Ebook Extracorporeal life support for adults: Part 1

(BQ) Part 1 book Extracorporeal life support for adults'' has contents: Physiology of extracorporeal life support; hypoxemic respiratory failure - evidence, indications, and exclusions; cardiogenic shock - evidence, indications, and exclusions,... and other contents.

Respiratory Medicine

Series Editor: Sharon I.S Rounds

Extracorporeal Life Support for Adults

Gregory A Schmidt Editor

Series Editor :

Sharon I.S Rounds

More information about this series at http://www.springer.com/series/7665

Editor

Extracorporeal Life Support for Adults

ISSN 2197-7372 ISSN 2197-7380 (electronic)

Respiratory Medicine

ISBN 978-1-4939-3004-3 ISBN 978-1-4939-3005-0 (eBook)

DOI 10.1007/978-1-4939-3005-0

Library of Congress Control Number: 2015950466

Springer New York Heidelberg Dordrecht London

© Springer Science+Business Media New York 2016

This work is subject to copyright All rights are reserved by the Publisher, whether the whole or part of the material is concerned, specifi cally the rights of translation, reprinting, reuse of illustrations, recitation, broadcasting, reproduction on microfi lms or in any other physical way, and transmission or information storage and retrieval, electronic adaptation, computer software, or by similar or dissimilar methodology now known or hereafter developed

The use of general descriptive names, registered names, trademarks, service marks, etc in this publication does not imply, even in the absence of a specifi c statement, that such names are exempt from the relevant protective laws and regulations and therefore free for general use

The publisher, the authors and the editors are safe to assume that the advice and information in this book are believed to be true and accurate at the date of publication Neither the publisher nor the authors or the editors give a warranty, express or implied, with respect to the material contained herein or for any errors

or omissions that may have been made

Printed on acid-free paper

Humana Press is a brand of Springer

Springer Science+Business Media LLC New York is part of Springer Science+Business Media ( www.springer.com )

Gregory A Schmidt, MD

Division of Pulmonary Diseases, Critical Care,

and Occupational Medicine

Department of Internal Medicine

University of Iowa

Iowa City , IA , USA

nurtured an outstanding program, exhibited remarkable vision in how to advance ECLS care, and opened my eyes to its new

possibilities

Extracorporeal life support (ECLS) consists of using an external gas-exchanging membrane to support oxygenation or carbon dioxide removal (or both), at times including circulatory assistance ECLS has been used in severe hypoxemic respira-tory failure (ARDS, pneumonia); diseases dominated by ventilatory failure such as status asthmaticus and COPD; cardiogenic shock; following cardiothoracic surgery complicated by circulatory or gas exchange failure; and as a bridge to lung trans-plant Historically, ECLS has been used sparingly, often as a last resort, and in few centers with the requisite expertise Three factors have combined to change this First, technological improvements in membranes, pumps, circuits, and cannulas have led to more effi cient and safer ECLS Second, the CESAR trial has shown that, for adults with severe ARDS, referral to an ECLS center improves outcomes Finally, the adverse consequences of conventional management of lung failure, including ventilator-induced lung injury, ICU-acquired weakness, and nosocomial infection, have become abundantly clear Some of these may be ameliorated by using ECLS in preference to conventional care As perceptions of the role of ECLS have evolved, more practitioners and more centers are developing ECLS capability

or positioning themselves to offer ECLS

The aim of this book is to deliver a concise, evidence-based review of ECLS for adult disease Adult medicine (rather than neonatal and pediatric disease, where ECLS has an established but limited role) represents the growth area for ECLS Chapters are devoted to describing the complex physiology and technology; the evidence base in varied clinical conditions; how to obtain vascular access; daily management of the circuit and patient; guidance regarding the weaning and decan-nulation process; and recommendations for crisis management and rehabilitation related to ECLS The text concludes with a fascinating historical review, showing just how far we’ve come

This text has been written for practicing physicians, nurses, perfusion specialists, therapists, and critical care trainees who are considering whether to refer their patients for ECLS, debating whether to offer ECLS capability to their patients, or are already providing ECLS but seek a practical reference to best practices and updated information It could never have been completed without the inspiration

from my colleagues at Iowa who strive daily to save the sickest patients; the trainees whose curiosity makes us all want to know more; my contributors who are at the forefront of a truly challenging fi eld; and our publisher at Springer-Link who pushed for this important book Finally, I recognize all those who do the hard work: the nurses, perfusionists, and therapists who dedicate their lives to the critically ill This

is an exciting time, ripe with change and opportunity We seek a path forward for the benefi t of all our patients

Iowa City, IA, USA Gregory A Schmidt, MD

3 Cardiogenic Shock: Evidence, Indications, and Exclusions 73 Nicolas Bréchot and Alain Combes

4 ECCO 2 R in Obstructive Diseases: Evidence, Indications,

and Exclusions 87 Lorenzo Del Sorbo and V Marco Ranieri

5 ECLS as a Bridge to Lung Transplantation 105

Christian Kuehn

6 Modes of ECLS 117

L Keith Scott and Benjamin Schmidt

7 Vascular Access for ECLS 133

Steven A Conrad

8 Circuits, Membranes, and Pumps 147

Bradley H Rosen

9 Ventilator Management During ECLS 163

Antonio Pesenti , Giacomo Bellani , Giacomo Grasselli ,

and Tommaso Mauri

10 Daily Care on ECLS 181

Giles J Peek

11 Crises During ECLS 193

Cara L Agerstrand , Linda B Mongero , Darryl Abrams ,

Matthew Bacchetta , and Daniel Brodie

12 Mobilization During ECLS 211

Gregory A Schmidt

13 ECMO Weaning and Decannulation 223

Sundar Krishnan and Gregory A Schmidt

14 The Story of ECLS: History and Future 233

J Ann Morris , Robert Pollock , Brittany A Zwischenberger ,

Cherry Ballard- Croft , and Joseph B Zwischenberger

Index 261

Matthew Bacchetta , MD, MBA, MA Division of Thoracic Surgery , New York- Presbyterian Hospital/Columbia University Medical Center , New York , NY , USA

Cherry Ballard-Croft , PhD Division of Cardiothoracic Surgery, Department of Surgery, University of Kentucky College of Medicine, Lexington, KY, USA

Giacomo Bellani , MD, PhD Department of Health Sciences, University of Milano-Bicocca, Monza, Italy

Department of Anesthesia and Critical Care , San Gerardo Hospital and

Milano-Bicocca University , Monza , Italy

Matthew J Brain , MBBS (Hons), FRACP, FCICM, DDU School of Public Health and Preventive Medicine, Monash University, Malvern East, VIC, Australia The Alfred Intensive Care Unit, Melbourne, VIC, Australia

Department of Medicine, Launceston General Hospital, Launceston, TAS, Australia

Nicolas Bréchot , MD, PhD Service de Réanimation Médicale , Hospital Pitié–Salpêtrière , Paris , France

Daniel Brodie , MD Division of Pulmonary, Allergy and Critical Care, New Presbyterian Hospital/Columbia University Medical Center, New York, NY, USA

Warwick W Butt, FRACP, FCICM ICU RCH, Department of Paediatrics UoM, Clinical Sciences Theme MCRI, Royal Children’s Hospital, Melbourne, VIC, Australia Paediatric Intensive Care Unit, Parkville, VIC, Australia

Alain Combes , MD, PhD Service de Réanimation Médicale, Institut de Cardiologie, Groupe Hospitalier Pitié-Salpêtrière, iCAN, Institute of Cardiometabolism and Nutrition, Paris Cedex, France

Steven A Conrad , MD, PhD, MCCM, FCCP Department of Medicine, Emergency Medicine and Pediatrics , Louisiana State University Health Sciences Center , Shreveport , LA , USA

Lorenzo Del Sorbo , MD Dipartimento di Anestesiologia e Rianimazione, Azienda Ospedaliera Città della Salute e della Scienza di Torino , Università di Torino , Torino , Italy

Inter-departmental Division of Critical Care Medicine, University Health Network, University of Toronto, Toronto, ON, Canada

Giacomo Grasselli , MD Department of Anesthesia and Critical Care , San Gerardo Hospital and Milano-Bicocca University , Monza , Italy

Sundar Krishnan , MBBS Department of Anesthesia , University of Iowa Hospitals and Clinics , Iowa City , IA , USA

Christian Kuehn , MD Department of Cardiac, Thoracic, Transplantation and Vascular Surgery , Privatdozent Dr med., Hannover Medical School , Hannover , Germany

Graeme MacLaren , MBBS, FRACP, FCICM, FRCP, FCCP, DipEcho ICU RCH, Department of Paediatrics UoM, Clinical Sciences Theme MCRI, Royal Children’s Hospital, Melbourne, VIC, Australia

Paediatric Intensive Care Unit, Parkville, VIC, Australia

Cardiothoracic ICU, National University Hospital, Singapore, Singapore

Tommaso Mauri , MD Department of Anesthesia and Critical Care , Fondazione IRCCS Ca’ Granda Ospedale Maggiore Policlinico , Milan , Italy

Linda B Mongero , CCP, BS Department of Clinical Perfusion , New York Presbyterian-Columbia University Medical Center , Locust Valley , NY , USA

J Ann Morris , BS Division of Cardiothoracic Surgery, Department of Surgery, University of Kentucky College of Medicine, Lexington, KY, USA

Giles J Peek , MD, FRCS CTh, FFICM Heartlink ECMO Centre , Glenfi eld Hospital , Leicester , UK

Antonio Pesenti , MD Department of Health Sciences, University of Bicocca, Monza, Italy

Department of Anesthesia and Critical Care , San Gerardo Hospital and

Milano-Bicocca University , Monza , Italy

Robert Pollock , BS Division of Cardiothoracic Surgery, Department of Surgery, University of Kentucky College of Medicine, Lexington, KY, USA

V Marco Ranieri , MD Dipartimento di Anestesiologia e Medicina degli Stati Critici, Ospedale S Giovanni Battista-Molinette, Università di Torino, Torino, Italy Dipartimento di Anestesiologia e Rianimazione, Azienda Ospedaliera Città della Salute e della Scienza di Torino, Università di Torino, Torino, Italy

Bradley H Rosen , DO Division of Pulmonary, Critical Care, and Occupational Medicine, Department of Internal Medicine , Carver College of Medicine, University

of Iowa Hospitals and Clinics , Iowa City , IA , USA

Benjamin Schmidt , MD Department of Surgery , Wake Forest University , Medical Center Boulevard, Winston- Salem , NC , USA

Gregory A Schmidt , MD Division of Pulmonary Diseases, Critical Care, and Occupational Medicine, Department of Internal Medicine , University of Iowa , Iowa City , IA , USA

L Keith Scott , MD Department of Anesthesiology , Wake Forest University , Medical Center Boulevard, Winston-Salem , NC , USA

Brittany A Zwischenberger , MD Division of Cardiothoracic Surgery, Department

of Surgery, University of Kentucky College of Medicine, Lexington, KY, USA

Joseph B Zwischenberger , MD Division of Cardiothoracic Surgery, Department

of Surgery, University of Kentucky College of Medicine, Lexington, KY, USA

© Springer Science+Business Media New York 2016

G.A Schmidt (ed.), Extracorporeal Life Support for Adults,

Respiratory Medicine 16, DOI 10.1007/978-1-4939-3005-0_1

Physiology of Extracorporeal

Life Support (ECLS)

Matthew J Brain, Warwick W Butt, and Graeme MacLaren

Introduction

Extracorporeal life support (ECLS) and related implantable circulatory assistance devices describe several advancing technologies with broadening scope that are being increasingly incorporated into management of critically ill patients

ECLS may be provided in several configurations to support or replace spiratory function (Fig 1.1) In veno-venous extracorporeal membrane oxygenation (VV-ECMO) the objective is to maintain systemic oxygen delivery by oxygenating venous blood returning to the right heart In veno-arterial mode (VA-ECMO),

cardiore-M.J Brain, MBBS (Hons), FRACP, FCICM, DDU ( * )

School of Public Health and Preventive Medicine, Monash University,

Malvern East, VIC, Australia

The Alfred Intensive Care Unit, Melbourne, VIC, Australia

Department of Medicine, Launceston General Hospital,

274-280 Charles St, Launceston, TAS 7250, Australia

e-mail: m.brain@iinet.net.au

W.W Butt, FRACP, FCICM

ICU RCH, Department of Paediatrics UoM, Clinical Sciences Theme MCRI,

Royal Children’s Hospital, Melbourne, VIC, Australia

Paediatric Intensive Care Unit, 50 Flemington Road, Parkville, VIC 3052, Australia

e-mail: Warwick.butt@rch.org.au

G MacLaren, MBBS, FRACP, FCICM, FRCP, FCCP, DipEcho

ICU RCH, Department of Paediatrics UoM, Clinical Sciences Theme MCRI,

Royal Children’s Hospital, Melbourne, VIC, Australia

Paediatric Intensive Care Unit, 50 Flemington Road, Parkville, VIC 3052, Australia

Cardiothoracic ICU, National University Hospital,

5 Lower Kent Ridge Rd, Singapore 119074, Singapore

e-mail: graeme_maclaren@nuhs.edu.sg

systemic blood flow is augmented by the extracorporeal blood pump, while VPA- ECMO describes augmentation of pulmonary arterial flow Both of the latter configurations can also incorporate support of oxygenation.

Configurations can also be classified by the site of vascular access, with cannulas being either peripherally placed via the great vessels or centrally placed via thoracot-omy Rotary pumps (without oxygenators) have been miniaturised, allowing develop-ment of left and right ventricular assist devices (LVAD and RVAD) respectively

A basic ECMO circuit consists of a blood pump and oxygenator connected by duits (Fig 1.1) Other components may be added to this basic configuration, in particu-lar other extracorporeal circuits such as renal replacement devices However, maintaining simplicity is important for safety, infection control and troubleshooting.Each configuration creates a unique interaction with the cardiorespiratory system Sound understanding of the physiology and limitations of each mode is

con-Fig 1.1 Schematic of ECMO configurations, circles represent pumps, diamonds represent

oxygenators VA-ECMO: veno-arterial extracorporeal membrane oxygenation demonstrating cavo- aortic flow VV-ECMO: veno-venous cannulation demonstrating cavo-atrial flow from the inferior vena cava to the right atrium via the oxygenator and pump VV-ECMO may also require a second cannula taking blood from the superior vena cava, or dual lumen cannulas that access blood from the inferior and superior vena cava, while returning blood to the right atrium VPA-ECMO: veno- pulmonary artery cannulation may be configured as atrial to pulmonary artery flow with or without oxygenation support LVAD: left ventricular assist device (usually implanted) taking left ventricular blood and returning it to the proximal aorta RVAD: a right ventricular assist device is not shown but may be implanted or external and can be configured identically to VPA-ECMO without an oxygenator, or may directly drain the right ventricle as per the LVAD Extracorporeal carbon dioxide removal (ECCO2R) is commonly performed with a VV-ECMO configuration, usually with a single dual-lumen catheter Intravascular membrane oxygenators have also been developed [ 1 ] but are not currently in clinical use

required to prescribe, manage and wean this support and recognise evolving complications of the therapy Although designed primarily to replace cardiorespira-tory function, the interaction of ECLS with several other physiologic systems must

be considered For example, most patients who require ECLS will have sustained a major insult such as severe sepsis, trauma or surgery, or have suffered from progres-sive cardiac or pulmonary disease The systemic inflammatory response syndrome (SIRS) may arise from the underlying disease or as a reaction to the non-biological material of the ECLS circuit The metabolic response to critical illness has direct implications for oxygenation and CO2 removal, as well as nutritional supplementation

to facilitate later weaning

In order to comprehensively understand ECLS and its effects on human physiology,

it is necessary to first review cellular metabolism and oxygen transport

Cellular Metabolism

The fundamental role of tissue perfusion is to provide sufficient substrate delivery

to match the metabolic demand of aerobic cellular metabolism While anaerobic metabolism can support cellular energy requirements for brief periods, only oxidative metabolism can maintain proper cellular and organ function

Cardiorespiratory physiology and any mechanical support must provide an adequate hydrostatic pressure gradient across capillary beds to support blood flow,

as well as maintain concentration gradients by which substrates, including oxygen, diffuse into the immediate environment of cells Likewise, a concentration gradient must be maintained from the cell to the blood path for the waste products of metabo-lism, primarily CO2, or lactate in the case of anaerobic metabolism These functions are interlinked as the waste products of energy production are generally weak acids and influence local perfusion and oxygen carriage

The quantities of substrate required per unit time will depend on the supported cell mass and its level of metabolic activity as influenced by demand (or stress), temperature, inflammation and hormonal regulation

Glycolysis and Aerobic and Anaerobic Metabolism

Glucose and other simple carbohydrates enter cells down a concentration gradient through glucose transporters that allow for tissue-specific behaviour such as prefer-ential basal uptake by the brain, concentration-dependent uptake by the liver, concentration- sensing by the insulin-secreting pancreatic β-cells and insulin- dependent uptake in skeletal muscle and fat [2]

Intracellular glucose is rapidly phosphorylated in the cytosol by hexokinases, after which it becomes the primary substrate for energy production or biosynthetic reactions including glycogen storage (Fig 1.2) Utilisable intracellular energy is

metabolised to two three-carbon pyruvate molecules (glycolysis) The conversion of pyruvate to acetyl-CoA, the tricarboxylic acid (TCA) cycle and oxidative phosphorylation only occur in mito- chondria and depend on oxygen to restore nicotinamide adenine dinucleotide to its oxidised form (NAD + ) for continued cycling The number of ATP generated depends on the source of reduction power; a single mitochondrial NADH produces 2.5 ATP; however, electrons from cytosolic NADH must be transferred to mitochondrial FADH2 which yields only 1.5 ATP each [ 3 ] Different amino acids can enter or be synthesised from the pathway at several points Acetyl-CoA is a key junction molecule providing the TCA cycle with two-carbon acetyl groups, not only from glycolysis but also from fatty acids and some amino acids In glucose excess, acetyl-CoA is the starting point for fatty acid synthesis and, in the starvation state, ketone body production when insufficient oxaloac- etate exists for acetyl groups to enter the TCA cycle Ketone bodies are produced predominantly

in the liver from fatty acid breakdown and constitute a glucose-sparing fuel for the brain and heart Humoral promoters and inhibitors of reactions are shown

stored in the phosphate bonds of adenosine triphosphate (ATP) and it is the breaking

of chemical bonds within glucose that powers ATP regeneration from adenosine diphosphate (ADP) and inorganic phosphate (Pi)

Glycolysis describes the fracturing of the six-carbon glucose molecule into two three-carbon pyruvate molecules with the net generation of two ATP molecules For glycolysis to continue, oxidative power (NAD+ concentration) must be continu-ally restored Under anaerobic conditions, this occurs by conversion of pyruvate to lactate Under aerobic conditions, pyruvate loses a carbon dioxide molecule to yield acetyl coenzyme-A This two-carbon acetyl group can be incorporated into fatty acids for storage or can enter the tricarboxylic acid (TCA) cycle to complete the chemical breakdown of glucose to CO2 This yields 38 ATP molecules, significantly more than glycolysis, but generates NADH in such quantities that a powerful elec-tron acceptor is required for efficient restoration of NAD+ so that the cycle can continue This electron acceptor is oxygen

Oxidative phosphorylation describes the process of restoring NAD+ to perpetuate the TCA cycle Although oxygen is utilised as an electron acceptor in many enzyme systems, its highest consumption is in this process Oxidative phosphorylation occurs in the inner mitochondrial matrix and it is to this intracellular destination that oxygen must diffuse in sufficient quantities to sustain ATP generation for normal cellular processes

When oxygen is not available in sufficient quantities, ATP generation from ADP can only continue in the cytosol by glycolysis This process is inefficient, as not only is less ATP produced but the resulting lactic acid is not as readily cleared from the tissues or body as carbon dioxide Lactic acid is thus a marker of glycolysis activity in a hypoxic environment and usually indicates inadequate tissue perfusion

or global hypoxemia of the organism

Carbon Dioxide Production and the Respiratory Quotient

The respiratory quotient (RQ) describes the ratio of the amount of carbon dioxide

( VCO2) produced per unit time to the amount of oxygen consumed ( VO2)

O 2 2

The respiratory quotient depends on the sources of fuel being used For glucose metabolism, the six carbon atoms result in production of six molecules of carbon dioxide while consuming six molecules of oxygen; it thus has a respiratory quotient

of 1 The reactions for oxidation of some amino acids and fatty acids (lipolysis) produce less CO2 (by not including the pyruvate to acetyl-CoA reaction, Fig 1.2) and hence have respiratory quotients of less than 1

In contrast, each acetyl-CoA molecule utilised for fatty acid synthesis (lipogenesis) results in production of a molecule of CO2 (from pyruvate to acetyl-CoA, Fig 1.2)

without increasing mitochondrial NADH As the rate of oxygen consumption depends on the mitochondrial concentration of NADH, lipogenesis results in CO2

production which exceeds oxygen consumption Some oxygen consumption still occurs as the synthetic reaction also consumes ATP; however, the RQ will be greater than 1 Examples of respiratory quotients based on theoretical stoichiometry include [4, 5]:

A normal adult has a whole body RQ measured by indirect calorimetry of around 0.8, reflecting utilisation of mixed fuel sources This value will alter in critically ill patients, depending on the nutrient availability and humoral control of metabolism While glycolysis reflects enzymatic processing of glucose, complete aerobic metabolism is coupled to TCA intermediate availability and, when carbo-hydrate loads are excessive (such as with glucose supplementation exceeding

4 mg·kg−1·min−1 [5.8 g·kg−1·day−1]), lipogenesis occurs with a respiratory quotient as high as 8 [5 7] resulting in a high CO2 burden

Metabolism in the Stressed State

Key hormones coordinate the response to nutrition supply and stress Insulin marks the fed state, promoting hepatic glucose uptake, glycogen and amino acid synthesis and conversion of acetyl-CoA to free-fatty acid production while in the peripheral tissues stimulating myocyte synthesis of contractile elements and adipocyte triglyceride deposition

Glucagon is secreted by pancreatic α-cells in response to low blood glucose levels and promotes glycogen breakdown and conversion of amino acids

(from muscle breakdown), lactate and glycerol Glycerol results from adipocyte triglyceride metabolism and the released free fatty acids are converted to ketone bodies by the liver for use as a secondary fuel source when glucose is scarce.The catecholamines epinephrine and norepinephrine are released in response to physiologic stress By increasing intracellular cyclic AMP, they promote glycoge-nolysis in muscles and catabolism of protein to release amino acids In the liver, epinephrine promotes gluconeogenesis, glycogenolysis and inhibits glycolysis These responses result in the hyperglycaemia that characterises the stress state and is exacerbated by exogenous administration of catecholamines and glucose The adverse effects of hyperglycaemia include osmotic diuresis, fat deposition in the liver and impaired immune function.

The metabolic profile of patients receiving ECLS is typical of the stressed state but the differences between this and the starvation state are important In starvation there is an overall decrease in energy expenditure with maximal use of triglycerides and ketoacids promoting conservation of muscle bulk The brain, heart and renal cortex adapt to utilising ketoacids for significant proportions of their metabolic requirements In contrast, the chronic stressed state is characterised by increased resting energy expenditure, accelerated catabolism of lean body mass—primarily amino acids from muscle catabolism [3]—and the immunosuppressive effects of hyperglycaemia and persistently elevated humoral mediators, including catechol-amines and cortisol

In those requiring ECLS, particularly those needing prolonged periods of heavy sedation, the combination of muscle catabolism, disuse atrophy, critical illness myopathy and myopathy associated with muscle relaxants can result in profound weakness The respiratory musculature is not spared from this process, with the result being prolonged weaning, a requirement for tracheostomy and the risk of secondary infection

Erythrocytes lack mitochondria and do not store glycogen and thus depend on anaerobic glycolysis of plasma glucose to lactate for ATP production However, glycolysis in erythrocytes is also utilised for reactions that do not produce ATP, such

as reducing power to correct oxidised haemoglobin (methaemoglobin carrying a

Fe3+ iron atom that cannot carry oxygen), glutathione production (protecting the cell membrane against oxidative damage) and the production of 2,3-diphosphoglycerate (2,3-DPG) that modulates the affinity of haemoglobin for oxygen [8]

The Rapoport–Luebering shunt (Fig 1.2) describes the pathway for 2,3-DPG synthesis from the glycolytic pathway In most cells, 1,3-DPG is rapidly converted

to 3-phosphoglycerate with the phosphate molecule transferred to ATP; however,

in erythrocytes up to 20 % of glycolytic flux occurs through the shunt, with the value dependent on ATP requirements [9] Oxygen depletion (resulting in fewer haemoglobin- binding sites for 2,3-DPG), acidotic conditions that inhibit 2,3-DPG synthesis and the accumulation of inorganic phosphate (Pi) which increases 2,3-DPG breakdown [8], result in decreased intracellular 2,3-DPG concentrations This is most relevant under conditions of red cell storage where lower glycolysis rates and accu-mulation of lactic acid can result in minimal 2,3-DPG concentrations at the time of transfusion Transfused red cells do not restore normal 2,3-DPG concentrations for some time and, given the relatively high transfusion requirements of patients receiv-ing ECLS, this effect may have significant implications for oxygen carriage The role

of 2,3-DPG will be further discussed below when considering oxygen carriage

Biophysics of Membrane Gas Exchange

Mitochondria can couple ATP production to NADH oxidation only if sufficient gen exists in the environment of cells Similarly, carbon dioxide diffuses from the mitochondria, through intracellular membranes, and away from the cell The flux of oxygen into the environment of cells and the reverse movement of carbon dioxide can be divided into two components:

1 Diffusion of gas molecules into and between liquid phases

2 The carriage of oxygen and carbon dioxide in blood

When considering pulmonary gas exchange a third component must be considered: the convective transport of the gas to the alveolar epithelium However, exposure of the extracorporeal membrane to fresh gas flow is somewhat simpler in ECLS and will be considered later in the context of carbon dioxide transport

Membrane Oxygenator Construction

Extracorporeal membrane oxygenators consist of a high surface area blood path separated by a membrane from a path for fresh gas flow (sweep gas) The devices are in continual evolution to optimise the efficiency of gas transfer, minimise untoward biological responses, reduce priming volumes, avoid plasma leakage and increase their simplicity and integration as systems Materials and construction of gas exchange membranes will be discussed in later chapters; however, a brief introduc-tion is important to understand their operation

Membranes may be arranged in folded sheets or, more commonly, as tubes known as hollow fibre oxygenators (Fig 1.3) The pores of earlier polypropylene

microporous membranes theoretically allow contact between plasma and the sweep gas; however, more recent materials such as poly-4-methyl-1-pentene utilise closed fibres and are thus considered true membranes [10] Most systems direct fresh gas through the lumen of the hollow fibres, while blood flows between the tubules (termed extra-capillary flow) The reverse configuration is also sometimes utilised, however, and overall characteristics such as total surface area for gas exchange, resistance to flow and trauma to formed blood components will be determined by factors such as membrane material, fibre diameter and length, fibre density and the velocity of the blood [11].

Fig 1.3 Schematic detail of hollow fibre oxygenator construction demonstrating extra-capillary

flow of blood around the gas-carrying hollow fibres Cross current flow exists between gas and blood Heated water tubules are also demonstrated The diffusion path for gas exchange is shown

(top-right) consisting of the porous membrane, and the boundary layer of adsorbed proteins

Parameters of effective diffusivity from Eq ( 1.9 ) are demonstrated with ε being the porosity—the

area of membrane occupied by gas, τ the tortuosity, an index of effective path length for gas to

traverse the membrane (a path length is shown but in reality will be unknown), and δ the

constric-tivity—the resistance to passage

Heat loss over the extracorporeal circuit into the environment can be substantial and heat exchangers are commonly incorporated into the oxygenator Figure 1.3 demonstrates one such design where the microporous membrane fibres are laid perpendicularly to impermeable capillaries that circulate heated water, allowing heat to be regulated.

Diffusion of Gas Molecules into a Liquid Phase

Concentration of Gases in Solutions

Unlike most solutes dissolved in body fluids that are quantified in moles, gas centrations are reported in units of pressure The universal gas law describes the relationship between the partial pressure of an ideal gas and its container, with ideal gas molecules best summarised as having minimal mass and intermolecular attraction:

The universal gas equation: P = the partial pressure, n = number of molecules of gas measured in moles, T is temperature in degrees Kelvin and V is volume of the con- tainer in litres R is the ideal gas constant which in SI units is 8.314 J·K−1·mol−1 or

in conventional units: 62.36 mmHg·K−1·mol−1

At a constant temperature, Eq (1.2) simplifies to Pµn / V Concentration is

defined as moles/unit volume, i.e n/V; hence pressure is proportional to gas

concen-tration, i.e the greater the number of gaseous molecules in a given volume, the more force those gas molecules will exert on the walls of the container The physical reaction of dissolving in solution is also proportional to the partial pressure of the gas above the solution, so that for oxygen dissolution [12]:

The rate constant KForward in Eq (1.4) describes the proportion of oxygen gas that

dissolves per unit time, while KReverse describes the proportion of the dissolved centration that leaves the solution to the gas phase When the system described

con-in Eq (1.4) is at thermodynamic equilibrium, the concentrations con-in the gas and liquid phases are stable and the constants may then be combined, resulting in

Eq (1.5), also known as Henry’s Law SC is the Bunsen solubility coefficient where

S = K /K and is gas- and solvent-specific S is affected by other dissolved

solutes and falls with increasing temperature (i.e KForward becomes smaller and

KReverse larger)

Due to difficulties in measuring the molar concentration of oxygen compared to measuring a volume of 100 % oxygen at standard conditions (STPD: 0 °C, 760 mmHg, dry gas), it is customary to report oxygen content in mL·dL−1 Under these conditions, oxygen approximates an ideal gas such that 6.02 × 1023 gas molecules (i.e 1 mol) occupy 22.414 L at 0 °C Quantification of human oxygen consumption is performed using STPD rather than BTPS (body temperature and pressure, saturated: Eq (1.7)) [13] as water vapour in the latter partially condenses with increasing pressure This vapour results in a significant deviation from an ideal gas and invalidates the relation-ship between the number of molecules and volume defined in Eq (1.2)

The Solubility of Respiratory Gases in Solution

The Bunsen solubility coefficient of oxygen is 0.003082 mL·dL−1·mmHg−1 and describes the measured solubility corrected to STPD Utilising this conversion, the solubility coefficient of oxygen in normal plasma is 1.38 × 10−3 mmol·L−1·mmHg−1

while carbon dioxide is nearly 22 times more soluble at 3.08 × 10−2 mmol·L−1·mmHg−1

[14] Thus, using Henry’s Law (Eq 1.5) in normal arterial blood the concentration

of dissolved carbon dioxide is nearly ten times that of dissolved oxygen:

(1.6)

It should be noted that this does not include oxygen and CO2 in chemical equilibrium with the dissolved gas such as that combined with haemoglobin or in reaction with water The significantly higher plasma concentration of dissolved carbon dioxide (Eq 1.6) resulting from its greater solubility allows for more rapid elimination by gas exchange membranes when compared to oxygen under the same flow conditions

In the gaseous phase, the partial pressures of individual gases combine to equal the total ambient pressure that the gas mixture exerts on its container, allowing each gas

to be reported as a fraction of the total For example, the partial pressure of oxygen in inhaled 37 °C air that is fully saturated with water vapour at 1 atm (i.e BTPS) is:

This summative requirement is only met when the solution is exposed to a gas phase

As solubility coefficients vary between gases, the number of moles of dissolved gas

in a given quantity of solution that is in contact with a gas phase has no such lent summation, i.e the summation of the partial pressures in solution will not equal atmospheric pressure

equiva-Since the partial pressure of a gas is proportional to the concentration in solution (at equilibrium), it may be used as a substitute for concentration even when no gas phase is present, as is the case for body fluids In this case it represents the partial pressure that would be required of a gas phase to maintain the existing concentra-tion of dissolved molecules in solution.

If a dissolved gas is consumed by chemical reactions in solution (e.g aerobic metabolism of oxygen), the equilibrium partial pressure required of a gas phase falls Upon exposure to a gas phase with a higher partial pressure, such as in the lungs or an oxygenator, gas molecules will dissolve, increasing the solution concen-tration until equilibrium is again reached This is the primary advantage of express-ing concentrations of dissolved gases in body fluids as partial pressures—apart from measurement techniques, it allows easy quantification of the concentration gradient from the site of gas exposure to the site of usage Its inconvenience comes when considering stoichiometric relationships such as described under respiratory quotient above

Biophysics of Membrane Oxygenation

The equations and constants introduced thus far describe a steady-state where a fixed quantity of gas is in equilibrium with a solution and assumes instantaneous reactions occurring in stationary homogenous mediums However, both in the human body and in the oxygenators used for ECMO, an exchange membrane is always interposed between the gas phase and the body fluids it dissolves in Even in the case of porous materials, a gas–blood interface is usually prevented by forma-tion of a biofilm comprising adsorbed blood proteins that forms after a short period

of operation These factors impose a time constraint in which gas exchange can occur and requires consideration of the mass transport of molecules into the body as

flux (J), defined as the passage of a quantity of solute per unit time.

The membrane flux of solute down a concentration gradient is described by Fick’s Law of Diffusion [15]:

mem-gradient across the membrane (ΔC, mmol·mL−1 or mmol·cm−3) and the area of the

diffusion front (A, cm2) The minus sign is mathematically required to describe flux from a high concentration to a low concentration [15] This universal statement of mass transport is applicable not only to the extracorporeal oxygenator, but also of oxygen moving from plasma to interstitial fluid and into cells

The diffusivity coefficient, D, is expressed as area over time (cm2·s−1) and describes

a unique constant of the gas, barrier, and solution under steady state conditions Higher numbers represent greater diffusibility with coefficients in gases being

orders of magnitude greater than coefficients in liquids Low-molecular- weight gases diffuse more quickly than higher-molecular-weight gases, and higher tem-peratures provide gas molecules with greater kinetic energy, increasing diffusion rates [11] This is in contrast to the solubility of the gas which decreases with higher temperature; however, it must be recalled that diffusivity specifies a transfer rate whereas solubility describes concentrations at equilibrium.

Describing diffusion in porous media—such as membranes in hollow fibre genators—requires more parameters to be incorporated into the constant, resulting

oxy-in effective diffusivity, DEff, that for an isolated membrane has the following parameters:

Under real conditions the value determined for the effective diffusion coefficient

is inseparable from the properties of any biofilm of adsorbed proteins or stationary layer of blood [11, 16] Furthermore the greater part of oxygen traversing the mem-brane immediately undergoes a chemical reaction with haemoglobin until the latter

is saturated This chemical reaction sustains the concentration gradient and is termed

an enhancement factor After incorporating haemoglobin, the DEff for oxygen depends not only on the membrane characteristics discussed, but is also a function

of haemoglobin concentration (the haematocrit, %Hct) Equation (1.10) strates an example of a term for the effective diffusivity of a membrane exposed to

demon-a turbulent bovine blood stredemon-am [18, 19]:

DEff =(2 13 0 0092 - ×%Hct)´10- 5 (cm s2 - 1) (1.10)

The Driving Force for Diffusive Transport of Gases

By combining Eq (1.5) (Henry’s Law) for solute concentration and Fick’s Law (Eq 1.8) an equation for diffusive membrane flux can be derived where k

is the permeability constant—the product of the solubility coefficient (SC, mmol·L−1·mmHg−1) already defined and effective diffusivity (DEff, cm2·s−1) having the units mol·cm−1·s−1·mmHg−1 [20]:

After a brief period of operation any extracorporeal gas exchange membrane exposed to blood will develop a film of adsorbed blood proteins and clotting factors (Fig 1.3) This is unlikely to be of uniform consistency and may be thicker in areas

of the oxygenator exposed to lower flows Similarly the total membrane area, A,

slowly decreases over the membrane’s operational life due to macroscopically visible fibrin deposition

In spite of this gradual decline in performance, over short periods of operating

time the average membrane thickness, l, can be considered to be constant and, under

steady-state conditions, can be combined with -k

O2 into a single constant, resulting

in a quantitative statement that the flux of gas is proportional to the driving force and is opposed by certain resistances [15] and should be familiar as equivalent to statements of resistance:

Mass transfer per unit area Driving Force

Resistance to Transportm

=

mmols cmwhere

Membrane Exposure Time

For a static liquid below a static gas, the rate of diffusion will decrease exponentially until equilibrium when the conditions of Eq (1.5) are met, i.e the rate of gas dis-solving into the liquid is equal to the rate of molecules leaving the liquid (Fig 1.4a)

It is clear from this figure that too short an exposure time will result in submaximal oxygenation

The more complicated relationship when haemoglobin is present is displayed

in Fig 1.4b Here the oxygen content of the blood displays a plateau due to the oxygen–haemoglobin dissociation curve (discussed below)

In ECMO, the time of blood exposure to the membrane is proportional to the length of membrane/hollow fibre traversed and inversely proportional to the blood flow rate As the length of microtubules adds to resistance to blood flow, determin-ing the optimal length for oxygen flux over the physiological range of blood flows likely to be encountered is an important design parameter With current hollow fibre oxygenators this length is around 4 cm for adequate oxygenation, with shorter lengths needed for carbon dioxide removal (see ECCOR below)

Relative Flow Direction

Figure 1.4a depicts a stationary blood and gas phase; however, a similar pattern exists for two phases moving in the same direction (co-current flow) and thus at low flow rates equilibrium will occur and diffusive flux will cease Inspection of the figure makes it apparent that replacing gas partly depleted of oxygen (where CO2

also contributes to the total partial pressure) will maintain the concentration ent Utilising countercurrent flow, where blood and gas flow in opposite directions, decreases the maximum concentration gradient at the blood inlet end of a hollow fibre but increases the gradient at the outlet, thereby maintaining a gradient over the entire fibre and allowing flux to continue along the membrane relatively indepen-dent of flow rates Cross-current flow is also utilised (demonstrated in Fig 1.3) resulting in differing gradients across the blood stream

Effect of Turbulence and Haematocrit on Local

Concentration Gradients

In an environment where oxygen exchange is occurring, the uptake of oxygen by haemoglobin maintains diffusion in plasma toward red cells [14] and augmenting haematocrit increases the flux of oxygen into the blood (Fig 1.4b) Creation of

Partial Pressure of Gas Phase

Partial Pressure vs Time for a static gas/liquid

O2 Content vs Membrane Exposure Time

Time (seconds) Time

14.00 12.00 10.00 8.00 6.00

O2 Content mL/dL 4.00

2.00 0.00

Partial Pressure Liquid Phase

Fig 1.4 Partial pressure vs time for a gas dissolving in a liquid: (a) Depicts partial pressures

approaching equilibrium for a static solution below a gas phase The slope of the curve (i.e the rate

at which equilibrium is approached) is proportional to the concentration difference (b) Describes

the oxygen content of serum containing red blood cells vs time exposed to 100 % oxygen across

a membrane Time has been adjusted to approximate transit times in current oxygenators Data derived from Katoh [ 18 ]

turbulent flow vortices more effectively purges tubule contents than laminar flow (discussed below) and brings erythrocytes into closer proximity to gas exchange membranes, increasing the local concentration gradient.

Resistances to Diffusive Transport

The total resistance to mass solute movement in Eq (1.12), RTotal, is the sum of

component resistances, which can be divided into gas phase resistance (RG),

mem-brane resistance RM, and blood side resistance, RB: RTotal = RG + RM + RB Of these RG

is negligible and the factors influencing RM have been extensively discussed The most

variable component is RB due to the formation of a stationary film on the blood side of the membrane (Fig 1.3 boundary layer)

Area of the Gas Exchange Membrane

It is not practical for manufacturers to specify a membrane surface area (A, Eq

(1.12)) that encompasses the complex microscopic geometry of a membrane and its

pores, the latter being incorporated into DEff as discussed [16] Furthermore, it not be guaranteed that any oxygenator design utilises the entire membrane area

can-evenly Thus, while flux per unit area (J/A) is a useful description of isolated

mem-brane performance, oxygenators are better characterised by their total flux by porating area into the equation for resistance:

Although lacking the precision of Eq (1.12) in defining properties of the membrane,

Eq (1.13) can be incorporated into monitoring gas exchange efficiency of an vidual oxygenator over time and will be discussed after oxygen carriage is consid-

indi-ered The value of JOxygenator is usually reported in oxygenator product specification sheets at varying blood flows

Ultrafiltration of Plasma Water Over the Oxygenator Membrane [ 15 ]

Analogous to the membrane flux of oxygen down a concentration gradient is the movement of water from the plasma to the gas phase The driving force here is hydrostatic pressure, rather than concentration gradient, and is generally defined as

a conductance (the inverse of resistance) termed the coefficient of ultrafiltration

(KUF, mL·min−1·mmHg−1):

=-

(1.14)

QF (mL·min−1) is termed the ultrafiltration rate and describes the appearance of fluid

within the gas containing hollow fibres, while the terms for pressure (PBlood and Pgas) describe the heights of a fluid column relative to atmospheric pressure in each com-partment Many of the factors already described for resistance to diffusion will be contained in the ultrafiltration coefficient and will not be discussed further The ultra-filtrate represents a homogenous fluid that will contain dissolved solutes from plasma proportional to the size of membrane pores, which are typically smaller than 1 μm

A major drawback of early microporous membranes was significant plasma age, as the open porous structure allowed significant ultrafiltration This has been significantly alleviated by newer closed-fibre membranes; however, some water flux still occurs under normal operating conditions The water then evaporates in the fresh gas flow and leads to an insensible water loss proportional to the fresh gas flow and may reach significance when supporting low bodyweight patients [10]

Oxygen Transport

As outlined above, oxygen has a limited solubility in plasma of 1.39 × 10−3 mmol·mmHg−1

or 0.0031 mL·dL−1·mmHg−1 at 37 °C For a normal arterial oxygen partial pressure

of 100 mmHg this equates to about 3 mL of dissolved oxygen per litre of blood

At that oxygen content, maintaining a nominal body oxygen consumption of

250 mL·min−1 would require a cardiac output of 80–120 L·min−1 and even breathing

100 % oxygen at normal atmospheric pressure would not sustain aerobic cellular metabolism, providing only 20 mL of oxygen per litre of blood [14]

12 and 16 g·dL−1 However, it is quite common for anaemia to be present in cally ill patients, and those receiving ECMO are often transfused to maintain a haemo-globin concentration around 10 g·dL−1, equating to 1.55 mmol·L−1 (assuming a molecular weight of haemoglobin of 64,458 g·mol−1) [21]

criti-Haemoglobin is a spherical molecule consisting of four globin subunits (two α and two β chains) with each globin containing a haem group in a peripheral molecu-lar crevice Each haem molecule consists of a central iron atom in the ferrous (Fe2+) state between two histidine amino-acids This structure allows the iron atom to bind oxygen without being oxidised to Fe3+, a change that would prevent further oxygen binding Haemoglobin demonstrates cooperative binding whereby the binding of oxygen to the iron moieties is enhanced if another binding site on the same molecule

is already occupied by oxygen As each haemoglobin molecule has four binding sites, it can exclusively be 0, 25, 50, 75 or 100 % oxygenated Haemoglobin satura-tion refers to the fractional occupancy of all the oxygen binding sites in a solution, and due to cooperative binding results in the sigmoid haemoglobin dissociation curve (Fig 1.5) [3]

Apart from the conformational change induced by oxygen itself, four other major factors influence the conformational state of haemoglobin—carbon dioxide, hydro-gen ion concentration (pH), 2,3-DPG and temperature By altering the affinity of haemoglobin for oxygen, each of these factors affects how saturated the haemoglobin

in a given quantity of blood is with oxygen at any partial pressure, and thus the gen content of that blood If all haemoglobin binding sites are occupied by oxygen

oxy-Oxygen Hemoglobin Dissociation and oxy-Oxygen Content

Partial Pressure Oxygen (mmHg)

SaO2 at pH 7.44 Temp 37.5° & BE 0 SaO2 at pH 7.32 Temp 37.5° & BE 0 Oxygen Content (ml/dL) for Hb 10g/dL

pH 7.44 Temp 37.5° & BE 0 Oxygen Content (ml/dL) for Hb 10g/dL

pH 7.32 Temp 37.5° & BE 0

Fig 1.5 The oxygen–haemoglobin dissociation curve as calculated by the Thomas modification

of the Kelman Eq ( 1.9 ) Also shown is the oxygen content for haemoglobin concentration of

10 g·dL −1 after applying Eq ( 1.15 )

(100 % saturation) the maximum oxygen carrying capacity is 1.39 mL per gram of haemoglobin in adults [13] or 1.312 mL per gram of foetal haemoglobin [14] Thus, the total oxygen content of adult arterial blood (CaO2) can be described by:

CaO2 =Hb SaO´ 2´1 39 0 0031 + ´PaO mLdL2( - 1) (1.15)The total oxygen content can then be multiplied by blood flow (cardiac output) to

give oxygen delivery, DO2 which has the units of flux: mL·min−1 Note the scaling factor of 10 to convert the units of oxygen content to mL·L−1

DO2 =QB´CaO2´10 (1.16)

Modulation of Haemoglobin’s Affinity for Oxygen

The Bohr effect describes alterations in haemoglobin oxygen affinity due to carbon dioxide and hydrogen ion concentrations Carbon dioxide binds to amino acids in the outer chains of haemoglobin to form carbaminohaemoglobin, stabilising the molecule with the ferrous elements in deeper crypts and facilitating release of oxy-gen from haemoglobin Similarly, increasing temperature and increasing hydrogen ion concentrations stabilise haemoglobin in the deoxygenated state Similar to oxy-gen, carbon dioxide binding is reversible and in compartments with a low CO2 con-centration the effect is reversed, promoting oxygen uptake (the Haldane Effect) This is of importance when considering oxygenation in systems designed primarily for CO2 removal and will be discussed later in the chapter

The glycolysis product, 2,3-DPG also decreases haemoglobin affinity for gen 2,3-DPG binds to deoxygenated haemoglobin, lowering the apparent affinity for oxygen by altering the electrostatic bonds that maintain the quaternary configu-ration [9] This has the most significance in transfused blood where, after 2 weeks

oxy-of storage, 2,3-DPG levels become negligible After transfusion oxy-of this blood DPG levels do not return to normal until nearly 48 h [22] In the absence of 2,3-DPG the affinity for oxygen is increased resulting in less oxygen delivery at the same periph-eral PO2 As will be seen, this is only likely to be a factor in the most severe oxygen-ation problems and the benefit of increased oxygen binding sites from transfusion will outweigh the transiently lower delivery (see Figs 1.8 and 1.9)

The Oxygen–Haemoglobin Dissociation Curve

The oxygen–haemoglobin dissociation curve is characterised by an upper plateau at higher partial pressures of oxygen where haemoglobin is between 90 and 100 % saturated Below this is a steep shoulder where the saturation of haemoglobin rap-idly decays as the partial pressure of oxygen falls Physiologically the factors that

shift the curve to the right (i.e to a lower affinity state) are local tissue metabolism and hence local oxygen consumption (Fig 1.5).

Several equations exist to model the normal oxygen–haemoglobin dissociation curve One of the most informative is the Thomas modification [23] of the Kelman [24] equation and its inverse [25], which calculate the haemoglobin saturation for any partial pressure of oxygen and allow for shifts of the curve due to temperature, [H+] and carbon dioxide

For convenience, the oxygen–haemoglobin dissociation curve is frequently described by the partial pressure at which 50 % of haemoglobin is saturated A nor-mal P50 for arterial blood is 26.3 mmHg Values higher than this describe a “right- shifted” curve, i.e a haemoglobin with lower affinity for oxygen Two curves of haemoglobin saturation are demonstrated in Fig 1.7, the only difference being the hydrogen ion content reflecting the higher carbon dioxide concentration in venous blood At high oxygen partial pressures consistent with arterial blood, the differ-ence in haemoglobin saturation is minimal In contrast, there is a significant differ-ence in the saturation of Hb at a partial pressure of 40 mmHg commonly found in venous blood

The implications of the oxygen Hb dissociation curve become clearer when oxygen content is also plotted on the same chart (Fig 1.5) At haemoglobin of

10 g·dL−1 the oxygen content at a partial pressure of 100 mmHg is 13 mL·dL−1 and minimally affected by the arteriovenous pH difference However, at an oxygen partial pressure of 40 mmHg, a 1 mL·dL−1 difference becomes apparent between the two content curves, being 10 mL·dL−1 at pH of 7.44 and 9 mL·dL−1 at pH of 7.32

In the tissues a drop in content can only occur if the oxygen is utilised by lism, so this “right-shifting” of the curve as the products of metabolism acidify capillary blood serves to bolster the partial pressure gradient for diffusion from capillary to cell

metabo-It should be noted that “right-shifted” oxygen–haemoglobin dissociation curves, while advantageous for unloading oxygen in acidotic tissues, may be counterpro-ductive at sites of oxygen uptake if abnormally low alveolar partial pressures exist Inspection of Fig 1.5 reveals that if oxygen uptake were to occur at a partial pressure

of 60 mmHg, a right shifted curve (pH 7.32) will carry 0.5 mL·dL−1 less oxygen This gap widens if uptake occurs at even lower partial pressures

Mixing Blood of Differing Oxygen Partial Pressures

The oxygen–haemoglobin dissociation curve is of importance when considering mixing blood streams with differing oxygen concentrations If the volumetric flow rate of both streams is similar then the haemoglobin saturation of the two streams can be averaged as a reasonable approximation of the resulting mixture However,

at differing flows, accurately calculating both the resultant oxygen tension and ration requires conversion to oxygen content and measurement of the flow rate of the two streams As mixing oxygenated blood is fundamental to ECMO, the steps of this process will be worked through (Table 1.1)

As demonstrated in Table 1.1, the average of the venous and arterial blood rations approximates the complete solution suitably when the flows are similar, but overestimates saturations when the volumetric flow differs The approximation worsens if the venous oxygen tension is reduced further In contrast, taking the mean O2 tensions of the unmixed samples massively overestimates the final partial pressures after mixing.

satu-Understanding this concept is important as it highlights a physical limitation on

systemic oxygenation: utilising a saturable oxygen carrier (haemoglobin) makes

oxygen delivery flow-limited Even if supranormal oxygen tensions are achieved via

an extracorporeal circuit, an inadequate ratio of circuit flow to cardiac output may result in suboptimal oxygen delivery

Veno-venous ECMO and Oxygen Transport

Having covered oxygen carriage and mixing blood with differing oxygen content, veno-venous (VV) extracorporeal membrane oxygenation can now be considered

As blood is accessed and returned to the venous system/right atrium (Fig 1.6),

Table 1.1 Mixing blood streams with differing oxygen concentrations

O2 Tension

O2 Content (mL)

Final partial pressure (mmHg) 51.7

O2 Tension

O2 Content (mL)

Final partial pressure (mmHg) 41

Two examples of mixing blood streams are given, one with matched input flows and one with fering input flows For simplicity only the volume of the blood stream is described; however, it can

dif-be assumed that the blood flow rate is this volume per minute In all calculations the haemoglobin

is assumed to be 10 g·dL −1 and is multiplied by the blood volume to give the total mass of globin In the first example the flows have been set to 1 dL·min −1 so that values for haemoglobin and oxygen content equate to values shown in Fig 1.5 The mean saturation and tension is shown for comparison to the result after converting to content

haemo-systemic oxygen delivery remains dependent on cardiac output, making this system simpler to quantitatively analyse.

Various access configurations are utilised (Fig 1.1); however, one of the most common techniques is to cannulate the common femoral vein with a cannula which has side- and end-fenestrations that allow blood to be drained from multiple points along the inferior vena cava This access cannula typically ends around 10 cm below the cavo-atrial junction, while the return cannula has a single terminal orifice in the right atrium If higher extracorporeal circuit flows are attempted, the inferior vena cava may collapse around the multistage cannula, intermittently obstructing flow and causing the external circuit to “shudder”, with consequent hypoxia If these higher circuit flows are necessary to achieve adequate oxygenation and this negative access pressure cannot be resolved by giving fluid, a second access cannula may need to be placed in the internal jugular vein

Figure 1.6 demonstrates a basic circuit It can be appreciated that the ECMO circuit is in parallel to venous return and thus a mixture of oxygenated blood from the ECMO circuit and deoxygenated blood from the venous return will enter the right heart To appreciate the implications of this parallel circuit, consider a young adult with severe acute respiratory distress syndrome fully supported by VV ECMO with both femoral and internal jugular access (Fig 1.6) Chest X-ray demonstrates bilateral “white-out” and it will be assumed the lungs are not contributing to systemic oxygenation

Fig 1.6 Schematic of a VV-ECMO circuit It will be assumed that both femoral and internal

jugu-lar access exists so all venous return is modeled with one vessel A mixture of fully oxygenated and venous blood enters the right heart and perfuses the pulmonary circulation before the left heart distributes it systemically (hence end-tidal CO2 will not reflect mixed venous PCO2) Varying with the cannula position and venous return (cardiac output), some amount of recirculation is very com- mon, reducing the amount of oxygen delivered Depending on recirculation, it may not be possible

to sample true mixed venous blood

The patient has the following parameters: cardiac output of 4 L·min−1, ECMO flow of 4 L·min−1, a mixed central venous haemoglobin oxygen saturation of 52 % and an arterial saturation of 99 % What will be the effect on his saturations if his cardiac output were to rise to 7 L·min−1 or fall to 2.5 L·min−1?

Apart from the non-contribution of the lungs, other assumptions include an oxyhaemoglobin dissociation curve with a normal P50 and good cannula position with minimal recirculation between the access and return lumens in the inferior vena cava To fully develop this model and illustrate several key points two other param-eters are required: the haemoglobin concentration and the patient’s total oxygen con-sumption (VO2) Initially, this hypothetical patient has a haemoglobin concentration

of 10 g·dL−1 and is at steady state consuming 250 mL·min−1 of oxygen with no ers of tissue hypoxia Thus, there are four independent variables: the cardiac output, the ECMO flow rate, the haemoglobin concentration and the target VO2

mark-While providing circulating oxygen is the goal of ECMO, it must be highlighted that the VO2 required to avoid anaerobic metabolism is a parameter that can only be achieved if the amount of oxygen delivered matches consumption If this required

VO2 exceeds tissue delivery, VO2 is then said to be supply-limited, initially resulting

in higher oxygen extraction with mixed-venous saturation decreasing first, followed

by clinical and biochemical markers of hypoxia such as confusion, oliguria and ing lactate as anaerobic metabolism ensues

ris-Figure 1.7a suggests that under these conditions, increasing the cardiac output from 4 to 7 L·min−1 will cause a drop in the arterial saturations from 99 to 84 %

In contrast dropping the cardiac output to 3 L·min−1 will not affect the arterial ration but will cause the central mixed venous saturation to fall from 52 to 25 %.The equations required to generate this model are the content and delivery equa-tions (Eqs 1.15 and 1.16) from which the principle of conservation of mass is applied

satu-to determine the oxygen content at each of the following points: the ECMO return cannula, the venous return, and the pulmonary artery (Fig 1.6) It should be noted that

in VV-ECMO the pulmonary artery is not the correct sampling site for mixed venous saturation Instead, the pre-oxygenator blood is the best approximation (but depend-ing on the vessels cannulated may not include superior vena cava flow)

The methods for mixing blood streams (Table 1.1) are used to determine the nary artery oxygen content and saturations In the absence of any lung function this is modeled as the systemic arterial oxygen delivery Modelling the solution requires an iterative approach to determine the highest achievable VO2 (if the required VO2 cannot

pulmo-be met) by altering the tissue oxygen extraction—this determines the venous oxygen flux to the right heart Example values from Fig 1.7a are shown in Table 1.2

Arterial Saturations Are Dependent on the Fraction

of Cardiac Output Captured

To explain the fall in arterial saturations with increasing cardiac output it should be appreciated that in this case the required VO2 did not change What did change was the total venous return, which increased by 75 % with the increased cardiac output

The ECMO flow remained at 4 L·min−1 so the ECMO “shunt”, i.e the fraction of deoxygenated venous blood in the right ventricle increased from 3 to 43 % The shunt is calculated as (Cardiac output—Oxygenated Blood flow)/Cardiac output where oxygenated blood flow equals the cardiac output if the cardiac output is less than ECMO blood flow The oxygenated blood volumetric flow will equal ECMO set flow minus any recirculation when cardiac output is greater than ECMO flow.

In summary, the resulting arterial saturation is analogous to the example of mixing blood streams of different content and the same effect can be achieved if cardiac output is left constant and ECMO flow is decreased (Fig 1.7b) In reality, increases

in cardiac output are likely to be accompanied by an increased VO2 and in this setting the mixed venous saturation will fall (Fig 1.7c)

Fig 1.7 Arterial (SaO2 ) and central mixed-venous oxygen saturations (SvO2) vs cardiac output in a patient fully supported by VV-ECMO Graph titles show ECMO flow conditions and haemoglo-

bin See text for discussion (a) VV-ECMO blood flow 5L.min–1 , body oxygen consumption 250ml min –1 and hemoglobin 10g.dl –1 (b) VV-ECMO blood flow 3L.min–1 , body oxygen consumption 250ml.min –1 and hemoglobin 10g.dl –1 (c) VV-ECMO blood flow 4L.min–1 , body oxygen consump- tion increasing with cardiac output, hemoglobin 10g.dl –1 (d) VV-ECMO blood flow 4L.min–1 , body oxygen consumption 250ml.min –1 and hemoglobin 7g.dl –1

Table 1.2 ECMO set rate (L

−1 ) Oxygenated blood flow (factors recirculation) (L·min

−1 ) Flux of fully oxygenated blood delivered by ECMO (mL·min

−1 ) Cardiac output/ venous return (L·min

−1 Oxygenated blood flow refers to the volumetric flow of oxygenated blood into the right heart For

−1 rather than mL·dL

Mixed Venous Saturations Are Determined by VO2

and Cardiac Output

To appreciate why the mixed venous saturations fall to the left of Fig 1.7a (where VO2

was held constant) when the cardiac output falls, recirculation between the access and return cannula must be considered Recirculation is an additional variable to the independent variables identified above

Several factors influence recirculation; however, this model incorporates only cardiac output: if the heart stops, there will be no venous return and recirculation might approach 100 % (if the great veins do not collapse), i.e blood will enter the ECMO circuit from the IVC and return via the right atrium, flowing in a retrograde fashion back down the IVC, resulting in no systemic oxygenation (amongst other adverse effects!) Similarly at low cardiac outputs, the blood not ejected will recir-culate; however, it becomes saturated with oxygen on the first pass and will be

unable to carry more This leads to an important point: when the ECMO flow falls

below cardiac output, the amount of oxygenated blood ejected by the right heart is cardiac output-limited Conversely, at high cardiac outputs and hence high venous return, recirculation will be minimal as the oxygenated blood will be “washed through” the right atrium into the right ventricle

Under conditions where the cardiac output is less than the ECMO flow, the blood entering the arterial circulation is fully saturated with oxygen However, to supply the required VO2, oxygen extraction from the low output has to increase, and hence, mixed central venous oxygen content and venous saturation fall

Figure 1.7c demonstrates falling mixed venous saturations with increasing

VO2 requirements at higher cardiac outputs This is caused by a combination of falling arterial oxygen content by the ECMO fraction of cardiac output mecha-nism described above AND increased extraction to achieve the higher VO2

requirement

The Effect of Oxygen Carrying Capacity

The parameters in Fig 1.7d are identical to Fig 1.7a except for a lower bin, so these graphs illustrate the effect of haemorrhage and transfusion The lower oxygen-carrying capacity means that at any cardiac output, the oxygen extraction must be greater, and hence, the venous saturation will be lower At lower cardiac outputs, the combined effect is enough that no further oxygen extraction can occur and the achieved VO2 falls lower than the target VO2—under these conditions signs

haemoglo-of tissue hypoxia will occur Thus in the setting haemoglo-of low cardiac output and borderline oxygenation, increasing the haemoglobin by transfusion may alleviate hypoxia while consideration is given to circulatory support