venous drainage. Perfusion 1999; 14: 419-423. 5. Lau CL, Posther KE, Stephenson GR et al. Mini-circuit cardiopulmonary bypass with vacuum assisted venous drainage: feasibility of an asanguineous prime in the neonate. Perfusion 1999; 14: 389-396. 6. Darling E, Kaemmer D, Lawson S, Smigla G, Collins K, Shearer I, Jaggers J. Experimental use of an ultra-low prime neonatal cardiopulmonary bypass circuit utilizing vacuum assisted venous drainage. JECT 1998; 30: 184-189. 7. Ahlberg K, Sistino JJ, Nemoto S. Hematological effects of a low-prime neonatal cardiopulmonary bypass circuit utilizing vacuum-assisted venous return in the porcine model. JECT 1999; 31; 195-201.
75
Appendix 1
8. R Berryessa, R Wiencek, J Jacobson, D Hollingshead, K Farmer, G Cahill. Vacuum-assisted venous return in pediatric cardiopulmonary bypass. Perfusion 2000; 15: 63-67. 9. Kirklin JW, Barratt-Boyes BG. Hypothermia, circulatory arrest, and cardiopulmonary bypass. Cardiac Surgery, 2nd edn. New York: Churchill Livingstone, 1993: 76. 10. Pedersen TH, Videm V, Svennevig JL et al. Extracorporeal membrane oxygenation using a centrifugal pump and a servo regulator to prevent negative inlet pressure. Ann Thorac Surg 1997; 63: 1333-39.
76
Appendix 1
Figure 1: Special constructed helix cannula.
77
Appendix 1
Figure 2: Experimental set-up
A: Set-up for pressure-flow relationship Pressure 1
Pressure 2
Transonic Flowmeter
Cannula Reservoir 1
DPT Weight balance
Pump Differerential pressure transmitter Reservoir 2
B: Set-up for gravity and VAVD Transonic Flowmeter
Reservoir 1 latex model Purse string Right atrium Pressure 1
Pump Cannula
Pressure 2
Reservoir 2
height difference mmHg
Differential pressure transmitter
Vacuum controller Venous reservoir
Switch between gravity and vacuum
78
Appendix 1
Table 1. Performance of the cannulas in the reservoir and in the inferior caval vein when using gravity drainage or VAVD. Cannula type
Gravity
VAVD
Reservoir 12 Fr straight 13.2 Fr straight 14 Fr straight 16 Fr straight 12 Fr angled 14 Fr angled 16 Fr angled Mean collapse pressure (mmHg)
Flow at minus 30 cm H20 (mL/min) 219 ± 20 285 ± 13 422 ± 11 728 ± 4 216 ± 13 454 ± 7 727 ± 35
Maximum flow before collapse of the vein (mL/min) 136 155 142 163 142 129 149 143 131 164 151
156 179 165
9.9 ± 1.2
12 ± 1.1
79
Appendix 1
Figure 3: Pressure – flow relationship.
Straight cannulae
Angled cannulae 30
30 12 French 14 French 16 French 13.2 French
25
25
20
20 Pressure at tip [mmHg]
Pressure at tip [mmHg]
12 French 14 French 16 French
15
10
15
10
5
5
0
0
100
300
500
Flow [mL/min]
700
100
300
500
700
Flow [mL/min]
80
Appendix 1
Figure 4:
Gravity straight cannulae
12 French 14 French 16 French 13.2 French
12 French 14 French 16 French 15 Pressure drop [mmHg]
Pressure drop [mmHg]
15
Gravity angled cannulae
10
10
5
0
5
0
25
50
75
100 125 150 175 200
Flow [mL/min]
25
50
75
100 125 150 175 200
Flow [mL/min]
81
Appendix 1
Figure 5.
VAVD straight cannulae 30
VAVD angled cannulae 30
12 French 14 French 16 French 13.2 French
25
20
12 French 14 French 16 French
25
Pressure drop [mmHg]
Pressure drop [mmHg]
20
15
15
10
10
5
5
0
0
25
50
75 100 125 150 175 200 Flow [mL/min]
25
50
75
100 125 150 175 200
Flow [mL/min]
82
Appendix 2
Hydrodynamical Comparison of Aortic Arch Cannulae
P.R. Verdonck, U. Siller, D. De Wachter, F. De Somer, G. Van Nooten
Int. J. Art. Organs, 1998; 21(11): 705 - 713.
83
Appendix 2
Abstract
The high velocity of blood flow exiting aortic arch cannulae may erode atherosclerotic
material
from
the
aortic
intima
causing
non-cardiac
complications such as stroke, multiple organ failure and death. Five 24 Fr cannulae from the Sarns product line (straigth open tip, angled open tip with and without round side holes, straight and angled closed tip with four rectangular, lateral side holes) and a flexible cannula used at the University Hospital of Gent (straigth open tip) are compared in an in-vitro steady flow setup, to study the spatial velocity distribution inside the jet. The setup consists of an ultrasound Doppler velocimeter, mounted opposite to the cannula tip in an outflow reservoir. An elevated supply tank supplies steady flow of 1.3 L/min of water. Exit forces at various distances from the tip are calculated by integrating the assessed velocity profiles. The pressure drop across the cannula tip is measured using fluid filled pressure transducers. The four sidehole design provide the lowest exit velocity (0.85 vs 1.08 m/s) and forces per jet (0.03 vs 0.15-0.20 N). The round sideholes are useless as less than 1 % of the flow is directed through them. Furthermore, the use of angled tip cannulae is suggested because the force exerted on the aortic wall decreases the more the angle of incidence of the jet deviates from 90°. Pressure drop is the lowest for the 4 side hole design and highest for the open tip and increases when an angled tip is used.
Keywords aortic cannula, in vitro hydrodynamics, sandblasting effect. 84
Appendix 2
Introduction
Atherosclerotic disease of the ascending and transverse aortic arch is an important risk factor for stroke associated with use of cardiopulmonary bypass (CPB) [1,2]. Detachment of atherosclerotic material from the aortic intima can be caused by external manipulation (such as cannulation and clamping) and internal disruption. Tissue erosion in the aortic arch is caused by the highvelocity jet emerging from an aortic cannula during CPB termed the “sandblasting effect” [3]. The high speed jet is caused by the relatively small cross section of the cannula tip which is around 8 mm in outer diameter for a cannula used on adults with average blood flows of 4 to 6 l/min. During the last decade lots of effort have been performed to design better cannulae. Already in 1986 the use of a long aortic arch cannula with its tip extending beyond the origins of the arch vessels was suggested because it could avoid the hazard of stroke by directing the high-velocity blood flow down the ascending aorta and away from the cerebral arteries [3]. Muehrche et al. made a different approach by designing a new arterial cannula with four side holes specifically to reduce the velocity of blood flow and the exit force in order to decrease the sandblasting effect and to produce a more gentle highflow perfusion [4]. Nevertheless because it is difficult to verify the position of the four jets inside the patient’s aorta one or more jets might still hit calcified material. Besides the above mentioned efforts cannula design needs to be improved both hydro- and hemodynamically to reduce the rate of perioperative problems. Influencing factors are multiple: pressure drop, flow rate, jet 85
Appendix 2
velocity, geometry of cannula, tip position in the aorta, shear stress, exit force and operation time. To improve the performance of cannulae on the long term it is necessary to evaluate the relationship between the hydrodynamic parameters in an experimental setup.
86
Appendix 2
Materials and Methods
1. Tested cannulae Five models out of the six cannulae tested in this study are selected from the 3M-Sarns product line and one is a self-made cannula, named “Gent Hospital” in this paper, used at the University Hospital of Gent. The cannulae distinguish each other by their geometries, dimensions and materials. The geometry describes the general shape of a cannula. Most models are a composition of three components: a connector, a tube and a tip. On one model of the tested group the distinction between the tube and the tip is not possible because they consist of the same piece (“Gent Hospital”). The entrance of the tip is defined as the first deviation from the general tube design. Figure 1 shows a schematic picture of a cannula. For all models the connector is a standard 3/8 inch (0.95 mm) Polycarbonate connector commonly used in the clinical practice. Tips are straight or angled fitted with or without side holes. The shape of these side holes is either round or rectangular. All tips except for one (“Gent Hospital”) of the tested samples are conic to gently decrease the diameter of the tube towards the end of the tip. Important dimensions to characterize a cannula are the diameter and length of the tip and the tube. The outer diameter of the tip is measured in French (circumference in millimeters). All samples in the test are 24 French (8 mm) cannulae. In all cases PVC is used as material for both the tip and the tube. Due to different additives (softeners) the stiffness of the tested examples at room 87
Appendix 2
temperature varies from soft (easy pliable) to stiff (not flexible at all). An overview of all tested models is summarized in Table 1 in which GOA represents the geometric orifice area calculated as π r2, with r the diameter of the tip, augmented with the area of side holes if present. The flow through a tube with inelastic walls depends on the velocity of the fluid and the effective outflow area EOA which does not necessarily equal the GOA. For the side hole cannulae the EOA will be determined experimentally from velocity flow measurements.
2. Experimental setup The setup consists of three parts: - a system of reservoirs, tubing and a centrifugal pump to provide a constant flow; - an ultrasound Doppler velocimeter to measure the velocities of the jet at various distances away from the tip and to visualize the contour of the jet; - a data acquisition system to assess the pressure drop over the tube and the tip as a function of the flow. Figure 2 gives a schematic overview of the experimental setup. The water is raised by a centrifugal pump from a tank to the upper reservoir where it enters a vertical tube of 1250 mm in height. The water column in the tube provides a constant bottom pressure of 95 mmHg because it is fitted with an overflow that leads any surplus water back to the tank. An array of two valves is attached to the outflow at the bottom of the reservoir to allow an accurate adjustment of the flow. The connection to the cannula is made by the same 3/8 inch tube that is used on extracorporeal circulation systems in the 88
Appendix 2
clinical practice. It is attached to a luer lockport that permits the introduction of two fluid lines into the cannula. The cannula is inserted into the outflow reservoir either through an opening in the wall (straight tip) or from the top (angled tip). Inside of the container it is fastened with a horizontally and vertically adjustable clamp. The variable support is necessary to position the tip of the cannula exactly opposite to the ultrasound probe. In order to have the option to measure the jet from the side and from below two extra openings for the ultrasound probe are intended at the bottom and on the sidewall respectively. The Plexiglas window on one side of the container makes it possible to see the tip of the cannula. A cylindrical reservoir is chosen to keep disturbing reflections of the ultrasonic waves from sharp edges low. The outflow reservoir offers an inner diameter of 200 mm and measures 590 mm in height. The water level inside of the container must stay constant to apply a positive back pressure on the tip of the cannula. This is realized by an overflow which is also connected to the tank by a plastic tube. By pulling the overflow pipe out of its socket the reservoir can be emptied quickly. The pressure is measured at two different positions inside the tested cannula with fluid lines connected to piezoresistive transducers. The pressure in the reservoir at the level of the tip can be computed by measuring the height of the water column above the center of the cannula. Knowing the pressure in three points (reservoir, connector and the beginning of the tip) for a given flow makes it possible to calculate the pressure drop across the tip ∆ptip, the tube ∆ptube including the connector ∆pconnector and the total length of the cannula: ∆pTube = pConnector - pTip 89
Appendix 2 ∆pTip = pTip - pReservoir ∆pCannula = pConnector - pReservoir The output of the pressure readings takes place numerically and graphically on the screen of the system with an accuracy of ± 0.5 mmHg within a range of ± 50 mmHg. The data acquisition software is developed at the Hydraulics Laboratory of the University of Gent. A clamp-on ultrasound flow probe (Transonic 3/8" Transonic Systems, Ihaca, New York) attached to the connection tube between the upper reservoir with the cannula is used to measure the average flow. To ensure that the flow is fully developed at the position of the flow probe (even for laminar flow conditions) the sensor is placed at a distance of one meter away from the origin of the tube. The velocities inside the jet are measured with (PWD) Pulsed Wave Doppler echography (Vingmed CFM 800). All measurements are performed in a detailed way and only high velocities (set arbitrarily above 0.9 m/s for all open tip and round side hole cannulae, and to 0.6 m/s for all rectangular side hole cannulae) are studied. The sampling volume is moved along a scan line running parallel to the symmetry line of the cannula (reference line). The first measurement is taken at this position which is still inside of the cannula. It is recorded as picture number one. Picture number two is located one cursor step to the left at the same distance away from the probe. The angle between two scan lines is 1°. The cursor is moved further to the left until the detected velocity is lower than 0.9 m/s or 0.6 m/s respectively. The part of the jet on the right side of the reference line is scanned in the same manner. The same procedure is carried out for all other distances from the transducer. The 90
Appendix 2
deepness is changed in intervals of 10 mm. Figure 3 shows the measuring points that are accessed by using the protocol described above. All measurements are performed for a constant flow of 1.3 l/min of water with 2 % cornstarch to improve the image quality. Besides PWD measurement also Color Flow Doppler (CFD) image are studied in two perpendicular planes, a horizontal and a vertical one.
3. Exit force The harmfulness of the sandblasting effect depends on the vector of the exit force of the jet that is perpendicular to the aortic wall (cosine term) (figure 4). This is in contrast with the shear stress, which is caused by friction force oriented laterally with respect to the wall. By definition the force is the product of the pressure p and the area A the pressure works on: F = pA cosα = ρ u2 A cosα where u represents the velocity and α the angle between the cannula and the horizontal plane (the angle of incidence). This can also be written as a differential equation: dF = ρ u2 cosα dA The force is a function of the angle of incidence and the radius and velocity of the jet. In addition the force depends on the distance from the point of the tip which makes it a three dimensional problem. To be able to integrate the formula it is assumed that the jet has a circular cross section which is a function of the radius (A = πr2 so that dA = 2πrdr). Furthermore an integration of the equation is only possible if a functional 91
Appendix 2
relationship for the velocity profile is given. For most situations the following profile is applicable assuming a turbulent velocity profile : u(r) = (1 - r/R)1/n Umax with R being the maximum radius and Umax the maximum velocity. Prandtl derived n to 7 from Blasius’ law of friction [5]. Rearranging the equations and inserting the expressions for the assumed velocity profile leads to: 2 /n + 1 2 /n + 2 2π ρ U 2max R r R r F(r) = + - 1 cos α r 1 - 1 - 2 /n + 1 R 2 /n + 2 R
The force at a certain depth is obtained for r = R (n = 7): 49 F = π ρ U 2max R 2 cos α 72
92
Appendix 2
Results
Figure 5 shows the comparison between the calculated geometric outflow area (GOA) and the measured effective outflow area (EOA) for all tested cannulae. These values deviate for cannulae with side holes. Figure 6 displays the high velocity core of the jet obtained with PWD in the horizontal plane (left panel) and the vertical plane (right panel) for the Sarns 9484. As mentioned before only a limited range of velocities is measured for each jet. Plotting these velocities in a chart according to their measuring position gives a good impression of the spatial distribution of the maximum of the jet which is referred to as the core of the jet. Distances measured in respect to the position of the ultrasound probe are drawn on the left hand side of the diagram (Fig. 6) and distances measured from the point of the tip are given on the right hand side with the tip being positioned at zero. At the boundaries of the jet flowing water starts to mix with the resting fluid in the reservoir. The core decreases in size with increasing distance from the exit of the cannula. Nevertheless it is observed that even at a distance of 90 mm away from the tip the maximum of the velocity in the center of the jet remains constant. This suggests that a velocity drop-off regarding the peak velocity is not yet present at that distance. These measurements together with the CF Doppler measurements show an axisymmetric jet indicating that the angled cut of the point of the tip does not effect basically the profile of the jet. Peak velocities of all the cannulae vary between 0.85 m/s (“Sarns 5847" and “Sarns 5774") and 1.08 m/s (“Gent Hospital”) for a steady water flow of 1.3 l/min (blood flow of 4 l/min). Figure 7 summarizes the measured peak velocities. 93
Appendix 2
For all cannulae it is observed that the peak velocities in the vertical plane are slightly higher compared to the values measured in the horizontal plane because reflections from the bottom of the reservoir and the water surface are higher in this case. An estimation of the maximum exit force of the tested cannulae is obtained from the average measured values for the velocities and the available cross section (Tab. 1). Figure 8 shows the variation of the exit force as a function of the radius. The peak pressure drop over the cannulae varies between 14.4 (“Sarns 5847") and 31.8 mmHg (“Gent Hospital”) for a water flow of 3 L/min which corresponds to a pressure drop of 132 - 291 mmHg for a blood flow of 9 L/min. Figure 9 shows the measured pressure drop for all cannulae. Besides the pressure versus flow chart (Fig. 10, left panel) the data is also represented in a dimensionless way where the Euler number is a function of the Reynolds number (Fig. 10, right panel). The Euler number is defined as Eu =
∆p and ρ u2
the Reynolds number as Re = ρ u D/µ where ρ represents the density, u the velocity, D the internal diameter and µ the dynamic viscosity. The graphs for the pressure drop over the tube, the tip and the cannula are drawn separately. As already can be seen from Figure 9 and 10 the total pressure drop across the cannulae is mainly influenced by the tip except for the model “Gent Hospital” where the tube plays the most important part because of its long and thin origin. Figure 11 summarizes the measured water flow across the cannulae for a
94
Appendix 2
constant pressure drop of 10.9 mmHg.
95
Appendix 2
Discussion
Sandblasting effect: peak velocities and exit force Two out of six cannulae (“Sarns 5847" and “Sarns 5774") tested in this study feature a new tip design where the blood flow exits through four lateral side holes and not through an open tip like on the other models which is believed to reduce peak velocities and exit forces [4]. To verify this statement the spatial velocity distribution of all cannulae is measured with an ultrasound Doppler velocimeter and the exit forces at certain distances from the tip are calculated out of the peak velocity and the diameter of the jet by means of an integration. Although aortic arch cannulae are routinely used during open heart operations the importance of their hydrodynamics is somewhat overseen. In literature a few papers are directly dealing with this topic. This is in contrast with the knowledge that in terms of clinically significant central nervous system dysfunction the most important embolic hazard of open heart operations in the current area is atheroembolism from the ascending aorta. The term “sandblasting effect” is used to describe the erosion process. In fact this is misleading because it is actually the high pressure exerted on the aortic wall rather than particles accelerated by a fluid (which would be the technical meaning of the expression) that causes the dislodge of material. Erythrocytes are too small and soft as to act like sand corns in the blood jet. Since the terminology “sandblasting effect” is found in all of the reviewed papers it seems to be an accepted phrase in the clinical field despite of its actual meaning. 96
Appendix 2
To quantify the significance of the sandblasting effect Grossi et al. measured intraoperatively the flow in the aortic arch of 18 patients undergoing CPB by by epiaortic ultrasonography [3]. All patients were cannulated in the ascending aorta, 10 with a short (15 mm) and 8 with a long (70 mm) cannula. The peak forward aortic flow velocities measured on the caudal luminal surface of the aortic arch were 0.80 m/s (± 0.23 m/s) when the CPB was turned off and 2.42 m/s (± 0.69 m/s) on CPB for the short cannula. Using the long cannula velocities of 0.53 m/s (± 0.20 m/s) and 0.18 m/s (± 0.10 m/s) off/on CPB were measured respectively. For all measurements a handhold probe was used connected to a Doppler velocimeter set to the continuous wave mode. Based on these measurements it was concluded that a long tip cannula should be used in patients with an atheromatous aortic arch because it confines the sandblasting effect to the descending aorta beyond the origins of the cerebral vessels. Grossi’s results are somewhat questionable because cannulae with different cross sections (long: 22 French versus short: 20 French) were compared. It is not surprising that the cannulae with the larger cross section (long) provide lower peak velocities. Furthermore the length of the tip has no influence on the quantity of the exit velocity assuming a constant flow for all cannulae but only the pressure drop increases with growing length. The accuracy of the velocity measurements which is inadequate in one case and it is peculiar that the long tip cannulae offer velocities that are even lower than the physiological values in the aortic arch (0.18 versus 0.53 m/s). This phenomenon is based on the fact that Doppler measurements are very 97
Appendix 2
direction sensitive (handhold probe) and that the continuous wave Doppler mode (integration of all velocities on one line) rather than the pulsed Doppler mode (local velocity measurements) had been chosen. Nevertheless our velocity measurements obtained with PWD deviates also from data published by Muehrche [4]. They reported very low peak velocities for a water flow of 2 l/min between 0.29 and 0.72 m/s. For example the peak velocity of the model RMI ARS 024C, which is a straight open tip cannula with an internal diameter of 6.1 mm, is measured at 0.57 m/s whereas a calculation would suggest a value of around 1.41 m/s. An explanation for this deviation could be probably found in an appropriate calibration of the laser Doppler anemometer. It is also questionable if the measured velocity drop-off obtained with an ultrasound velocimeter reflects the actual situation. Due to the limited lateral resolution of ultrasound Doppler velocimeters it is likely to underestimate peak velocities if the width of the jet is approximately of the same size as the width of the sampling volume. The calculated velocity drop-off appears too low in this case. All open tip cannulae offer equal peak velocities and diameters of the jet resulting in the same exit forces. The four side hole cannulae provide a larger EOA which produces lower peak velocities and therefore reduces exit forces.
Pressure drop The pressure drop across a cannula for a given blood flow is of concern in the clinical practice because it adds to the total pressure loss of the CPB and needs to be taken into account to adjust the roller pump of the extracorporeal system previous to the operation. 98
Appendix 2
Due to this interest pressure versus flow charts are recorded for each of the tested cannulae. Losses are measured for the tube and the tip of the cannulae seperately to prove that the tip dominates the total loss. All pressure versus flow charts are also presented in a dimensionless manner (Euler versus Reynolds number) to advertise the benefits of dimensionless numbers. For the flow of interest one has to compute the Reynolds number; look up the corresponding Euler number in the graph and compute the resulting pressure drop as ρ u2 * Eu, with u the mean velocity, which is the flow rate divided by the cross-sectional area. There is also no need to rescale the graphs for blood although the measurements are performed with water. This is of great advantage when hemodilution and hypothermia are present, because they alter the dynamic viscosity and therefore the pressure flow relationship. However with the dimensionless numbers Re and Eu the graph is normalized for a Newtonian fluid of any viscosity. Montoya et al. proposed a standardized system to describe pressure versus flow relationships in vascular access devices e.g. aortic arch cannulae [7]. Their request is that catheters are usually characterized by the French number and length only. This description does not provide any information about the pressure-flow relationship of the catheter nor does it allow for performance comparisons between catheters. Their system allows to characterize any vascular access device by a single number denoted as “M” which may be determinated from the geometry or from simple in vitro pressure-flow measurements. M is defined as log (LDC4.75
) where L represents the length and DC the characteristic diameter of the
cannula. The system can be used by surgeons who wish to choose an 99
Appendix 2
appropriate catheter when size or pressure limitations are given or by manufacturers who may supply M as a specification which will allow for performance comparisons between catheters. However the M number does not provide any new information because it could be replaced by two already existing dimensionless numbers: the Euler and the Reynolds number. Euler is defined as Eu = ∆p/ρu2 = λL/2DC for a straight tube with λ a dimensionless friction number. Inserting Blasius’ equation for λ [5] and substituting the velocity by the flow Q, gives the relationship Eu/Re-0.25 = 0.158 L/DC which is constant for a given geometry. If all pressure and velocity values were provided in a dimensionless manner in terms of Reynolds and Euler this approach would be an alternative to the M number which might have some difficulties to become widely accepted.
100
Appendix 2
Conclusions
In summary the four side hole designs show a superior hydrodynamic performance in the in vitro study compared to the end hole cannulae. However the situation might look different in an in vivo setup. E.g. the amount of the exit force exerted on the aortic wall depends very much upon the angle of incidence. The jet of the straight open tip cannulae hits the aortic wall almost perpendicular resulting in a high impact on possibly calcified tissue whereas the jet of the angled tip cannulae hits the wall at a flatter angle which results in a lower force on the aortic wall. The four side hole designs are difficult to judge in this respect because one or more of the four jets is likely to hit the aorta at a right angle. It must be taken into account that the exit force is much lower compared to the other cannulae but since the threshold value to erode calcified plaque is unknown it remains questionable if the design offers a large advantage compared to the angled open tip cannulae in an in vivo situation. It is suggested to determine the threshold value for tissue erosion in an in vitro setup before starting an in vivo investigation of the flow patterns of the cannulae to be able to judge the force that needs to be applied to erode calcified material. Meanwhile it is advised to use angled tip cannulae to direct the blood flow away from the aortic wall reducing the impact on the aortic intima.
101
Appendix 2
References
1. Katz E.S., Tunick P.A., Rusinek H., Ribakove G., Spencer F.C., Kronzon I. Protruding aortic atheromas predict stroke in elderly patients undergoing cardiopulmonary bypass : experience with intraoperative transesophageal echocardiography. J. Am. Coll. Cardiol. 20:70-7, 1992. 2. Ribakove G.H., Katz E.S., Galloway A.C. et al. Surgical implications of transesophageal echocardiography to grade the atheromatous aortic arch. Ann. Thorac. Surg. 53:758-93; 1992. 3. Grossi E.A., Kanchuger S., Schwartz S., McLoughlin D.E., LeBoutillier M., Ribakove G.H., Marschall K.E., Galloway A.C., Colvin S.B. Effect of cannula length on aortic arch flow : protection of the atheromatous aortic arch. Ann. Thorac. Surg. 59:710-2, 1995. 4. Muehrche D.D., Cornhill J.F., Thomas J.D., Cosgrove D.M. Flow characteristics of aortic cannulae. J. Card. Surg. 10:514-519, 1995. 5. Streeter V.L. Handbook of fluid dynamics. Mc Graw-Hill Book Company; 1961. 6. Guiot C. et al. Continous and pulsed Doppler power spectral density in steady flow : an experimental investigation. Med. & Biol. Eng. & Comput. 35:146-159, 1997. 7. Montoya J.P. et al. A standardized system for describing flow/pressure relationships in vascular access devices. ASAIO Transactions. 37:4-8, 1991.
102
Appendix 2
Table 1. Geometries, dimensions and materials of the tested 24 French cannulae
103
Appendix 2
Figure 1. Schematic drawing of a cannula (a connector, a tube and the tip).
Connector
Tube
Tip
104
Appendix 2
Figure 2. Experimental in vitro setup.
Upper Reservoir
Flow Meter
3/8" Tube
∆H
Data Aquisition System Valves D
A
Computer Pressure Transducers
Ultrasound Probe
Flow Probe
Cannula Ultrasound Machine
Outflow Reservoir
Luer Lock Pump
Lower Reservoir
105
Appendix 2
Figure 3. Schematic representation of measured sample volume in the jet.
20,00 15,00
Width (mm)
10,00 5,00 0,00 -5,00 -10,00 -15,00 -20,00 0
20
40
60
80
100
120
140
Depth (mm)
106
Appendix 2
Figure 4. Calculation scheme for the exit force on the aortic wall.
Tip
Aortic wall Jet
Core of jet Flow profile
α A
d
um
107
Appendix 2
Figure 5. Comparison between geometric outflow area and effective outflow area of all tested cannulae.
80 70 60 50
GOA
40
EOA
30 20 10
Sarns 5774
Sarns 5847
Sarns 4401
Sarns 165264
Sarns 9484
0
Gent Hospital
Outflow cross section (mm 2)
90
108
Appendix 2
Figure 6. Measured velocity core of the jet for a “Sarns 9498" in a horizontal plane (left panel) and a vertical plane (right panel).
1,20
1,20
1,15
1,15
1,10
1,10
1,05
1,05
1,00
1,00
0,95
0,95
90
0,90 -7 -6 -5 -4 -3 -2 -1 0 1 2 3 4 5 6 7
20
1,20
1,20
1,15
1,15
1,10
1,10
1,05
1,05
1,00
1,00
0,95
0,95
80
0,90 -7 -6 -5 -4 -3 -2 -1 0 1 2 3 4 5 6 7
30
1,20
1,20
1,15
1,15
1,10
1,10
1,05
1,05
1,00
1,00
0,95
40
0,95
70
0,90 -7 -6 -5 -4 -3 -2 -1 0 1 2 3 4 5 6 7
40
1,20
1,20
1,15
1,15
1,10
1,10 1,05
1,00
1,00
0,95
0,95
60
0,90 -7 -6 -5 -4 -3 -2 -1 0 1 2 3 4 5 6 7
50
1,20
1,20
1,15
1,15
1,10
1,10 1,05
1,00
1,00
0,95
1,10 1,05 1,00 0,95
40
0,90 -7 -6 -5 -4 -3 -2 -1 0 1 2 3 4 5 6 7 1,20 1,15 1,10 1,05 1,00
Distance from sensor (mm)
1,15
Distance from tip (mm)
-7 -6 -5 -4 -3 -2 -1 0 1 2 3 4 5 6 7
60
1,20 1,15 1,10 1,05 1,00 0,95
70
1,20 1,15 1,10 1,05 1,00 0,95
30
0,90 -7 -6 -5 -4 -3 -2 -1 0 1 2 3 4 5 6 7
80
1,20
1,20
1,15
1,15
1,10
1,10
1,05
1,05 1,00
0,95
0,95
20
0,90 -7 -6 -5 -4 -3 -2 -1 0 1 2 3 4 5 6 7
90
1,20
1,20
1,15
1,15
1,10
1,10 1,05
1,00
1,00
0,95
0,95
10
0,90 -7 -6 -5 -4 -3 -2 -1 0 1 2 3 4 5 6 7
100
1,20
1,15
1,15
1,10
1,10 1,05
1,00
1,00
0,95 -7 -6 -5 -4 -3 -2 -1 0 1 2 3 4 5 6 7
1,20 1,15
0,95 0,90 Width (mm)
1,10 1,05 1,00 0,95
-10
120
0
0,90 -7 -6 -5 -4 -3 -2 -1 0 1 2 3 4 5 6 7
1,15
1,00
-7 -6 -5 -4 -3 -2 -1 0 1 2 3 4 5 6 7
110
1,20
1,05
120
0,95
0
0,90
1,10
10
0,90 -7 -6 -5 -4 -3 -2 -1 0 1 2 3 4 5 6 7
1,20
1,05
110
20
0,90 -7 -6 -5 -4 -3 -2 -1 0 1 2 3 4 5 6 7
1,05
100
30
0,90 -7 -6 -5 -4 -3 -2 -1 0 1 2 3 4 5 6 7
1,00
90
40
0,90 -7 -6 -5 -4 -3 -2 -1 0 1 2 3 4 5 6 7
0,95
80
50
0,90 -7 -6 -5 -4 -3 -2 -1 0 1 2 3 4 5 6 7
Velocity (m/s)
Distance from sensor (mm)
0,95
50
0,90 1,20
70
60
0,90 -7 -6 -5 -4 -3 -2 -1 0 1 2 3 4 5 6 7
1,05
60
70
0,90 -7 -6 -5 -4 -3 -2 -1 0 1 2 3 4 5 6 7
1,05
50
80
0,90 -7 -6 -5 -4 -3 -2 -1 0 1 2 3 4 5 6 7
Distance from tip (mm)
30
90
0,90 -7 -6 -5 -4 -3 -2 -1 0 1 2 3 4 5 6 7
0,90 -7 -6 -5 -4 -3 -2 -1 0 1 2 3 4 5 6 7
Velocity (m/s)
20
-10
Width (mm)
109
Appendix 2
Figure 7. Measured peak velocities.
1,07 1,00
1,07 1,00
1,051,08
Sarns 165264
Sarns 4401
Gent Hospital
0,85 0,75
Sarns 5774
0,80
0,85 0,75
Sarns 5847
Peak velocity (m/s)
1,00
1,05 0,97
Sarns 9484
1,20
0,60 0,40 0,20 0,00
110
Appendix 2
Figure 8. Calculated exit forces (N) for all tested cannulae.
0,3
0,25
Exit Force (N)
constant 0,2
n=8 n=7 n=6
0,15
parabolic linear
0,1
0,05
0 -0,01
-0,005
0
0,005
0,01
Radius (mm)
111
Appendix 2
Figure 9. Pressure drop (mmHg) for constant water flow of 1.3 l/min.
25 20 Tip 15
Tube
10 5
Gent Hospital
Sarns 4401
Sarns 165264
Sarns 9484
Sarns 5774
0
Sarns 5847
Total pressure drop (mmHg)
30
112
Appendix 2
Figure 10. Pressure-flow (upper panel) and Euler-Reynolds (lower panel) relationships for a “Sarns 9484".
18 16
Pressure drop (mmHg)
14 12 Cannula
10
Tip 8
Tube
6 4 2 0 0
0,5
1
1,5
2
2,5
3
Flow (l/min)
3
2,5
Euler
2 Cannula 1,5
Tube Tip
1
0,5
0 0
1000
2000
3000
4000
5000
6000
7000
Reynolds
113
Appendix 2
Figure 11. Water flow (l/min) across the tested cannulae for a constant pressure drop of 10.9 mmHg.
4 3,53 3,5 2,51
2,56
Sarns 9484
Sarns 5774
2,33
Sarns 165264
2
2,31
Sarns 4401
2,5 1,66
1,5 1 0,5
Sarns 5847
0
Gent Hospital
Flow (l/min)
3
114
Appendix 3
Comparison of two dissimilar designs of paediatric aortic cannulae
D. De Wachter, F. De Somer, PR Verdonck
Int. J. Art. Organs, 2002; 25(9): 867 - 874
115
Appendix 3
Abstract
Any extracorporeal blood treatment requires an adequate and safe connection to the circulation. For cardiopulmonary bypass procedures, aortic and venous cannulas are utilized. However, the performance of these cannulas is not only dependent on their size (diameter), but merely on their complete geometric design. In this paper two aortic cannula designs are evaluated haemodynamically for two different sizes (8, 10 Fr) with both aqueous fluids as well as with blood. Using the novel concept of equivalent diameter, a new performance parameter, and the theory of dynamical similarity the results obtained with different fluids can be compared. Data points of one cannula can be fitted with a parabolic equation. There is a significant performance difference between the two 8 Fr cannulas. The 10 Fr cannulas differ non-signicantly except when water is used. Equivalent diameters obtained with water in the turbulent region are significantly higher than those obtained with fluids that have a higher viscosity (blood and aqueous glycerine mixture). The latter fluids have comparable viscosities and render an equal equivalent diameter. The coefficients of their proper parabolic fit lines can be easily recalculated into each other. This provides a simple method to quickly determine pressure drops over cannulas in the operating room.
116
Appendix 3
Introduction
During open heart surgery procedures, the heart is arrested. To cover heart and lung functions during this period, an extracorporeal shunt is established over the heart (cardiopulmonary bypass), which includes a blood pump to pump the blood through the extracorporeal circuit and back into the main circulation, and an artificial lung for gas exchange. Blood is collected on the venous side by means of a plastic cannula and drained by gravity into a collecting reservoir. On the positive pressure side of the pump a plastic aortic cannula is employed to inject the oxygenated blood into the intracorporal circulation. Pediatric aortic cannulas are made from soft plastic and on the one end they allow a connection to PVC tubing, while at the other end they have a small tip, suitable for introduction in the aorta of children. This tip is either in hard plastic or it is a wire-reinforced. The dimensions of a paediatric cannula are a compromise between technical requirements (blood flow vs. pressure drop) and practical limitations (aortic diameter, small incision). Tip dimensions are commonly stated in french (Fr); 1 Fr corresponds to an outer diameter of 0.33 mm. However inner diameters (ID) may differ significantly (Table 1). The purpose of this study is to characterise the hydraulic resistance of paediatric aortic cannulas on the one hand with an aqueous solution, similar as has been done for adult size cannulas [1] and secondly with blood. The second objective is to test whether these values can be used to determine actual flow resistance in the operating room where the cannulas are perfused by blood with different haematocrits and temperatures. 117
Appendix 3
Materials and Methods
4 different cannulas are included in the study (Table 1). All cannulas are from DLP (Medtronic®). Of each type 2 specimens are studied. These cannulas can be grouped in two ways: according to size (10Fr or 8Fr) or according to tip design (series 750xx or 77xxx). The 750xx series cannulas have a short light blue stiff plastic tip with a sudden diameter reduction (Figure 1). The inner diameter (ID) of their tip is smaller than the ID of a 77xxx series cannula of comparable French size. The latter series of cannulas have a long wirereinforced flexible tip with a smooth diameter reduction halfway the cannula. The length of the 750xx series is shorter than the 771xx series, mainly because the flexible tip of the latter can be introduced downstream into the aorta. Pressure-flow relationships are assessed using three different fluids: pure water, an aqueous glycerine solution (35 vol. %), both at room temperature (20-25°C) and bovine blood with a hematocrit of about 32% (at 20°C and 37°C). The test set-up consists of a tubing set, a variable speed rollerpump and a small temperature controlled reservoir (Figure 2). The fluid is pumped from the reservoir through the cannula and back in the reservoir. The tip of the cannula is positioned in a long large bore tube (length: 50cm, ID=1/2 inch), to minimise afterload influence. In fact this tube serves also as a sort of reservoir in which kinetic energy of the fluid jet, that is propulsed from the cannula’s tip, can be dissipated. Flow rate is measured with an ultrasonic transit-time clamp-on flow meter (Transonic®, Ithaca NY, USA), that was previously calibrated for each fluid by a volumetric method. Pressures are measured with 118
Appendix 3
fluid-filled piezo-electrical transducers (Ohmeda, Gent, Belgium), two-point calibrated before each test run. To determine a single data point, flow rate and pressures are recorded during a finite time, through a computer data-acquisition card (PC74, Eagle Technology, Cape Town, South Africa) and suitable software, typically at a sampling rate of 200Hz. Due to roller pump operation, these curves exhibit a periodical pattern. Actual data points are then determined by averaging these values over an integral number of periods. All averaged data points are used to fit a parabolic equation through the origin, stating the pressure drop (∆P) flow rate (Q) relationship of a cannula. The parabolic equation is obtained by polynomial regression (Sigmaplot, SPSS Inc., Erkrath, Germany). Data obtained with different fluids render different parabolic equations. However using the theory of dynamic similarity [2] (see Appendix), the coefficients of these parabola can be recalculated and compared. Since during clinical use, these cannulas are perfused with blood, the most useful conversions are those to blood. The simplest method is to directly convert pressures (P) and flow rates (Q) according to these ratios: Qb ρ f µ b = Q f ρb µ f
2 Pb ρ f µ b = Pf ρ b µ 2f
(1)
where subscripts b denote blood and f any other fluid (water, aqueous glycerine, ..). ρ and µ are respectively the density and the dynamic viscosity. A second method rescales the coefficients of the fitted parabolic equation. If the parabola is defined with ∆P the pressure drop and Q the flow rate: ∆P = aQ 2 + bQ
(2)
119
Appendix 3
Then the new coefficients are determined by rescaling with ratios of density and dynamic viscosity: ab = a f
ρb ρf
bb = b f
µb µf
(3)
The last method consists of the utilisation of dimensionless numbers, which are independent of the fluid’s physical properties. The Euler number (Eu) is a dimensionless measure of the pressure losses and the Reynolds number (Re) of the flow rate: π 2 De4 ∆P Eu = 16 ρQ 2
Re =
4 ρQ πDe µ
(4)
The equivalent diameter De is determined from the effective diameters. The effective diameter Deff is defined as the internal diameter that a circular tube with the same length (L) as the cannula should have to exhibit the same pressure drop as the cannula under study at a particular flow rate. At fully developed turbulent flow (Re > 4000; water measurements), it is determined from the Blasius equation (top of eq. 5). At laminar flow (Re < 2300; blood & glycerine measurements) it is derived from the well-known Poiseuille equation (lower part of eq. 5). Deff = 19 0.0541ρ 3 µL4 Deff
128µLQ =4 π∆P
Q7 ∆P 4
(5)
Deff should be independent on the fluid’s density (ρ) and dynamic viscosity (µ), as their effects are cancelled by the flow / pressure drop ratio. However, when using the Poiseuille equation, Deff depends on the flow rate because of special pressure losses in the cannula that are not linearly proportional to the flow 120
Appendix 3
rate. As a reasonable approximation, it can be assumed that Deff varies linearly with Re. Therefore the actual equivalent diameter (De) is obtained as the Deff at Re=1000 on the linear regression line that is fitted through all Deff in the laminar range for blood and aqueous glycerine and as the average of the effective diameters determined from all measurements in the turbulent flow range (Re> 4000) for water. The Blasius equation is utilized for water since most water measurements fall in the turbulent flow regimen because of its much lower viscosity compared with blood. Dynamic viscosity of the fluids (µ) is either determined from literature data (water [3], blood [4]), or measured in a viscometer (aqueous glycerine & blood). Blood viscosity µb can be described by exponential functions, with µp: plasma viscosity, T: absolute temperature (K) and Hct the fraction of red blood cell volume [4]: µ p = exp(−5.54 + 1800 T )
(6)
µ b = µ p exp(2.35 Hct )
121
Appendix 3
Results
Viscosities of the fluids are respectively: 3.66 mPa.s for bovine blood at 37°C; 5.66 mPa.s for blood at 20°C; 3.36 mPa.s for 35%-65% glycerine-water mixture at 20°C and 1.00 mPa.s for tap water. In Table 2 the equivalent diameters of the four cannulas are presented. They range from 2.5 mm for the 75008 8 Fr cannula to 3 mm for the 10 Fr cannulas. The equivalent diameter obtained with water measurements is 6% (75xxx series) to 10% (77xxx series) higher than for measurements with blood. The coefficients of the parabolic pressure equation (eq. 2) are listed in Table 3. The quadratic coefficients (a) are greater than the linear coefficients (b). Both coefficients, but especially the linear coefficient (b) are generally greater for the aqueous glycerine compared to the blood measurements, while for water they are generally smaller. Also, the ratio of b/a is generally smaller for water measurements. This ratio is a measure of the flow rate at which the influence of the special pressure losses in the tip overhaul the pressure losses in the cannula tube, the latter at these low flow rates being linearly related to the laminar flow rate of the fluid. In Figure 3 the pressure-flow relationships of all cannulas as measured with bovine blood at 37°C are plotted along with their fitted parabolic regression line proper. If the total pressure drop over each cannula would be limited to 200 mmHg, the maximal blood flow rate through the different cannulas is respectively 0.78 L/min (77008); 1.05 L/min (77008); 1.60 L/min (77110) and 1.64 L/min (75010). 122
Appendix 3
In Figure 4 the Eu-Re relationship is shown for cannula 77008 on a semilogaritmic plot, with their respective transformed parabolic regression lines. As is observed, the type of fluid does not influence to a great extent the position of the plots that are in the laminar region (Re
![[PDF]Download Cardiopulmonary Bypass and ... - Google Sites](https://m.moam.info/img/260x300/pdfdownload-cardiopulmonary-bypass-and-google-site_6477c955097c4737708c1649.jpg)
